PDA

View Full Version : Project Finally under way



timc
12-19-2007, 03:04 AM
The swedish beauties have finally arrived. Matsj have provided me with a pair of gorgeous H9800 clones. The 2435HPL i purchased earlier.

Hooked them up yesterday just for some prelimnary testing and they sound great. I'm crossing them over at 1khz and using the 126A woofer in my L65's for the lower octaves. I still have to do some sensitivity adjustment to the horn and tweak the crossover point a bit, but for a first try I think it sounds quite good. I wont have time to do anymore before over the hollidays but i think the rest of the winter will be very exciting. Hi-Fi wise at least:applaud:


-Tim

Joe Alesi
12-19-2007, 03:21 AM
The swedish beauties have finally arrived. Matsj have provided me with a pair of gorgeous H9800 clones. The 2435HPL i purchased earlier.

Hooked them up yesterday just for some prelimnary testing and they sound great. I'm crossing them over at 1khz and using the 126A woofer in my L65's for the lower octaves. I still have to do some sensitivity adjustment to the horn and tweak the crossover point a bit, but for a first try I think it sounds quite good. I wont have time to do anymore before over the hollidays but i think the rest of the winter will be very exciting. Hi-Fi wise at least:applaud:


-Tim
Hello Tim,

You have conquered the hard part- getting started:D. Sounds like a great project- good luck and please share some progress pics when possible. I just noticed you are a K2 s9800 fan. Me too. Any chance of a big MTM for you along the same lines- using the H9800?

timc
12-19-2007, 04:00 AM
Hi.

Yes. The hardest part is indeed getting started ;)

Im not going MTM. When i can afford it ill try to get a pair for 1200FE woofers. My inspiration is Greg Timbers Personal system. But without subs.

I dont have a digital camera myself, but i will try to borrow one over the hollidays.

The 126A is just a temporary solution.


-Tim

timc
01-17-2008, 02:03 AM
Hi.

Im going to put together a temporary passive crossover now, and im a bit confused about polarity.


Im going to use a 2nd order butterworth at 1000hz. The schematics says to inverse the polarity on the tweeter. The compression driver says to put the negative wire on the positive connector for positive compression.


Does this result in that i should put the positive wire on the positive terminal?


Edit: Have changed to 2nd order.


-Tim

Robh3606
01-17-2008, 08:32 AM
Hello Tim

Nice set-up!! If positive compression is towards the throat then hook them up that way. The 126A should be JBL polarity which is opposite the rest of the world. Just take a 1.5volt battery and verify the polarity of the woofer.

Rob:)

timc
01-17-2008, 09:30 AM
Hello Tim

If positive compression is towards the throat then hook them up that way.

Rob:)



How do i know if this is the case? Positive compression in this context could mean both ways. Is there a standard for compression drivers?


-Tim

Robh3606
01-17-2008, 10:23 AM
If you look at any of the contemporary compression driver data sheets they all have the same note about + voltage applied on the negative cause movement towards the phase plug. If you have a diaphram move towards the phase plug you will get a positve preasure wave into a horn. Same as a woofer diaphram moving forward.

The phase of the driver can also be influenced buy the lenght of the horn. Go with opposite phase but also try in phase as well. Use what sounds best or if you have an RTA what gives you the smoothest response through the crossover region.

Rob:)

timc
01-17-2008, 10:24 AM
Thanks for the help Rob. I will use RTA when completing the setup. Just wanted to know how the stuff works.



-Tim

Robh3606
01-17-2008, 10:30 AM
Hello Tim

Your welcome. If you have an RTA your are set. You can easilly see what's going on and make adjustments as required.

Rob:)

timc
01-21-2008, 02:14 AM
A small update.


Hooked up a passive crossover this weekend. 12db slope at 1khz. In addition I'm using the timbers filter, and some EQ above 14khz.


The transition between woofer and horn was quite good, considering this was just a theoretical filter and I dont have a spec sheet for the 126A. The horn plays a bit too loud so I'll borrow a Copland DRC205 from work to determine how much more damping is needed. This will also give me an accurate in-room response measurement.


-Tim

timc
01-22-2008, 11:52 AM
I have now done some measurements. It doesnt look too bad. I was surprised that the horn actually played a bit to low. I have to increase the output by 2-3 db. And then use the EQ to flatten 2 bumps.

-Tim

Jakob
01-23-2008, 05:40 PM
Hi Tim!

Nice project you have going there!

May I ask what amplifiers You are using?

I also wonder how You mounted your drivers to the horns. I've been trying to find UNC 1/4" machinescrews with countersunk heads, but it's impossible. I've searched the net for any info regarding retailers that ship internationally, but for no good...

Regards: Jakob (who also will get a pair of horns from Mats! :D )

Zilch
01-23-2008, 07:18 PM
http://www.mcmaster.com/

Search "Machine Screws," then "Flat Head"

$7.70 for 50 in 18-8 stainless, 1/2" long....

timc
01-24-2008, 02:06 AM
Hi Tim!

Nice project you have going there!

May I ask what amplifiers You are using?

I also wonder how You mounted your drivers to the horns. I've been trying to find UNC 1/4" machinescrews with countersunk heads, but it's impossible. I've searched the net for any info regarding retailers that ship internationally, but for no good...

Regards: Jakob (who also will get a pair of horns from Mats! :D )


Hi.

I used skrew with normal heads and drilled a hole in the back of the horn to sink them into. (very bad english here.......)


At the moment im using a Crown XTi-1000. It sounds very bad in the mid/highs but the bass is incredible. My plan is to buy a set of Quicksilver Audio Horn mono amplifiers for the horn and use the XTi on the woofer. I will also need a 4+ channel preamplifier because i want the digital crossover to have a high input signal. That means placing it right after the player.

Now all thats missing is $$$ :(

-Tim

macaroonie
01-24-2008, 02:24 AM
Jakob these fellows are here in the UK but I am sure they will help. I've had a couple of deliveries from them both fast and accurate.

http://www.modelfixings.co.uk/fixings.htm

Jakob
01-24-2008, 06:45 PM
Thank You guys, and Tim: didn't mean to highjack your thread, sorry!
I'm impressed with your "using-ordinary-screws-and-drill-holes-in-the-horn"-solution, why didnt I come up with that? Guess I know now why I'll never get a Nobel-prize...:o:

I'm also very interested in Quicksilvers amps. When I first heard about them (from someone on this forum, Widget maybe???) a couple of years ago they were not allowed to sell me any amps because the amps were not tested and approved by the european electric security department (bad english but you get the drift...) but now I see they have distributors around the globe, including Germany, so thats great news!!!

Macaroonie:
Love your country!!! Spent last summer driving around the highlands and visited Skye and Glen Coe amongst a lot of other places and Edinburg: what a lovely city! And the people are so friendly and humouristic. Guess that's why the vikings so often turned their longboats in that direction ;)

matsj
01-24-2008, 10:03 PM
I have some.

mats

timc
01-25-2008, 02:08 AM
Jakob: No problem. There is always room for "science" :D


Good luck with your project.




-Tim

timc
01-29-2008, 12:59 PM
126A and 2435HPL/H9800

timc
01-29-2008, 02:33 PM
Some more pics

timc
04-28-2008, 09:27 AM
Ok.

Im now moving forward again.


The measurements seem ok but for one thing.


I measure a big peak at 1250Hz on the 2435h/H9800. It's about 7-8db in amplitude.


Any idea what causes it? I'm using the HF compensation filter and a crossoverpoint at about 1khz with a 12db butterworth slope.


Any suggestions are welcome :)



-Tim

4313B
04-28-2008, 09:40 AM
I measure a big peak at 1250Hz on the 2435h/H9800. It's about 7-8db in amplitude.


Any idea what causes it?Yes. It's normal. Rather than go into what to do (my main computer is down right now along with all that data) you can examine Greg's networks such as the K2-S9800 network and Everest II network. Other networks to examine would be the SK2-1000 and 4338. You're going to have to come up with your own particular solution for your specific application. I believe I also posted the schematic for Greg's home system which you might want to look at too.

timc
04-28-2008, 09:47 AM
Yes. It's normal. Rather than go into what to do (my main computer is down right now along with all that data) you can examine Greg's networks such as the K2-S9800 network and Everest II network. Other networks to examine would be the SK2-1000 and 4338. You're going to have to come up with your own particular solution for your specific application. I believe I also posted the schematic for Greg's home system which you might want to look at too.


Thats a relief to hear really.


Im using Gregs Compensation filter. I thought that should take care of things?

Are we talking about a resonance compensation that i should apply?



Thank you very much for the reply :)


-Tim

4313B
04-28-2008, 09:59 AM
Are we talking about a resonance compensation that i should apply?I had it all worked out for Project May but I'll have to dig for all those files once I get that machine back up and running.

timc
04-28-2008, 10:01 AM
Yes. I had it all worked out for Project May but I'll have to dig for all those files once I get that machine back up and running.


Thanks. Then i have recognized it from the S9800 schematics. I'll have to check my drawers if i have the components for it :)


You'r very helpfull, and i really appreciate it. Thumbs up!


-Tim

4313B
04-28-2008, 10:09 AM
Here's something readily available that I ran a few weeks ago:

435Be vs 476Be on H9800 horn (raw curves, no networks/compensation)

The 435Be needs a couple more notch filters to smooth it out. Both sound really nice to some people like Greg and myself.

vernb
04-28-2008, 03:16 PM
Hello Tim,

You have conquered the hard part- getting started:D. Sounds like a great project- good luck and please share some progress pics when possible. I just noticed you are a K2 s9800 fan. Me too. Any chance of a big MTM for you along the same lines- using the H9800?

Hi Tim
What is MTM ?
Cheers
Vernb

timc
04-29-2008, 01:00 AM
Not sure about the exact phrase, but i think it plays on the D'appolito configuration with a tweeter in the middle of two midranges. Midrange-Tweeter-Midrange.


In the JBL case it should maybe be WTW.............



-Tim

pos
09-25-2008, 04:39 AM
Hi Tim,

How is your project going on?

did you use the compensation filter Zilch described here?
http://www.audioheritage.org/vbulletin/ ... stcount=84 (http://www.audioheritage.org/vbulletin/showpost.php?p=168664&postcount=84)

and curves:
http://www.audioheritage.org/vbulletin/ ... stcount=79 (http://www.audioheritage.org/vbulletin/showpost.php?p=168641&postcount=79)

I am curious on how these turn out on matsj's H9800 clones

timc
11-09-2008, 12:37 PM
Sorry for not answering you Pos. Havent checked the thread for while.

An update. I have now done some more work with the crossover. I started out with GT's notchfilter and the additional paralell notch that Zilch implemented, and then altered some values untill i got some half decent results.

