[linuxworks]

I'm wondering how audible the class-A improvement might be on a pimeta or ppa amp?

also, is there any chance the standard pimeta or ppa class-A circuit would cause problems?

I'm hearing some high pitched noise (I think, but I'm not 100% sure yet since I have not put it on the scope yet) that seems to modulate with the music. when there is a lot of HF activity, I hear some 'noise gating' kind of behavior almost like the cymbals are being replaced with keyed noise bursts. something like that.

I'm trying to see if this is oscillation or my dac or even my headphones needing replacement (hd580 that are almost 10 yrs old). lots of variables to try to narrow down. but I've been building the amps with the class-A circuit in them (2 Q's and an R - quite simple) and I wonder if I should just leave them out for now or even if they are helping at all.

anyone ever found that removing the class-A mod HELPED things? or am I perhaps blaming the wrong circuit element here?

 

[NelsonVandal]

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif I'm wondering how audible the class-A improvement might be on a pimeta or ppa amp? also, is there any chance the standard pimeta or ppa class-A circuit would cause problems? I'm hearing some high pitched noise (I think, but I'm not 100% sure yet since I have not put it on the scope yet) that seems to modulate with the music. when there is a lot of HF activity, I hear some 'noise gating' kind of behavior almost like the cymbals are being replaced with keyed noise bursts. something like that. I'm trying to see if this is oscillation or my dac or even my headphones needing replacement (hd580 that are almost 10 yrs old). lots of variables to try to narrow down. but I've been building the amps with the class-A circuit in them (2 Q's and an R - quite simple) and I wonder if I should just leave them out for now or even if they are helping at all. anyone ever found that removing the class-A mod HELPED things? or am I perhaps blaming the wrong circuit element here?

You shouldn't hear noise from any amp using high impedance phones like HD580, and it should definitely not follow the music (unless it's some kind of psychoacoustic phenomenon where you hear the noise as the music goes down). A DAC-problem maybe? Haven't you tried the DAC with any other amp? And haven't you tried to connect your phones to another amp/DAC/DAP?

Regarding the class A thing - try it live, or try it blind letting a friend change it without you knowing wich is wich, switching between class A and "AB". I'd be amazed if you hear any difference at all. All/most opamps are class A up to a certain load, and the diamond buffer is a very light load, and the opamp is probably class A even without the class A tweak.

By the way, what opamps are you using?

 

[linuxworks]

I've swapped dacs until my ears have bled. well, almost

I have old dacs (audio alchemy era stuff) and some new stuff (noodle dac and bantam dac). doesn't seem to matter!

for opamps, I've swapped a few things around but lately I'm running those $18 BB chips on a dual browndog DIP adapter. I had those BB chips from years ago (wow!) and have that as my dual amp. for simple amp (pimeta) I'm using the 8610 on a browndog. I've also tried the 8620 instead of the 'expensive pair' (LOL) and saw no diff to my ears.

I may have to have someone else (other ears) have a listen. I'm starting to have some tinnitus (I think) as I get older and its harder to tell the finer things ;(

I was curious, though, how much of an improvement the class-A mod typically gives and if its *EVER* been trouble to anyone. is it a 'just install it and don't question it' kind of mod or could it ever cause problems?

 

[linuxworks]

Quote:

Originally Posted by NelsonVandal /img/forum/go_quote.gif Regarding the class A thing - try it live, or try it blind letting a friend change it without you knowing wich is wich, switching between class A and "AB". I'd be amazed if you hear any difference at all. All/most opamps are class A up to a certain load, and the diamond buffer is a very light load, and the opamp is probably class A even without the class A tweak.

it almost sounds like the class-A section is very very optional and most people would not even notice if it was missing? is it really a last 1% kind of thing?

if most people would not even notice it in or out and its there really just for getting better paper-specs, I'm all for leaving it out.

 

[panda]

battery driven - no class a
ac - always use class a when possible

 

[sacd lover]

Quote:

Originally Posted by NelsonVandal /img/forum/go_quote.gif You shouldn't hear noise from any amp using high impedance phones like HD580, and it should definitely not follow the music (unless it's some kind of psychoacoustic phenomenon where you hear the noise as the music goes down). A DAC-problem maybe? Haven't you tried the DAC with any other amp? And haven't you tried to connect your phones to another amp/DAC/DAP? Regarding the class A thing - try it live, or try it blind letting a friend change it without you knowing wich is wich, switching between class A and "AB". I'd be amazed if you hear any difference at all. All/most opamps are class A up to a certain load, and the diamond buffer is a very light load, and the opamp is probably class A even without the class A tweak. By the way, what opamps are you using?

