5
\$\begingroup\$

I have the following super simple op amp circuit that makes use of the AZV321 single op amp (https://www.diodes.com/assets/Datasheets/AZV321.pdf) in order to balance the voltage on a pair of supercapacitors in series.

The op amp and capacitors are receiving 5V and the circuit works for a period of time before the op amp fails and drives the C2 voltage toward 0V. After that, the op amp is no longer functional.

I have checked the specifications of the op amp for the current that might be sourced or sinked by the op amp and nothing seems out of line but the problem occurs repeatedly.

The battery source Vin, is very stable without any fluctuations and even when the R3 resistor is completely removed so no current is externally driven, the op amp fails. This op amp specs indicate that it can handle 5 volts.

In case there was a problem sinking or sourcing too much current, I even removed R3 so that the voltage would simply match the output of the voltage divider. However, the problem still occurs.

I am now wondering about the resistance of the voltage divider, but could not see anything that might describe what is happening.

Please ignore the pin numbering on the op amp as the diagram was just made with an available drawing package. Pin 7 is connected to 5V and pin 1 to ground. I have observed the death of the op amps both using the battery and an external power supply.

The capacitors were precharged and connected to the circuit with a switch once they had the exact same voltage as existing circuit. No inrush current or voltage transients were observed with a scope.

enter image description here

\$\endgroup\$
8
  • \$\begingroup\$ Power pins 1 and 7 are not connected in your schematics. Please show actual connections. \$\endgroup\$
    – Ale..chenski
    Commented Jan 31, 2019 at 0:10
  • \$\begingroup\$ Exactly what else is connected to this circuit? Is the op-amp power connected across the battery? Are there any other components at all? You're very close to the abs max supply voltage of the op-amp. \$\endgroup\$
    – Spehro 'speff' Pefhany
    Commented Jan 31, 2019 at 0:11
  • \$\begingroup\$ Also, how "completely removing R3" might possibly reduce /eliminate output current? \$\endgroup\$
    – Ale..chenski
    Commented Jan 31, 2019 at 0:13
  • 1
    \$\begingroup\$ Did you get the part from a reputable source? Have you used ESD precautions when building your circuit? \$\endgroup\$
    – crj11
    Commented Jan 31, 2019 at 0:55
  • 1
    \$\begingroup\$ With 5F capacitors and an op amp capable of driving only 20-160mA of current, oscillation seems an unlikely candidate to me. Can you explain your reasoning? \$\endgroup\$
    – Jason K.
    Commented Jan 31, 2019 at 2:32

1 Answer 1

Reset to default
1
\$\begingroup\$

I suspect this Op Amp has no business being in this design.

It has near rail to rail output using a push=drain<>pull=collector rail to rail driver when you need a Darlington complementary Emitter-Follower for stability. If/ when it oscillates from certain loads not specified in the datasheet such as yours, it may suffer from oscillation and shoot-thru failure on the low side as it has an NTC tempco BJT collector output which rise in current gain and BW with temp, making it a good candidate for unity gain oscillation.

Upon reading the Phase margin plots for this it is clear the phase margin drops to zero with rising capacitive load and even though you have 200 Ohms in series, I expect this IC was running hot with > 1/4 W and burnt out in a matter of minutes with thermal runaway on the low side as current gain and current limit rises with temp thus raising the oscillation power consumption.

enter image description here

Solution use any buffered BJT Op Amp, not a Rail-Rail FET type with a BJT low side drive. You do not need RRIO type OA. If you want more current drive, add a push-pull emitter-follower inside the unity gain loop.

enter image description here

This shows Phase margin reducing with rising capacitive load, a feature not found normally found in unity gain stable BJT Op Amps with active current limiting.

\$\endgroup\$
4
  • \$\begingroup\$ Thanks for the suggestion. Can you explain how oscillation would manifest itself in a situation where we are balancing a couple of 5F capacitors? I would have thought that these large capacitors would prevent any significant oscillation. A previous comment indicated the same idea, but didn't explain how it would be occurring. \$\endgroup\$
    – Jason K.
    Commented Feb 23, 2019 at 15:16
  • \$\begingroup\$ THe compensation doesn't work well for excessive load current . See page 6 bottom Gain and Phase vs. Frequency Load Resistive and Capacitance and extrapolate Maybe 500 Ohms might work or 1K but why not use a BJT Op AMP that is stable with 0 Ohms + 20 mA + current limit or more \$\endgroup\$
    – D.A.S.
    Commented Feb 23, 2019 at 15:23
  • \$\begingroup\$ So, I take it that you are thinking that at some point the circuit is getting more current than it is designed for and basically burning out the op amp? So I would either have to limit the current significantly or use an op amp with better characteristics. And when R3 is removed, are you seeing the oscillation causing the issue? I will try a different op amp and get back. \$\endgroup\$
    – Jason K.
    Commented Feb 23, 2019 at 18:31
  • \$\begingroup\$ I see that the Op Amp can handle DC short circuit current but not when it is oscillating at unity gain frequency with both high and low side drivers cross-conducting at 2.5V out of phase with high SOA power burning out the strongest driver (low side) But it is the lack of phase margin that is the root cause with this unique RRO topology. even with a low series R to a short circuit lagging current of a supercap \$\endgroup\$
    – D.A.S.
    Commented Feb 23, 2019 at 18:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.