BA output caps

Support for JLM Audio Kits

Moderator: Joe Malone

qu4ntum
Posts: 8
Joined: Wed Mar 05, 2008 1:45 pm

BA output caps

Post by qu4ntum »

Hi,

if i am using the jlm output transformer with my ba build, can i omit the dc blocking caps on the BA output?

Thanks for your help,

wil.

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BA output caps

Post by Joe Malone »

Hi,

if i am using the jlm output transformer with my ba build, can i omit the dc blocking caps on the BA output?

Thanks for your help,

wil.
:shock: NO. The output of the opamp is sitting at +24v roughly and omitting the output cap would kill the opamp and maybe the output transformer instantly.
Joe :-)
JLM Audio
Capturing Audio without Injury

chrisp
Posts: 184
Joined: Sun Mar 04, 2007 11:46 am
Location: Sydney, Australia

Re: BA output caps

Post by chrisp »

To add to Joe's reply, remember that the BA design is a single supply (48V+ and 0V) rather than a bipolar supply (24V+ and 24V-, with 0V sitting halfway in between). This makes the power supply easy, but to get any opamp to work with a single-sided supply, you add ("bias") the opamp + input with 1/2 V+ (24V in the case of the BA), and then use a blocking cap to remove this DC voltage from the output.
Chris P
---------------------------------------------------------
I do lots of things. I believe eclectic skills are best.

qu4ntum
Posts: 8
Joined: Wed Mar 05, 2008 1:45 pm

Re: BA output caps

Post by qu4ntum »

*smacks head*

is 470uf the minimum to pass all audio freq?

thanks

qu4ntum
Posts: 8
Joined: Wed Mar 05, 2008 1:45 pm

Re: BA output caps

Post by qu4ntum »

i guess ill figure this out on my own, thanks.

greenmanhumming
Posts: 117
Joined: Tue Apr 10, 2007 1:19 am

Re: BA output caps

Post by greenmanhumming »

I would also like to upgrade or omit the output caps, and the one in the feedback loop

I recently built some valve mic pres and the BA's don't really compete, I think the opamps and transformers in the BA's should be fine so I suspect that output caps are the weak link.

seems to me if the output is not required to drive 600ohm loads, then into a typical 10k A/D converter input a 4.7uF should be sufficient for the output caps, but I may have missed something. I will try it when I find some suitable film caps.

I would also be interested in building a true +/- 24v supply to omit the caps altogether, can anyone instruct me on what is required? will it work just with a basic +/-24v supply?

chrisp
Posts: 184
Joined: Sun Mar 04, 2007 11:46 am
Location: Sydney, Australia

Re: BA output caps

Post by chrisp »

Err, Joe may wish to step in here if I'm wrong, but my understanding from other posts is that the output cap is still needed even if the supply is bipolar +/-24V, because there is still some DC bias in the output due to the nature of the opamp design. (I really could be wrong about this next part) The output is actually a transistor terminal that carries a DC bias for class A operation, to avoid the crossover distortion inherent in transistor switching at 0V. This is a known issue in all transistor class A designs.

What was the valve pre project?
Chris P
---------------------------------------------------------
I do lots of things. I believe eclectic skills are best.

greenmanhumming
Posts: 117
Joined: Tue Apr 10, 2007 1:19 am

Re: BA output caps

Post by greenmanhumming »

hmm, ok, so upgrading the caps might be the way to go.

can anyone recommend a brand of caps that does no compromise 50v film caps of high value? I've heard of MIT caps as being this but can't find them for sale...

the valve ones I have are 630v and HUGE. i'm not expecting to fit the caps on the board obviously, but size is still a consideration in the 1U case...

the valve pres I've made were 2 of my own design , 2 of new york dave's mila, and 2 of his one bottle, the mila is great, my own design with 2 12at7 gain stages and a parallel cathode follower is great, the one bottle is not so great yet (about on a par with the BA's) but i'm tweaking it.

while I'm asking, anyone know a cheap source of lundahl transformers? I have some 1538s which are great and available, but I want to try 1578 and 7903 but they are pricey...

evan38109
Posts: 4
Joined: Wed Mar 18, 2009 4:18 am

Re: BA output caps

Post by evan38109 »

For what it's worth, I got curious/bored a while back and decided to play around with the caps in a BAD build (dual 99V, Carnhill in, Jensen out). If memory serves, I tried a bunch of things and eventually wound up replacing the last cap (C12) with a 1000uf Nichicon KZ, along with replacing the two 100uf bipolars (C5, C6) in the middle of the signal path with Nichicon ES. These are pretty big caps, so they're hanging off the back of the board in a rather ungainly way. I'll try to post a photo at some point.

Like I said, I tried several options - and neither I nor others could hear any substantial difference. I had two channels side by side: one with the experimental cap setup and the other stock. No one I knew could reliably pick out one from the other in an A - B test. Maybe some golden-eared guru could, but I generally feel that if I can't hear a difference, it's not going to matter in a finished song.