I found out that my school has an older version of Clio. That speeded things up a bit.

The horn now has respons of +- 2.5dB from 1KHz to about 15KHz. Im satisfied with that for now. Ill post a graph in a day or two. Next step is to match the Woofer to the horn.

However I am wondering how Zilch and 4313B manges to get the measurements they get. No matter how i tweaked the filter, i just couldnt get it flatter. Thats just probably me not grasping it fully, but anyway :dont-know. I will go another round with it at a later point.

Could it be due to small differences between the Clones i got from Matsj and the original H9800 Zilch did the measurements with?

A question for the knowledgeable to finish: I initially intended to get the 1200Fe woofer for my project, but due to price and availability I might reconsider. What would be your recomended woofer for this 2 way system? LE14-h3? I'm not after the extreme bass extension, but rather a very quick, precise sound. An bassextension down to about 35Hz is enough for me. I will add a sub at a later point.

Thats it for now. I'll try to get both filters up and running during the next week. Needless to say I'm very excited :D




-Tim

4313B
11-09-2008, 01:03 PM
You might want to go with the newer LE14H-4 instead.

timc
11-09-2008, 01:07 PM
How is price and availability for the Le14-h4 ?


-Tim

4313B
11-09-2008, 01:11 PM
I'll send you a PM.

pos
11-09-2008, 01:48 PM
Hi Tim,


Could it be due to small differences between the Clones i got from Matsj and the original H9800 Zilch did the measurements with?

I think that is a major part of the problem (maybe also having the horn sitting on top of an enclosure instead of mounted on a baffle?...)

I think our clones are different than the real H9800 in the throat.

I have attached a picture of the throat of my clone, and you can clearly see the difference with the picture of Widget's clone in from the project may blog:
http://www.audioheritage.org/images/projectmay/blog/04-05/dismounted.jpg

Widget has integrated the metal plate inside the horn, whereas in our clones the mounting plate is simply added, with an additional cylindrical wooden throat. It is more like we are having a 3mm tube between the driver and the throat of the horn.
I think this is a problem when considering that the 2435 has a very short phase plug that was perfecly matched with the H9800 as can be seen here:
http://www.jbl.com/home/k2_story/k2_story_pics/S9800_draw_large.gif
as well as on a picture Rob posted of one of his 2435 without the mesh (I did not find the thread).

The second attached picture shows the situation we have on our mounting plates, compared to the correct approach that Widget used for Project May

I am thinking of removing this wooden part of the throat, and clamp the driver directly using speaker clamps (I am not really a fan of the metal mounting plate anyway, as it always needs adjustments to perfectly fit the aperture.

timc
11-09-2008, 01:53 PM
Hi Pos. This clarified quite a lot.

Consering the Wooden tap you mount the driver to. Wont you loose low end response when shortening the length of the horn?

How will you center the driver without it?


-Tim

pos
11-09-2008, 02:07 PM
Wont you loose low end response when shortening the length of the horn?
I hope that beside this wooden tap the rest of the horn geometry is correct. So if we remove it and find a way to correctly mount the driver we should have a correct H9800 clone, with the correct length.


How will you center the driver without it?
This is already quite difficult with the mounting plate (at least in my version), as the plate is quite loose around the driver.

Using speaker clamps the adjustment could at least be done directly with the driver mounted on the horn, by looking down the throat from the mouth of the horn.
And there is enough room to fix the clamps around the driver (this would be more difficult on a 435Be, but it also has 4 indentations for the screws that could be used to attach the clamp).

Another solution would be to use a very thin metal mounting plate and to reduce the length of the wooden tap, to minimize the problem.

The best solution would be to do as Widget did for project may:
http://www.audioheritage.org/html/projectmay/blog/2004-05.htm
(21 may)
But this is quite difficult and risky...

What do you think?

timc
11-09-2008, 02:19 PM
The thing Giskard did looks great, but i dare not try it without a spare horn.

If you remove the taps on your horn, let me know what you find. For now i dare not do anything more too them.


-Tim

pos
11-09-2008, 02:34 PM
I will start a new thread to ask for advices on this issue

timc
11-10-2008, 02:14 AM
Here is a graph of my measurements. I have no idea what happens above 10KHz. Could be just because i didn't use a proper setup for the Mic.

Note that the smoothening is only 1/12 :)

And please have in mind that this is my first real project, ever. So I'm kind of a noob really.......

Robh3606
11-10-2008, 05:29 AM
Do you have a measurement on the horn with no compensation, an impedance run on the horn and a schematic??

Rob:)

timc
11-10-2008, 05:33 AM
I didn't take a measurement without the compensation. and nor did i do an impedance plot.

I have the schematics but not on the computer, i will post it tomorrow.

I used the filter zilch made for the H9800 and just altered some of the values. The layout is identical.


-Tim

Robh3606
11-10-2008, 05:41 AM
I used the filter zilch made for the H9800 and just altered some of the values. The layout is identical.

Are you crossed over with an active crossover?? Did you measure both drivers?? How do they match up??

Rob:)

timc
11-10-2008, 05:43 AM
Hi.

No. i use a 2. order passive butterworth.

15uf in series and 1mH in paralell if i recall corectly. The filter placed before the compensation network.

I might have to change the values, but i just tried a few combinations and it gave a nice higpass rolloff.

I did not measure both drivers this time around. I have done measurments on both with a Copland Drc205 at home earlier, and they look quite similar.


-Tim

Earl K
11-10-2008, 07:02 AM
I used the filter zilch made for the H9800 and just altered some of the values. The layout is identical.

Tim ,

- Its best to avoid insulting those who have actually done the "heavy-lifting" work of making this driver/horn combo perform up to its potential .

- In my best recollection , Zilch merely implemented the compensation filter that Giskard gave to him ( & as originally presented by Greg Timbers ) .

- Giskard has done a lot of work making the combo work down to a 800 hz crossover point ( maybe even lower ) . He's posted those results ( as 4313B ) with response comparisons to the 476Be ( on the same horn ) . I believe he even posted schematics with these comparisons .

- Zilch ( by contrast ) is seemingly still enamoured with higher crossover points for these drivers ( @ 1.6 Khz ) and has ( as yet ) not grasped the necessary mechanics in forcing this horn/driver combo to perform below its' natural cutoff .
- ( Not to pick on Zilch since he has done a lot of good & public design work / but he's presently travelling a different road, compared to what you need to follow here ) .

<> cheers

timc
11-10-2008, 07:21 AM
I see your point Earl, but remember that I am using the 2435hpl and not the 435Be. As i understand the 2435 don't like to be used as low as the 435, due to higher Fs. I do want the crossoverpoint as low as possible, but from what i have read it should be at minimum 1KHz for the 2435.

The filter posted by 4313B was also designed to work with active crossover, and he suggested that some tweaking would be needed with a passive system.

If you read Pos's post about the difference in the throat of the horn, you'll see that i have the same problem. I might remove this bit of the throat at a later point to see if i can get the same results as Widget and 4313B.

For now however ill just try to get them working decently.

Thanks for your input :)

-Tim


Edit: Now i see what you meant by insulting. The reason i mentionend Zilch in this sentence, was that he implemented an extra paralell notch, wich my horn really needed. I'm fully aware that this is GT's design and would of course not rob him of the honors. Nor 4313B that took the time to explain and post it.

pos
11-10-2008, 08:30 AM
Earl, here is Zilch's filter for the 2435/H9800 combo :


did you use the compensation filter Zilch described here?
http://www.audioheritage.org/vbulletin/ ... stcount=84 (http://www.audioheritage.org/vbulletin/showpost.php?p=168664&postcount=84)

and curves:
http://www.audioheritage.org/vbulletin/ ... stcount=79 (http://www.audioheritage.org/vbulletin/showpost.php?p=168641&postcount=79)

4313B
11-10-2008, 08:30 AM
Do you have a measurement on the horn with no compensation, an impedance run on the horn and a schematic??

Rob:)Rob is right. We need you to post the .mls file of the horn and driver combo without any filter in place (use an 80 uF or 100 uF cap in series for protection, it's a ~ 4 ohm driver). Make sure the resolution is 1/12 and phase is set to minimum. Also make sure you are reversing the leads to the 2435 when you take the measurement so that red is to black, etc. Take the measurement at 1m, preferably with the horn facing up into the air. Also post the .sini file. Rob or I can export what data we need from these two files.

timc
11-10-2008, 08:33 AM
Will do :)

Think i can manage to get it done tomorrow.



-Tim

pos
11-10-2008, 08:33 AM
I see your point Earl, but remember that I am using the 2435hpl and not the 435Be. As i understand the 2435 don't like to be used as low as the 435, due to higher Fs. I do want the crossoverpoint as low as possible, but from what i have read it should be at minimum 1KHz for the 2435.
Tim, I don't think there is anything that indicate that the 2435 cannot go as low as the 435Be.
http://www.audioheritage.org/vbulletin/showpost.php?p=57023&postcount=2

4313B
11-10-2008, 08:38 AM
Tim, I don't think there is anything that indicate that the 2435 cannot go as low as the 435Be.The original statement that the 2435 wasn't to be used below 1 kHz came from JBL Pro (the designer of both drivers). They do not recommend using it below 1 kHz. (Also realize that they are using waveguides in most instances.)

Consumer does things a bit differently; I think the 2435/435Be has been used as low as 650 Hz, 750 Hz for sure. But these are acoustic crossover frequencies. I can't say for sure where the high pass filters were actually set in each instance. You can figure that out by looking at the various schematics.

4313B
11-10-2008, 09:04 AM
Earl, here is Zilch's filter for the 2435/H9800 combo :I plugged Zilch's modification into LEAP with my measurement files and the result is the blue curve. I'm not saying it is wrong. I am saying what I originally said here: #22 (http://audioheritage.org/vbulletin/showpost.php?p=208357&postcount=22)

You're going to have to come up with your own particular solution for your specific application.

My point is, Zilch's modification might not work with my H9800 and 435Be. I'm not completely happy with the graphical response of Greg's unmodified filter (red) in this instance either but I would have to actually listen to it to determine whether or not it sounded bad.

timc
11-10-2008, 09:12 AM
Just a sidequestion. If you have a raw measurement of a driver in LEAP, you can just add the crossover network and it calculates a theroetic output?


If so i really need to get som $$$


-Tim

Guido
11-10-2008, 11:48 PM
I can only second what Rob says.
My networks where spot on from the first try!

But don't expect LEAP to create a network by a fingerprint. You still need quiet a bit of network experience.


David has stated that with LEAP a decent filter designer can crank out a world class filter during his or her lunch hour

4313B
11-11-2008, 09:12 AM
How about just posting the files?

