0
\$\begingroup\$

I'm working with a device which employs the Microchip TC4422A to drive a series LCR load in the low kHz range, PWM, 50% DC. The device actually has several of these driver circuits, for several loads, each controlled independently.

The system exhibits no failures when all loads are connected (several hundred systems in the field). In configurations where there is an unused driver circuit (load not connected) there are many failures of the unused TC4422A device. The unused driver output is connected to a trace on the PCB which is several inches long, all the way to the PCB header connector. The circuit is floating, nothing pulling it up or down, no capacitors, no steering diodes.

I believe the reason for failure is the floating output of the unused driver. I believe the otherwise floating circuit is being coupled to external radiated EMI, causing the voltage on the driver output pin to exceed maximum ratings, causing latch-up of the driver.

If there was an issue with the power supply/ground to the drivers, they should all be at equal risk for failure, however the unused driver is the only one failing.

What I don't understand is precisely how this is happening. I can't reproduce the failure on command. There are many variables in the final installation.

Any insight is greatly appreciated. My apologies for the general nature of the information, it's a sensitive situation and don't want to divulge too much.

EDIT: Added schematic

U2 with the load connected will not exhibit any problems. U1 with no load connected, output floating, fails at an alarming rate. The two devices share power supply, input driven from the same uC. U1 is not driven by software in this configuration.

TC4422A Example Schematic

\$\endgroup\$
10
  • 2
    \$\begingroup\$ Welcome to EE.SE. You could divulge a link to the datasheet, I'm sure, and a clue to what kind of chip the TC442A is (although you do give a few clues in the tags). You should probably give some sort of schematic sketch of the important bits otherwise we're all guessing. \$\endgroup\$
    – Transistor
    Feb 10, 2018 at 18:13
  • \$\begingroup\$ Floating output also equals ESD target.... \$\endgroup\$
    – Trevor_G
    Feb 10, 2018 at 19:48
  • \$\begingroup\$ "The unused driver output is connected to a trace on the PCB which is several inches long, all the way to the PCB header connector." - Can show us the PCB layout, or a photo of the board? Does the device have any form of EMI shielding? What EMI sources do you think might be present? \$\endgroup\$ Feb 10, 2018 at 23:03
  • 1
    \$\begingroup\$ "U1 is not driven by software in this configuration." - what exactly does this mean? \$\endgroup\$ Feb 10, 2018 at 23:09
  • \$\begingroup\$ If the unused Driver input, from the MCU, is tri-stated then that 9amp power driver may oscillate. Ensure the input does not float; add a 100Kohm pulldown. \$\endgroup\$ Feb 11, 2018 at 3:32

1 Answer 1

2
\$\begingroup\$

It turns out that Vdd was being raised beyond the maximum ratings, up to 23-24V. The mechanism was unexpected and took some time to identify.

Thanks everyone for your help.

\$\endgroup\$

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.