I am testing isolated gate driver IRS10752LPBF on the bench and I have a result I don't understand.

The only change I made on my test rig compared to the typical application proposed in the datasheet is that I don't have a bootstrap for the high side but an independent isolated power supply as I want to be able to run at 100% duty factor later on.

The gate output pin is floating for these tests (no MOSFET driven).

Screen shot attached: channel 1 yellow is the PWM signal while channel 3 purple is the gate driver output pin measured against the high side reference voltage.

As it can be seen on the screenshot, the output runs at 1/4 of the input frequency.

I have never seen that before. How is it possible? I am running at relatively low frequency. Is it a possible reason? My controller is limited at 1 kHz on this rig. (I don't see why but I am clueless on this one ...)

Any insight on what could cause this will be welcome Added a second screen shot after a suggestion was made in a comment: The first is the original at 50% duty factor while the second is at 25%

At the request of a commenter I am adding a crude schematic of the rig. It is very simple. Channel 1 is measured between pin 3 and 2 and channel 2 ( or 3 as I changed during the tests) between pin 5 and 4 schematic [25% duty factor][2] 50%% duty factor Thanks

  • \$\begingroup\$ do you get different output when you attach a reactive load? \$\endgroup\$
    – vicatcu
    May 2 at 20:28
  • 1
    \$\begingroup\$ Where is your bootstrap diode? \$\endgroup\$
    – JonRB
    May 2 at 21:09
  • 2
    \$\begingroup\$ You sure you can operate a bootstrap chip like this? \$\endgroup\$
    – JonRB
    May 2 at 21:19
  • 1
    \$\begingroup\$ I'm sort of wondering if the circuit needs some kind of connection, any connection between Vs and COM as would normally appear in such a circuit. Like through a 10K resistor or something because it's not actually a galvanically isolated IC. \$\endgroup\$
    – DKNguyen
    May 2 at 21:28
  • 1
    \$\begingroup\$ Can this work with 2 completely isolated power supplies? Look at the blockdiagram in the datasheet. Can the low side communicate with the "HV Well" without some ground return? (Normally trough the load to shared ground) \$\endgroup\$ May 2 at 21:33

Thanks to @ DKNguyen and @ Unimportant I found the problem. You are right: This particular chip apparently requires a connection between Vs and COM. On the rig I could short the 2 pins and the weird signals became normal. I then put a 1 meg resistor between Vs and Com and it still worked so it can be transformed in something useful as shorting VS and Com can't really work on a high side gate driver.

  • 1
    \$\begingroup\$ Good. Don't forget to accept your own answer to indicate to the system that the problem has been solved. Otherwise it keeps popping it back onto the home page looking for an answer. You may have to wait some hours or a day before you can accept. I can't remember how long. \$\endgroup\$
    – Transistor
    May 2 at 21:36
  • \$\begingroup\$ As @Transistor kindly mentioned, that enforced delay before self-accepting is possible, is 48 hours from the time the question was asked. So for this question, 2021-05-04 at 20:20:08Z is the earliest when accepting this self-answer becomes possible. \$\endgroup\$
    – SamGibson
    May 2 at 21:46

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.