Then other forum members can look at the data and we can go on from there?

timc
11-11-2008, 10:00 AM
I had to return the measuring equipment today. so i didn't get a chance to do anything more. I'll see if i can lend it again this weekend. Will post files as soon as i can.

Thanks for all the willingness to help.

4313B. I fully agree that you need to know what you are doing, but everyone have to start out somewhere :)


-Tim

4313B
11-11-2008, 10:02 AM
I had to return the measuring equipment today.Ah! It's cool that it is available to you. :)

Jakob
11-11-2008, 07:02 PM
Good to see that you are making progress Tim. I'm following this topic with great interest since I also own a pair of H9800 clones and 2435's. Haven't had the time yet to start experiment though.

A tip regarding tight sounding woofers would be 2234, E145 and ME150. But as a matter a fact, the LE14H family all sound tight to me...

Guido
11-12-2008, 06:11 AM
A tip regarding tight sounding woofers would be 2234, E145 and ME150. But as a matter a fact, the LE14H family all sound tight to me...

The ME150 is an outstanding woofer. I still not 100% convinced that I'll keep my 1500AL. There is still a chance that the ME150 will return into the cabinets.
The 1500AL sounds more "hifi" and the ME150 sounds more "JBL" if this describes it for you.

And if you find 2227 cores the ME150 can be done as the ME150 cone kits are available (at least here).

4313B
11-12-2008, 07:58 AM
The 1500AL sounds more "hifi" and the ME150 sounds more "JBL" if this describes it for you.I suppose.

timc
11-12-2008, 12:47 PM
The ME150 is an outstanding woofer. I still not 100% convinced that I'll keep my 1500AL. There is still a chance that the ME150 will return into the cabinets.
The 1500AL sounds more "hifi" and the ME150 sounds more "JBL" if this describes it for you.

And if you find 2227 cores the ME150 can be done as the ME150 cone kits are available (at least here).


If and when you decide to get rid of one of the pairs, send me a message. If i have the cash, and haven't already bought woofers, ill buy them ;)


-Tim

timc
11-20-2008, 07:09 AM
I have now done a raw driver measurement. The MLS file looks ok, but i get some weird dips with the sinus sweep. This is most likely because i measure in a room wich is not really good for this kind of stuff. But...i cant do anything about that bit at this point.

Where do i get that .sini file? Is it the sinus sweep measurement? If so it is just called .sin in the version of CLIO i have avaiable.


-Tim


Remove the .pdf from the filename and unzip it.

Robh3606
11-20-2008, 08:11 AM
OK good. Here's what the raw files look like.

Rob:)

4313B
11-20-2008, 08:50 AM
I have now done a raw driver measurement. The MLS file looks ok, but i get some weird dips with the sinus sweep. This is most likely because i measure in a room wich is not really good for this kind of stuff. But...i cant do anything about that bit at this point.It isn't a slam dunk affair. If it was really easy Harman wouldn't blow millions on measurement facilities. ;)

Where do i get that .sini file? Is it the sinus sweep measurement? If so it is just called .sin in the version of CLIO i have avaiable.Probably just .sin

***

Rob, what is the impedance file suffix again? I can't remember off hand. I thought it was .sin or .sini

Robh3606
11-20-2008, 11:24 AM
Hello 4313B

They are the sini files

Rob:)

timc
11-20-2008, 12:21 PM
Ah...my mistake...i should have guessed that you were talking about impedance measurements. Will see if i can get one done tomorrow.

Just scrap the .sin file i posted. I recon the MLS is more accurate.


-Tim

timc
12-03-2008, 05:59 AM
Here is the impedance run finally. I stopped it at 200Hz. not sure if the driver would like to go lower.

The file is not a zipfile. just remove the .zip from the name.


-Tim

timc
01-05-2009, 12:28 PM
Bump.

Hoping to get some reply to this soon :)


-Tim

4313B
01-22-2009, 12:06 PM
Bump.

Hoping to get some reply to this soon :)I was hoping to get some time this afternoon but that isn't going to happen now. I'll try over the weekend. :yes:

timc
01-22-2009, 01:31 PM
Thanks 4313B. Take your time :)


-Tim

4313B
02-04-2009, 03:17 PM
The impedance file imported into CLIO just fine but the amplitude file is incompatible. I opened it in notepad and noticed

AUDIOMATICACLIO 6.27 s ISA

in the header and my version is 7.13

Can you go ahead and export the file using your version of CLIO? Three exported files - 1/12th, 1/6th and 1/3rd octave smoothing, 2048 points, minimum phase

timc
02-05-2009, 02:14 AM
Here are the exported .txt files.

Hope i have done correctly now.



-Tim

timc
02-22-2009, 08:43 AM
Finding anything? All looks ok?


-Tim

4313B
03-12-2009, 04:14 PM
I plugged your files into LEAP today and here's what showed up.

Ignore the Project File name. I just used that file to import your data into for a quick visual.

timc
03-13-2009, 02:06 AM
Look similar on my screen in Clio :)

Does it look ok?



-Tim

4313B
03-13-2009, 06:46 AM
Comparing it to the stock horn and driver measurements I took awhile back it looks pretty similar. It might have a bit more energy around 1 kHz than stock but that's about it.

timc
03-13-2009, 07:55 AM
Comparing it to the stock horn and driver measurements I took awhile back it looks pretty similar. It might have a bit more energy around 1 kHz than stock but that's about it.


Thats good news :)


-Tim

pos
03-13-2009, 09:22 AM
compared to 4313B's mesuerements:
http://audioheritage.org/vbulletin/showpost.php?p=217843&postcount=23

Why is the impedance curve so different?


2435HPL on H9800 clone:
http://audioheritage.org/vbulletin/attachment.php?attachmentid=37925&stc=1&d=1236899718

435Be on H9800: (green)
http://audioheritage.org/vbulletin/attachment.php?attachmentid=33779&stc=1&d=1218805546


2435HPL on H9800 clone:
http://audioheritage.org/vbulletin/attachment.php?attachmentid=37924&stc=1&d=1236899704

435Be on H9800: (green)
http://audioheritage.org/vbulletin/attachment.php?attachmentid=33777&stc=1&d=1218805519

4313B
03-13-2009, 09:36 AM
Why is the impedance curve so different?Different scale maybe. Different driver maybe. If you think these drivers have similar impedance curves they sometimes don't, primarily due to the ferrofluid.

JBL and I discussed this a couple years ago with respect to these compression drivers. That's why Greg was so excited that all six of his 476Be's were pretty well matched. When you get drivers that similar you don't mess around with them. You don't take them apart to see how they work or any of that kind of goofiness. You just bolt them up and count your lucky stars. Consistency is a huge thing with JBL, always has been, unfortunately it can be fairly elusive. (This is exactly why various engineers at JBL have said time and again with respect to DIY - MEASURE YOUR OWN SPECIFIC DRIVERS. DON'T RELY ON PUBLISHED DATA.)

Here is Tim's compared with the one I was looking at most recently. Mine (grey) has a wee bit more data points. I believe Tim has 2435HPL's while I have 435Be's.

timc
03-13-2009, 09:41 AM
Pos. That green curve you have there is with filter. My measurement is a raw curve

-Tim

Earl K
03-13-2009, 10:28 AM
timc,


How about just posting the files?

Then other forum members can look at the data and we can go on from there?

Some Clarification please ;
- Is the following response curve with or without a filter in place ?
- Regardless , the response between 1200 to 2200 hz is about the flattest I've seen for this driver/horn combo . Does your other combo measure as nicely as this ?
- The 4db "bump" at 1K is quite symmetrical , therefore I'd try flattening it with a series LCR ( notch ) filter ( as part of your overall crossover filter ) . >< cheers

http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=37924&stc=1&d=1236899704

timc
03-13-2009, 10:31 AM
Hi Earl.

I have not measured the other driver/horn. This measurement was done without any filter/compensation. Only a fairly large capacitor in series for protection. About 60uF if i recall correctly.



-Tim

Earl K
03-13-2009, 10:46 AM
http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=37924&stc=1&d=1236899704

Thanks for that info .

- That raw curve ( above ) is about the best that I've seen / originally I thought it displayed with some compensation in place ( albiet , a slightly more droopy version of GTs' compensation filter ) . It looks pretty close to the Plane-Wave response ( above 1K ) .

- I guess that explains how you were able to get + 3db ( above 10K which isn't typical ) using GTs' original HFcompensation along with your regular LC before the Lpad ( BTW, Zilches' notch filter isn't applicable for use by your driver ) .

http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=35223&stc=1&d=1226308442

- That's the good news / the bad news is this driver will need a HF compensation circuit that is a bit milder from what is typically used .

- If your second driver/horn combo measures more like the others that I've seen / it will need something more aggresive ( for HF compensation ) than whatever this first one ends up getting ( you need to measure the other combo soon to see where it's at ) .

>< cheers

4313B
03-13-2009, 10:49 AM
- Regardless , the response between 1200 to 2200 hz is about the flattest I've seen for this driver/horn combo . Does your other combo measure as nicely as this ?Here's the 1200 Array with the 045Be removed and the 435Be run all the way out using the following compensation filter. I have the active filter schematic that goes with it somewhere but it's been buried in hundreds of these files.

This curve sounds unbelievably good!

timc
03-13-2009, 10:49 AM
Yes, i might have to make separate filters :(

I will finish one speaker before starting on the next. I have too many project going on atm.


-Tim

timc
03-13-2009, 10:51 AM
4313B: That Array curve there between 1k and 5k looks amazing!


-Tim

4313B
03-13-2009, 10:52 AM
4313B: That Array curve there between 1k and 5k looks amazing!It sounds much better than it looks.

timc
03-13-2009, 10:54 AM
It sounds much better than it looks.


HAHA! Is that even possible.......... I believe you. Is that your private system?


-Tim

Earl K
03-13-2009, 11:02 AM
Here's the 1200 Array with the 045Be removed and the 435Be run all the way out using the following compensation filter. I have the active filter schematic that goes with it somewhere but it's been buried in hundreds of these files.

This curve sounds unbelievably good!

Nice ! ( I'm envious )

I wonder if a curve like that ( with that driver ) would work for me . At age 55 ( & with my M.I. past ) I can't hear much above 13K ( maybe this is why I don't understand Widgets' need for a tweeter ) .
Typically for a HiFi curve, anything above 10K I'll boost by at least a few db .

timc
03-13-2009, 11:29 AM
A friend of mine will help me with driving tomorrow, so i will be able to post the results for the second driver/horn tomorrow or on monday. Hope it looks the same */Crossing fingers.


-Tim

4313B
03-13-2009, 11:45 AM
HAHA! Is that even possible.......... I believe you. Is that your private system?Those are Greg's files of his former 1200 Array system.

