Hi
I have 3 big electrolytic caps 220uF for the power supply of my tube amp and I would like to add some foil capacitors in parallel. How do I determine what capacitance is ideal?
The guy that sold me the amplifier and built it (I think) recommended me to do that by using 0.1-0.47uF capacitors. But how do I determine what exact value? Does it even matter?
I would very much like to avoid testing that, because it just would be a lot of work to desolder and resolder everytime,.
It's a single ended heaphone amplifier.
I have 3 big electrolytic caps 220uF for the power supply of my tube amp and I would like to add some foil capacitors in parallel. How do I determine what capacitance is ideal?
The guy that sold me the amplifier and built it (I think) recommended me to do that by using 0.1-0.47uF capacitors. But how do I determine what exact value? Does it even matter?
I would very much like to avoid testing that, because it just would be a lot of work to desolder and resolder everytime,.
It's a single ended heaphone amplifier.
Last edited:
Paralleling a film cap with electrolytic AT the power supply invites problems. It's been discussed several times and one of the DIYAUDIOr's has produced graphs which show multiple "peaks" in impedance. Perhaps take a look at this application note from Analog Devices: https://www.analog.com/MT-101?doc=cn0304.pdf
Are you referring to this thread here? Just found it now https://www.diyaudio.com/community/threads/is-bypassing-psu-capacitors-effective.126697/
It's very confusing to me. The reason for that is we do it to improve sound and if what Eva says is true then it would deteriorate sound quite a bit. A lot people swear by bypass caps, surely such a deterioration in sound should have been noticed?
Tbh, I think that Eva greatly exaggerated the impact of a tiny bit of wire more.
I am about to have a meltdown, ngl. Also some people talk about 1/100th rule of thumb that would be 2.2uf so 4x as much as 0.47.
I might just order really cheap ones in 3 different values and compare but man that's gonna be a pain. Especially considering they have 5-10% tolerance, so I would need to order a big bulk and match them all.
"Some People"(on the internet?)........ who want to make others nuts.
In general, paralleling should be avoided (as with anything, there are exceptions, like in microwave circuits), and paralleling capacitors having a ratio greater than ~100 is always detrimental.
No paralleling is best, but if it itches you anyway, stay below 50.
Some more info here:
https://www.diyaudio.com/community/...ps-with-electrolytic-caps.106648/post-2257381
No paralleling is best, but if it itches you anyway, stay below 50.
Some more info here:
https://www.diyaudio.com/community/...ps-with-electrolytic-caps.106648/post-2257381
Here are some examples -- fewer than Elvee's demonstration linked above, but illustrating placing the bypass right at the opamp. compared to the electrolytic is bypassed immediately next to the bulk capacitor. The little bit of inductance and resistance tames the impedance:
Attachments
Look up French Audiomat tube amps .Their reference models differed from regular ones with added uuuge film caps parallel to main electrolytic PSU Caps . It wasn't anything like 1/10 , more like 50-100 uF of a cap. Audio tends to be quite subjective with objectivists usually having most atrocious sounding systems they swear by, so use 1/10 ,5/10 and half the capacitance and make your executive decision. Best of lucks
Nowadays, to me that seems a bit redundant.
Because you might aswell just use a big foil capacitor which actually exist for "as low" as $50 for a 220uF capacity. Heck, I just checked you can by 900uF film capacitors for $70 https://www.mouser.ch/c/passive-components/capacitors/?capacitance=900 uF&product type=Film Capacitors
At this point I see no point in using electrolytics and then add 50-100uF capacitors in parallel if you can just use film ones directly.
The whole point of using film capacitors in parallel to electrolytes is that a 220uF electrolyte is going to be $8 a small film capacitor $1. $9 total compared to $50+ big difference.
Maybe, I am missing something tho.
Could you link me where you found the information about the Audiomat amplifiers and maybe their official website?
It depends very much on the actual configuration: if you just place the 100n directly at the terminals of the big cap (it doesn't matter whether it is 100µ, 1000µ or else), you will get nasty resonances and gain nothing in terms of lowering the global impedance.
