Printable View
I killed it. i found the issue. it was a phasing issue. everything was messed up. i decided to use Ian's indications and look at the schematics not the wiring. Long story short the sound is now clearer and crisper is like i purchased another pair of speaker. It sounds as it should.
I am sooooo happy! Thank you Ian, thank you Ivica, thank you Bo for your phasing suggestions.
i will make new measurements to see if there is a difference now.
Happy Easter everybody.
Bogdan
That's good news.
It will only keep getting better and better now.
That is true with this homemade crossover. Typically, vintage JBL networks are solid colors to Red speaker posts; anything black stripes connects to Black posts.Quote:
Originally Posted by Ian Mackenzie
That's terrific, Bogdan! :hurray:Quote:
Originally Posted by lordoflight
Dammit - this was just getting interesting!Quote:
Originally Posted by Ian Mackenzie
Oh well, off to the next JBL triage...:wave:
Quote:
Originally Posted by boputnam
We're not done yet!
Hang around Pal we need your wisdom with these hand held rta thingies.
I suggest a program for peak performance:
1. Test and learn rta measurements and Calibration (Bo)
2. Test & Confirm filter functionality (lan)
3. Adjust and trim L .Pads (lan & Bo)
3. Evaluate overall in room response with tra (Bo)
4. Install 2308 lense and re test
Well, those of you Dayton Plug-in Mic fans, here's the calibration file that came with Ian's recent purchase - I stuck it into Excel, just for fun.Quote:
Originally Posted by Ian Mackenzie
Attachment 71277
There's a total of 256 filters, varying from 0.55 Hz to 534.1 Hz in width. Safe to say, I've never (ever...) seen a correction approached like this. I shudder to think of the phase response...
Now, it's up to Ian to figure out #2 - to measure that his on-board mic is actually flat!
Now I need indications on:
How to tune the l-pads
and
How to tune the bass ports (i made mine adjustable. One cadboard tube sliding in another cardboard tube)
How can i determine the best frequency for the best bass reaponse on these speakers? is there a simple way using a multimeter or based on box resonance calculators?
As I know the volume is hard to determine given the padding inside the box. Please help.
I will purchase the 2308 lenses andinstall them and the dayton mic to repeat tests.
Also i realized that the left mid speaker was badly repaired and has a ton of glue at the neck of the membrane where it meets the spider on the outside. I am afraid to look below the cap. Thing is i will have to purchase another 2123 dont know if it worth reconing with aftermarket recone kit.
WinISDQuote:
Originally Posted by lordoflight
I used this a lot, back-in-the-day. There may be something more current (Widget ?), but this used to do the trick...
Okay one thing at a time.Quote:
Originally Posted by lordoflight
There is no point adjusting he L pads until we get the bass tuning right.
The tuning of these woofers (2235) is generally in a net internal volume in the order of 5 cu ft 3 and tuned to 34 hertz.
We can simulate the tuning quite accurately if we know some basic information.
So l can calculate the port length can you please advise :
The internal enclosure dimensions?
The external dimension of the midrange enclosure?
The internal diameter of your ports?
(If you have calculated the net internal enclosure volume for the woofer please advise)
The enclosure stuffing should be around 2.54 cm on sell internal surfaces except the front baffle.
I plan to use a precision analyser, the LMS to measure the driver curves and simulate the system in Leap. I will do the iPhone FFT and if we can match that with LMS curves uand with our friend's system l then the implementation is good.Quote:
Originally Posted by boputnam
In greater depth the mid filter in the 4344mk11 is entirely different to the 3155 as is the 2123 from the 2122.
So we need to ensure the phase response at the mid and horn crossover point is optimal.
The crossover can look like it sums flat but phase can be sub optimal.
I can do that remotely.
The RTA is not loudspeaker design tool in that it cannot driver phase/filter phase transform plots as far as l am aware.
As stated the intent is to evaluate the 4344m11 and blended Giskard 3155 networks.
I will report on the findings and take the next steps.
As far as flatness is concerned I think it's about the adjustment of the Lpads absolute levels of the Lf, Mf, Hf and UHF given its a mixed bag.
Graphic or Parametric eq is out of scope so we are dealing with the accuracy of SPL measurements boost or cut.
Assuming the Lpad plate of the 4344mk11 is realistic and given the 2235 is about 93 db ref 1 watt versus the 4344mk11 ref sensitivity of 95 db (the mid Lpad needs to be adjusted down to -minus or cut 3.5 db from full.)
