Pierre's 170W in 1600W out Looped Very impressive Build continued & moderated

[T-1000]

A diode can clamp the voltage to a safe level. The current then can be directed to somewhere else or be dissipated as heat in the resistance of the coil. A Zener clamp will dissipate the power as heat in the Zener diode. Be sure the Zener diode can handle that power.

More on snubber protection: https://wiki2.org/en/Snubber

Regards
Ole

The classic snubbers mostly are wasting energy while protecting discrete elements. Here we are also trying to recover energy back to the input which makes task a bit more challenging.

[onielsen]

The classic snubbers mostly are wasting energy while protecting discrete elements. Here we are also trying to recover energy back to the input which makes task a bit more challenging.

Then an active snubber is needed instead of the simple passive ones. Zener diodes are totally wasting power when their clamping voltage is reached. An active snubber is an extra switch (transistor) that diverts the power back to the power supply or perhaps to the load. The energy can be temporarily stored in a capacitor that then is discharged to somewhere else where the energy can be useful.

Looking at Pierre's device dissipating heat in the heat sink while delivering power to itself and also giving off power to an external load it really makes enough power to use the cheaper and simpler passive snubbers. First make it work like Pierre did and then perhaps optimize it or stay with a cheap model even though it may not be as efficient as a more advanced model using active snubbers. After all it's not like a circuit running on a battery where every possible waste of power has to be minimized for extending the period between charging the battery.

Regards
Ole

[pmgr]

Thanks for the update L192... I was wondering if you were still at it.Not much new to report on my side other then when I re-positioned all my stator coils to 5 slot space I forgot to also reposition all my H-Bridges, so they were still at 6 slot spaces. So you can imagine that did not do well and have since found 9 of the H-Bridges that are fried. So I can no longer test.
I have someone that will buy 30 of these for me to continue: https://rover.ebay.com/rover/1/711-53200-19255-0/1?ff3=4&toolid=11800&pub=5574900467&campid=5338252492&mpre=https%3A%2F%2Fwww.ebay.com%2Fitm%2F200950936475%3FrmvSB%3Dtrue
It will be a week or more before I receive them.
So without a doubt this confirms the L298N are not going to be able handle the current no matter if you have one or even two per coil.
Regards
Luc

Luc, at this point I think it is safest you stick with relays instead of trying another H-bridge configuration. The flyback diodes need to be able to withstand a large reverse voltage and also need to carry a large forward current. If you have taken out the old H-bridges, you should check the flyback diodes on them to see if they are burned out (check in both forward and backward direction). It wouldn't surprise me if they are dead as well.

Stick with relays and get the version that will let you select whether it switches on a low or a high so you won't need to design any additional inverter logic and can drive two relays with one Arduino output. Additionally, you will need to make a flyback diode board like Pierre has (I note he uses two diodes in parallel to increase current capacity). You could ask him for the part number so we can check what current capacity and  reverse voltage they need to be able to withstand. Additionally, their reverse current should be as small as possible as they will draw current from the capacitor bank in reverse and thus waist energy.

PmgR

[listener192]

The classic snubbers mostly are wasting energy while protecting discrete elements. Here we are also trying to recover energy back to the input which makes task a bit more challenging.

Well actually the diodes on the boards worked just fine. The problem is just that the L298N cannot handle the current required for this device.
I can report no diodes failed.
Originally when I tabled the idea of using the L298N, I thought the coils were being pulsed individually.

Regards
L192

[T-1000]

Well actually the diodes on the boards worked just fine. The problem is just that the L298N cannot handle the current required for this device.
I can report no diodes failed.
Originally when I tabled the idea of using the L298N, I thought the coils were being pulsed individually.

Regards
L192

You may try circuit similar to my design to handle load  - https://i.imgur.com/R3H77fv.png (single module for 4 coils in schematics, multiply to needed amount) and offload coils handling to MOSFETs.

Hopefully that is helpful...