Jump to content

Crossover Capacitors and Crossovers In General


Recommended Posts

14 minutes ago, Curious_George said:

I have a pair of 833A's that are waiting for a future SE projects! Looks like you are on my design team Cappy...

 

YAY!!!

 

I've been dying to build one but don't have the speakers for them yet. Mine will be a tad different and cut down on a huge expense, the output transformer.

 

My end game system includes something I have been wanting to do for a very long time but I just do not have the time. When I retire is when I'll start the process. I want to build my own large panel electrostatic speakers and fill the low end in with two subs. I figure then I can run a few kV's to the plates and bias the panel directly off the plate. I'll probably need plate chokes which is another part I have been wanting to do, wind my own. Then I can start winding output transformers also once I get decent but a choke should be easier to start with.

Link to comment
Share on other sites

10 hours ago, Crankysoldermeister said:

You mean like, take a stock pair of AA's, and just replace the T2A with one of the others -- no, I have never done that. See, I have left work for you guys to do.

 

Got a feeling you knew I'd react to this 🙂

 

Here are LTspice plots of the voltage across the squawker (in this case a 14.6 ohm resistor) in an AA crossover.

The green is tap 3 of a T2A, the red tap 3 of a 3636.  The lower inductance of the T2A contributes to the difference in slopes.   Noticeable???

 

I have always found it interesting that the impedance of the tweeter circuit attached to tap 0 increases the attenuation to the squawker by nearly 1.5db.

 

 

Screenshot (154).png

Link to comment
Share on other sites

According to the specs, the T2A's total inductance is 45.6mh (11.4 *4).

 

The inductance of the 3636's is pretty high, and I think that messes with my LCR meter, so I measure the inductance between taps 0 and 3, then multiply that by 4.  I end up with one at 88.8mh and the other at 76.8mh.

 

Link to comment
Share on other sites

6 minutes ago, mboxler said:

According to the specs, the T2A's total inductance is 45.6mh (11.4 *4).

 

The inductance of the 3636's is pretty high, and I think that messes with my LCR meter, so I measure the inductance between taps 0 and 3, then multiply that by 4.  I end up with one at 88.8mh and the other at 76.8mh.

 

 

Yup that sounds about right. Crites said larger coil and core so one would assume it had more inductance across the entire winding.

 

You can see the plots start to differ more at low frequencies because the inductance reactance is getting lower and starting to play more of a role with the reflected impedance of the mid-horn since they are in parallel.

 

Interesting.

Link to comment
Share on other sites

I remember reading this thread started by @Crankysoldermeister when Dean was having trouble measuring the 3636 inductance.  Both Al and Bob chimed in, but their point was that the inductance meant nothing...the level of attenuation is all that matters.  I can understand Al's point only because his resistor "swamps" out the autoformer inductance anyway.

 

It's my understanding that the large inductance between taps 0 - 5 may make measurements difficult.  Another way to measure is across taps 0 - 1, which is 25% of the total windings.  That inductance times 16 should be the inductance across 0 - 5.   I do worry that my numbers are off no matter how I measure.

 

 

 

 

  • Like 1
Link to comment
Share on other sites

3 hours ago, mboxler said:

I remember reading this thread started by @Crankysoldermeister when Dean was having trouble measuring the 3636 inductance.  Both Al and Bob chimed in, but their point was that the inductance meant nothing...the level of attenuation is all that matters.  I can understand Al's point only because his resistor "swamps" out the autoformer inductance anyway.

 

It's my understanding that the large inductance between taps 0 - 5 may make measurements difficult.  Another way to measure is across taps 0 - 1, which is 25% of the total windings.  That inductance times 16 should be the inductance across 0 - 5.   I do worry that my numbers are off no matter how I measure.

 

The 3636 goes in 1db steps correct? I don't have any experience with it but that's what it says on Crites. I think measuring from 6db points which is an easy 1/2 turns 1/4 inductance/impedance (-6db) or 2x turns and 4x (+6d) impedance/inductance can make the math easy. But 6db is not exactly double 3db. So you can't just do another 2x (.5x) turns and 4x (.25) impedance inductance.

 

3db steps is 1.4x turns 1.96x (+3db) impedance/inductance or .7x turns .49x impedance (-3db)

 

You'll just get more accurate values using the 6db measurements and doing whole number maths is easier.

 

Anyway, Yes in theory if inductance was infinite all that matters is the turns and ratio of winding to attenuate. Just like the primary of an output transformer many just go by the impedance ratio reflected by the secondary and ignore inductance, the inductance reactance is in parallel with the reflected load impedance so when the inductance is low enough to reduce the reflected impedance and load down the tube, hence inductance is important factor of an output transformer. With a large enough inductance in the autoformer the same thing can happen. In theory if the inductance is large enough it can be ignored and attenuation is made by the amount of turns. But if the impedance gets low, the inductive reactance is still going to be in parallel with whatever the reflected resistance will be and can increase the amount of attenuation which in our case making a filter works for us.