Earl K
03-13-2009, 12:25 PM
Tim,

- If you're PC based, you should download ARTA . (http://www.fesb.hr/~mateljan/arta/)
- This powerful measuring software does most things that'll you'll need it to do ( no it won't automatically spit out crossover designs ) .
- It's cost is very minimal ( less than 100 Eu since it's shareware ) for the fully unlocked system . The help file is goofy ( & minimal ) since it's written by someone who obviously thinks we are all ( his ) Grad students .

Anyways ;
- With a good USB sound card ( such as M-Audios' "Fast Track Pro (http://de.m-audio.com/products/de_de/FastTrackPro.html)" ) & a decent test mic you can have a powerful ( yet relatively inexpensive ) audio measuring system . This sound card has nice preamps ( along with phantom power & 20 db of padding ) which are important features helping to avoid having to buy a small mixer ( just to power the microphone ) . I own the larger "Fast Track Ultra" version .

- ARTA does great coloured CSD plots ( about the best I've seen FWIW ) in a few different flavours .

>< cheers

ps ; Ignore the horn type noted in the first file ( it's actually an RCF H9041 ) .

timc
03-13-2009, 12:47 PM
Looks like a nice piece of kit Earl. I cant afford anything at the moment. Spening all my cash on cycling equipment for the moment.

I'm using ClioWin 6 ISA for the measurements. With the Clio spec microphone.


-Tim

Zilch
03-13-2009, 01:36 PM
I guess that explains how you were able to get + 3db ( above 10K which isn't typical ) using GTs' original HFcompensation along with your regular LC before the Lpad.What L-pad?


(BTW, Zilches' notch filter isn't applicable for use by your driver ) .And why, specifically, would that be?


Here's the 1200 Array with the 045Be removed and the 435Be run all the way out using the following compensation filter. I have the active filter schematic that goes with it somewhere but it's been buried in hundreds of these files.That's the filter I started with two years ago.

It's for H4338, not H9800, correct?

4313B
03-13-2009, 01:44 PM
That's the filter I started with two years ago.

It's for H4338, not H9800, correct?H4338, SAM1HF or 1400 Array horns. Greg has said the H9800 and Everest II horns aren't too far off. I think I posted their curves some time ago. They are all delt with in fundamentally the same way.

Zilch
03-13-2009, 01:49 PM
H4338, SAM1HF or 1400 Array horns. Greg has said the H9800, Everest II and S9900 horns aren't too far off. I think I posted their curves some time ago.That's certainly what I found. I tweaked it a bit with a tank filter and got the results I posted earlier on the factory H9800 you loaned me and Widget's clone.

I never resolved the issue of the DCR's in the filter, as you originally posted it here. Do we know those? :dont-know

timc
03-13-2009, 01:55 PM
I first tried with the original filter posted by 4313B. In original form it gave a coarser response than i posted earlier in this thread.



-Tim

4313B
03-13-2009, 01:58 PM
That's certainly what I found. I tweaked it a bit with a tank filter and got the results I posted earlier on the factory H9800 you loaned me and Widget's clone.

I never resolved the issue of the DCR's in the filter, as you originally posted it here. Do we know those? :dont-knowNothing special, they are the usual ~ 0.6 ohm or less.

4313B
03-13-2009, 02:01 PM
I first tried with the original filter posted by 4313B. In original form it gave a coarser response than i posted earlier in this thread.Well that's a bummer and one of the pitfalls of DIY without lots of tools. You'd probably do well to simply clone the K2-S9800 filter unless you don't like that sound, which is entirely possible.

timc
03-13-2009, 02:03 PM
I did actually try the original 9800 filter for the midrange. Looked darn similar to the graph in the tech paper you have posted.

However, when removing the LP section of it, things went crazy. I guess ill have to just start to make it from scratch. Step by step......:(


-Tim

Earl K
03-13-2009, 02:04 PM
What L-pad?


Hi.

No. i use a 2. order passive butterworth.

15uf in series and 1mH in paralell if i recall corectly. The filter placed before the compensation network.

- The last time I looked, GTs' compensation filter had a pad built into it . Sure there's a coil buied under the conjugate resistor and a capacitor paralleled to the build-out resistor / but those two resistors still fulfill the classical defination of an Lpad . The padding just happens to be "frequency-dependant" due to the extra parts.





Originally Posted by Earl K http://www.audioheritage.org/vbulletin/images/buttons/viewpost.gif (http://www.audioheritage.org/vbulletin/showthread.php?p=244291#post244291)
(BTW, Zilches' notch filter isn't applicable for use by your driver ) .

And why, specifically, would that be?

- Because the values used within your tank-filter, give too much notch for this specific driver ( which measures quite a bit flatter above 2.5K, than most 2435H specimens ) .

>< cheers

pos
03-13-2009, 02:04 PM
Pos. That green curve you have there is with filter. My measurement is a raw curve

-Tim
I dont' think so. It is a raw mesurement (see link)

4313B
03-13-2009, 02:08 PM
However, when removing the LP section of it, things went crazy.I'd need to see exactly what you removed...

timc
03-13-2009, 02:09 PM
This is the datasheet for the 2435H : http://audioheritage.org/vbulletin/showthread.php?t=14369


The values i have measured looks quite similar. But yes, when reading the thread it seems to be a raw measurement 4313B has posted.


-Tim

timc
03-13-2009, 02:12 PM
I'd need to see exactly what you removed...


The 0.3mH inductor in series with the driver and the 6uF capacitor mounted in paralell. I might have misinterpreted the design, It was done in kind of a hurry that bit. But i thought that theese components made a second order LP.


-Tim

4313B
03-13-2009, 03:07 PM
If you still have those networks put the 0.3 mH inductor back in, leave the 6 uF shunt out and pull out the third LCR trap. See how that sounds. It really doesn't lend itself to a decent 2-way high pass. That's still going to be something like 6 dB down at 12 kHz or so.

If you no longer have those K2-S9800 networks then don't bother with the above suggestion.

I'll play around with your specific data after you measure the second driver/horn set and when I have time.

Zilch
03-13-2009, 03:54 PM
O.K., assuming anybody gives a whit, here's where 4313B first posted it:

http://audioheritage.org/vbulletin/showthread.php?p=136480#post136480

Here's where I parsed it:

http://www.audioheritage.org/vbulletin/showpost.php?p=145641&postcount=73

Here's where I first built it, charge coupled and tested it on PT waveguides:

http://www.audioheritage.org/vbulletin/showthread.php?p=143808&#post143808 (http://www.audioheritage.org/vbulletin/showthread.php?p=143808&highlight=key+filter#post143808)

Here's where I modded it:

http://www.audioheritage.org/vbulletin/showthread.php?p=146690&#post146690

And built the modded version:

http://www.audioheritage.org/vbulletin/showthread.php?p=149792#post149792

Here's where I tested that on H9800s, factory and clone:

http://www.audioheritage.org/vbulletin/showthread.php?p=168641&#post168641 (http://www.audioheritage.org/vbulletin/showthread.php?p=168641&highlight=key+filter#post168641)

And showed what the notches do:

http://www.audioheritage.org/vbulletin/showthread.php?p=168701#post168701


Yes, i might have to make separate filters. :(

I will finish one speaker before starting on the next. I have too many project going on atm.That would be "excessive." If they're sufficiently different to warrant that, find the reason and fix it.... :yes:

4313B
03-13-2009, 07:12 PM
O.K., assuming anybody gives a whit, here's where 4313B first posted it:Hell yes I give a whit!
It means I don't need to get involved with this! :dancin:

Zilch
03-13-2009, 09:54 PM
It means I don't need to get involved with this! :dancin:Not so FAST, pal!

:rotfl:


John W reported a "swayback" response similar to that Timc showed earlier:

http://www.audioheritage.org/vbulletin/showpost.php?p=168659&postcount=83

It's unclear where he ultimately ended up with that, or what filter he was actually using. Same's the case here; Timc hasn't posted his schematic.

That, in combination with your sims and Earl's analysis, is all pointing toward my supplemental notch as being "inappropriate."

There's not much I can do other than verify my findings, as I don't have any of these later clones, and no LEAP, either.

If it'd help to have raw FRs and/or impedances on a bunch of drivers on the Widget clone, those I can acquire and attach here.

Does Mr. Widget still have a factory H9800? If so, I can loan him the same drivers to measure on that, assuming he's willing.... :dont-know

timc
03-14-2009, 01:31 AM
Zilch. It is your work on theese, and the postings by 4313B concerning GT's home system that got me going in the first place. I have read through what i can find on the 2435/435

About schematics. I started with the original compensation filter. That did not sound correct, nor measure correct. My guess is because GT is using active filtering, and im not.

Then i tried the Zilch modified version. Looks similar to the curve i posted, but with a deeper dip in the middle. The one posted has the same layout but with minor changes in the values.


-Tim

timc
03-14-2009, 07:55 AM
Now im very confused and frustrated.

I did a MLS measurement on the second horn and driver. And did a new measurement on the previous one. They look the same within error margin of the measurement setup, wich to be honest aint ideal.


The weird thing is that this time the peak at 1Khz is not there...... else it looks about the same.

Any ideas or suggestions. I will do additional testing on monday, to see if there is some error somewhere.


-Tim

pos
03-14-2009, 09:30 AM
did you get your drivers cleaned/refilled?
I sent mine to Guido for this, and they also received some aquaplass in the process.

But even then I have got some differences between the two drivers.

the green and purple curves are the two 2435, without any horn.
There are up to 3db differenc around 2.5khz, and 5db around 12khz.

For comparison the red and blue cruves are two 2407 (without horn either), that look nicely matched.

Robh3606
03-14-2009, 09:33 AM
Here let me help you get even more confused. I took your 1/6 and sini file and threw it into the basic filter. Changed the shunt inductor value and added your high pass filter. This is what LEAP shows with that. I can't figure out what your actual filter is. You need to post a schematic. Trying to do this without one is very difficult at best. Don't loose hope once you get them right your are going to be happy.

The Red curve is with the .100mH and the Blue is the .300mH posted schematic. That coil value changes tilts the curve. Not saying this is good but you can look at it and once you post a schematic we can plug it in.

Rob:)

timc
03-14-2009, 09:46 AM
The schematic for the current filter is at school (where the test equipment are.) I will post it on monday.


-Tim

Zilch
03-14-2009, 10:12 AM
Any ideas or suggestions.No one has posted a passive highpass for use in combination with the Timbers filter, though 4313B has suggested where to start. He's also going to see if he has info on the DIY active one Greg uses. I suggest you purchase an inexpensive active like Behringer CX3400 for use in development work to get on the same page here. Those are ~$150 with the requisite cables, and they also have a less expensive two-way plus sub version for under $100.

