diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   Class A biasing of opamps (http://www.diyaudio.com/forums/solid-state/6983-class-biasing-opamps.html)

DrewP 21st October 2002 11:37 PM

Class A biasing of opamps
 
I read in a diy article years ago that you could class A bias an AD711 by running a 6k resistor between the output and the negative supply rail. (assuming that elsewhere you were taking care of the offset)

2 questions:

1) will this work also for an NE5534
2) If using the NE5534 in a phono stage with a feedback loop, will the offset screw up the feddback system?

I can't see how you can have a high gain opamp and feed back a DC offset from the output without it bolliksing things up.

Advice?

BTW, I'm using the 5534 due to it being both cheaper and lower noise than the 711 in this application but would like to class-a it if possible. It's reported to make a significant positive improvement.

TIA

Drew

peranders 22nd October 2002 06:40 AM

1 making the output stage going in class A is nothing unusual.

Check here

http://www.linear.com/prod/datasheet.html?datasheet=202

2 Don't know exactly what you mean but you don't makes things worse with this arrangement, only warmer or even hotter chip.

Note that this type of arrangement is only wise if the load is small (not heavy). Typical application is between a buffer like in the example above. The RIAA-network is rather heavy load if you want low noise so the advantage will probely be little. NE5534 has very small distortion figures without this class A thing. I have seen values of down 0.001%.

djk 22nd October 2002 10:07 AM

Feedback will keep the difference between the inputs at 0V. If you put a 5K pull down resistor to a -15V rail all that will happen is the output transistor on the +15V rail will turn on a little harder to keep the output at 0V. How much harder? 15V/5K=3mA. how much heat? (15V ^2)/5K=45mW.

peranders 22nd October 2002 10:43 AM

Quote:

Originally posted by djk
Feedback will keep the difference between the inputs at 0V. If you put a 5K pull down resistor to a -15V rail all that will happen is the output transistor on the +15V rail will turn on a little harder to keep the output at 0V. How much harder? 15V/5K=3mA. how much heat? (15V ^2)/5K=45mW.
I ment heating the chip! A couple of milliamps isn't much if you include your load but this current can just about enough to reduce distortion if you use a buffer.

The best solution is to use a current generator, a JFET or a BJT. Check my QSXM3 RIAA amp for an example of this.
http://home5.swipnet.se/~w-50674/hif...0schema_p1.pdf

The Saint 22nd October 2002 12:44 PM

Putting an Opamp into class A with a resistor on the o/p to any supply rail was originally shown by Walt Jung of Audio Amateur Magazine fame.
An active current source works great as well.
The concept works and it does make an improvement...

Crossover distortion See'ya.....

DrewP 22nd October 2002 09:44 PM

Many thanks folks.

It seems that if I have a PSU that'll handle it and coupling caps to remove any dc offset then it should work fine.
And, of course, it's nearly free.

Can always stick a heatsink on the chip if needed.

drew

Fred Dieckmann 22nd October 2002 10:22 PM

For the truly serious......
 
http://www4.head-fi.org/forums/showt...current+source

Fred:2c:

hifiZen 22nd October 2002 11:18 PM

lower noise eh?
 
Could be, I haven't looked at the '711 datasheet recently. But opamp noise is a fickle thing, and often misunderstood... ei, or the input voltage noise is most often quoted when comparing opamp noise specs. However, one must be aware that opamps also have current noise at their input terminals, and depending upon your circuit impedances, the total noise of the circuit could vary dramatically from the opamp's ei, since this does not necessarily dominate the noise of a circuit.

If I recall correctly, the output voltage noise will be something like:

en = sq.rt.[ ei^2 + (R*in)^2 + 2RkT ]

where:
en = output noise voltage
ei = input voltage noise of opamp
in = input current noise of opamp
R = source impedance seen by input (should be matched at each input for best performance)
k = Boltzmann's constant
T = temperature in degrees Kelvin (approx. Celcius + 300)

That's all just off the top of my head, so I could be off by a little, but that's the basic form of the equation. So, next time you're checking out opamp specs, pay attention to the current noise as well, and consider the impedances which will be in your circuit. If the datasheet doesn't list current noise, you can get a ballpark idea by looking at the input current spec. Most bipolar opamps will have input currents in the pA range, while FET input opamps will often have input currents in the fA range!

Anyway, 5534 is pretty low noise, but I hope some DIYers out there will find this info useful next time they're trying to build a low-noise circuit.

As for output stage biasing, a FET current source is probably your best bet... costs a few cents more than a plain resistor, but should sound better in most cases.

hifiZen 22nd October 2002 11:21 PM

Oh, something I forgot... that noise gets multiplied by the gain of the circuit, so you have to account for that too!

djk 23rd October 2002 12:38 AM

"coupling caps to remove any dc offset then it should work fine"

Adding the pull down resistor does not change the DC off set.


All times are GMT. The time now is 12:21 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2