paralleling power supply caps.

[stretch2011]

More specifically a large electrolytic and a smaller film.
I've done some reading over on ax84 and wanted to hear what you all have to say about it.

The jazz of lowering the esr sounds good, but I'm more interested in using a 600v+ film cap then a electrolytic of lower voltage (yet of sufficient voltage) to protect from the start up surge voltage.

Seeing a film cap is 10x cheaper than a high power electrolytic this may be a cheap solution to surge problems.

[David Root]

I now use 600V electrolytics for plate and screen supplies, since they have become reasonably priced. (Only for voltages above 400V though). No more 350V caps in series. The smaller footprint is very useful.

I wish it were possible to do this with can caps but AFAIK there are no 600V multi cap cans.

I have tried bypassing a first stage 20uF node with a .022uF film cap, but could hear no difference.
TAG member Jana, who is very knowledgeable, does this and can discern an improvement in clarity, but found no benefit other than in the first stage node.

[stretch2011]

Thanks David,
I seen just radios has 600v caps!

[Leo_Gnardo]

More specifically a large electrolytic and a smaller film.

It's long been an idea that's taken hold in the diy hi fi community. Some pro manufacturers have built it into their gear as well. Presumably electrolytic caps aren't really so good at bypassing high frequencies, that's where the parallel film cap helps out. In a power supply you're not just filtering DC, you're decoupling various amp stages, and you don't want them to "talk to each other" via the power supply. I first saw reference to this practice in late 70's - early 80's Audio Amateur magazine.

600V caps - are these new ones like the old "dynamite sticks" found in Ampeg and Sunn gear, made of two caps in series without balancing resistors and packaged in one cylinder? Has anybody dissected one to see whether it's a re labeled 500V cap or some such impostor? Sure it would be great if they're well engineered, well built improvements, but you know me Leo the Skeptic.

[David Root]

No, these are not the old dynamite sticks. These are modern e-caps, etched of course. Stretch ID'd the ones I buy, and the source, brand is MIEC, 105 deg C, good quality caps, no failures yet and I have used quite a few. Made in Taiwan, not China. There is a difference IMO.

[Littlewyan]

I read in Gerald Weber's book that etched caps aren't good at all for high voltages. He states they will test fine at low voltages but at high voltages they won't have the same microfarad value and so the filtering won't be as good. He recommends we avoid any capacitors made in Taiwan. Apparently even the LCR Caps Marshall used weren't very good.

With regards to bypassing the electrolytic, Marshall did this in their little 1Watt amps. They bypassed every filter cap with a 0.1uF Cap. Odd as I don't think they've done this in any other amp they make, only the 50th anniversary 1Watt amps.

[David Root]

Gerald Weber isn't exactly a generally accepted authority on anything.

I have used both MIEC Taiwan caps and LCR cans, nothing wrong with either in my experience.
Most if not all modern e-caps are etched. That is how the size has been reduced so much in recent years, to meet major market demands (we don't count as a major market).
The only disadvantage of etched construction vs. plain film that I'm aware of is that plain film is capable of higher current flows than etched. This may be an issue with a motor start cap but not I think with the up to 200mA or so that we deal with in power tube plates.

[stretch2011]

So to restate the originally intended question.

For an example of the power supply:
In the layout, film cap first then the electrolytic.
400v b+

If you parallel a .002 630v film cap with a 20 uf 450v electrolytic ( its been recommended that the film cap be 1000x smaller to prevent them turning into an inductor),will the film cap protect the electrolytic from the surge voltage?

Reason for question: Potentially improve the circuit while reducing cost.

[stretch2011]

Another question since I don't have an oscilloscope to check this out.

Has anyone probed around in the power supply to see exactly what all frequencies are present?

[Leo_Gnardo]

So to restate the originally intended question.

For an example of the power supply:
In the layout, film cap first then the electrolytic.
400v b+

If you parallel a .002 630v film cap with a 20 uf 450v electrolytic ( its been recommended that the film cap be 1000x smaller to prevent them turning into an inductor),will the film cap protect the electrolytic from the surge voltage?

Reason for question: Potentially improve the circuit while reducing cost.

The hi fi power supply goo roos recommend a 100:1 ratio, 0.22 film the nearest common value. Again, reason being power supply electrolytics do their best filtering at lower frequencies DC to a couple thousand Hz, & aren't ideal caps iow not so good about filtering higher frequencies a couple KHz on up.