So as a starting point the iPhone needs to be able to resolve a 3.5 db difference with the SPL meter function. All we are doing is turnings up the mid LPad full then cut 3.5 db.
The HF and uhf LPads can be referenced to the 4344mk11 voltage drives from the input voltage to the network.
This will get the system dialled in.
Given the nature of the horn response plots it's difficult to determine even from smoothed curves the actual level ie do you take the average or the peaks of the horn to compare with the mid? I cannot imagine doing this with pink noise on an RTA. We are talking 1/2 db
Window pass / fail.
So l prefer the SPL difference calculation and rely on the safe harbour of the prior engineers to get the mid , horn and slot to match.
This is a close starting point for the mid/LF level match.
I just ordered the Dayton IMM6 mic. When I will receive it we're in business.
Meanwhile I will take again the measurements of the box to give them to Ian for the ports calculations.
When we start to set out the L-pads I am expecting step by step instructions not to miss anything.
I will be your puppet Ian. :)
big box inside dimensions are
125 cm
46 cm
33 cm
please assume 1.5 to 2 cm padding back and laterals
small box exterior dimensions
23.5 cm
28 cm
30 cm
vents interior diameter 10 cm the inside cardboard tube 10.5 the outside cardboard tube.
OkayQuote:
Originally Posted by lordoflight
You have 170 litres.
The effect of the stuffing is to make the box appear 10% larger and then we allow for the space occupied by the woofer, bracing and the ports say 10% and the box is 170 litres net or 6.00 cu ft3.
For large boxes this is a generalised assumption but is +-5%.
The ratio of woofer Vas to the enclosure volume is such we can live with 5%.
If you put more than simple shelf braces in your enclosure let me know.
You advise you have a passive crossover so the woofer has a small dcr resistance in the series inductor.
The simulator will do the rest
Okay adjust the length of both ports to 27 cm in length. (28 Hz)
Try that and see how it goes.
If you feel you prefer a more aggressive bass adjust the ports to 22 cm (30 Hz)
Either Tuning will yield a response down to 32 hertz.
Much of this will depend on your room.
In the short term set up your iPhone for SPL with the signal generator set to 800 hertzQuote:
Originally Posted by lordoflight
Place the mic immediately in front of the mid cone (10-20 cm) held steady with a stand.
Start the test generator.
With the Lpad of the mid up full adjust the Lpad until the reading falls by 3.5 db.
Keep the total spl under 90 db to avoid clipping the meter.
You will find any movement in the vicinity of the mid cone will effect the reading so it will take a few tries to get this right.
If your SPL does a digital read out then it will be more accurate.
Take note of the actual SPL value and distance of the mic from the baffle.
Do exactly the same for the other speaker.
Raise the mic until it's centred in front of the horn at the same distance
Adjust the sine wave frequency to 4000 Hz.
With the HF Lpad up full turn down the Lpad until the reading is the same SPL as the mid cone.
Do the same for the other speaker.
Repeat for the slot at 9000 Hz
This will give you ball park adjustment and both loudspeakers will be exactly the same.
You will really notice the precision of this approach with the sound stage as you have 4 drivers you are trying to adjust. It will jump right out at you!
It's futile to try and do this by ear without a meter to measure and record the changes.
Given the mic is not calibrated yet its not completely accurate but it's in the ball park.
Attempting to adjust the Lpads with pink noise at 2 metres is not anywhere near as accurate as spot frequency net SPL measurement in the near field for adjusting the actual levels in small amounts. If you look at the scale of your RTA is was 6 db per division and the room is effecting the overall measurement.
We need 1.0 Db resolution when adjusting the Lpads and preferably 0.5 db resolution.
Can you imagine buying a phono preamp with -+ 1.00 db variation on the frequency response! That would be unacceptable.
So why why would not want to adjust your loudspeakers so at least both left and right are identical.
(The 2308 lense will require re adjustment of the horn).
Once you get the Dayton mic we can do some other measurements and you will be more informed about how the loudspeaker is performing.
thank you ian. I have solid bracing inside the box. I tried to replicate the 4346 from krs. double backbone and double struts between the baffle and the back above the small box.
my room is the worse kind of room, approx 60 sqm room with plasterboard walls which absorb the bass. Can i improve something here?