Current limit hit with encoder connected

[Garagehermit]

I have several potential threads of issues ,but I am going to mind-dump it all out and hope someone can scan through and ID the real culprits worthy of their own threads .

I’m having trouble getting motor to spin with encoder connected. I have blown up two IGBTs, and each time, the failed IGBTs were connected to a gate driver without a current sensor associated with it, i.e. Gate drivers #1 and #2 were associated with current sensors #1,#2. I have had to replace several components on gate driver board #3 (TVS diode, DcDc converter, etc), but I am hopeful some of this is mostly coincidence and occasional shoddy soldering. I will circle back to this a little later.

My motor is rated at 23amps, 234V. My IGBTs are rated at 50amps. With command “start 2”, the motor is able to spin without hitting current limit when ocurlim>80. With fslipspnt = 7 & ampnom>50, the motor spins slowly (guessing around 4 rev per sec). The speed increases when I increase ampnom and when I increase fslipspnt (obviously). When I connect single channel encoder(proximity sensor) high current limit is tripped even at ocurlim=150. Same happens when I connect encoder while motor is spinning. Most of the time the current limit is hit without any visible motion. I am not confident I should keep increasing ocurlim.

Time for some side notes.
A few months ago, I ran the motor successfully with a used inverter GE AF650-GP to try to rule out motor issues.
Around the same time, With a quad encoder connected, I spun motor very slowly with full throttle. I then changed a few settings and maybe switched encoder wires A and B but I didn’t think to write any of it down because the motor spun wildly fast when I throttled up. I thought this was wrong so started rewiring, resoldering and through a string of bad decisions eventually blew up an IGBT. One of my oversights maybe worth adding to the ‘don’t try this at home’ for other newbies, I had current sensor #1 associated with gate driver #2 and had sensor #2 associated with driver #1. Does this matter? It happened by accident when I reversed two wires to get motor to spin opposite direction.
After reading a post about encoder issues, I realized the motor was supposed to spin wildly fast when motor is free spinning. I haven’t been able to recreate that outcome with single channel encoder. For the quad encoder setup, since my motor shaft only extends 3mm beyond the bearing, I epoxied a magnet with a small 3/8” shaft extension/adapter... but I don’t trust that setup in all-weather driving conditions which prompted the move to single channel. I just had to drill and tap a hole in the gear box to access gear teeth. Lucky me, if only I had thought about this in the beginning.

Before moving forward with troubleshooting, I need to fix the driver board components and get a few new 50amp IGBTs. I have 100A IGBTs but seems like overkill for a 4kW motor. I might be in need of a few new gate driver boards since I can’t keep desoldering parts forever on the same boards.
Is it possible to order just bare driver boards and a voltage sensor board or could they be added to a new kit order? It may come to a point where it is necessary to buy a new kit. I made this one last almost a full year of joy, discovery and heartache.

[johu]

I think you need to restart from this step: https://openinverter.org/wiki/Schematic ... te_drivers
I can send you some bare driver boards for postage, let me know your address.

[Garagehermit]

Previously, I modified gate driver to increase gate resistor to 23Ohm (added resistor in parallel with 10ohm gate resistors). The IGBT spec sheets recommended higher gate resistance for 1.5us dead time. After replacing TVS diode and shortening the connectors from driver boards to IGBT, I started testing drivers again.
With zero bus voltage, I set deadtime=135 to get 1.8us deadtime. Connected 48v bus voltage and got similar result. When I scoped signals of each driver board, the voltage span (from Vmax to Vmin) of GTOP was roughly twice the voltage span of GBOT. I didn’t worry much since observed same for each driver board. There was 1.5mA current from battery to IGBT bus with fslipspnt=1, ampnom=1, pwmfrq=8.8KHz. When I changed to 4.4kHz and deadtime=160, the current from battery was 0.80mA. I thought this might be a sign of shoot-through but got 0.7mA reading from new IGBT module (3 half-bridges in one module). I can post pics of scope GTOP,GBOT if needed. When comparing GTOP to GBOT, the amplitudes are different and have slightly different shapes. But scoping each driver board provides the same differences in signals.