Additional bypassing at the point of load is a completely different story, as Jack demonstrated: when the power transmission line has a significant length with associated resistance and inductance, it makes sense to improve the local bypassing. Not only does it make sense, but it is essential in many cases. That's the case with a board with a number of fast opamps, or old-style, two layer logic boards having a criss-cross design: each IC was individually bypassed. Such a distributed decoupling worked well, because of the parasitics and because all the caps had a similar value, in general 100n.
Doubling up a large Ecap with a small ceramic, or worse film cap is counter-productive.
That said, this kind of bad practice rarely has directly noticeable adverse consequences: the overall supply impedance generally remains low enough to avoid serious problems (when a resonance happens to fall at some critical frequency, like a clock multiple or sub-multiple, it can be catastrophic, but it is rare), but it can cause issues of some kind, like unexplained reduced performance, and the first cure designers try is to increase bypassing, by adding some more parallel caps, which at best does nothing, or aggravates the problem
A point worth noting: if you think that adding a small cap in the middle of a longish line will improve matters, think twice: the cap will be alone in-between two inductive path, and will resonate like mad.
It will behave like an undamped LP filter of high Q, which will be beneficial for the ultimate HF rejection, but will wreak havoc at lower frequencies due to the multiple and violent resonances at lower frequencies, where it really matters
It will behave like an undamped LP filter of high Q, which will be beneficial for the ultimate HF rejection, but will wreak havoc at lower frequencies due to the multiple and violent resonances at lower frequencies, where it really matters
I think these resonant circuit considerations are over-rated. To make this "resonate like mad" it needs to be excited by a very sharp current transient - far above the capabilities of real-world audio amps. Film caps parallelling bulk caps don't do any harm in practice, they just do nothing and can be omitted like all other snake-oil.
They're filtering signals on the order of 60-120Hz and their harmonics. Eva's reasoning and measurements are deeply flawed, a tiny bit of wire doesn't have that big of an impact on the impedance let alone on the percieved sound.
I see no other way of trying it because the internet seems full of contraddicting informations regarding this. We have gone from 100nf to 100uF for an ideal value of a parallel capacitor. There is jut no real common consesus and it's all just speculation at this point.
People just like number and measurements regardless how inaccurate they are.
Turns out Decware offers a mod for their amps with bypass capacitors on the power supply. https://www.decware.com/cgi-bin/yabb22/YaBB.pl?num=1553391822
Whenever a manufacturer offers "mods" for a previous product they sold, it's to convince the buyer to spend even more money for something with built-in flaws, or for keeping the "cash flow" coming in.
I stay away from products like that.
Sneaky marketing tactic.
Doesn't have to be. Also in radio circuits, there often existed the practice of wiring a 0.1μf capacitor across the large electrolytic in order to de-couple them from high frequency.
IMO, bypasses across electrolytics are a missed shot. Doesn't matter if "everybody" does it, it's not supported by theory. If it's going to accomplish anything is needs to be local to the circuit. There's another factor, and that's RF noise. It's typically generated at the bridge and is the reason bypass caps and snubbers are used across the diodes. RF problems tend to be ignored by audio people because they can be subtle and because many don't have the equipment to look for them. Sure, you can't hear RF, but the transistors can react to it and the result can be bias shifts and worse audio performance. By the time you get to the filter caps, it's way too far from the problem to effectively fix it. If you look at the combined impedance of any reasonable size film bypass across an electrolytic it doesn't help. Parallel caps in RF and logic circuits can work (see Jim Williams) but pitfalls abound without measurement.
If this were the 1970's I'd suggest taking a AM transistor radio with a pull-out antenna and waving it over your power amp with the cover off. If you hear any change in the hiss or maybe a tone, you have an RF problem.
If this were the 1970's I'd suggest taking a AM transistor radio with a pull-out antenna and waving it over your power amp with the cover off. If you hear any change in the hiss or maybe a tone, you have an RF problem.
I apologize maybe I overreacted. I just think that measurements don't tell the whole story in general with audio, has been my experience at least.
I might just do some testing and see if it does anything.