class A biasing kills battery life?
Dec 7, 2007 at 3:42 PM Thread Starter Post #1 of 20

ishtob

100+ Head-Fier
Joined
Oct 3, 2007
Posts
158
Likes
10
I built a battery powered PIMETA and noticed that after I put in the JFETs for class A bias, it seems to be draining the battery much faster than before the bias.

anyone knows why?
 
Dec 7, 2007 at 3:44 PM Post #2 of 20

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,405
Likes
31
It shouldn't. The class A bias setup should only result in a couple more mA drain on the battery.

I think we need more info about the amp to give more advice. So if you could lay out all the specs that would be helpful.
 
Dec 7, 2007 at 4:16 PM Post #4 of 20

Bjornboy81

100+ Head-Fier
Joined
May 18, 2007
Posts
282
Likes
10
If the correct FET's were used in the correct setup than battery life shouldn't be affected that much...some, but not much. What FET's are you using (did you buy them matched from Tangent)?
 
Dec 7, 2007 at 4:33 PM Post #5 of 20

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,405
Likes
31
Quote:

Originally Posted by thrice /img/forum/go_quote.gif
Adding those jfets increases the amount of current your amp pulls from the power supply, thus lowering your battery play time. Tangent has articles on his site discussing the use of that particular part of the circuit.


Doesn't it only do like 3mA of bias though? I wouldn't have thought that the add'l 6mA total would have made much in the way of a difference.
 
Dec 7, 2007 at 4:48 PM Post #6 of 20

thrice

Member of the Trade:Rockhopper Audio
Joined
Mar 25, 2004
Posts
2,103
Likes
10
Ok so let's say he's using an average 9V that gets about 625 maH of current:

Quiessant current without bias: ~17mA (if I remember correctly) Run-time: 36.75 hours

With Bias: say the 3mA you suggest

Run-time: 31.25

That's over 5 hours loss from just 3mA of current.

Look at the difference with the Mini3 in battery life:

High Performance pulls about ~25mA with a 270maH batt: 10.8 hours

Battery saving version pulls ~10mA with 270maH batt: 27 hours

But look at this: 13mA with the same 270maH batt: 20 hours.....3maH and we lost almost 25% battery time

Or look at this:

20mA with a 625mah Batt: 31.25 hours

23mA with a 625maH batt: 27.17 hours

We lose ~13% battery run time.

So I think the main question is, how much batt time is he losing and what's the quiescent current of the amp?
 
Dec 7, 2007 at 5:29 PM Post #7 of 20

NelsonVandal

1000+ Head-Fier
Joined
Aug 27, 2006
Posts
1,023
Likes
11
It's 3 x 3 mA = 9 mA extra current draw (depending on the IDSS of the JFETs it could be less or more). This is more than needed. I guess 0.5 to 1 mA is enough, probably more than enough. The BUF634's imput impedance is very high. You have to add resistors to lower the current, if you can measure current with your multimeter. The question is if the class A biasing do any good at all. Some opamps seem to sound better this way and others don't. You have to listen to find out yourself. If you can't hear the difference - don't use class A biasing. It's as easy as that!
 
Dec 7, 2007 at 5:50 PM Post #8 of 20

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,405
Likes
31
Quote:

Originally Posted by NelsonVandal /img/forum/go_quote.gif
It's 3 x 3 mA = 9 mA extra current draw (depending on the IDSS of the JFETs it could be less or more).


3x? Are you saying the ground channel gets the Class A bias as well? That wasn't my understanding of things, not that I haven't been wrong before.
wink.gif
 
Dec 7, 2007 at 6:11 PM Post #9 of 20

NelsonVandal

1000+ Head-Fier
Joined
Aug 27, 2006
Posts
1,023
Likes
11
Oops, sorry you're right n_maher. I should've looked at the schematic first. It's of course 2 x X mA. I was thinking of my own amps.

Maybe I'm wrong about the bias current as well, but why does it have to be so high? 2N5484 is specified to 1 - 6 mA IDSS.

Most of the opamps I've class A biased haven't improved an inch. I'm not really sure any of them improves with class A bias. It could be placebo, but I think OPA134 sounds better this way. What's your oppinon? It's easy to do an A-B comparison with a switch.

