Bipolar Op amps in single ended circuit

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Potato Cakes

Well-known member
Joined
Jul 1, 2014
Messages
2,258
Location
Nashville, TN
Hello,

I have a +24VDC single ended circuit using a SMPS and I am needing to integrate some INA137s which require +/-VDC. The best example I know how to implement this is the JLM Baby animal design which I own several. From what can I see in this schematic and some examples I've found online, +VDC from the PSU goes to the V+ connection of the op amp, and 1/2VDC from a voltage divider goes to the In+ connection of the op amp and the V- connection goes to 0V/ground. I have no clue how this works even though I know that it does as all of my BA preamps do. With the INA137s being used as differential line receivers to unbalance incoming signal, I would be putting 1/2VDC on pin 2 of the incoming XLR for all inputs that are being unbalanced in this fashion. I know that incoming VDC from PSUs are manipulated elegantly and easily for situations like this all of the time, but when I have to figure this out for myself I get mild brain swelling and the learning process hits a brick wall as after all of this time do DIY I still can't comprehend PSUs fully. Anyone who does understand this madness, would you be as so kind to help me understand what is going on so I can move forward and finish this current project that ails me?

Thanks!

Paul

Baby animal schematic:

http://www.jlmaudio.com/Baby%20Animal%20Dual%20Mic%20Pre%20Schematic.pdf
 
It is all a question of that good old fashioned concept 'bias'.

In split supply designs, the inputs are biased from 0V simply because this is half way between +V and -V. This means the outputs are also biased close to 0V. The reason for doing this is to ensure the maximum available voltage swing of the outputs.

In single supply systems (single ended means something else), you still need to ensure the maximum possible output swing is possible and you do this by biasing the circuit to half the supply voltage. The obvious difference is that your analogue ground reference is now half the supply volts below the dc level of the inputs and outputs. The answer is simple, you use input and output coupling capacitors to block the dc.

Note, your comment that the + input is connected to half the supply voltage is not strictly correct. The input is biased to half the supply voltage via a resistor.

Cheers

Ian
 
ruffrecords said:
It is all a question of that good old fashioned concept 'bias'.

In split supply designs, the inputs are biased from 0V simply because this is half way between +V and -V. This means the outputs are also biased close to 0V. The reason for doing this is to ensure the maximum available voltage swing of the outputs.

In single supply systems (single ended means something else), you still need to ensure the maximum possible output swing is possible and you do this by biasing the circuit to half the supply voltage. The obvious difference is that your analogue ground reference is now half the supply volts below the dc level of the inputs and outputs. The answer is simple, you use input and output coupling capacitors to block the dc.

Note, your comment that the + input is connected to half the supply voltage is not strictly correct. The input is biased to half the supply voltage via a resistor.

Cheers

Ian

Thanks for the assist, Ian! I have been studying more and it's slowly becoming less murky territory. I think I now have a starting point to try at least. I think for what I am doing I will be able to get away with a passive approach. If not I do have one rail splitting IC on hand to try.

Thanks!

Paul
 
Here is how to hook up a THAT1646 in a single rail circuit.
http://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=6&t=261&p=5775
 
I don't understand either of the circuit references.

How can that "baby animal" circuit even work if there's only max 14mA of phantom power and each op amp probably draws at least 5mA. The phantom voltage will be drawn down to nothing.

And in that THAT1646 circuit, why is there an op amp there? It would work equally well (probably better actually) if it was just a voltage divider and cap. The input impedance of the ground pin is 5K.
 
squarewave said:
How can that "baby animal" circuit even work if there's only max 14mA of phantom power and each op amp probably draws at least 5mA. The phantom voltage will be drawn down to nothing.
The 48V is not Phantom Power
 
squarewave said:
... It would work equally well (probably better actually) if it was just a voltage divider and cap.
Why do you think so? The quoted schematic is obviously taken from THAT. 
 
moamps said:
Why do you think so? The quoted schematic is obviously taken from THAT.
Even if it is from THAT (and I don't see why that is obvious) I can think for myself and I don't see why an op amp would be needed there. Maybe the THAT1646 equivalent circuit diagram isn't showing something critical but the impedance of the ground pin looks to be pretty high compared to even a mediocre electrolytic. I have actually run THAT1646 from a virtual ground that only used an electrolytic to bypass and it worked fine.

The only reason I can think of that would necessitate the op amp is if the virtual ground was used for something else too and it had to sink / source DC. Or maybe there's some corner case high frequency CMRR "out-smarts" mumbo jumbo that I'm missing. Otherwise, I don't see why it's needed.
 
>>     The quoted schematic is obviously taken from THAT.
> Even if it is from THAT (and I don't see why that is obvious)


"Image courtesy of THAT Corporation"

> I can think for myself and I don't see why an op amp would be needed there.

Yes, there is a 5k which may be a summing junction (not a virtual "ground"). But I also see a couple 20k from the output back to "Gnd". And these seem to bridge/null a balanced output to an un-balanced input. Output balance/CMRR is specced about 60dB, 1000:1, so I would speculate that a 0.1% or 5 Ohm added resistance might spoil the intended performance.

OTOH, Wayne's plan for multiple 1646 shows 47r to each. "... <<47R impedance at LF to realize maximum output balance". The 0.1uFd won't do that over most of the audio band.

I do agree some sufficiently large cap (2,000uFd?) will nail the "Gnd" pin. It also appears that <<47r may be fine even to an experienced designer.

But it is not an expensive experiment on the breadboard. The real cost is finding a worst-case situation which really stresses output balance and rejection. We all know it works good at home and at the first dozen gigs, then the 13th gig buzzes like a bastard.
 

Attachments

  • THAT1646-guts.gif
    THAT1646-guts.gif
    10.2 KB · Views: 15
PRR said:
Yes, there is a 5k which may be a summing junction (not a virtual "ground"). But I also see a couple 20k from the output back to "Gnd". And these seem to bridge/null a balanced output to an un-balanced input. Output balance/CMRR is specced about 60dB, 1000:1, so I would speculate that a 0.1% or 5 Ohm added resistance might spoil the intended performance.
Yeah, I suppose the impedance of an electrolytic at LF can be pretty high (100 ohms?). So maybe mains hum might not be rejected as well as it could. That could be worth an op amp.
 
Back
Top