So serious. I just gave it as an example. Actually, I do like a surface mount ceramic in that position right next to the electrolytic best right now. I use the 0.01 styrene down stream right next to the pass transistor. Still in parallel with the 'lytic, just not right next to it.
Last edited:
For simple power amp rails I tend to do 15,000uF electrolytic, plus 47uF electrolytic, plus 1uF film closer to the output devices. Seems to work and sound fine.
Well, there is a big difference between 1uF a stacked film cap and a 0.01uF polystyrene cap. There is no need to speculate, you can put the equivalent circuit into LTSpice and see what happens on a load step.
It would probably work just fine without the 47uF or the 1uF honestly depending on location and layout. Working fine and sounding fine is a really low bar to clear when we're talking about power supply reservoir caps. It helps to know why you are using 15000uF and why there is a 47uF and a 1uF cap.
Try the biggest mica cap value you can fit right next to the raw voltage entry point of the pass regulator element (if you have any, otherwise maybe the finals or the front end). Give that mica cap a dedicated thick solid core wire return to the circuit ground entry point.
Last edited:
Especially when you start looking at the lead/package inductance on power amp output transistors, bypassing to the pin with a large'ish x7r (10 uF, overrate the voltage a lot) generally yields the widest bandwidth PSU bypass. 100 nF on the pins of opamps is common, and good practice.
But that's engineering, so not appropriate here.
But that's engineering, so not appropriate here.
In my best designs a value that high in any other material sounds too bright and/or harsh. Lost of listening to analog sourced material brought me to this.
PS other aspects of the design have to be up to snuff to really reveal these things
Back to work for me now. Orders to fill!
Surely this must be some sort of joke. What is the function of a shunt capacitor most likely under 10nF attached to some length of wire...
We know what an ideal power supply is. Why do you not just find out what brings your power supply closer to that instead of throwing **** at the wall and hoping it sticks as a design philosophy?
Last edited:
Especially when you start looking at the lead/package inductance on power amp output transistors, bypassing to the pin with a large'ish x7r (10 uF, overrate the voltage a lot) generally yields the widest bandwidth PSU bypass. 100 nF on the pins of opamps is common, and good practice.
But that's engineering, so not appropriate here.
Yeah, I was going to find more links but this is a good one from Gerhard:
http://www.hoffmann-hochfrequenz.de/downloads/experiments_with_decoupling_capacitors.pdf
Would you suggest I make no claims to improve THD, or validate them when I can despite knowing the minimum?
Distortion isn't a measure of subjectivity. I stand by my claims for the quality of the sound.
He didn't ask questions because he wanted answers, he just wanted extra excuses to reject it, tell me off. It was very insulting to claim we do guess work and don't make "HIFI" (not the word choice I'd use, but understood). We measure everything we can, and more important measurements were taken than a THD measurement - measurements that validate the entire project working as intended.
I guess we could simply not release it publicly, but that defeats the mission statement.
Yes, not trusting one's own or others hearing then T&M is all that's left in such case.
This brings the subject of what to measure ?....Ultimate FR ?, yes and no...IMD ?, yes most certainly....THD ?, yes and no....Noise Floor ?....yes and no.
Ultimate FR.....this needs to be balanced in both directions, ie ultrasonic response needs to be balanced by near DC response.
Reducing UFR in either direction individually causes an 'imbalance' that sounds 'wrong'.
Reducing UFR in both directions can sound ok, it just presents as 'benign' information loss. Telephony depends on this.
IMD.....this causes discordant masking and spoils the party.
THD.....the spectral nature of the produced harmonics is mission critical. 2nds and 3rds add excitement and presence, higher order harmonics are poison. This still holds down to vanishingly low levels of THD.
Noise Floor.....the nature of this underlying noise drives the subjective sound of the system.
This system intrinsic noise dynamically modulates system produced distortions and signal, causing 'signature' sound.
I believe this is a subjective differentiator between OL and GNFB designs.
The 'gotcha' in audio design is that as the measured numbers improve the remaining diminishing errors become even more apparent (subjectively).
This is the conundrum that is lost on the T&M only camp.
As Hugh Dean commented recently, just accept that systems will likely never be measurement perfect....understanding and working with the errors and making them subjectively acceptable/pleasing is the path to good sounding audio.
Dan.
No joke. Glad I don’t have to work with attitudes like yours any more.
Back to filling orders!
It's a good'un. Plenty of app notes from all the major manufacturers out there that discuss this aspect, albeit in terms of high speed circuits. So it's daunting to recommend just one.
As to the point I was making:
http://www.nteinc.com/specs/2900to2999/pdf/nte2920.pdf
Lead inductance for this to3p is quoted at 5 nH on the drain, and source at 13 nH. Common cited number is 10 nH per lead for something like a to-220, so in the general range we'd expect.
Here's another one to explain in clear terms that the lead inductance will dominate any tiny capacitor's effect. (from a gentleman who lives in infamy on this thread, haha)
LM3886 chip amp supply decoupling.
Last edited:
That's not very friendly, why would you say that, what does it mean? (rhetorical question). What is a 'better engineer'? (rhetorical question) and why did you add 'almost'. Maybe you can point out who these 'better engineers' are, so we may direct our attention in there direction.
Ideally the accurately reproduced harmonics in a good recording should be sufficient. If the recording is not exciting enough for you simply add seasoning to taste, one advantage of this approach is that you can better control your excitement levels.
Texture can be weak with some stereos, but a dash of 3HD will often bring it back to a more solid feel.
How do you 'know' if you haven't measured?
I am sure you believe that, but that would make you a better analog audio engineer than JC and he staked claim to that title yesterday
It didn't read that way, unless you two have prior.
- Status
- Not open for further replies.