Model 19 network mod for K145

[89-300ce]

I'm thinking of ordering components to build my networks.

K145, 511b, BMS 4552nd.

I'm looking at using a model 19 (N1201-8a) network

From WinISD

416-8b , Re=6.9, Z= 8
K145 , Re=8.8, Z= 16

WinISD uses Z for calculating crossover components, and using Z I would have to double the inductance and halve the capacitance.

Since L1 in the crossover also affects the HF circuit I would modify only L2.

Stock L1 (.3mh) and L2 (2.7mh) = 3.0 mh
I need

L1 (.3mh) and L2 (5.7mh) = 6.0 mh

Is my thinking correct on this? I know real world will differ but I'm looking for a reasonable start point to begin experimenting.

Thanks

Jorg

[rs237]

yes , your thoughts are correct , only change C3 from 21µF to 10µF.
Or buy Altec 416-8b and give me your K145 :D

regards
juergen

[Earl K]

Hi Jorg,


Is my thinking correct on this?

- The problem with your thinking is that it appears that you assume that a woofer has an AC impedance curve that resembles a flat line.
- They don't / they all resemble big happy smiles .

- You should run an AC impedance curve for your K145s and then plot them out before doing too much else .



- K145 , Re=8.8, Z= 16

- WinISD uses Z for calculating crossover components, and using Z I would have to double the inductance and halve the capacitance.

- A DCR of 8.8 ohms implies a nominal AC impedance of @ 11-12 ohms ( somewhere on the curve ) / though it's a certainty that a Z of 16 ohms will be reached at some point on the curve.

- FWIW, JBL referred to this woofer as an 8 ohm driver / while a similar or identical voice-coil as found in the le15a was referenced as 16 ohm .
- As mentioned , I'd go with 12 ohms as my starting reference "Z" (for transposing values of an exisitng crossover ) . This means multiplying existing coil values ( on the LF section ) by 1.5 and multiplying capacitor values by .6667 ( the reciprocal of 1.5 ) .





Since L1 in the crossover also affects the HF circuit I would modify only L2.

Stock L1 (.3mh) and L2 (2.7mh) = 3.0 mh
I need

L1 (.3mh) and L2 (5.7mh) = 6.0 mh



I know real world will differ but I'm looking for a reasonable start point to begin experimenting.

- Off the top of my head ( when being experimental ) / I'd add a Zobel ( RC conjugate placed across the woofer terminals ) that forces ( averages down ) all impedances above say 300 hz into a 12 ohm value ( maybe even 8 ohms ).
- This approach will have an advantage of allowing the use of smaller ( less expensive ) coils . A disadvantage ( if it can be called as such ) is that the lower impedance Zobels will burn off a couple of db in efficiency . OTOH, this wider spread in apparent efficiencies could be used to eke out a bit of extra UHF ( with a redesigned HF circuit / offer Zilch money to help :) ) .

- BTW, the Altec N1209-8a topolgy uses a Zobel on the LF section . This was undoubtably included into the design so that this network could be effectively matched to a bunch of different Altec woofers . ie; It was meant for more general purpose duty .

- Something to remember when playing ( experimenting ) with crossover designs / you can always create a larger value coil by wiring two smaller valued coils in series to each other ( the mH values will add directly if the coils are kept apart from each other ). One needs to separate them so they don't mutually couple ( ie; don't stack them ) . If you couple or stack them / you'll need to measure their new inductance with an inductance meter .

- Capacitors can be paralled to each other to create larger values.

- If you don't already have it / get something like TruaAudios' TrueRTA so you can measure the acoustic results of your experiments ( the Cdn $$$, being at an all-time high to the US $$$, means now is an opportune time to purchase ).


<> Earl K

[89-300ce]

yes , your thoughts are correct , only change C3 from 21µF to 10µF.
Or buy Altec 416-8b and give me your K145 :D

regards
juergen

If you knew what I had to go through to get these drivers you'd understand why I'm not going to give them up. That, plus I got needs. Needs that only 40 lbs of AlNiCo seem to satisfy.:)

Jorg

[89-300ce]

Hello Earl,



- The problem with your thinking is that it appears that you assume that a woofer has an AC impedance curve that resembles a flat line.
- They don't / they all resemble big happy smiles .

