Small function generator doesn't work (OpAmp overheats)

Hello,

I tried to make a simple function generator using the MAX038 IC.
Schematic: https://www.circuitlab.com/circuit/mx397v/max038-function-generator



the OpAmp is a LT1364.

When the MAX038 is NOT connected yet, I'm getting + and - 15V, as expected.
For what should be + and - 5V, I'm getting +5V and -6V. When I connect a load to the output of the 7905, the voltage goes to -5V. I think this is not a problem.

The supply voltage is 29V
When I connect the MAX038 IC, the -+15V changes into +25V and -3.5V for some reason, and the OpAmp's temperature goes to smoking hot in about 10 seconds.
How is this possible?! how can the output of the OpAmp suddenly change, while it's inputs stay the same with those 10k resistors?

Kind regards



EDIT:
additional information:
The voltage across the 10k voltage divider at the + input of OpAmp A, stays nicely at +-15V while the output goes to +25V and -3.5V.
When overheating, the whole circuit draw 170mA of current.
Without the MAX038, when not overheating, the circuit draws 27mA.

Is it possible that the OpAmp can't deliver enough current to the MAX038?
OpAmp can deliver 60mA, MAX038 I+ is 35mA, V- is 45mA.

I'm not sure how to calculate what current will flow trough the OpAmps output, since the positive voltages is sinked and the negative is sourced, right?
Do I have to add or substract the two currents?

But still, if the MAX038 requires more current than the OpAmp can deliver, is this the result you get? that the output caps to one side?



Other things I tested:
- Leaving V- of the MAX038 IC unconnected: This resolves the problem. I'm getting a nice +-15V and nice +-5V.
- V- still not connected, but used a resistor to add 50mA load to the -5V. (to simulate the "opamp output being loaded too much"): No problem, nice +-15 and +-5V.
Conclusion: the OpAmp is able to supply the required current.
- Leaving V+ of the MAX038 IC unconnected (to test if lower load current would at least make the negative side operate properly): Problem still occurs, +25V and -3.5V is generated.
Conclusion: Lowering the current which the OpAmp needs to deliver does not fix the problem.
Conclusion: Nothing indicates that the OpAmp is being loaded too much. This is probably not the problem.

Overall conclusion: I think the MAX038's V- pin is broken. Maybe it creates a short circuit? -> measured it's resistance from V-pin to each pin I connected to GND: Result is always infinite resistance.
-> no short? so what is wrong then??

The "virtual ground generator" would mainly provide the difference between MAX038 positive and negative supply + signal amplifier output current. So it can basically work.

Nevertheless it's a bad power supply design that should be replaced by something more state-of-the-art. It's a sufficient reason against it that it will make fault diagnostic really hard. LT1364 also isn't designed to run with 2x10 µ output short, it's expected to be stable only by capacitor ESR, but beware to place too much low ESR bypass capacitors. Unfortunately you prefer ceramic bypass Cs with high MHz amplifiers...

ah I see. But without the MAX038 connected, the capacitors don't give a problem, and with resistors connected,
giving the same load as the MAX038, also no problem, so I think I should search in another direction?

I really don't understand how it's possible for the OpAmps output to be giving a wrong value while it's input stays the same :/ It must mean that the negative side is being pulled towards ground right?


UPDATE: I replaced the OpAmp with the same model; I had a spare one. Problem still the same. I hope I didn't also break the new one.
I can't try replacing the MAX038 since it is a $10 IC and it would take weeks for me to get a new one..

You should measure the MAX038 supply currents. May be it's a connection fault.

Iv+ 110mA Datasheet: 45mA typ
Iv- 0.27mA Datasheet: 35mA typ

V+ 5V Datasheet: +5V typ
V- -1V Datasheet: -5Vtyp

Not understandable at first sight why the negative supply drops although the positive is overloaded...

I suggest to test the MAX038 circuit first with +/- 5V lab supply.

well I'm thinking it is normal for the V+ to get overloaded.
When the V- side gets loaded too much, the opamps output will be pulled towards V-, which automatically increases the voltage from output to V+, since the total voltage must always be 30V.

Supply voltage is 29V
V+ = +25V and V- = -3.5V
if you also take in account the 0.6V voltage drop across the diode
then it kind of adds up

Good suggestion for the lab supply, but I don't have any :/
and I also don't have spare resistors to connect to the inputs and outputs for the test.
(or am I allowed to just ONLY connect the supply pins, and leave the rest disconnected, or to ground?)

For example, there is a pin that requires to be 12kOhm to ground, when not used. But I don't have spare resistors :/

- - - Updated - - -

Alright, I used what I had laying around, which was a 5V ~2A supply.
I left the MAX038 in it's socket, disconnecting only the V- pin.

I used the external 5V supply to power the V- pin. Supply voltage drops from 5V to 4.8V and maximum supply current goes trough the pin; 2A.
?! There are no sparks or anything when I connect it. The IC does get hot after about 2seconds.

So I'm guessing the IC was a broken one? If not, it now definitely is ..

- - - Updated - - -

I now also took out the IC, and applied the -5V directly to GND and V-. Same thing happens; 2A and heats up.
As a test, I hooked up the +5V directly to GND and V+, the same way as the previous test for V-, to check if that would work and it gives me a nice 35mA like what should happen.

I guess, that the device was defective before. To be on the safe side, I would power the MAX038 with a current limited (e.g. 50 mA or 100 mA) power supply in first test.

Double checking the circuit for possible wiring faults is another point.

Ok. Thanks for all the help.

I'm going to buy a new MAX038 IC.

There are 5V regulators just before the MAX038 so I think it's 100% safe for the new IC?
Is there any chance the circuit causes the MAX038 to be destroyed?

I don't want the IC to also go broken from the moment I plug it in.

Update:
I have finally received my new MAX038 IC and guess what?
It works like a charm!!
I'm so happy that I no longer need to get frustrated on finding the problem; just a faulty component

thanks for all your help!

There is still some noise on the signal, but that's probably because I didn't add any of the suggested small capacitors on the IC. I now can go further in optimizing it!