If the op amp is not fully biased in Class A the op amp is running in Class A/B. True class A operation would destroy an op amps in no time. Class A up until a certain load is still Class A/B.

 

[linuxworks]

I had to study that class A, B, AB stuff when I took my ham ticket, some 30 yrs ago. I forgot most of it, though, since I don't do this for my day job (I'm a software guy by profession).

so you're saying that by using the cascode trick, you're still not really class-A but really its AB ? I didn't know that. does that equate to less 'bragging rights' (effectively) ?

 

[TzeYang]

no, cascode trick turns off the PNP device and you'll get Class A single ended output stage.

sacd_lover, class bias depends on the load. A buffer usually means a very high impedance and it does not "load" the stage before it at all (ideally). Which means even the internal pre-set bias is good enough to stay at Class A region.

 

[linuxworks]

Quote:

Originally Posted by panda /img/forum/go_quote.gif battery driven - no class a ac - always use class a when possible

on paper, I think I agree with you.

but can this way (cascode) ever ever EVER cause problems?

its very cheap, cost wise. there is ZERO reason to not install it.

but if I'm doubting things or trying to reduce complexity or possible causes of 'problems' is it worth it to disable it and possibly not even turn it back on (ie, reconnect a hanging R lead to complete the circuit) again?

I like simplicity and I'm not a huge fan of 'the last 1%'. if this is really inaudible to 95% of the people (here) then I'd just assume leave it out unless it just cannot (at all, ever) cause problems.

hope that's not overasking my question too strongly

 

[espressogeek]

Class A on a Pimeta is cheap. Why not would be a question I would want to know. Battery power is a good reason I suppose but other than that....

 

[tangent]

I don't recall ever hearing of anyone with a problem amplifier fixing it by removing the class A biasing parts. Sometimes it's suggested anyway just to pare the circuit down to the simplest thing that still works in an effort to remove variables, but that's just standard troubleshooting practice.

It's easy to test, though: lift R10. Problem still there, or no?

 

[MisterX]

How close is the mains transformer to the amplifier circuit board?

 

[linuxworks]

ok, everyone has convinced me - it seems 'harmless' on the pimeta and can only help.

is that 1k (r10) gospel or is it at the edge of what's allowable?

I read that the 200ohm r11 (on the buf634) may also 'want' to be higher to add stability. so I wonder if the r10 1k is a 'set and totally forget' value or if I should consider playing with it?

on the ppa's there is a trimmer in series, so that's why I ask.

 

[majkel]

Class A op amp biasing makes slight to no difference, in many cases for the worse. Class A makes sense at the output stage and then it's audible as definite improvement.

Biasing the BUF634 into class A is a trade-off between it's current efficiency and preserving class A for low impedance headphones at high volumes. To be in the middle, set the bias current to roughly 120mA.

 

[tangent]

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif is that 1k (r10) gospel or is it at the edge of what's allowable?

It's nearly an arbitrary value. The higher the value, the better it isolates the op-amp's output from the input capacitance of the top JFET. It can't be arbitrarily high, however, as that would add noise and push the voltage drop high enough to perhaps cause a problem.

But, this is not the part you are looking for. <significant Jedi stare>

Having too low a value here isn't going to cause a high-pitched whine.

Quote:

I read that the 200ohm r11 (on the buf634) may also 'want' to be higher to add stability.

Er, no. Perhaps someone was confusing that with the 200 ohm input impedance on the buffer itself? In some other designs you will find around here, you'll see values as high as 1K between the op-amp output and the buffer input. This is doing essentially the same thing as R10 in the PIMETA: isolating the op-amp output from the input capacitance of the buffer. My view is that Burr Brown put 200 ohms there for a reason.

Again, I don't know of anyone who has ever fixed a balky amp by putting more resistance here.

Bottom line, I think you're haring off looking for exotic fixes, when there are a whole lot of other, more common things to look at.

You think it's instability. OK, what's the idle current draw of the amp, and does it stack up to what you expect? If so, it's much more likely to be interference than oscillation.

Quote:

on the ppa's there is a trimmer in series, so that's why I ask.

The trimmer on the PPA is in a totally different place than R10 in the PIMETA. The PPA has an R10, too, in exactly the same place as R10 on the PIMETA. (On purpose.)