The reason the above capacitors are in there right now is that they were the last ones I tried. Since it didn't make a lick of difference, I didn't bother to change after that.

greenmanhumming
Posts: 117
Joined: Tue Apr 10, 2007 1:19 am

Re: BA output caps

Post by greenmanhumming »

did you try a quality film cap though? I can well imagine that changing one good quality electrolytic for a higher quality electrolytic made no difference.

In valve circuits caps make more difference to the sound than the valves themselves, or even the circuit topology within reason. they are second only to the transformer in shaping the tone, so i figured they must matter here too.

evan38109
Posts: 4
Joined: Wed Mar 18, 2009 4:18 am

Re: BA output caps

Post by evan38109 »

I didn't have (and probably couldn't afford) a film cap that big. I have transformer outs on almost all my BAs and BADs and I do drive 600 ohm loads (sometimes less...I'm prone to mult abuse) so I don't want to go low on the cap value. I did try bypassing with some film caps and had pretty much the same feelings about the sound.

I don't want to discourage you in the least. Try a bunch of things, experiment and have fun!

greenmanhumming
Posts: 117
Joined: Tue Apr 10, 2007 1:19 am

Re: BA output caps

Post by greenmanhumming »

I've replaced the two output caps with cheap 4.7uF 250v polyester film caps, and added a film bypass cap on the feedback cap.

I think there is a small but worthwhile improvement in the sound, and the 4.7uF caps don't seem to cause any bass loss into my 10k inputs.

when using a 99v the -ve output has 0v on both sides of the cap. I presume it would be safe to just short these and avoid the cap here as long as using a 99v.

when using the opa2604 the second output cap is required.

so, concept proved I think, now i need to find some really good film caps to use that aren't too huge... any suggestions? the polyester ones I have in there are the cheapest and least bulky I have. I have been using claritycap SA in the valve pres, but I have none spare and I'm not sure they'll fit in the case, and they are 630v which is more than 10x whats needed....

there are loads of nice caps for valve stuff out there, but i'm lost looking for premium caps for low voltage stuff

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BA output caps

Post by Joe Malone »

greenmanhumming wrote:I've replaced the two output caps with cheap 4.7uF 250v polyester film caps, and added a film bypass cap on the feedback cap.

I think there is a small but worthwhile improvement in the sound, and the 4.7uF caps don't seem to cause any bass loss into my 10k inputs.
Making the output caps smaller will raise the signal to noise and also introduce ripple into the low mid and low region if feeding a output transformer even if the output transformer is only feeding a 10k. 4.7uF even direct into 10k will make a phase roll in the low end. One cap that is new isn't going to change anything much when masked by much higher THD in transformers etc. You mention your valve pres sounding better this is probably more due to the way higher THD and slowed transient speed and has nothing to do with one cap inline.
when using a 99v the -ve output has 0v on both sides of the cap. I presume it would be safe to just short these and avoid the cap here as long as using a 99v.
Doing this will cause the gain to go to DC so any offset on the output of the opamp will be multiplied by the gain amount so the 99v with its 1 to 2v volt offset would end up with its output at 48v and clip with almost no input level
when using the opa2604 the second output cap is required.

so, concept proved I think, now i need to find some really good film caps to use that aren't too huge... any suggestions? the polyester ones I have in there are the cheapest and least bulky I have. I have been using claritycap SA in the valve pres, but I have none spare and I'm not sure they'll fit in the case, and they are 630v which is more than 10x whats needed....
I would say the concept is far from proved with a little listening and no testing :-) Remember that the ESR goes up with voltage so the higher the volts the worse this will get. I have lost count of the number of extreme hifi products i have had in here that were having problems due to being modified by amateurs with huge poly caps so crammed in next to each other that they display inter capacitance between each other causing oscillation at worst or reducing the noise floor as these caps pickup power transformer flux in some cases. Turning 90% back to normal made them sound and work better.
there are loads of nice caps for valve stuff out there, but i'm lost looking for premium caps for low voltage stuff
No the really is that if it is big in size and expensive it will usually test worse than a standard new cap selected for the right conditions. I did a lot of testing for a very expensive hifi gear manufacturer and found that if the part looked amazing the performance was usually the opposite :wink:

If you were really looking to make the gain cap and output cap there very best you would not be going higher volts or lower cap values as both of these increase ESR. You would check to see what the highest DC voltage you see at any gain setting is on the output pin of the opamp and make the cap that voltage or slightly higher. So usually you could use 25v types (we use 50v standard so the caps can handle any fault conditions and not be damaged.)

For the output coupling cap you want the super lowest ESR so using 5 x 100uF 25v tants in parallel would be unbeatable (except by 10 x 47uF 25v tants etc).

The gain cap needs to be super low leakage type so again 5 x 100uF 25v tants would be perfect.

Only problem with tants is that if a fault occurs putting higher DC voltage across them they fail and go short circuit and this will usually kill the opamp and maybe your transformer.
Joe :-)
JLM Audio
Capturing Audio without Injury

greenmanhumming
Posts: 117
Joined: Tue Apr 10, 2007 1:19 am

Re: BA output caps

Post by greenmanhumming »

Hi Joe, thanks for the detailed reply, I think we are a bit at cross purposes on some points though...