You can't do this without measurements, and you've got to get that aspect to the point that you can do that with accuracy and precision. Do you have the D'Appolito book? The raw curves you've posted here look very much like we're all measuring the same driver/horn combination, but you must standardize your methodology according to the known requirements for obtaining good, repeatable results confidently. This is going nowhere worthwhile until you get that part down; Rob and 4313B have the tools to help you with this, and they are willing, obviously, but your have to give them reliable data to work with.

Everyone must ALSO be on the same page with respect to what filter you're working with. Forget my charge-coupled variants for now, and build according to the original Timbers schematic 4313B posted at #83, above, here. There are enough members working with this, now, that it's worthwhile to define a design for everyone to use in this, and the Timbers filter is a logical standard for a baseline.

You say your system is working +/- 2.5 dB and then post response curves showing +/- 4.0 dB, on which basis I, for one, have little confidence in your statement that the Timbers filter produced "rougher" response results. Similarly, when you say you tweaked the circuit such that there are now "minor" variations, I'm also concerned. Get the variables under control.

With respect to your system in general, invert the L-65 on a furniture dolly or other suitable base; the woofer is spaced too far from the horn to integrate well in your present prototype. That's kind of off-topic at this point, though. Concentrate on optimizing the horn performance. I would not be too concerned about revising the horn/driver interface, either, unless/until your measurements reveal just how that is affecting the actual performance. You're not there yet, but what you've shown here so far certainly looks promising.... :thmbsup:

Robh3606
03-14-2009, 10:18 AM
No one has posted a passive highpass for use in combination with the Timbers filter,

Hello Zilch

The curves I just posted has his passive highpass if I got it right. It's going to change anyway with different woofers and how they are rolled off. You can come up with a generic put you are going to have to tweek it. Look at the second schematic, the first one is just a place holder.

Rob:)

timc
03-14-2009, 10:26 AM
Zilch:

I only said +-2.5dB between 1Khz and 10Khz. I agree with you about standardisation and all that. The problem is the enviroment i take the measurements. It's just a room. I don't have something even close to a 2pi anechoic enviroment. If i had, it all would have been easier. On monday i will try to do another set of measurements and post what info i can find.

Also remember that this is my first serious loudspeaker project. It might be overambiscious but. Time will tell. I have already learned alot about speaker design in general, and how filters work. So im happy about that, even if the project itself takes a long time.

About the HP filter. It a basic second order butterworth at approx 1Khz electrical.

I have tried what you said about the L65 box, and integration wise i agree. The problem is that the integration with my room goes haywire. However i placed it it just sounded boomy. My room is a little bit small for 12" really. Luckily i will only be here for another 6-8 months :)

Thanks for all your input.


-Tim

Zilch
03-14-2009, 10:30 AM
The Red curve is with the .100mH and the Blue is the .300mH posted schematic. That coil value changes tilts the curve.L22 in the R + L shunt, Rob?

Robh3606
03-14-2009, 11:00 AM
L22 in the R + L shunt, Rob?

Yes that's it. The value of those two let you lift or drop/tilt the curve for lack of a better description. Here it is with no R+L compared to the original curve with 100uh.

Rob:)

Earl K
03-14-2009, 11:12 AM
Here it is with no R+L compared to the original curve with 100uh.

Could we please see a curve without the 100uH coil / but with the shunt resistor still in place , ?

( I don't think many people realize what burying a coil into an lpad actually accomplishes )

Thanks >< Earl

Robh3606
03-14-2009, 11:39 AM
Hello Earl

The blue is with no inductor. Red is standard curve with .100mh. Pretty useful they have the same type shunt in the L250 Jubilee on the 050 tweeter.

Rob:)

Earl K
03-14-2009, 01:31 PM
Thanks Rob !

timc
03-15-2009, 06:03 AM
I have now done a new series of measurements. The nice flate response between 1.2k and 2k is not there this time around for whatever reason. I use a different room so i guess that might be the reason. Same equipment.

First here is a post of the schematic. The () indicates the real values i used. I couldnt replicate Robh's schematic exactly due to lack of components. And the program didnt have all the values i used. So look at the () where there is any. The last 8ohm resistor indicates the driver, so just ignore it.

timc
03-15-2009, 06:10 AM
I did measurements building up the filter from left to right. Here are the exported datafiles:

-Tim

timc
03-15-2009, 06:10 AM
Last one.


Edit: Been litening a bit to it now. Downmixing to mono and running the horn only. By far the best sound so far. :)

Earl K
03-15-2009, 06:40 AM
( last november )Hi.

No. i use a 2. order passive butterworth.

15uf in series and 1mH in paralell if i recall corectly. The filter placed before the compensation network.

I might have to change the values, but i just tried a few combinations and it gave a nice higpass rolloff.

I did not measure both drivers this time around. I have done measurments on both with a Copland Drc205 at home earlier, and they look quite similar.


-Tim

http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=38006&stc=1&d=1237122291

Tim,

- What's happened to the 1mH coil ( that you say you previously used ) ? It was right after the 15uF capacitor .
- Did you actually decide to leave it out this time ( or did you just forget to add it into your drawing ) ?

>< cheers

timc
03-15-2009, 06:41 AM
lol.....Just a miss in the drawing. It is just after the 15uF cap.


-Tim

Earl K
03-15-2009, 06:52 AM
I have now done a new series of measurements. The nice flate response between 1.2k and 2k is not there this time around for whatever reason. I use a different room so i guess that might be the reason. Same equipment.

- But did you use the same measuring methods, with the same settings ?
- You should be able to make repeatable measurements above 1K , fairly easily . Even if it means measuring at half the distance to the horn .

- Perhaps rereading the CLIO CLINIC (http://www.audioheritage.org/vbulletin/showthread.php?t=8370) thread will help you .

>< Earl K

timc
03-15-2009, 06:57 AM
- But did you use the same measuring methods, with the same settings ?
- You should be able to make repeatable measurements above 1K , fairly easily . Even if it means measuring at half the distance to the horn .

- Perhaps rereading the CLIO CLINIC (http://www.audioheritage.org/vbulletin/showthread.php?t=8370) thread will help you .

>< Earl K


The settings is identical with one exception. This time i managed to place at a precise loacation. 1m from the mouth of the horn. Last time i hold the mic in my hand, at about 60cm.

Personally i have more faith in the new measurement setup.

You suggest doing it again at 50cm distance?

Give me a min and ill try.


-Tim



Edit: Reducing the distance made no difference.

Earl K
03-15-2009, 07:42 AM
The settings is identical with one exception. This time i managed to place at a precise loacation. 1m from the mouth of the horn. Last time i hold the mic in my hand, at about 60cm.

Personally i have more faith in the new measurement setup.

You suggest doing it again at 50cm distance?

Give me a min and ill try.


-Tim



Edit: Reducing the distance made no difference.

- Okay, I too have greater faith in your new measurements .
- You shouldn't hold the measurement mic. . The proximity of your body & arm will skew ( contaminate ) the results . This means your earlier MLS snapshots can't be trusted .

- Is the circuit maker software that you're displaying, on your home computer or on the computer at school ? ie; You need something like this at home so that you can try out different ideas ( like notch filters ) on your own . Having software like that around will help you develop an understanding of what the various elements in JBLs' newer crossovers actually accomplish .

>< cheers

timc
03-15-2009, 07:45 AM
I have the Multisim software booth at home and at school.

the problem is that the simulation is not really that accurate, compared to the real measurements. Im not sure how to make a good loadsimulation of the driver/horn.


-Tim

Earl K
03-15-2009, 07:52 AM
I have the Multisim software booth at home and at school.

the problem is that the simulation is not really that accurate, compared to the real measurements. Im not sure how to make a good loadsimulation of the driver/horn.

Yes , I understand that problem , I also don't know how to create a good "load simulation" .

- Still , software like that is very useful at showing the trends . With it ( Multisim ) and a decent home based measuring system ( like ARTA ) you can learn to dial in quite accurate changes into your passive networks .

>< cheers

timc
03-15-2009, 07:54 AM
Yes. Multisim has been very helpfull in understanding what different filters do.

When i get a job, and start earning some $$$ ill save up to buy LEAP i guess.



-Tim

Robh3606
03-15-2009, 08:57 AM
Hello Tim

I have a disconnect on thus end. Which file has a measurement using 300uh?? From what I am seeing in CLIO the Full Filter and the 2Mh/100uh are essentially the same. The first plot compares the 2 measurements in CLIO. The next plot is actual vs measured in LEAP using the Final Measurement. Predicted is with 100uh and your parts as you posted them in the thread. The last plot is LEAP with a comparison of the final in Red and the Predicted in Green using the 300uh value. It matches up really well with 100uh not the 300uh. Did you measure with the 300uh??

Rob:)

timc
03-15-2009, 09:31 AM
Hi.

I did measure using 300uH as well, but the response was much worse than with 100uH so i didn't save it.

I also wonder why the full filter is almost identical to the on without the last series notch. What is it suppose to do?


Im impressed how close simulation is to my actual measurements.


-Tim


Edit: Ah i see i placed the 300uH in the schematic by mistake. Sorry

timc
03-15-2009, 09:52 AM
Here is the response of the filter with a 8ohm resistive load.

With and without the last RLC notch.

Robh3606
03-15-2009, 10:36 AM
Hello Tim

Oh good so then there was only a mistake in the schematic. What series notch filter?? Can you use the schematic designations?? Do you mean L3 C5 R3. Thats a notch in the 1-2K range.

What's different now?? It sounds better?? You didn't build the original filter??

It's not really 8 ohms try 6. Look at your impedance plot that's the actual that you use. Use the average not a peak.

Rob:)

Zilch
03-15-2009, 10:39 AM
I did measure using 300uH as well, but the response was much worse than with 100uH so i didn't save it.Compare Cyan to Violet here:

http://www.audioheritage.org/vbulletin/showpost.php?p=145641&postcount=73

You can see it pulling down 1.8 kHz here:

http://www.audioheritage.org/vbulletin/showthread.php?p=168701#post168701

Here it is in 4313B's sim:

http://audioheritage.org/vbulletin/showpost.php?p=144967&postcount=12

timc
03-15-2009, 10:49 AM
Oh good so then there was only a mistake in the schematic. What series notch filter?? Can you use the schematic designations?? Do you mean L3 C5 R3. Thats a notch in the 1-2K range.



Yes i meant that one. The MS simulation shows what it does. Didnt think of doing that before after i posted.






What's different now?? It sounds better?? You didn't build the original filter??


Rob:)

What original filter? I'm just saying it sounds better than the filter creating the graph i posted earlier. That one sounded a bit harsh. Now it sounds fairly smooth.


Lowering the load to 6ohm increases the attenuation slightly.