Some hi fi eggspurts go so far as to add yet a third parallel cap, that's your 0.002, should filter well into the RF range.

Film cap first, then electrolytic? They're in parallel, there is no first and second.

"Prevent them from turning into an inductor?" I dunno about that.

No way the film cap will protect the elytic from surge voltage. Wonder who thought that one up. (In "Chipmunks" voice): Gerald!!!!

In short, I can see the case for a parallel filter, more or less 100:1 ratio, it isn't critical. You go Jana! Personally I've never made a habit of it. But if crustomers demand, who am I to say no, add labor & markup on film caps.

Something that hasn't been mentioned yet, location of the extra film caps. Of course, right there across the filters, you would think. There's been a long running practice in printed circuit making, where "local decoupling" caps are located remotely from the main power supply, sometimes dozens of small caps scattered around a big complex board. With this in mind you might consider adding similar small value film decouplers at the location of various preamp & drive stages.

Now what about cathode bypass caps, small film caps there too? If this catches on I could make a fortune loading up all the Fenders & Marshalls & etc that pass thru the workshop. All we need is a couple of rock stars who insist on them and blab to Guitar Holder & other magazines & websites about it.

[stretch2011]

Man I feel like a dumba$$. Of course they are in parallel they charge at the same time.

The 1000x times smaller and the inductor statement was things I read over on the ax84 forum ( not saying its right or wrong because I don't know any better either way).

You make a good point on film cap location! Something on the other forum they mentioned was using ONLY a film cap of the same value instead of an electrolytic on the cathodes.

SO here's another question. What if you used a film cap then a small resistor ( we'll say small enough to not drop hardly any voltage) then your electrolytic?

I don't know if any of this is really practical. I don't have a long experience with electronic or any theories. So I'm just asking questions to GAIN KNOWLEDGE. Sounds like a good quote huh?

[Leo_Gnardo]

Man I feel like a dumba$$. Of course they are in parallel they charge at the same time.

No worries, I'm sure we've all ... etc.

The 1000x times smaller and the inductor statement was things I read over on the ax84 forum ( not saying its right or wrong because I don't know any better either way).

Who knows where the poster on ax84 got their info, possibly a hand me down from some Audio Amateur reader, got "played telephone" with, added an extra zero, it happens, no big deal. BTW your 0.002 uF is 1:10,000 with respect to 20/22 uF so moving decimal places, adding extra zeroes, misteaks happen all the time. again.

You make a good point on film cap location! Something on the other forum they mentioned was using ONLY a film cap of the same value instead of an electrolytic on the cathodes.

Using non polarized electrolytics as Ck's as well has been proposed. If you think there's some deficiency to using electrolytics for audio, then wouldn't two el's back to back be twice as worse? Hmmm... All in all I'd favor films. But small-voltage films in values you would use as bypasses are expensive by comparison to el's, and also physically large. If through experimentation you hear an advantage, no worries, you're going to build or mod only a couple of amps. You can see why manufacturers making tens of thousands opt for the cheap solution. So far I continue to use ordinary electrolytics as Ck's.

SO here's another question. What if you used a film cap then a small resistor ( we'll say small enough to not drop hardly any voltage) then your electrolytic?

I don't know if any of this is really practical. I don't have a long experience with electronic or any theories. So I'm just asking questions to GAIN KNOWLEDGE. Sounds like a good quote huh?

Now we've spent a bunch of time & key flapping, you could spend as much time soldering in a couple of film caps & find out what your ears tell you. I've been slagged for this a couple times "Oh Leo, you're such an EMPIRICIST!" Doesn't worry me much, comes a time you have to call a halt to the conjectures, break out the soldering iron & find out if any of these pipe dreams hold water, to mix a metaphor with a canoe paddle. My recommendation: solder. Then listen. Then report if you hear an improvement. Or not. Remember these are guitar amps, if you get too hi fi, could lose the funk & get sterile sounding. Improvement or no? Experiment: that's how you, and all of us, will really gain knowledge. More conjecture, break out the bong , we can dream up some more, some ideas may work in the real world, some not so much.

[stretch2011]

Leo...
You're the man haha.

Looks like its getting time to break out the soldering iron!