- You should run an AC impedance curve for your K145s and then plot them out before doing too much else .



Wouldn't I need to do the 416-8b's as well for it to be meaningfull in this case?




- A DCR of 8.8 ohms implies a nominal AC impedance of @ 11-12 ohms ( somewhere on the curve ) / though it's a certainty that a Z of 16 ohms will be reached at some point on the curve.

- FWIW, JBL referred to this woofer as an 8 ohm driver / while a similar or identical voice-coil as found in the le15a was referenced as 16 ohm .
- As mentioned , I'd go with 12 ohms as my starting reference "Z" (for transposing values of an exisitng crossover ) . This means multiplying existing coil values ( on the LF section ) by 1.5 and multiplying capacitor values by .6667 ( the reciprocal of 1.5 ) .






- Something to remember when playing ( experimenting ) with crossover designs / you can always create a larger value coil by wiring two smaller valued coils in series to each other ( the mH values will add directly if the coils are kept apart from each other ). One needs to separate them so they don't mutually couple ( ie; don't stack them ) . If you couple or stack them / you'll need to measure their new inductance with an inductance meter .

- Capacitors can be paralled to each other to create larger values.



That's a good approach. I'll buy some quality coils and caps and combine them with some cheaper caps and coils for testing. Once I find a good setting I can upgrade these cheaper components as well.




- If you don't already have it / get something like TruaAudios' TrueRTA so you can measure the acoustic results of your experiments ( the Cdn $$$, being at an all-time high to the US $$$, means now is an opportune time to purchase ).


I'll look into this.

Thanks

Jorg

[Earl K]

Hi


- You should run an AC impedance curve for your K145s and then plot them out before doing too much else .


Wouldn't I need to do the 416-8b's as well for it to be meaningfull in this case?

- No, not really / not when you are going to "make it up as you go" to find the "best fit" for those K145 woofers .

- I don't really see what advantage or insight you would immediately gain ( in having the original impedance curves ) / apart from furthering your education on the "hows & whys" in some of these design matters.

- Having said that, do you have a 416-8b to measure ?

- Having a well appointed box of spare parts ( coils, caps & resistors ) , coupled with a capacity to immediately "see" the results of different part-substitutions ( as on a RTA ) will speedup your design process immensely .
- The coupling of basic electrical theory with some good empirical ( trial & error ) approaches will achieve decent results .

<> :)

[rs237]

If you knew what I had to go through to get these drivers you'd understand why I'm not going to give them up. That, plus I got needs. Needs that only 40 lbs of AlNiCo seem to satisfy.:)

Jorg


I can understand you well. Here in Germany it is certainly to be gotten still heavy K145. I possess unfortunately only one.

regards
juergen

[89-300ce]

Hi

- Having said that, do you have a 416-8b to measure ?



No, unfortunately. I'd love to someday have a collection of transducers, but just starting out in this DIY hobby means I have to satisfy myself by just listening to other peoples stuff. But, that's been a lot of fun too.



- Having a well appointed box of spare parts ( coils, caps & resistors ) , coupled with a capacity to immediately "see" the results of different part-substitutions ( as on a RTA ) will speedup your design process immensely .
- The coupling of basic electrical theory with some good empirical ( trial & error ) approaches will achieve decent results .

<> :)

These changes are too subtle to evaluate by ear?
I haven't had a chance to look at this RTA setup yet.

Jorg

[Earl K]

These changes are too subtle to evaluate by ear?

Yes.


:)

[89-300ce]

When shopping for parts how close should my "part values" be to the design values.

eg. design calls for 4.2 mh

do I use 4 mh or 4 mh + .25 mh ?

I've found reference to < 2% on caps
< 7% on coils

Is this reasonable?

Thanks

Jorg

[Earl K]

- Try to keep LCR values to within 5% tolerance .

- You can always buy air-cored coils that are slightly larger in value than they need to be / & then unwind a few windings to get to the exact mH value .

- This means you'll need to own a LCR meter ( @ $100.00 at parts express )
- The meter is also handy for zeroing in on cap & resistor values .