-Tim

Robh3606
03-15-2009, 11:04 AM
What original filter? I'm just saying it sounds better than the filter creating the graph i posted earlier. That one sounded a bit harsh. Now it sounds fairly smooth.


Ok I don't know what filter that was?? Glad it sounds better to you.


Didnt think of doing that before after i posted.


I did it after this stuff takes some time.

Rob:)

Zilch
03-15-2009, 11:14 AM
Hi, Rob.

What will happen if Tim puts the correct values in that final notch?

Those are fairly critical in my experience.... :yes:

Earl K
03-15-2009, 11:37 AM
What will happen if Tim puts the correct values in that final notch?

- Calcs show that the Centre Frequency ( Fo ) of the Notch will move downwards by @ 46 hz ( from about 1798 to 1752 hz ) .

- FWIW, 10% LC parts tolerances will shift that notch point around just as much, if not more.

><

Robh3606
03-15-2009, 11:50 AM
Hello Zilch

Here it is with and without the notch correct/Gregs values.

Rob:)

timc
03-15-2009, 11:51 AM
Any suggestion on how to reduce the dip in the middle? 2.3 to 5k.



-Tim

Earl K
03-15-2009, 12:12 PM
Any suggestion on how to reduce the dip in the middle? 2.3 to 5k.

- You can't easily elevate that area passively / but you can lower the areas on either side to better match the dip .

- I'd add another almost one octave wide notch centered around 7K ( to lower the 5000 to 9500 hz band, just like Zilch did ) .
- RE; the existing 1798 notch, I would see what it does by centering it at 1750 hz & then widening it out to grab the 1250 to 2150 hz area ( or so ) and increasing its' depth a bit ( by using a smaller valued resistor ).

- You should be able to play with your MultiSim program to arrive at the necessary ( ballpark values ) .


- Hint ; increasing the coil/inductor size narrows the notch // so to widen it start by halving the coil size ( 1.5mH to .75 or .8 mH ) and doubling the capacitors size . The resulting notch could be a bit wide / but it'll be a good learning exercise to play around with the values .
- The value of "R" in an LCR notch filter will help dictate the depth of the notch .

><

ps - EDIT ( my "hint' instructions were originally backwards )

Zilch
03-15-2009, 12:27 PM
Any suggestion on how to reduce the dip in the middle? 2.3 to 5k.Déjà vu.... ;)

Robh3606
03-15-2009, 12:32 PM
Look to add additional notch filters would be the easiest thing to do. The 9800 had 3 of them if I remember right.

Rob:)

timc
03-15-2009, 12:35 PM
Déjà vu.... ;)



Yeah :P


Robh. Could you throw in the following notch at the end of the filter and post the results?

L: 0,15mH
C: 2.9uF
R: 8,2Ohm


-Tim

Earl K
03-15-2009, 12:55 PM
Déjà vu.... ;)

Yea, aint it the truth .

Zilch , where's your parallel notch centered ? ( or do I have to get out my calculator and circuit design book ? )

- Tim, it may be time to let this thread rest for another 3 months while you figure out how to do some of this stuff for yourself ( afterall , that's the defination of DIY ) .


>< cheers

timc
03-15-2009, 01:00 PM
Hehe. You have a point Earl. I'm just tired of not having sound in them.


Music on my Klipsch RF-62 is just not the same.


I do think i have understood how this thing work now, i just have to learn how to calculate the values. The good thing about robh running simulation in leap, is that i dont have to build and test so much. It takes a very long time. Problem is that i don't have the measuring equipment at home. Have to go to the school for testing.


-Tim

Zilch
03-15-2009, 03:19 PM
Zilch , where's your parallel notch centered ? ( or do I have to get out my calculator and circuit design book ? )It'd be good if you did that, actually; I never calc'd it, as I developed it empirically.

Frankly, as I recall, it's the same as I had used earlier to flatten HL91 response; I just stuck it on, and it worked. Broadband, low Q was the requirement as I saw it, so I used a tank filter.

http://www.audioheritage.org/vbulletin/showthread.php?p=168701#post168701

Edit: I get 6721.15 Hz from 4.3133 uF, 0.13 mH, and 6731.54 Hz w/4.3 uF in the non-CC version. Confirm, please.... :thmbsup:

Earl K
03-15-2009, 04:14 PM
Edit: I get 6721.15 Hz from 4.3133 uF, 0.13 mH, and 6731.54 Hz w/4.3 uF in the non-CC version. Confirm, please.... :thmbsup: - Yep, those are the frequencies that I get .

- As it turns out ( for both "series or parallel notch filters" ) the math is the same when calculating the notch frequency center .
- Conversely, calculating the "Q" for each class of filter is different ( & IME nearly impossible to get right the first time ) .

- So a recap of the math ;

f = 1 / [ 2*pi*{ the square root of ( L*C) }]

- "f" is the notch center frequency value , given in hertz
- with "L" being expressed in Henrys ( divide mH values by 1,000. )
- and "C" being expressed in Farads ( divide uF values by 1,000,000. )

>< cheers

Zilch
03-15-2009, 04:29 PM
Thanks, Earl, I feel better now.

I was pretty bummed for a while in this; it seemed as though I had my head up my ass when I originally did it the hard way.

[Does happen.... ;) ]

Jakob
04-15-2009, 08:05 AM
So Tim, how are things going? Did you listen to the sweet sound of Beryllium during easter holiday or did you have to settle for the Klipsch's?

Regards: Jakob

timc
04-17-2009, 12:45 PM
So Tim, how are things going? Did you listen to the sweet sound of Beryllium during easter holiday or did you have to settle for the Klipsch's?

Regards: Jakob


Neither actually. I use the L65 in its original state atm. I have gotten a pair of 2213 woofers and all i need is a cabinet for it. Im looking for someone to cut the MDF since i dont have any equipment. I friend on a norwegian forum have helped me with a filter with looks flat. In simulation at least.

So basically i don't know whent things will be up and running. It's a pain to have most of the parts lying around and not beeing able to enjoy them :(

On the other hand. Summer is on the way and hifi takes a backseat for a while. I spend as much time as i can out on my bicycle during the summer.


-Tim

timc
11-30-2009, 01:06 PM
Right.

Time to move on. I have created a 3D drawing for a protoype box. It will house the 2213 woofer and a port. Volum is about 95 liters. The H9800 will sit on top for easy time allignment.

So to something i have been pondering and not found much info about. Where is the best place to have the reflex port. I have it on the front, but is there a rule of thumb between distance from woofer center and the tuning frequency of the port?


Best regards Tim

timc
12-04-2009, 11:18 AM
With some help from someone at a norwegian hififorum i have done some simulation on my horn and think i have come up with something usable. Comments are welcome. Ignore the peak at 1000Hz for now.


-Tim

Earl K
12-04-2009, 11:50 AM
The simulation certainly looks flat enough .

What's the insertion loss of this filter ?

<> :)

timc
12-04-2009, 12:06 PM
The simulation certainly looks flat enough .

What's the insertion loss of this filter ?

<> :)


Hi Earl. I'm unfamiliar with that term. Do you mean the how much attenuation i get?

If so it is about the same as the original Timbers filter. Maybe 1-2 dB more.


-Tim

Earl K
12-04-2009, 12:12 PM
Do you mean the how much attenuation i get?

Yes, that's what I meant .


If so it is about the same as the original Timbers filter. Maybe 1-2 dB more.

Okay, ( now, I'll need to track down that info ;) )

Were you planning on biamping ? ( without reviewing your project file, I'll admit that I have forgotten )

:)

timc
12-04-2009, 12:19 PM
No. There will be no biamping to start with. I will mate it to a 2213 woofer. Then when i get some more income, I'll start looking for a permanent woofer solution.

I will try various crossover points between 800Hz and 1000Hz.


-Tim

Earl K
12-04-2009, 12:26 PM
No. There will be no biamping to start with. I will mate it to a 2213 woofer.

- Ahh, Okay .

- Are you going to orient the H9800 sideways ( ie; vertical, as in the 1400 Array ) to see ( hear ) if you can improve the horns' "imaging" capabilities ? ;)

<> :)

4313B
12-04-2009, 12:49 PM
Ain't no turning Everest II's sideways! :banghead:

timc
12-04-2009, 01:06 PM
Ain't no turning Everest II's sideways! :banghead:


Actually that could be interesting. I have seen a custom Tannoy speaker with a similar bass configuration but with vertival orientation. Designwise i wouldn't mess with perfection ;)


No, there will be no sideways horn for me. It might be better, but to me the Arrays and similar just look weird. Its an aesthetic thing. Anyways, for me something close to the S9800 is more than good enough.

Sidenote: If i DID put the horn vertically. Would i not get in trouble with the different dispersion pattern at the crossover point?


-Tim

Earl K
12-04-2009, 01:49 PM
No, there will be no sideways horn for me. It might be better, but to me the Arrays and similar just look weird. Its an aesthetic thing.
Anyways, for me something close to the S9800 is more than good enough.

- Many of us in the SR ( sound reinforcement ) field wouldn't agree about statements suggesting "a weirdness in the look" .
- In fact, I think it all looks pretty natural to these eyes .

- FWIW, turning a 90 by 40 degree horn sideways can give one a "cheap & cheerful, look-see" into the "DD" ( Defined Directivety ) phenomena .
- Try it , you may actually hear something you like . :)

- Here's a link to one of Dr. Geddes creations ( the Summa ) . This speaker is well noted for it's imaging capabilities / I feel it's worth studying the polar patterns and pondering the "why ?" .
- Geddes is well aware of the E2 ( & the E1, & the "DD family" of horns ) / he also knows why these devices image well .
http://www.gedlee.com/images/summa_FR.jpg

- Here's another waveguide which, ( by the same defining parameters ) should have strong imaging capability ( though the good Doctor would be loath to agree because of the inclusion of a diffraction slot in the throat ) .

http://www.audioheritage.org/photopost/data//500/XT-1086_SPL_polar_horizontal.PNG



Sidenote: If i DID put the horn vertically. Would i not get in trouble with the different dispersion pattern at the crossover point?

( I already think you're in trouble ;) ) ,,, but it is a viable question ,, ie ;
- Why bother sending all that HF energy into the floor & ceiling ? ( & I don't have a reasonable answer for you )


<> cheers

Earl K
12-04-2009, 01:59 PM
Ain't no turning Everest II's sideways! :banghead:

Why would one do that ? There's no need ( obviously ) to mess around with such a successful design .

<> cheers

4313B
12-04-2009, 02:39 PM
Why would one do that ? There's no need ( obviously ) to mess around with such a successful design .

