I'm an electrical engineering student at TAMU working on a project for my first internship: a signal recorder for an experimental helicopter. For context, here are some of my requirements:

  • 19 DC analog inputs ranging from 0V to 6V (but could be as high as 24V)
  • Must be powered by 24V supply (5V=Teensy, 5V=TLC1543, +/-24V = LM324)
  • Must NOT effect the input signal in a noticeable way

To do this, I've designed an Op-Amp Buffer using LM324s and a voltage divider to bring the (potentially) 24V signal into the 5V range. I'm also using a PDM1-S24-D24-S and a SPU01N-05 to try to power the +/-24V for the Op-Amps and the 5V for the Teensy and ADCs, respectfully.

Unfortunately, my design has a major problem: The +/-24V DC-DC converter explodes after a few moments connected to power.

I think it is because the capacitors within the PDM1-S24-D24-S cannot discharge and eventually explode. When I check the resistance between the positive and negative 24V output pins on my circuit, my multi-meter oscillates between a small value (~500 ohms) and an "open loop" message - suggesting that either the resistance is massive or that there is no connection at all.


simulate this circuit – Schematic created using CircuitLab

Above is a rough schematic of what I've currently designed, for brevity I've omitted the other 17 Op-Amps but the pattern of connection continues.

It seems to me that the loads are in parallel, and therefore should have a smaller value (potentially too small, i.e. a short), but my measurements suggest otherwise. I tried adding the 10k resistors to solve this, but the results were the same. Can someone help me identify the problem causing my DC/DC converter to explode? And suggest a solution?

I recognize there are other problems with my design at the moment, but I will ask for help with those on a separate question. But I will happily hear any suggestions that aren't related to the question.

An additional question: Is it bad practice to use a different DC-DC converter to supply the ADCs that will read the signals here? Could it also cause problems with the +/-24V DC-DC?

  • 5
    \$\begingroup\$ +/-24V won't do the LM324s any good. The smoking remains probably don't do your DC/DC convertors much good either. \$\endgroup\$ – Brian Drummond Jul 8 at 19:59
  • 5
    \$\begingroup\$ @DehLeprechaun Well, now you do. Did you guys actually look at the LM324 datasheet before applying voltages to it? Who was it that decided it was okay to apply a 48V difference across the power pins of the LM324 to begin with? The LM324 is designed to accept a maximum of 32V OR +/-16V, in other words a maximum 32V differential between the power pins. That other circuit should be re-designed too because it could fail at any time and a repair with the same part won't necessarily be successful. \$\endgroup\$ – DKNguyen Jul 8 at 20:15
  • 3
    \$\begingroup\$ You should be running at the recommended ratings, not trying to push the absolute maximum ratings. \$\endgroup\$ – DKNguyen Jul 8 at 20:21
  • 2
    \$\begingroup\$ Right there on the front page : Single supply ... to 32V. Dual supply ... to +/-16V. Where do you think it says +/-30V? Page/figure/table please. \$\endgroup\$ – Brian Drummond Jul 8 at 20:22
  • 3
    \$\begingroup\$ The "absolute maximum" power supply voltage is +/- 16, or 32 V single-supply. There is no promise the circuit will work normally under these conditions. It only means that any higher voltage is likely to damage the chip. 48 V is higher than 32 V, so the datasheet is clearly telling you the chip could be damaged by your design. \$\endgroup\$ – The Photon Jul 8 at 20:22

This datasheet says that the recommended input voltage is 30V, but doesn't specify that it should be 60V across the whole thing. It does say that it can take a +/- 30V input, however, so we assumed that was acceptable.

Right next to the line that says +/- 30 V is the maximum applied to the input pins is a line that says +/- 16 V is the maximum to be applied to the power supply pins.

enter image description here

Given you had a spec telling you exactly what you're allowed to apply to the power pins, there's no reason at all that another spec about what you can apply to the input pins should be relied on to predict what's allowed on the power pins.

The absolute maximum voltage to be applied to the power pins is +/- 16 V.

You applied +/- 24 V, so you shouldn't be surprised at all that your chips blew up.

Also be aware that absolute maximum ratings are damage thresholds. They don't promise normal operation or performance. You should operate the chip at the recommended operating conditions ratings if you want to be sure of the circuit performance.

| improve this answer | |
  • \$\begingroup\$ +/-30V is the max DIFFERENTIAL input, i.e. 30V on one input and 0V on the other ... or vice versa. Allowed voltage on either input is essentially the supply range. \$\endgroup\$ – Brian Drummond Jul 8 at 20:46
  • \$\begingroup\$ This is definitely an important detail that we misread while going over the data sheet. Strangely, we haven't seen any problems from the LM324s, it's the power supply that typically goes first. Any idea why that is? \$\endgroup\$ – DehLeprechaun Jul 8 at 20:49
  • \$\begingroup\$ @BrianDrummond, my main point is if you want to know the maximum power supply voltage you should look at the line about the maximum power supply voltage, not the line about input voltages (whether differential or on an individual pin). \$\endgroup\$ – The Photon Jul 8 at 20:50
  • 2
    \$\begingroup\$ @DehLeprechaun, Your DC-DC has 21 mA maximum output current. And each op-amp takes 0.7 mA. Plus whatever current it supplies out its output (up to about 1 mA, if I read your schematic right). If you have all your sensors near their maximum output at the same time, you could easily be overloading the DC-DC. \$\endgroup\$ – The Photon Jul 8 at 20:51
  • 1
    \$\begingroup\$ The op-amps might also be pulling more than the specified 0.7 mA, since you're powering them with too high a voltage. \$\endgroup\$ – The Photon Jul 8 at 20:53

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.