[grumpy]

-
- The meter is also handy for zeroing in on cap & resistor values .

:yes: & also handy for matching, which I'll simply suggest might be of equal or more value
in the implementation process (L vs R tracking). -grumpy

[89-300ce]

Another question.:)

What wattage of resistors do I buy. PartsExpress sells them rated up to 20 watts. The BMS 4552 is rated to 60 watts. I've got 200 watts/channel in the amp.

If I need 3 ohm do I run 3 of the 20 watt 3 ohm resistors in parallel to achieve the 60 watt rating ?

Thanks

Jorg

[danwilliams]

I have a pair of L150's that I refoamed two years ago. Are there any upgrades members would suggest-new capacitors, different drivers, different internal bracing? Best amps tube or transistor for this unit?

Dan W.

[Baron030]

Hi Jorg

If you run three 20 watt 3 ohms resistors in parallel the power rating is going to be 60 watts. But, the impedance is only going to be 1 ohm.

I am assuming that you are talking about "R6" in the schematic, which has a value of 3 ohms.
And I am going to assume that the BMS 4552 driver has an impedance of 8 ohms and is rated 60 watts.
Since, the "R6" resistance value is lower then the driver's impedance, the resistor is not going to dissipate as much heat as the driver will.

Using the online L-Pad calculator listed below, I plugged in a 3 ohms series resistor value and a dummy shunt value 100,000,000 ohms and then I played around with the input power level until I came up with 60 watts going to the driver.
The calculated resistor wattage rating was only 22.5 ohms.

Since, it is very unlikely that you will be running your BMS 4552 driver at full power for sustained periods of time, even a single 3 ohm 10 watt resistor is not very likely to blow in this application. But, if you want to make absolutely sure that "R6" will never pop, then I would suggest using two 6 ohm 10 or 12 watt resistors wired in parallel or a single 20 watt 3 ohms resistor for the "R6" value.

L-PAD Calculator Link:
http://www.lautsprechershop.de/tools/index_en.htm?/tools/t_schwingkreis_en.htm

Baron030 :)

[89-300ce]

Hi Jorg



Using the online L-Pad calculator listed below, I plugged in a 3 ohms series resistor value and a dummy shunt value 100,000,000 ohms and then I played around with the input power level until I came up with 60 watts going to the driver.
The calculated resistor wattage rating was only 22.5 ohms.

Baron030 :)

Hello Baron030,

I can follow along with the example but I'm not sure I understand completely. When the speaker pushes 60 watts the resistor disipates 22.5 watts of heat?

How would I model/acount for R2 , R4, and R5 in this circuit?

When I substitute 5.9 ohms ( R2 + R5 ) for the parallel resistor in the "Level Calculator" I get some VERY high wattage ratings for the resistors.

Also, does the cap C4 negate R2 and R4's effect on R6?

Jorg

EDIT:

I guess if I'm going to set the LPad (R3) to full by-pass I shouldn't try for 60 watts at the driver :o:

[Zilch]

I don't know why you guys are even talking about 60 watts to the driver. You'll never put more than one watt to it, most likely.

Look at the stock crossover. The're using 10-watt resistors there, I believe.

Yes, C4 bypasses certain frequencies around R2 and R4. R3 (+ R5) determines how much of the rest of the bandwidth (already limited by C1, L3, C2) gets through....

[Baron030]

