Converter In-Rush Thermistors etc, UPDATE -4

BFL13 wrote:

ken white wrote:
............


If we look at the worse case current for the 5 ohm thermister, it would allow approximately 170 volts peak divided by 5 ohms or 34 amps of initial current to flow.


If we look at the worse case current for the 2 ohm thermister, it would allow approximately 170 volts peak divided by 2 ohms or 85 amps of initial current to flow.
.

I know there was the single cycle value being different from actual current (but can't remember in what way ) but how about that 34a limit for the 5R compared with this spec for the converter at 40a?

•MAX INRUSH CURRENT, SINGLE CYCLE: 40 AMPS

Could the 40a translated into real amps be too much for the 34a 5R and ok for the 85a 2R?

The single cycle (or multiple cycle for that matter) inrush current spec is at least partially determined by the thermistor characteristics. (My guess is signficantly affected by the choice of thermistor, but then it's only a guess for whatever that is worth)

So with the factory OEM thermistor they spec'd a single cycle inrush current.

Now put in a different thermistor, and without measuring, or having a set of schematics handy to analyze, you wouldn't know what the single cycle inrush current actually is.

And I think in your use, you were ending up with the worst possible hot restart conditions. The current needed during a hot restart is determined by several things, a dominant one is how fast the input caps discharge. And the higher the output load on turnoff, the faster the caps discharge. With a open circuit output, it could take considerable time to discharge the input caps, and a hot restart may not result in high inrush current.

But with a 100A output demand from a dead battery when you turn off the converter, the input caps are going to discharge very very fast. Probably fully discharged faster than you could turn the switch on again. And because your were drawing max input current, the thermistor is very hot. So, it now takes about 3 minutes for the thermistor to cool down to where it's resistance is 1/2 of the cold resistance. (From the data sheet on thermal recovery time constant)

So.... turning it back on in short order is really subjecting everything to very high current peaks. And it is NOT doing anything good to the power switch contact either! the high inrush current on contact closure, especially if there is any switch contact bounce (almost guaranteed), can really degrade switch life with contact arcing. Either ending up with a switch with high contact resistance, or occasionally even welding the contacts closed.

BTW, at the company I worked for, the common practice on power supplies, knowing customers would turn them off and back on again in short order, was to add a nice fix for this problem. A common approach was this. (There are other solutions as well.) In parallel with the thermistor was a relay, either mechanical or solid state. Relay was driven by the power supply with a very short time delay. A few cycles of 60Hz, up to maybe 30 cycles. After that time delay the relay closed, thermistor shorted out, everything works fine, and the thermistor never had chance to really heat up much. So if you hit the power switch quickly, the cycle would repeat, but on startup the thermistor was still at or closes to it's cold resistance, limited inrush current, then out of the circuit again.