 

If inductance makes zero difference either way why are we seeing different low frequency behavior from LTSpice? Lower the inductance even more and see what happens, it will roll off at the low end even faster. So yes, inductance does matter, but once it reaches a certain level, and that depends on the reflected impedance of the mid-horn and the inductance becomes large enough the roll off won't change. like 15 ohm with infinity in parallel is still 15 ohms, but 15 ohms with say 65 ohms in parallel will lower the 15 ohms to 12 ohms. And we all know reactance is reduced with frequency.

 

Play around with inductance values in spice. You'll see that going huge with inductance you'll get very little changes in the roll off characteristics because the turns and reflected load dominates, but you'll see when the inductance is lowered and the lower you go the more the inductance will actually move the low frequency roll off characteristics in our favor. I.e. more attenuation of lows. Again going back to the output transformer primary inductance, going with a very large amount of inductance the reflected load impedance will always dominate, but since inductive reactance is in parallel, if you choose too low a primary inductance at low frequencies it will start to effect the reflected load by reducing it and loading down the tube.

 

Make sense?

 

 

Link to comment
Share on other sites

21 minutes ago, captainbeefheart said:

Make sense?

 

I figured out how to simulate autoformers in LTspice a while back.  At that point I realized that the inductance does matter, unless you use the "swamping" resistor.  My point is that I wish I could have been a part of that conversation knowing what I know now.

 

Another case in point is the 5mh value of the shunt inductor on the AK-3.  It's in parallel with the 7.3mh inductance between taps 0 - 3 of the T4A, resulting in a 3mh shunt inductor across the K55M.   On the AK-2, the inductance between taps 0 - 3 is 11.4mh, therefore a 4mh shunt inductor was used to reach that 3mh target.  Coincidence???

  • Like 1
Link to comment
Share on other sites

4 minutes ago, mboxler said:

Coincidence???

 

Definitely not a coincidence. 3mH is about the perfect amount of inductance where any lower and the roll-off would be too low but it's just at the right value to not waste anything, cost in size of the autoformer for getting higher inductance is a waste since it doesn't change anything.

Link to comment
Share on other sites

I love following this thread, btw now its more the coils than the caps. If i could find my original AA type somewhere in the basement. 
Unfortunately i dont have the original coils for the bass anymore. But i still have the original autoformers for the K400 horn. A layman's question. I have had AA crossovers from Crites since 2009 but with polyester caps for a year, not the real stuff from JEM but ok for now, even though for me 13 = 2x 6.8 and 2 is 2.2.
Would it make a difference if I use the Autoformer from the original Xover? Were its properties very different from those on the Crites AA switch from 2009? I'm not compulsively neurotic about authenticity, but still very curious about what the original sound was. Of course, you can say, just try it. But I don't have currently that much time and if you say it doesn't make a difference then I won't do it.

Link to comment
Share on other sites

26 minutes ago, KT88 said:

I love following this thread, btw now its more the coils than the caps.

 

It's all about caps, taps, and coils 🙂

 

26 minutes ago, KT88 said:

A layman's question. I have had AA crossovers from Crites since 2009 but with polyester caps for a year, not the real stuff from JEM but ok for now, even though for me 13 = 2x 6.8 and 2 is 2.2.
Would it make a difference if I use the Autoformer from the original Xover? Were its properties very different from those on the Crites AA switch from 2009?

 

If your Crite's AA came with a 3636, then there will be a subtle change as shown in the plots I attached.  A little more voltage to the squawker between 300 and 500hz and a little less after that.  I doubt I would hear that.

 

The T2A inductance values on the spec sheet are +/- 15%.  It would be interesting if you could measure the inductance on both T2A's just to see how far off, if any, they are from spec.  Again, I'd measure between taps 0 - 3 and multiply that by 4, or between taps 0 - 1 and multiply that by 16.

  • Like 2
Link to comment
Share on other sites

1 hour ago, KT88 said:

I have had AA crossovers from Crites since 2009 but with polyester caps for a year, not the real stuff from JEM but ok for now, even though for me 13 = 2x 6.8 and 2 is 2.2.

 

6.8+6.8 = 13.6

 

That's going to give you much more of a sonic impact vs any difference between any two polyester capacitors at the same value. I.e Switching to JEM caps will get the speakers sounding more original not because of any special properties other than it's closer to 13uF for the correct output transfer.

 

If you want to parallel capacitors I would parallel 8.2uF and 4.7uF, they are common values and fairly easy to find.

 

8.2+4.7 = 12.9

 

Honestly when I test caps of say 10% tolerance they are often fairly spot on but if they aren't perfectly spot on they are on average on the higher side than lower side so the two tolerance differences will probably get you to as close to 13uF as a real 13uF 5% tolerance cap.

 

Same with the 2uF. Using 2.2uF will slightly change the output transfer and that will be more of a sound difference than any two polyester caps compared to one another of the same value exact value. The easiest and least expensive way to get perfectly 2uF is use two standard 1uF values in parallel. Easy peasy and no special  non-standard caps need to be found.

