How to determine fly-back diode properties

I've bought a second hand car, which only had 1 electric window that was still working. It took some effort but they are all working now, with a combination of 2 relays and an ATtiny. The main difficulty was of course to actually connect them to the switches that were already present but i've managed. Now i did of course add fly-back diodes for the relays and i used 1n4148 for that which i am confident will suffice for that, but i actually need a fly-back diode for the actual motor that i am controlling. Now the question is of course, what would suffice for that ? I put another 1n4148 for that, but i somehow doubt that that will suffice, though i may be wrong with that. The part driving the relays is actually fairly well protected, but i am also quite curious as to how to determine what would do. When i measured the resistance of the coil of the motor it was about 4.5 Ohms, i didn't actually measure the current going through, but i figure it should be less than the 2.3 A that should allow for at 12 Volts, but again i may be wrong. Anyway, anyone who can shed some light.

From what I remember... When you break the connection the collapsing magnetic field tries to maintain the same current so a diode rated at 3A (or more) should work.

1 Like

The diode should be rated to carry at least the motor stall current. The voltage rating should be at least the supply voltage. As with most things add some margin for safety, perhaps 25% or more.

Yes, the inductor tried to maintain the current at the moment it is interrupted, so if the maximum current is 5A the diode could be called on to carry 5A.

I would have thought that the motors are reversed by reversing the polarity of the supply, in which case you can't just put the diodes across the motor. Search for H bridge circuits and copy the way the diodes are connected.

Yes they are, when both relays are in rest state, they connect both poles to GND, by turning on one or the other the polarity is inverted. The fly-back diode can be just before the relays, which are a mechanical switch not an electronic one.

How long for ? would 3A peak forward surge for 8.3ms suffice ?

I suggest not. The contacts open and 300V or more appears across them, causing an arc and damaging the contacts. A millisecond or 2 later they close and the current flows through the diode, but then it's too late. You need 4 diodes as per a H bridge.

Edit
If you can guarantee that only one relay at a time will operate then 2 diodes will do, one across each normally closed contact.

So you need four diodes. The neatest way is a bridge rectifier, "-" to ground, "+" to 12 V and the two "~" to the motor leads.

No point rating to the peak specs of the rectifiers though it is probably sufficient, just use the continuous rating and be completely safe. You will however not need it to be heatsinked.

Ok, well that is clear, probably what i'll do.

This is a little less clear to me.

Well i am going to 2x 1N4007 since that is what i have lying around. When i pull the 1n4148 off that is on there now, should i expect it to be damaged ?

This may be interesting for you.

Coil Suppression Can Reduce Relay Life.pdf (48.9 KB)

Nothing new of course. Slowing down the relay operation (specifically, contacts opening) leads to more arcing, especially on DC.