<> cheersI mentioned using my pair of H4338 horns in the same orientation as the 4338 a few weeks ago and it was suggested at the time to try vertical ala Array. The suggestion was followed up by an email a few days later so I suspect I should probably at least try it. :p

I just haven't been able to come up with a design that looks neato. I doubt I would do something as nice as John W did a few years ago.

Zilch
12-04-2009, 03:02 PM
- Here's another waveguide which, ( by the same defining parameters ) should have strong imaging capability ( though the good Doctor would be loath to agree because of the inclusion of a diffraction slot in the throat ).No actual slot, rather, a similarly loathsome diffraction "device" in the contoured flare.... :)

Earl K
12-04-2009, 03:10 PM
I doubt I would do something as nice as John W did a few years ago.

There's the rub / John W is in a very exclusive club .

<> :)

timc
12-04-2009, 03:13 PM
There's the rub / John W is in a very exclusive club .

<> :)


IMO John W is not IN a club. He IS the club.......


All humble respect to the other woodworkers in here.



-Tim

Earl K
12-04-2009, 04:24 PM
No actual slot, rather, a similarly loathsome diffraction "device" in the contoured flare.... :)

Yes, you're right , I shouldn't have used the word "slot" in my description .

<> cheers

Zilch
12-04-2009, 04:33 PM
Unlike a true slot, you have to look REAL hard to find the "culprit."

I believe we beat Earl back pretty well on these issues, Rob and I, with respect to PT waveguides. In the end, he admitted, "Better than ordinary diffraction horns...." :D

timc
12-05-2009, 11:30 AM
Ok. I have done some tweaking to the filter to make it compatible with avaible components. Not as smooth in simultaion, but close.


So. I have the 2213 woofer. My friend is currently looking for a high performance 12" woofer to mate to his driver/horn. We are currently looking at the JBL 2206H and the Fostex W300 AII.


Is the 2206 totally unsuited for this application without subwoofer?



-Tim

4313B
12-05-2009, 12:30 PM
Ok. I have done some tweaking to the filter to make it compatible with avaible components. Not as smooth in simultaion, but close.Just so it sounds good is all you care about.
Is the 2206 totally unsuited for this application without subwoofer?Probably. The 1200FE has about a half octave more bottom end and it can sound a little thin with some media. The 2204H falls in between but it has a larger box requirement than either the 2206 or 1200FE.

You might end up liking the old 2213 just fine. :dont-know

timc
12-05-2009, 12:38 PM
Just so it sounds good is all you care about.Probably. The 1200FE has about a half octave more bottom end and it can sound a little thin with some media. The 2204H falls in between but it has a larger box requirement than either the 2206 or 1200FE.

You might end up liking the old 2213 just fine. :dont-know


I hope ill enjoy the 2213. The alternative is for a friend who have joined in on the project. So we are making 2 pairs. The second woofers are for him.

Is the 1200Fe-8 still avaiable for purchase? I got the impression it was not avaiable anymore.

If it still is ill call him. He prolly want them asap ;)

-Tim

badman
12-07-2009, 05:05 PM
The 2213 is a very nice woofer, with good frequency exension and smooth response. It's dramatically less efficient than many other JBL woofers, being more of a home-hifi beast. This is achieved by a slightly less powerful motor and a significantly heavier, coated cone. It wants a very big box, but will go very deep in 4 ft+ sealed. Vented would want to be ginormous. The top end (goes to 5k, fairly flat) is rough (sonically, not FR), as you might expect from such a driver. The good news is, it's easily acoustically attenuated. I put a 5X4 rectangle of 1/8" acoustic F11 felt on the inside of a grille over the driver and got a 6dB/oct rolloff at 2kHz, fairly consistent both on and off axis. This matches the phase well, so with a zoebel, you can overlay it as if it were an additional pole on an electrical filter (without the zoebel, the impedance curve will throw off your electrical poles).

timc
12-09-2009, 02:25 PM
Thanks for the tip Badman. I will try a closed box as well.


Soldered the filter for the horn together today and did a test run with them together with the woofer of my l65. I used the Copland DRC to compensate for unoptimized crossover between the woofer and horn. It was close, but no cigar.

OMG! The sound from the horn is awesome. There are none of the peaks and dips that bugged me while listening to the previous design. And to be honest, the L65 woofer did a fair job too.

I will get the measured curves from the Copland tomorrow, but i could not hear a big difference between before and after from the horn alone. That is a good sign :)

So, I'm a very happy camper now.


-Tim

4313B
12-09-2009, 03:49 PM
OMG! The sound from the horn is awesome.

Where'd the "horn sound" go?

Oh, we got rid of that years ago...

:cheers:

Ian Mackenzie
12-15-2009, 01:54 AM
Where'd the "horn sound" go?

Oh, we got rid of that years ago...

:cheers:

Yeah I thought the E2's sounded too Hifi.

There are some interesting and radical ideas floating around about all horn system like The Living Voice Vox Olympian. (Vitavox and Tad drivers). Its is acost no object loudspeaker

All horn system are on the way back in hi end circles.

This article is worth a read in terms of how they came up with the particular components and the configuration. See the LG2 Bronze over aluminium for the HF horn and time alignment of the horns.
http://www.6moons.com/industryfeatures/livingvoice2/olympian.html

timc
12-15-2009, 04:05 AM
Yeah I thought the E2's sounded too Hifi.





Are you kidding, or serious?

If serious, what was it you didn't like?


-Tim

4313B
12-15-2009, 07:10 AM
Yeah I thought the E2's sounded too Hifi.I prefer Greg's 1200 Arrays. E2's are too wide and dominating for my environment. They are seriously fantastic though. I've got the single woofer version.

All horn system are on the way back in hi end circles.I'm not a big horn fan and I'm certainly not in the high end circles. I have no interest in those goings on.

Which reminds me, did you see alot of Tannoy while you were in Japan? They're supposed to be more popular there than JBL. I wouldn't know anything about that.

Are you kidding, or serious?

If serious, what was it you didn't like?


-TimI would imagine that he didn't like the fact that the airlines considered them too big for carry on... that's about the only thing wrong with them that I can figure.

Ian Mackenzie
12-21-2009, 01:30 AM
Are you kidding, or serious?

If serious, what was it you didn't like?


-Tim


I don't recall saying I did not like it them...see my review on the Japan / JBL thread cc 2008

You have to hear for yourself and compare these SOA systems to the classic vintage systems then form your own impressions.

I guess if the best of the vintage systems were up there with the E2 they would sound more Hifi as well.

As a side note I have messed with this stuff fair bit over the past 5 years with various owners and the reality is if you are keen enough the best of the bunch does a lot of things very very well for a lot less $ but they will never be E2s.

The Tannoys are certainly in the Japanese hi end audio market but I seriously doubt they are more popular when you see whole floors of serious audio shops dedicated to the E2.

pos
01-25-2010, 05:27 AM
Any update on your project Tim?
I understand you got some 1200Fe? :D
How better is it now than with the 122a ?

timc
01-26-2010, 02:10 AM
Any update on your project Tim?
I understand you got some 1200Fe? :D
How better is it now than with the 122a ?


Hopefully an update will come in the next 2 weeks. We are in the process of making enclosures for the 1200Fe. For now We have just tested that they work, and I will burn them in during the weekend.

Compared to the 2213 i have no idea. Have not really compared them yet. First impression is that the 2213 goes deeper but lacks the clearity in the midrange that the 1200Fe delivers. This is VEEEEERY prelimnary.


-Tim

timc
02-07-2010, 02:28 PM
Hey all.

Things are moving forward. This weekend we have made a 110L enclosure for the 2213. We will stuff this with solid material for experimenting with the 1200Fe.

We hope to be able to put some sound through them during the week. Tuesday maybe :)

Then i will have to come up with a LP crossover for the 2213. I have already simulated a prelimnary crossover for the 1200Fe based on the S5800 schematics. I just "transposed" it from 6ohm to 8ohm nominal impedance. Hope it works out ok.

-Tim

timc
02-08-2010, 01:24 PM
Ok.

Fired up the 2213's today in a CA 110 litre closed box. No damping material today, will get some during the week.

They go really deep, considering how high they also go. Very full rich sound, but not mushy and smeared. A little echo can be heard because of the missing internal damping, but hey. This sounds very promising ;D


-Tim

JuniorJBL
03-26-2010, 06:48 PM
Update?:D

timc
03-27-2010, 12:51 AM
Update?:D


Hell yea!

We have made second set of boxes for the 1200Fe. The transition between horn and woofer is fairly smooth, but not perfect yet. Almost there. Same with the 2213 woofer. Almost homerun now.

The 2213 is crossed over at 800Hz, 2. order. The 1200Fe uses the same filter as in ths S5800, but values change to match the different impedance. Not sure what to call this filter. It is two 2. order filters in series, with different cutofs. Can it be called a quasi 4. order?

The 2213 sounds smooth and extended in the lower bass, but lacks a LITTLE bit of pace and punch. The enclosure is now properly sealed with a fair bit of damping inside. Maybe a tiny bit too much. This needs some experimenting.

The 1200Fe.....Has the most scary realistic voice reproduction i have ever heard. I will go so far at to say that it betters the 1500AL in the S9800 on this. A little bit cleaner, and a little bit more relaxed. It also has a lot better punch and pace than the 2213. Of course the 1200Fe does not have the authority and impact of the 1500AL, but the midrange.,,,,,oh the midrange.....

Because of this i have started looking for alternative woofers, since the 1200Fe are not mine. Tips are welcome. I have looked at various Monacor woofers. Eminence Kappalite, PHL 4530, TAD-TL1201 and B&C 12TBX100. Any advice is welcome on this subject. So far the TAD's are most desirable but by faaaaaar the most expensive.



Another good news is that i have got a job at the acoustics department at NTNU (norwegian university of technology and science, Often called NUTS) here in Trondheim. That means i now have access to B&K measuring equipment. WinMLS and LEAP measuring software. Anechoic chamber (down to about 100Hz) and a 2Pi anechoic chamber (down to about 400Hz)


In my crystall ball i see alot of speaker building in the future.

PS: Pictures will come next week.

-Tim

pos
03-27-2010, 01:42 AM
Hi Tim

Great to see some updates!

You can try the AE TD12M (or maybe H).
It is said by Duke LeJeune to be on par with the TL1102.

I have a pair of TL1101H here that I plan to try in my setup (ME150H + TL1101H + H9800 with 2435Aq or 2452Be).

If you go for the 1200Fe or similar, are you planning to use subs?

Are you doing something for the "tap" in the H9800 clone throat?

timc
03-27-2010, 01:58 AM
Hi Tim

Great to see some updates!

You can try the AE TD12M (or maybe H).
It is said by Duke LeJeune to be on par with the TL1102.