Hi Jorg

In my previous posting, I calculated the total power leading into "R6" and the driver combination at 82.5 watts total. That is 60 watts being absorbed by the driver and an additional 22.5 watts being absorbed by "R6". Also, worth noting is that the impedance of the "R6" and the driver combination has been raised to 11 ohms. And that the "R6" resistor attenuates the driver by 2.8 db. Now, if we ignore effects of the "C4" cap for a moment, and just concentrate on "R4" (3.9 ohm) value, then the "R4" resistor can be calculated as being in series with an 11 ohm driver load. Again, to fake out the online L-Pad calculator, I used a dummy parallel shunt value of 1,000,000 ohms and with an input level of 111.8 watts. And the calculated damping of the "R4" resistor works out to be 29.3 watts. And that the "R4" resistor attenuates this 11 ohm driver impedance by an additional 2.6 db. At this point, if we have 111.8 watts of input power, 29.3 watts are being absorbed by "R4"; 22.5 watts are being absorbed by "R6", and finally 60 watts going to the driver. Also, worth noting is that the impedance of the "R4", "R6" and the driver combination has now been raised to 14.9 ohms. Actually, the "C4" cap will cause this impedance drop down some from this 14.9 ohms value and that the "C4" will also reduce the amount of power that "R2" and "R4" has to a handle. Remember "C4" is being used to bypass the "R2" and "R4" resistors to provide a boost at the high frequencies in this circuit and its impact is less at lower frequencies. Now, moving onto the "R2" (3.9 ohm) value, again ignoring the effects of the "C4" cap for a moment, think of "R2" resistor as being in series with two loads. One load is coming from the "R4", "R6" and the driver combination, which has an impedance of 14.9 ohms and the other parallel load coming from the "R3" and "R5" combination. Since, "R3" is a variable resistor; it is a lot harder to calculate the exact amount of power that "R2" resistor needs to carry. But, if we play around with into the online L-pad calculator a while, plugging in 14.9 ohms as the driver's impedance, and 3.9 ohms ("R2") as the series resistance and the "R3" and "R5" combination values as the parallel resistance. You can start to see that the "R2" resistor does require a lot more power handling then any of the other resistors.

With "R3" set to its highest value of 35 ohms, it would require 214.4 watts of input power to drive the BMS 4552 driver to 60 watts and that "R2" resistor would have to dissipate 57.6 watts. And the "R3" and "R5" combination would have to dissipate 45 watts.

It should be noted that these power rating are for a sustained constant input level of 214.4 watts. In really, most of the power is being used to reproduce music is in the lower frequency bands. And that a resistor can handle short bursts several times their power rating without failure. So, you can safely use lower resistor power ratings then I just described above, without worrying about blowing up anything.



Quoted from Zilch:
I don't know why you guys are even talking about 60 watts to the driver. You'll never put more than one watt to it, most likely.


Oh, you are so right, Zilch.
One watt is pretty loud. :rotfl:

Baron030 :)

[89-300ce]

Thanks Baron030,

I'm just new at this and your clear explanation of the logic makes sense.
If I use the BMS4552 efficiency of 113db@1w/1m and set an arbitrary goal of 120db I'm looking at a driver load of around 8w. When I apply this load to your example the R3/R5 parallel resistance can see near 12watts while everything else is comfortably below this.

At this point I'm going with single Mills 12W 1% Non-Inductive Resistors for R2, R4, and R6. I'll use parallel Dayton 10W 2% Non-Inductive resistors for R5.


Thank you for your help.

Jorg

[89-300ce]

In hindsight, now that all my parts have arrived, I have to ask this question.

I've got two 8 ohm Lpads per crossover. To achieve the 36 ohm rating of the Lpad shown as R3 will I likely have to add fixed resistors in series by trial and error to get the Lpad into the operating range I like?
I've only seen a 16 ohm Lpad available. Should I order a pair of these?

Thanks

Jorg

[Zilch]

Measure the resistance between terminals 1 & 2 as shown on the schematic as you vary the control.

What's that read?

[Baron030]

Hi Jorg

All L-Pads are actually made up of two rheostats.
If refer to the schematic picture below you can see that one of the rheostats is in series with the speaker and the other in parallel with it.
I think if you measure the resistance between pins 2 and 3 (the series rheostat). You should find that resistance will fall into a range between 0 to 8 ohms, depending on the L-Pad setting. While, the resistance between pins 1 and 2 (the parallel rheostat) will fall into much wider range. Hopefully, this parallel rheostat will fall in a range between 0 and 35 ohms or more, depending on the L-Pad setting.

In most crossover networks, the two L-Pad rheostats are wired together just like in the picture below.
But, in the case of the Altec Model 19 network, the two L-Pad rheostats are not wired together, hence the need for two complete L-Pads per crossover network.

Baron030 :)

[89-300ce]

1-2 reads .5 - 38 ohm
2-3 reads .5 - 8 ohm
Whew!

It's all new to me:)

Thanks all.

Jorg