millwood said:
Of course you can (for most amps). People usually float the ground to 1/2 of rail, and use a cap on the output to block the DC offset (not at 1/2 of rail).
it has been done many many many million times over,
I know.
But try to use a virtual ground with two resistors on a power op-amp.
You have to reference a mid-point between + and ground on the input.
That's what those two 100k resistors are doing there.
And that's what we can call a single supply implementation.
With a virtual ground it's not a single supply anymore.
Yes.millwood said:
I never said you can't.
I just said that the implementation is different.
You can't just throw V+ and Ground to most op-amps and expect them to work.
Those two 100k RA resistors have to be there.
With double supplies you don't need them.
Neither the output cap.
Compare the normal implementation with the single voltage, on the LM3875 datasheet. It's a good example.
With large currents (like with power op-amps) the PSU voltage will get unbalanced (-1v, Ground, +25v).millwood said:
You may use two resistors to make a virtual (PSU) ground with signal op-amps.
millwood said:
Sorry, you are wrong. The biasing has to do with current, not voltage. The NPN device (on the + rail) will be conducting when current needs to flow out of the output pin; the PNP device (on the - rail) will conduct when current needs to flow into the output pin. For the case of a simple resistive load, if there is always a +ve DC voltage on the ouput, then there is always current flowing out of the output pin, and of course, the NPN device will always conduct, so the opamp will essentially be operating in class A. However, real world loads are not purely resistive. The current through a reactive load does not depend purely on voltage, so maintaining a positive voltage does not guarantee a positive current. Also, in the real world, we cannot have that huge DC offset. And placing a DC-blocking capacitor in there will block the DC current, defeating the purpose (no more class A).
The best way to bias an op-amp into class A is to put a current source to -ve rail so that current always flows out of the output pin. This forces the NPN device to always conduct and the PNP device to never conduct. Take a look at Linear's LT1115 datasheet. They use a current source (a FET and resistor) to the -ve rail in many of the example circuits, although they don't actually mention why.
This thread is really flipped out. How did bridge connection come in to this?matjans said:
Anyway, matjans, check
Meta42 and PPA and myQRV-04
You must first consider that many opamps, especially those for audio have very special bias networks for class A or AB and some of them is impossible to do with discrete parts. I suspect that some opamps really can be helped with this external class A circuit and some or not. The question is which one? LM324 can REALLY be helped! A good example of what this current generator does!
My personal experience is that the differnence is hardly worth the efforts. My ears can't detect any difference at all with AD8610 and OPA627 but this isn't the same thing as this is the truth for everybody.
If you just want to test, use a plain resistor to V- or a JFET + resistor (current generator).
Re: Re: biasing opamps in class a
My experience shows that biasing it with a resistor can make the opamp sound worse (depending on the opamp). I experienced it when playing with the output stage of a CD player. It compressed the sound, took away all the liveliness and free flow from it.
peranders said:
My experience shows that biasing it with a resistor can make the opamp sound worse (depending on the opamp). I experienced it when playing with the output stage of a CD player. It compressed the sound, took away all the liveliness and free flow from it.
hi guys,
i tested an opa2132 with bias currents from 0.4 to 8mA and to tell you the truth i don't think i can really hear any difference. As it's a battery powered headphone amplifier i'll just leave it in class ab. disdtortion is very low with this chip (0.00008 or something) anyways,
It might be my somewhat cheap sennheiser hd477's.
I hooked it up to my gainclone and guess what, it sounded pretty good (although gain was far too high), a lot more control in lower freq's. Soundstaging is also better.
So, my next project is an opamp based pre for my gainclone. Maybe i'll give class a biasing another shot there (although the opa627's have an even lower distortion figure).
Viva the opamp! (don't shoot me for saying this!)
i tested an opa2132 with bias currents from 0.4 to 8mA and to tell you the truth i don't think i can really hear any difference. As it's a battery powered headphone amplifier i'll just leave it in class ab. disdtortion is very low with this chip (0.00008 or something) anyways,
It might be my somewhat cheap sennheiser hd477's.
I hooked it up to my gainclone and guess what, it sounded pretty good (although gain was far too high), a lot more control in lower freq's. Soundstaging is also better.
So, my next project is an opamp based pre for my gainclone. Maybe i'll give class a biasing another shot there (although the opa627's have an even lower distortion figure).
Viva the opamp! (don't shoot me for saying this!)
matjans said:
I'm tired of saying this everywhere, but here it goes again:
The best sound I have is from my (buffered, now LM3886+OPA627) IGC power-amp working with my OPA627+BUF634 stand-alone pre.
I made several other GCs (IGC, NIGC, different chips, etc.), integrated, and none gets to this outrageous performance.
matjans said:
And this proves that a high-end op-amp can sometimes be good for the sound.
This defeats the ultra-minimalism theory.
Sometimes I
listening to some well-recorded music.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.