3
\$\begingroup\$

I'm designing a moderately powerful motor controller (integrated into a product). The board runs on 24V and has a 100uF electrolytic input cap, as well as 2 10uF ceramics near the switcher.

Whenever I connect this board to power, there is a hearty spark. It hasn't noticeably damaged anything yet, but it seems like something that could.

How do you guys handle this on your PCBs? MOV? TVS DIODE? Nothing?

It's a desktop product so it will get connected and disconnected occasionally.

\$\endgroup\$
2
  • \$\begingroup\$ Anecdotal answer: I often see this behavior when I plug in PCBs to a power supply that is powered on. To avoid it, I turn the power off every time I plug in or unplug PCBs. Often theorized it as arc current or some kind of instantaneous short, but have never delved into it deeply enough to tell you the why. \$\endgroup\$
    – InBedded16
    Commented Oct 11, 2023 at 16:42
  • \$\begingroup\$ @InBedded16 At this voltage, it will be some arcing, but mostly contact vaporization of the connector materials as they approach extremely closely, tunneling current begins to flow, materials heat up due to the high current density, and the cloud of vapor can then support an arc. Thus, contact wear occurs. The peak current can be quite large (compare +V / (electrolytic ESR)) but the energy in this case isn't tremendous. \$\endgroup\$ Commented Oct 11, 2023 at 20:21

2 Answers 2

3
\$\begingroup\$

If the power is coming from a power brick or some other supply that can be turned off and on and is designed to be turned off and on while connected to a load, then you might consider putting a note in your instruction manual to connect the power supply first then turn it on. I would also make sure that the connector you're using safely contains the spark within the connector cavity. Contact erosion shouldn't be a huge issue if it's infrequent but who knows what the end user is going to do; adding a note to the effect that plugging and unplugging the device at the 24V input is not the correct way to turn it on or off might be prudent.

The real problem is whether or not the transient you're seeing results in an input voltage spike above your maximum allowable input. See LTC app note 88.

\$\endgroup\$
1
  • \$\begingroup\$ That app note perfectly answers my question. Thanks! \$\endgroup\$
    – Drew
    Commented Oct 12, 2023 at 14:21
1
\$\begingroup\$

It sounds like you are hot plugging the DC supply (powering up the DC supply first, then plugging it into the electronics). There are two issues with hot plugging. The first is a large current inrush as you power up your capacitors (the arcing you are seeing), which can exceed the power ratings of components or traces. Hot plugging sometimes also causes voltage spikes which can exceed the maximum voltage electronics can handle. For example, we've seen 48 V and higher spikes when hot plugging a 28 V DC supply. Neither is good.

At the minimum, a TVS diode will help clip the voltage spike. Having the electronics turn on a second or two after the connector is plugged in allows for the initial spike to dissipate through the TVS diode. To limit the inrush, slowly ramp up the input voltage to the electronics. To do all this, we put a surge stopper IC between our input connector and the rest of the electronics. For example, the Analog Devices LTC4364:

https://www.analog.com/en/products/ltc4364.html

The IC provides over-voltage, under-voltage, and reverse voltage protection. It will also shut down in the event the downstream electronics exceeds some value of current. Page 1 of the data sheet shows the IC limiting voltage to the electronics for a fast transient spike (like hot plugging)

https://www.analog.com/media/en/technical-documentation/data-sheets/ltc4364-1-4364-2.pdf

enter image description here

For the specific problem of hot-plugging, you can time delay the turn on of the output voltage by adding capacitors to the UV and OV (under voltage and over voltage) pins as shown on page 13 of the data sheet:

enter image description here

We typically use a second of delay to allow the initial voltage spike to pass before turning on the downstream electronics. The capacitor on the TMR pin allows you to control the ramp of the output voltage to reduce inrush due to powering up your capacitors.

\$\endgroup\$
1
  • \$\begingroup\$ Note that the electrolytic probably has enough ESR to address the inrush voltage overshoot issue, though a TVS may still be desirable say in automotive applications, or if the power cable is long, either way something subject to surge. A load switch, hot-swap or wired-OR controller is the cheaper version without surge suppression function. \$\endgroup\$ Commented Oct 11, 2023 at 20:18

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.