Help with Biasing an Op-Amp

Thread Starter

madscientistdan

Joined Mar 6, 2012
36
Hello,
I need a bit of help with a bandpass filter circuit. I used Texas Instruments software to design a 4th order bandpass filter
using the THS4032 op amp chip.

I made a prototype PCB that lets me switch jumpers to run in dual supply mode or single supply mode. I have some
applications for both so I wanted to try them both out. I've attached a schematic for the TI design and the actual circuit
I designed.

I designed a PCB and assembled the circuit and tested in dual supply mode and it works perfectly. When I try to run in
single supply mode I cannot bias the op amp input to 6V. The power source for single supply is +12V. For dual supply
mode I set all jumpers between pins 2 and 3. For single supply I set all jumpers between pins 1 and 2.

In single supply mode I measure 12.13 volts at the right side of R15, but 8.60 volts across it. I measure 3.48 volts across
R14 and about 3.5 volts with respect to ground right in between R14 and R15 (I want this to be 6volts). If I remove the
jumper from JP5 I see exactly 6 volts at pin 1 of JP5 and 6 volts across R14 and R15 as I would expect.

I see the same type of behavior with the bias point between resistor R26 and R27.


Does anyone have an idea what could be wrong? In a different version circuit for a high pass filter I use the same
technique for biasing and it works perfectly in single or dual supply mode. I've double checked all my PCB design and
components and they check out fine. I've built two of these circuits and they both behave exactly the same.

I just wanted to know if anyone sees anything I'm doing wrong from the design itself.



 

Attachments

Last edited:
Looks ok to me.
So long as you fit all the right jumpers it should work.
I might have used 100uf instead of the 10uf as some of the filter resistors are only 3k9.
 

Thread Starter

madscientistdan

Joined Mar 6, 2012
36
Thanks. I still don't understand why I am not seeing the voltage between the two 100k resistors at 6V to properly bias the op amp when running in single mode.
 

ifixit

Joined Nov 20, 2008
650
Hi Dan,

In single supply mode the current in the supply virtual ground divider resistors should be at least equal to the value of the quiescent supply current used by the op-amps. Each op-amp in the package uses typically 7.5 mA, so that's 15 mA total per package. R14 + R15 should be set to 11-3V/15mA or 753Ω. Use 360Ω each.

In the circuit you have, any small difference in quiescent current between the -ve and +ve rail will cause an offset because the 100K resistors cannot supply the current to keep the virtual ground point at midway. Any DC output from the op-amps has to be supplied from one rail, or the other, not both at the same time.

Having said this, you can likely get away with R14 and R15 set in the 5 to 10K range so long as there is little, or no DC output current needed from the op-amps. Experiment.

The smoothing caps on the virtual ground point should selected such that there is low ripple for the lowest frequencies expected. Your look like they will be good.

Regards,
Ifixit
 

Thread Starter

madscientistdan

Joined Mar 6, 2012
36
Hi Dan,

In single supply mode the current in the supply virtual ground divider resistors should be at least equal to the value of the quiescent supply current used by the op-amps. Each op-amp in the package uses typically 7.5 mA, so that's 15 mA total per package. R14 + R15 should be set to 11-3V/15mA or 753Ω. Use 360Ω each.

In the circuit you have, any small difference in quiescent current between the -ve and +ve rail will cause an offset because the 100K resistors cannot supply the current to keep the virtual ground point at midway. Any DC output from the op-amps has to be supplied from one rail, or the other, not both at the same time.

Having said this, you can likely get away with R14 and R15 set in the 5 to 10K range so long as there is little, or no DC output current needed from the op-amps. Experiment.

The smoothing caps on the virtual ground point should selected such that there is low ripple for the lowest frequencies expected. Your look like they will be good.

Regards,
Ifixit
Thanks so much for the help! I'll try your suggestion right now.
 
Top