0
\$\begingroup\$

schematic

simulate this circuit – Schematic created using CircuitLab

I'm a programmer and self-taught digital designer, and I thought I could deal with an op-amp even though I haven't for over 20 years.

Quick version: I'm taking an analog 0-3.3 V signal and turning it into a 0-5 V signal using a single-supply op-amp in non-inverting mode with a voltage follower after that, powered at 10 V.

All worked fine, and then suddenly the voltage follower output died.

More detail: I have an analog signal (not PWM, true ADC) coming out of an ARM processor at max. 3.3 V. I'm using a 14-pin, four op-amp chip, in this case a TLC2274 which I bought to replace the 30 year old MC3403 I had laying around, which would not go down to the negative supply as my memory and the spec said it would.

I'm going into one of the op-amps on the chip and using a 4.7 kΩ resistor from the output to the negative input and 5.6 kΩ and a 5 kΩ trimpot to GND to convert to 5 V.

I then go into two other op-amps on the chip as voltage followers, one to each of two outputs. One is just going to an analog meter for "visualization", the other is right now doing the same but is intended to go over a long line to a motor controller.

Still in testing, I have the "main" output going to an analog mA meter with an appropriate resistor, and a scope.

All was fine and then suddenly the "main" op-amp output died. The input looks good, the output is 0 V, as is of course the negative input. No noticeable overheating, or "Magic Smoke".

I don't know the impedance of the motor controller, and the 100 ft cable will add its own problems, but I'm sitting testing on a bench and it failed, how can I prevent that?

Does one normally use a resistor, resistor plus Zener, or something else to protect an op-amp output? I'm not seeing that in "Art of Electronics" or any of my (old) op-amp circuit books. If it dies on the bench, what do I do in the real world? The spec says the chip is short circuit protected, could this have been static?

So, questions:

Is there a way to improve this circuit?

I am assuming the problem is from static, but not sure if it could be something else, so what could be the reason for the failure since I have no voltages available beyond the supply and GND, to which the part is protected from shorts.

Is there a way to protect the op-amp I'm using? One comment already suggested a 100 Ω resistor, but that's to compensate for the cable that has not yet been connected.

\$\endgroup\$
1
  • 6
    \$\begingroup\$ A schematic will keep many more of us interested in the question. There's a CircuitLab button on the editor toolbar and when you use "Save and Insert" an editable schematic is saved in your post. You don't need a CircuitLab account, no screengrabs, no upload. Double-click a component to set its properties. R, H and V to Rotate and flip Horizontal and Vertical. Welcome to EE.SE. \$\endgroup\$
    – Transistor
    Jun 26, 2020 at 17:43

2 Answers 2

1
\$\begingroup\$

Op-amps and power supplies need bypass capacitors for stability and surge protection. Generally about 100nF on the power supply rails, and larger capacitors on the voltage regulators.

Voltage sources should be connected to op-amp inputs with resistors, at least 1k, and matching the impedance of the other input.

You show a 3.3V 10 Hz sine wave input on the non-inverting input of OA1 and OA3. This may be just a schematic error, where there should be a 3.3V DC signal.

\$\endgroup\$
1
\$\begingroup\$

You replaced a quad 741 type with a quad CMOS type which has rail-to-rail outputs.

CMOS is rugged in good design but can fail on the bench from a slip of usage.

  • This could be from driving a breadboard with a low impedance signal and controlling Vdd separately from the signal or from ESD exceeding the protection level.

  • The SCR latchup failure-effects may occur when either input exceeds the supply rail and delivers more than 5mA or the output is back-driven more than the current limit.

  • However, the output is current-limited and tolerates shorts to either rail.

  • a 100 ft cable with approx 20 pf/ft will represent 2nF load which is generally unstable on CMOS outputs unless you add a 50 to 100 Ohm series resistor.

Stresses beyond those listed under “absolute maximum ratings” may cause permanent damage to the device.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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