I work on my master thesis and use the ADUM4135 guys to create a 3-phase bridge for BLDC motor. Unfortunatelly two upper drivers are continuously in fault state. When I open lower transistors to charge up the bootstrap capacitors and power up the upper drivers, the READY pins associated to the upper drivers are set active, but the FAULT pins are active too. I thought it is cause of the desaturation detection, so I short the caps between the desat. pins and VSS2 pins (see attached figure), disable the desat. detection. Well this didn't help.
Strange thing is that only two upper drivers are wierd, the last one is ok and works fine. By the way, I also tried replace those two upper drivers with brand new but nothing. The problematic guys are the first two on the left in the schematic.
Can anyone help me please?
thanks a lot for replay ADI!
The first thing I notice is the decoupling/bootstrap capacitance on the output sides of the drivers are a bit small. We usually recommend 10 µF + 0.1 µF on VDD2 to GND2/VSS2. At a minimum, you should increase the 0.1 µF capacitors to at least 2 µF. If that doesn't fix things, try the below suggestions.
The READY and nFAULT signals get transferred to the input side on the same isolation coil. The fact that READY is working means we have communication on the back channel. Let's first isolate whether the bootstrapping is affecting operation.
1) For this test, don't power up the bus voltage. Try powering just one of the affected ADuM4135 ICs that are showing the problem by powering directly with a bench supply across VDD2 and GND2/VSS2. Power up VDD1 as normal, and see if the nFAULT signal is asserted (low).
If that IC works properly when powered by the bench supply, the problem might be in the speed the bootstrapping charging is occurring at, or possibly ringing seen on the VDD2 pin during startup. One thing to try is to place a series resistance in series with the bootstrap diodes. A value of 0.5 to 1 Ω is usually enough.
2) In the same setup as above, power VDD1 normally, and power VDD2 of an affected part directly with a bench supply. If the nFAULT pin is asserted (low) again, try bringing nRESET low for 1 µs, then bring nRESET back up to VDD1. If this clears the fault, then the fault could be being asserted at startup.
Let me know how these tests go.