BA283AV, gain, and bipolar PSUs

[Potato Cakes]

Hello, everyone,

I had posted several questions about the BA283AV circuit on various threads and I found answers to said questions and I wanted to share them here.

Last time I was here, I was trying to figure what/if anything needed to be changed to get a BA283AV circuit to work with a bipolar PSU as this would drastically simplify the power scheme when wanting to incorporate things like ICs and a headphone amp all in the same design. Tonight I finally got around to  doing some testing with a 500 series rack PSU. I connected the V- from the PSU to the 0V on the board. The common and chassis connections for the PSU were both tied together and connected to ground. Shielding for the input transformer and XLRs were also tied to ground. The result is a very quiet circuit that has plenty of gain for mics. No hum or buzz, even with board out on the bench and near things that usually would cause interference. I changed the power scheme back to the standard +24V/0V produced the hum that has plagued many a build based on this circuit. I swapped back the V+/- wiring on the BA circuit and it was back to quiet. And this was with board out of a case and near the PSU. 

There was plenty of gain for microphones as well when the circuit was wired with the proper input and output transformers. Both stages on the BA circuit can have potentiometers/trimmers installed for gain adjustment.

The circuit still needs to be biased, but maybe I can find a pair of resistors that would work in place of the trimmer. I haven't made it that far. If I could find a way to use fixed resistors in place of the bias trimmer is the last thing I would like to figure out. I don't know if this has already been done and is now common knowledge.

Thanks!

Paul

 

[Bo Deadly]

You can use any voltage for "ground" in the BA283 as long as it's vaguely +24V relative. You just need to be careful that you don't accidentally connect that "ground" to the 0V of the enclosing circuit because that would effectively short out the negative rail. So for example, if you simply connect the the gain control terminal T to your enclosing circuit 0V, you will be shorting 12V / 1K8 = 6.7mA from 0V to the negative rail. So you have two options:

1) Tweak the BA283 board to seek out and disconnect all resistors on the outside of coupling caps and and then reconnect them to 0V of the enclosing circuit. However, this introduces a new problem which is that the coupling caps could be accidentally reverse biased. Meaning if the bias in the BA283 is below 0V of the outer circuit, the electrolytic cap will be reverse biased and it could be damaged.

2) Add more coupling caps to the input and output to completely isolate the BA283.

As for hum and noise, there should no reason why a single supply should be noisy. Your particular single supply power supply is probably just noisy compared to your bipoloar one.

 

[Rocinante]

There are plenty of transformers out there, that have a 24v and 5v or 12v winding which you could run through a separate bipolar power supply to connect your ic's, led's, etc...

To hop on this though, can one just use the +24v and 0v coming from the bipolar psu to the ba283 and then use the +-24v to power the other ic's, headphone amps, etc... For instance; opa2604 can handle +-24v.

 

[mjrippe]

There are plenty of transformers out there, that have a 24v and 5v or 12v winding which you could run through a separate bipolar power supply to connect your ic's, led's, etc...

To hop on this though, can one just use the +24v and 0v coming from the bipolar psu to the ba283 and then use the +-24v to power the other ic's, headphone amps, etc... For instance; opa2604 can handle +-24v.

Or use further regulation to bring it down to +/-15v.  The ADR Scamp rack had +/-24v power rails but many of the modules had 7815 and 7915 regulators on board.

 

[Winston OBoogie]

When I've run a circuit designed around a single-sided supply (such as a Neve or Langevin) from a Bi-Polar +/- 16V supply,  I've left  the +16V alone and just added a negative reg to take the -16V down to about -10V and the run the circuit between the + and -.
This puts the centre bias point at about +3V  which usually, but not always, means all my caps are still properly oriented.
Being Neve too, there is something to be said for a positive bias on those tantalums. 
I dare say some might flame me for that last remark but, whatever, been there,  measured it, bought the t-shirt. 

 

[Potato Cakes]

When I've run a circuit designed around a single-sided supply (such as a Neve or Langevin) from a Bi-Polar +/- 16V supply,  I've left  the +16V alone and just added a negative reg to take the -16V down to about -10V and the run the circuit between the + and -.
This puts the centre bias point at about +3V  which usually, but not always, means all my caps are still properly oriented.
Being Neve too, there is something to be said for a positive bias on those tantalums. 
I dare say some might flame me for that last remark but, whatever, been there,  measured it, bought the t-shirt. 

I think if any assertion around here is validated with measurement, then the person making those claims should be safe from the slings and arrows of internet comment combat.

 

[Winston OBoogie]

I think if any assertion around here is validated with measurement, then the person making those claims should be safe from the slings and arrows of internet comment combat.

I agree.  I've none of my measurents at hand unfortunately as they were done eons ago with analogue test gear that required me taking a picture to display the data and, well...      Geoff Tanner (ex Neve) has been vocal online in the past regarding distortion vs bias voltage on dry tantalum caps.
A better (more credible than I?)  source might be the tests done by Cyril Bateman and published in 'Electronic World' in the early 2000's.  However, his initial testing of tantalum with a  0.3V audio signal produced  10 times more distortion than the electrolytics under test and so he excluded them from further tests.  His summary was thus:
"Distortion of this Tantalum bead capacitor, is ten times worse than found with similar value and voltage aluminium electrolytic capacitors.
Distortion does reduce slightly with application of DC bias."

Attached is his test result showing the general nature of the distortion from tants.