  • Like 2
Link to comment
Share on other sites

3 minutes ago, captainbeefheart said:

 

6.8+6.8 = 13.6

 

That's going to give you much more of a sonic impact vs any difference between any two polyester capacitors at the same value. I.e Switching to JEM caps will get the speakers sounding more original not because of any special properties other than it's closer to 13uF for the correct output transfer.

 

If you want to parallel capacitors I would parallel 8.2uF and 4.7uF, they are common values and fairly easy to find.

 

8.2+4.7 = 12.9

 

Honestly when I test caps of say 10% tolerance they are often fairly spot on but if they aren't perfectly spot on they are on average on the higher side than lower side so the two tolerance differences will probably get you to as close to 13uF as a real 13uF 5% tolerance cap.

 

Same with the 2uF. Using 2.2uF will slightly change the output transfer and that will be more of a sound difference than any two polyester caps compared to one another of the same value exact value. The easiest and least expensive way to get perfectly 2uF is use two standard 1uF values in parallel. Easy peasy and no special  non-standard caps need to be found.

Thanks for reply. Again two Layman Qs. Is it no issue to put two so different values like 8,2 and 4,7 together in parallel? Secondly, no ESR issues concerning the deviation against a „one cap solution“ when using two 1 uF in parallel? I thought that ESR regarding e.g. polyester vs. polypropylene caps were one of the issues when using polyprops? But with two polyesters do I not reach the behave of one polyprop?

Link to comment
Share on other sites

49 minutes ago, mboxler said:

 

It's all about caps, taps, and coils 🙂

 

 

If your Crite's AA came with a 3636, then there will be a subtle change as shown in the plots I attached.  A little more voltage to the squawker between 300 and 500hz and a little less after that.  I doubt I would hear that.

 

The T2A inductance values on the spec sheet are +/- 15%.  It would be interesting if you could measure the inductance on both T2A's just to see how far off, if any, they are from spec.  Again, I'd measure between taps 0 - 3 and multiply that by 4, or between taps 0 - 1 and multiply that by 16.

Thanks for reply, so no worries to keep the Crites Autoformers. I will have a look if they are 3636 types. Just for fun I can measure the inductance of both T2As when we are back home from vacation to see if they are off from spec.

  • Like 1
Link to comment
Share on other sites

35 minutes ago, KT88 said:

Thanks for reply. Again two Layman Qs. Is it no issue to put two so different values like 8,2 and 4,7 together in parallel? Secondly, no ESR issues concerning the deviation against a „one cap solution“ when using two 1 uF in parallel? I thought that ESR regarding e.g. polyester vs. polypropylene caps were one of the issues when using polyprops? But with two polyesters do I not reach the behave of one polyprop?

 

ESR = tan / 2πfc

 

When two caps are parallel the ESR is also in parallel

 

If you crunch the numbers with capacitors that have the same tangent (loss angle) and you put them in parallel you'll see ESR hasn't really changed from just one single capacitor.

 

And no there is no issue paralleling any capacitor values no matter how far apart they are in value.

 

Tangent of say .06 for a film

 

8.2uF and 4.7uF vs 13uF

 

13uF

 

.06 / 2*π*1000*.000013 = .7 ohms ESR

 

8.2uF

 

.06 / 2*π*1000*.0000082 = 1.16 ohms ESR

 

4.7uF

 

.06 / 2*π*1000*.0000047 = 2 ohms ESR

 

Since the two ESR's are in parallel, that's 1.16 ohms and 2 ohms in parallel

 

1 / ((1/1.16)+(1/2)) = .73 ohms ESR

  • Like 1
Link to comment
Share on other sites

42 minutes ago, captainbeefheart said:

 

ESR = tan / 2πfc

 

When two caps are parallel the ESR is also in parallel

 

If you crunch the numbers with capacitors that have the same tangent (loss angle) and you put them in parallel you'll see ESR hasn't really changed from just one single capacitor.

 

And no there is no issue paralleling any capacitor values no matter how far apart they are in value.

 

Tangent of say .06 for a film

 

8.2uF and 4.7uF vs 13uF

 

13uF

 

.06 / 2*π*1000*.000013 = .7 ohms ESR

 

8.2uF

 

.06 / 2*π*1000*.0000082 = 1.16 ohms ESR

 

4.7uF

 

.06 / 2*π*1000*.0000047 = 2 ohms ESR

 

Since the two ESR's are in parallel, that's 1.16 ohms and 2 ohms in parallel

 

1 / ((1/1.16)+(1/2)) = .73 ohms ESR

Both your answers make me feel very positive, CBH. Thank you for the detailed and technically high-quality explanation of the ESR for parallel caps. Practically, both your answers in their combination mean that I can achieve the correct capacitance in sum with standard values in parallel connection. (The blessing of JEM components will follow later).

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...