Ishtob, if you want to have your cake and eat it too, you can use AD744, AD829 or NE5534 comp pin out - bypassing the output stage, using only the class A first stages. I don't recommend NE5534, because it's really a lesser chip, very low resolution.
 
Dec 7, 2007 at 7:42 PM Post #10 of 20

balou

500+ Head-Fier
Joined
Apr 7, 2007
Posts
563
Likes
10
well, I wouldn't recommend the AD829. Without very careful balancing of the inputs, DC offsets will be sky high. Unless you have a separate DC offset nulling circuit - but the standard pimeta doesn't.
 
Dec 7, 2007 at 8:43 PM Post #11 of 20

NelsonVandal

1000+ Head-Fier
Joined
Aug 27, 2006
Posts
1,023
Likes
11
Quote:

Originally Posted by balou /img/forum/go_quote.gif
well, I wouldn't recommend the AD829. Without very careful balancing of the inputs, DC offsets will be sky high. Unless you have a separate DC offset nulling circuit - but the standard pimeta doesn't.


I meant used in the ground channel, like in LISAIII. There's no problem with it in this position, and it sounds very good.
 
Dec 12, 2007 at 9:35 AM Post #12 of 20

jsmithepa

500+ Head-Fier
Joined
Oct 16, 2007
Posts
508
Likes
10
Well I dunn know if current A designs are different but Classic A amplifiers, the transitors/FETs whatever are ON all the time, that's why it drains batt. B Designs don't turn on until there is signal, but problem with the .6v bias of a transistor (.3v of Mosfets?) so a compromise for A/B (most household amps) with base(?) bias of .6/.3 volts so transistors/FETs on a little bit.
 
Dec 12, 2007 at 11:29 AM Post #13 of 20

majkel

Headphoneus Supremus
Joined
Jun 1, 2007
Posts
2,783
Likes
59
Most modern op-amps don't respond to class A biasing. One of the reasons is because they have diamond buffer outputs (AD845, AD811, some others). All you can achieve is a slightly different and probably worse sound. The only op-amps I found responding to class A biasing were CA3140E (it's a joke to use it for audio) and AD825 which sounds different with no class A bias (I would say best), another way when biased to sink into class A, and the other way when sourcing from class A bias.
Using compensation pin as output is a typical forum hoax. It's sounds terribly thin and bright, much worse than normal output of the AD744. If it improves overall sound then it means you have real problems with your rig's sound signature being far from neutrual/transparent. There might be many reasons like: bad source, bad amplifier, high output impedance (OTL tubes, resistors in the path) driving low impedance phones, poor quality interconnects, etc.
Summarizing:
3.4mA class A biasing for AD825 might be interesting... same as this chip alone.
 
Dec 12, 2007 at 12:31 PM Post #14 of 20

00940

Headphoneus Supremus
Joined
Nov 6, 2002
Posts
4,493
Likes
47
Quote:

Originally Posted by majkel /img/forum/go_quote.gif
Using compensation pin as output is a typical forum hoax. It's sounds terribly thin and bright, much worse than normal output of the AD744. If it improves overall sound then it means you have real problems with your rig's sound signature being far from neutrual/transparent. There might be many reasons like: bad source, bad amplifier, high output impedance (OTL tubes, resistors in the path) driving low impedance phones, poor quality interconnects, etc.


The origin of using the compensation pin is an Analog Devices paper called "opamps application", edited by Walt Jung and Walt Kester. See this pdf : http://waltjung.org/PDFs/ADI_2002_Se..._Drivers_I.pdf , page 20. Not your common "forum hoax"...
 
Dec 21, 2007 at 12:50 AM Post #15 of 20

villekille

Head-Fier
Joined
Nov 22, 2006
Posts
84
Likes
10
All the amps on tangent's site which have class a biasing are buffered amps. The mini3 doesn't have class a biasing AFAIK but it doesn't have a buffer.

Is there anything to gain with class a biasing if using only a high current output opamp without any buffer? The bias current should be higher I guess.
 

Users who are viewing this thread

Top