I have a pair of TL1101H here that I plan to try in my setup (ME150H + TL1101H + H9800 with 2435Aq or 2452Be).

If you go for the 1200Fe or similar, are you planning to use subs?

Are you doing something for the "tap" in the H9800 clone throat?

I have looked at the AE TD12M, just forgot to write it. There was a thread on another forum, where the Eminence Kappalite was prefered over the AE TD12M. But of course, that is just one mans opinion. I have decided to save up and buy 2-4 types of woofers, and just see what i prefer. I have dealer status on Eminence and Monacor, so i get those for a nice price. The B&C and TAD is jolly expensive.

I will not use subs to begin with. The 1200Fe gives nice respons to about 35Hz. Thats enough for now. But ofc some 15"-18" subs are desirable, but need a larger space than i got. The other options simulate nice response to around 40Hz. Adding roomgain, and I think it is enough.

I have thought about that "tap" in the horn throat, but it measures and sounds very nice like it is. Maybe later. I can't really be arsed to mess around more with the compensation network at this point.


-Tim

yggdrasil
03-27-2010, 03:31 AM
You could also look at LE14H-3/4.

I have a pair of TAD TM1201, which of course is another beast. Will be used with 2245 for low's.

timc
03-27-2010, 03:58 AM
You could also look at LE14H-3/4.

I have a pair of TAD TM1201, which of course is another beast. Will be used with 2245 for low's.


Hehe. I have looked long and hard at the LE14H-3, but i have decided to stick with 12".

The TM-1201 would be my choice if making it 3ways. But I want to make a 2way, that can be augumented with a lower bass/sub in the future.

I might also try a pair of 2206 on each side. Have heard both pros and cons about that one, but i will have to judge for myself. Big pluss with the 2206 is the high sensitivity.

-Tim

4313B
03-27-2010, 06:47 AM
The 1200Fe.....Has the most scary realistic voice reproduction i have ever heard.Yeah they do! Jerry hit a home run with that design. 1200FE and SAM1HF for the win.
I will go so far at to say that it betters the 1500AL in the S9800 on this.Greg "fixed" that in the S9900. ;)
It also has a lot better punch and pace than the 2213.The 1200FE has some serious Bl. The 1200FE is quite simply a night and day better transducer.
So far the TAD's are most desirable but by faaaaaar the most expensive.I'd imagine that you'd get what you pay for. Does TAD make anything bad?
Another good news is that i have got a job at the acoustics department at NTNU (norwegian university of technology and science, Often called NUTS) here in Trondheim.Grats! :applaud:

JuniorJBL
03-27-2010, 07:16 AM
Another good news is that i have got a job at the acoustics department at NTNU (norwegian university of technology and science, Often called NUTS) here in Trondheim. That means i now have access to B&K measuring equipment. WinMLS and LEAP measuring software. Anechoic chamber (down to about 100Hz) and a 2Pi anechoic chamber (down to about 400Hz)

-Tim

That is wonderful news Tim!!

timc
03-27-2010, 08:36 AM
Grats! :applaud:


That is wonderful news Tim!!

Thanks :)

Earl K
03-27-2010, 09:09 AM
Another good news is that i have got a job at the acoustics department at NTNU (Norwegian university of technology and science, Often called NUTS) here in Trondheim. That means i now have access to B&K measuring equipment. WinMLS and LEAP measuring software. Anechoic chamber (down to about 100Hz) and a 2Pi anechoic chamber (down to about 400Hz)


Tim,

- That's great news about your new job ! :applaud:

- I can see by this new job that you're in this audio thing for the long haul / so again , Congratulations !

<> Earl K

PS ; I learned from watching this years Winter Olympics that you Norwegians are a bunch of over-achievers ( total medal count per population ) . Good on you all ! :rockon1:

timc
03-27-2010, 09:16 AM
Tim,

- That's great news about your new job ! :applaud:

- I can see by this new job that you're in this audio thing for the long haul / so again , Congratulations !

<> Earl K

PS ; I learned from watching this years Winter Olympics that you Norwegians are a bunch of over-achievers ( total medal count per population ) . Good on you all ! :rockon1:


Thank you Earl. Not to brag, but we also have the Nr. 1 in the world at Chess. :)

Mr. Widget
03-27-2010, 09:48 AM
Congratulations on getting what sounds like a very cool gig!

As far as I know, TAD is no longer making any woofers smaller than 15"... at least for sale as raw components. There will be a new dealer price list April 1st and it may show new product, but over the past few years the list has gotten shorter and shorter as the prices have gone up and up. :(

I have used a PHL 4530. In direct comparisons to the TAD TL-1102, it was significantly punchier sounding... very dynamic and exciting. It didn't seem natural though. It felt a bit like a photograph where the color saturation was jacked up. Kind of like a post card compared to a normal photo. I think they would be great for home theater where an extra bit of excitement is cool, but I wouldn't use them in my reference audio system.


Widget

timc
03-27-2010, 09:53 AM
It didn't seem natural though. It felt a bit like a photograph where the color saturation was jacked up. Kind of like a post card compared to a normal photo.


Great comparison!

Choosing a woofer is hard. Im going to look at a fairly large appartment next week. About 70 Sqm. If i go for it (get it) i might scrap the 12" and go for a 15". That will make the choice easier. TAD 1601A/B i guess.


-Tim

timc
04-06-2010, 10:14 AM
Finally finished one of the 1200Fe enclosures. Glued a moderate amount of damping materiel in them, and gave them a whirl in the anechoic chamber.

Without filter, the response i got, was more or less a carbon copy of the JBL datasheet. :D

Very good. Both for me and JBL's credibility. A+

Mr. Widget
04-06-2010, 10:17 AM
I wish I had an anechoic chamber to play in. :bouncy:

Lucky bastard!;)


Widget

timc
04-06-2010, 10:25 AM
I wish I had an anechoic chamber to play in. :bouncy:

Lucky bastard!;)


Widget

Thank you sir! I do indeed fell very lucky having been given this opportunity to work amongst several highly educated acousticians.

I'll ease your pain a little. It is only true anechoic to about 100Hz or so. But there are no nasty dips/peaks, so it actually produces a fairly accurate result. But mathematically,,,only to about 100Hz (depth of the wedges)

Harryup
04-20-2010, 12:34 PM
As far as I know, TAD is no longer making any woofers smaller than 15"... at least for sale as raw components. There will be a new dealer price list April 1st and it may show new product, but over the past few years the list has gotten shorter and shorter as the prices have gone up and up. :(

Widget

I talked to Andrew Jones at an HiFi show in Stockholm in February and he said that they were looking into the raw speakers. I asked if the 11" in Reference One was a development och 1102 and it was of course different since they needed a lower frequency response and higher SPL but they had used some ideas from the 1102. I asked if that speaker is a 1103 and he did not confirm or denied that since they had not looked into the raw speakers so much. But they are looking into starting a monitor line if they can see a demand and if they have the time to do it. They certainly have the speakers and the tradition.

/Harryup

badman
04-20-2010, 03:25 PM
I found that a 1/4" thick square of f11 wool felt about 5x5, mounted on a grille in front of the 2213 gave it a lowpass characteristic of about 2k first order, where normally they play up to and a little above 5k. An additional bonus is that when you do acoustic filtering, harmonic distortion is reduced (the high frequency harmonics are filtered just like the high frequency fundamentals).

timc
04-20-2010, 11:43 PM
Are you suggesting that I put a piece of wool between myself and the woofer?

Won't that screw up imaging, and midrange clarity?

timc
10-26-2010, 11:03 AM
Around 4 years after it started, this project nears it conclusion. I know the thread is only 3 years, but my brain was spinning at something like this a year beforehand.

Here is the description of pair number 2 (wich unfortunately is not mine). Mine is not finished yet, but the design is done.

The speakers are up and running. The 1200Fe driven by a Crown XTi-1000. The XTi does the filtering and EQ of the woofer. The signal is sent thru the amplifier to a dBX234XL, wich cuts the LF from the horn at around 500Hz. Above this we only use the acoustic rolloff of the horn/driver. The acoustic crossover is around 900Hz. The horn is driven by a VTL ST-85. In-'room response is around +-2dB in the listening position. Disregarding the room nodes. Work is in progress for resonators to kill them.

So, was it worth the time and effort? Without a doubt. I have learned loads of stuff about designing loudspeakers, and dealing with compensation nettworks for horndrivers. It has been fun, although frustrating at times.

So, does it sound ok? Oh indeed. The 2435HPL with aquaplas (thanks Guido) and the H9800 lookalike (thanks Matsj) sound wonderful. Smooth, and detailed, but with a rythem and transient power to scare the devil. The 1200Fe is a wonderfull driver. It manages to bring forth the musicality in instruments. It is not just a sound, but it sounds like you imagine it was meant to be. I talk alot about the "vibration" you get from a bass guitars. Almost like the tone is shivering. This is something the 1200Fe manages in spades. Even more impressive is the part of the midrange reproduced by the woofer. I actually think it is sligthly better than in the S9800. A tad smoother, and start/stop of weak sounds are easier to pick up. What i miss is the enourmous impact you get from the 1500Al, so in the future, there will be subs. Maybe based on 2235 or 2245.

I will post some pictures later.

Finally I want to thank all who contributed, and helped along the way. An extra thanks goes to Greg Timbers, for designing the original, inspiring with his homebrewed "Array system", and doing all the technical stuff around the compensation network available. Also thanks to 4313B for help with aquiring the 1200Fe woofers. Matsj for providing the Horns. And also all the people who have been working with the 2435's and posting data, and thoughts. You know who you are! :applaud:


-Tim

pos
10-27-2010, 04:27 AM
This is great! :bouncy:
Do you have responses curves of the two drivers with these filters?
What slopes are you using?
Have you tried the H9800 vertically?

timc
10-27-2010, 05:15 AM
This is for the horn.

Have not saved any meaningfull measurement of the woofer.

Slopes are 48db/oct linkwitz for the woofer, and 24dB/oct for the horn. Only acoustical rolloff down to 500Hz, where the active filter kicks in.

timc
10-18-2011, 01:36 PM
Time for an update. Its about time really.

Today I brought the speakers home after some final adjustment in the anechoic chamber. I decided to scrap the whole passive idea, and they are now running fully active. For now i use a Behringer DCX2496 crossover wich does all the EQ and crossover function. 4th order linkwitz for the woofer and 2nd order for the horn. Acoustically this gives a 4th order slope for the horn.

A Crown XTI-1000 is driving the woofer, and I have borrowd a small miniwatt tube amplifier for the horn. Im not fully satisfied with the performance of the 2213 woofer. It's good, but there is room for improvement. Im currently looking at alternatives. Best candidate so far is RCF 15N451 or AE TD-15H. Suggestions are welcome.

-Tim