Reverse Diode Overheating in 4-Phase 3kW 400V DC SRM Motor With IGBT Switching

Hi Guys,
I am playing with a 4 phase 3 Kw 400 V DC SRM motor with IGBT switching. I keep trying different circuits and keep finding that the reverse diode accross the winding overheats quickly. I used 4 x 25F120 in parallel on a C Dump circuit, attachments is added.
At 50 V DC there is quite a bit of heat at 20 KHZ, which is the carrier frequency and the duty cycle per phase ( 4 phases) is obviously 25 %. My experiments underneath was with a single IGBT and coil (100 VA) running without stopping, just to make sure it works.
I prefer a 1 IGBT topology because in an SRM, forward reverse is not done by reversing the current through the coil, but by switching the right coil on at the right angle.
What advice can you give me please to get the best efficient and effective solution. I do not have much experience with IGBT switching but I am determined to get it right. Obviously the target would be not to absorb the reverse energy in the coil, but to give it back to the supply. I have tried R Dump, very hot, C Dump, less heat but still a lot. The fact that I need 4 diodes in parallel is baffling.
Can anybody advise me?
jdver

Hi Cobus-I think you will need a half bridge per phase, with an IGBT pulling up and another pulling down. I  also think that the real-world motor will behave much differently than the test jig you have set up.  Infineon (among others) has some good design resources for IGBT drivers for switched reluctance motors.  Look HERE.  Although these examples show a 3-phase motor, it is not that hard to add more phases.-Rick

Hi Rick,Thank you for your advice. I checked it and some other sources as well and came up with this.However it gives me a 30 Volt above supply V spike with 40 DC supply. The spike lasts 250nSec, which I assume is the slow switch on characteristic of the Diode.I will try with Schottky diodes as well, but need to get rid of this spike because with a 400 V supply this spike on a SRM coil will most likely peak at 300 V above supply. ie 700 V.Because the SRM does not operate on bipolar pulses, so I understand, this circuit saves one IGBT.Any recommendations etc?Regardsjdver

If you do a Google search on Switched Reluctance Motor Schematic you'll find lots of examples of almost the same schematic- and they all use two transistors per phase.  You might try one of those circuits.-Rick