* I'm not using an output transformer. in either the valve pres or the BAs

* on the schem I have the -ve output is not even connected to the 99v.... the 99v is not a balanced amp, and the other output is just grounded through the cap... so I don't understand what you are saying about missing out that cap, besides I've done it and it works, so there must be a misunderstanding there.

* I'm not using higher voltage caps because I want to, i'm using them because I can't find large value film caps at 50v (I know you are saying don't bother with film caps, but that's another issue)

* I'm not going small value because I want to especially either, i just need to find the best compromise between physical size, quality and capacitance value, and the 4.7uF were what was available to try.

* are you saying that 4.7uF into 10k will cause a phase roll even in a valve circuit? I presume so. I will consider increasing those...

* I agree about large poly caps causing coupling - it can be a problem, and they have to be placed and if necessary shielded appropriately.

* I disagree that large poly caps sound worse, in my valve circuits the choice of caps makes a huge difference, its not just the fact of using valves, its also the choice of capacitors. changing from cheap generic polyesters to claritycap SA in my valve circuits is an improvement of a similar magnitude to going from an OEP transformer to a lundahl. My valve circuits don't sound much better than the BA's if I put cheap caps in them. (tho different obv.)

* however there may be some reason why cap choice is less important (or just different) for SS circuits, i'd like to know why though if that is the case.

* although I like the slower transients and higher THD of valves, i also like the speed and cleanness of the BA's, however I feel a perceived lack of midrange weight and a sort of harshness in the top end (not just speed) that I am hoping to improve with better caps. If better caps are not the route to improving this what is? (something to do with the NFB maybe????)

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BA output caps

Post by Joe Malone »

greenmanhumming wrote:Hi Joe, thanks for the detailed reply, I think we are a bit at cross purposes on some points though...

* I'm not using an output transformer. in either the valve pres or the BAs

* on the schem I have the -ve output is not even connected to the 99v.... the 99v is not a balanced amp, and the other output is just grounded through the cap... so I don't understand what you are saying about missing out that cap, besides I've done it and it works, so there must be a misunderstanding there.
I have not even mentioned anything about the -out cap. I only talked about the +out cap and gain cap. The -out cap can be left out and only slightly reduce the balance of the noise cancelling since the phase shifts on +out and -out will be slightly different.
* I'm not using higher voltage caps because I want to, i'm using them because I can't find large value film caps at 50v (I know you are saying don't bother with film caps, but that's another issue)
Also didn't mention anything about the film cap which would have been tried as the bandwidth cap from the opamp out to -in I presume. The 99v opamp is a low gain type design so most times no bandwidth cap is needed so the 47pF or 100pF can be left out especially if no output transformer is fitted.
* I'm not going small value because I want to especially either, i just need to find the best compromise between physical size, quality and capacitance value, and the 4.7uF were what was available to try.

* are you saying that 4.7uF into 10k will cause a phase roll even in a valve circuit? I presume so. I will consider increasing those...
You want to keep the roll off point to less than 5Hz preferably.
* I agree about large poly caps causing coupling - it can be a problem, and they have to be placed and if necessary shielded appropriately.

* I disagree that large poly caps sound worse, in my valve circuits the choice of caps makes a huge difference, its not just the fact of using valves, its also the choice of capacitors. changing from cheap generic polyesters to claritycap SA in my valve circuits is an improvement of a similar magnitude to going from an OEP transformer to a lundahl. My valve circuits don't sound much better than the BA's if I put cheap caps in them. (tho different obv.)
Not saying that at all polypropylene is better than polyester for sure. But the extreme high end polyprop caps tend to add extra voodoo (read as series resistance or inductance) to almost make sure they do have a sound. I do not know the claritycap so i am not suggesting this off them. But caps marketed to High End Hifi that I tested years back got worse with the more they cost.
* however there may be some reason why cap choice is less important (or just different) for SS circuits, i'd like to know why though if that is the case.
I am more saying to think outside of the obvious polyprop especially when dealing with low volts. Tants when there is a polarising voltage like in the BA are great. They get a bad wrap due to so much use as coupling caps with no DC voltage as then they add distortion.
* although I like the slower transients and higher THD of valves, i also like the speed and cleanness of the BA's, however I feel a perceived lack of midrange weight and a sort of harshness in the top end (not just speed) that I am hoping to improve with better caps. If better caps are not the route to improving this what is? (something to do with the NFB maybe????)
The 99v is low gain like tube circuit but still higher NFB than most valve circuits. Your value pre is unlikely to have a output impedance less than 1k so will be softening the transients at high and low end. Worth trying the tants but also a series resistor or pot up to 5k with the output to reduce the output damping and I think this may do what you want mid range wise maybe.
Joe :-)
JLM Audio
Capturing Audio without Injury

Post Reply