The circuit below is part of a design to simulate natural fluctuations geomagnetic recordings for test purposes. The ELF sinewave thus needs to be made irregular in frequency and amplitude by introducing a random modulation, in this case low pass filtered noise.

I had planned on using an AD633 quad multiplier for this purpose because it will (supposedly) handle low frequency. However, in doing so it is preferred to retain the single supply.

I understand that the AD633 is intended to operated from a +/-15V supply, with pin 6 to negative. But it was suggested that 12V may work depending upon the x y voltages. So far I have had no success. The output just goes either to rail or 0V.

Can anyone please advise what I might try using this chip, or suggest an alternative approach that will produce a similar multiplied product?

 

The AD633 Power suggested supply voltages can range from ±8 V to ±18 V (data sheet 1st page) so, from that, the minimum single supply voltage would be 16V.

And for single supply operation all signals should be offset to +8V (pseudo-ground).
The maximum signal levels are only within 5V of the rail voltages, so your signal range with a single 16V supply would be about +5V to +11V.

LTspice example simulation below (PG is 8V pseudo-ground).
Is that sufficient for your purposes?

 

I will try your helpful circuit suggestions this evening and report back. Get it working at 16V with the 8V offset and then see if I can reduce it a bit. I assume the LM324 specified in your diagram is just SPICE a placeholder for the TLC272. The output of your simulation is exactly what I am looking for.

I noticed a mistake in my diagram. Pin 8 links to the positive rail only and is not linked to Y1.

 

I assume the LM324 specified in your diagram is just SPICE a placeholder for the TLC272.

Yes, it's just a simple op amp inverting circuit with gain so the actual model used is not critical.

Note that if you need the output centered a 0V, you can couple it through a series (DC block) capacitor followed with a resistor to ground.

 

I built the circuit exactly as shown in your schematic and checked to ensure pin 4 was at half rail. However, when I plugged in the AD633, R4 became hot as did the chip. Not smoking but too hot to keep my finger on. R3 remained cool. The output at pin 7 goes to rail.

Obviously there is a problem. My guess is that current from the pseudo ground is shorting internally to actual ground via pin 5. Is there a solution?

I am hoping the AD633 has not been damaged.

 

What is the PG voltage when the AD633 is conneced?
I would double-check all the wiring against the schematic.
Make sure V+ only goes to the proper nodes.

 

PG sits at 8V until I connect pin 5 of the AD633, then it drops to 0.9. At a loss to understand why. Reistance between pins 4 and 5 unpowered is 520 ohms. I have checked the wiring numerous times. Is there another approach to multiplication that I can try that will work with a single supply and no divider?

 

PG sits at 8V until I connect pin 5 of the AD633, then it drops to 0.9. At a loss to understand why. Reistance between pins 4 and 5 unpowered is 520 ohms. I have checked the wiring numerous times. Is there another approach to multiplication that I can try that will work with a single supply and no divider?

You could try a different model multiplier, but I don't think that will help your problem.
It would appear that the AD633 may be zapped.
Do you have another one to try?

C1 may have generated a large current into the multiplier when the power was removed.
To prevent that, connect a 1k resistor at the output of C1 to the multiplier inputs 2, 4, 6, and the (+) pin of the op amp before you try again.

 

Apart from part failure, I do not understand the cause of the fault. These chips are too expensive to keep frying as test subjects, and a replacement would need to be ordered in.

Unless someone reading here is willing to debug this for me for a small fee, or propose another approach to multtplication, I will have to shelve it for the time being.

Of course, it bugs me not to find the solution.

 

I do not understand the cause of the fault.

I believe it was surge current from C1 when the power was removed from the circuit, as I suggested.
I see no other obvious causes the failure, other than random parts failure, which is possible but not generally likely.

a replacement would need to be ordered in.

Generally it's not good practice to have only one item of a part for breadboard testing, but if it's your personal money than I understand.