Active pre-charge/inrush control

[Arlo1]

I got it. I used a relay to connect the 12v supply to the controller and the relay is connected strait to the 12v supply to turn it on.
I had 2 blown mosfets but now I got it all sorted. It works awesome! I don't have heat sinks on the mosfets but I turned it on and off a lot of times and they don't get warm at all.
As well there is no sparks now that it works right. I guess this is because the negative from the DC/DC is not connected until the pre-charge is done.

 

[Arlo1]

OK, updated schematic below. With the fuse in this position, if there is a short in the controller or overload, the fuse will blow and prevent the contactor from closing.

View attachment 1

Fetcher you copied the wrong one lol here you go.

 

[fechter]

OK, I went back and deleted the wrong schematic.

Glad it's working OK now. The FETs should not need heat sinks in this application if you're using big TO-220 sized ones.

If you look at the controller volage while you precharge with a scope, you should see a nearly linear voltage rise.

What is the time it takes between closing the switch and the main contactor closing?

 

[Arlo1]

OK, I went back and deleted the wrong schematic.

Glad it's working OK now. The FETs should not need heat sinks in this application if you're using big TO-220 sized ones.

If you look at the controller volage while you precharge with a scope, you should see a nearly linear voltage rise.

What is the time it takes between closing the switch and the main contactor closing?

I used a .47uf cap and it seems to be about .6 seconds from when I turn it on till I hear the contactor.

 

[fechter]

OK, sounds about right. You should be OK with that value. I calculated around 1 second for a 1uF cap depending on pack voltage.

 

[Arlo1]

Ok so I worked my ass off before flying out to my brothers wedding. I got the bike ready for a test ride a few wires need cleaning up but other then that and further controller development its a done bike. I went to turn it on and it would not turn on the contactor...
So I think what happened last time is the caps in the controller were still charged while I unhooked the power supply and the voltage was dropping and I turned the switch back on the contactor made its multiple pulse thing (this is with no power hooked to the battery wires.).
I think this is when the q1 (4110) popped. So rather then rushing I left for the flight and I will have a couple hours till my first ride as a finished bike when I get home.

I wonder if there is a way to prevent this in the future. But for the most part this is a scenario that should never happen.

 

[fechter]

Hmmm... not sure how that could happen. Which one blew?

It would be possible if the zener diode or the diode across the relay coil was bad. You might want to try testing those to make sure they're still OK.

 

[Arlo1]

Hmmm... not sure how that could happen. Which one blew?

It would be possible if the zener diode or the diode across the relay coil was bad. You might want to try testing those to make sure they're still OK.

q1


It happened when I turned the switch back on with no power hooked to the battery wires and the caps we still charged up.

 

[fechter]

With the supply disconnected, there is no current path through the drain-source so about the only thing left is if a spike got past the zener diode and blew the gate. The zener diode is supposed to clamp any spikes before damage. If the zener turns on too slow, I could see how a spike could possibly get through.

To prevent this, two things might help: Replace the zener with a 12v TVS diode, which is designed to turn on very fast. Other approach would be to put a small MLCC cap across the zener. Something like 0.1uF. With a cap, the gate should be protected against very fast transients. A larger cap, like 1.0uF might work OK too, but will slow down the precharge a little. The cap will protect both FET gates. If you use a cap, you probably don't need the TVS diode.

A third approach would be to use a DPDT switch and use the other half to short the gate to source on Q1 when the switch is off. This would totally protect the gate when the switch was off.

 

[Arlo1]

OK I think my Zener is a 17v... I don't think I have a 12v in stock. I will do a Zener inventory when I get to it and see what I can do.
I have a few un packaged zeners I will google what they are and I will most likely put a digikey order together soon.

 

[fechter]

OK I think my Zener is a 17v... I don't think I have a 12v in stock. I will do a Zener inventory when I get to it and see what I can do.

Ah, that's probably the problem. Absolute maximum for the gate is around 20v. The FET will be fully on at 10v, so most anything between 10v and 15v is usually OK. If the zener is too close to the breakdown voltage, then it's possible a short transient would be enough to punch through the gate.

From what I can tell, the turn-on time for even a cheapo zener diode should be pretty fast, like nanoseconds. None of the datasheets seem to have this parameter listed.

It won't hurt to put a 0.1uF across the zener in any case.

 

[Arlo1]

OK I think my Zener is a 17v... I don't think I have a 12v in stock. I will do a Zener inventory when I get to it and see what I can do.

Ah, that's probably the problem. Absolute maximum for the gate is around 20v. The FET will be fully on at 10v, so most anything between 10v and 15v is usually OK. If the zener is too close to the breakdown voltage, then it's possible a short transient would be enough to punch through the gate.

From what I can tell, the turn-on time for even a cheapo zener diode should be pretty fast, like nanoseconds. None of the datasheets seem to have this parameter listed.

It won't hurt to put a 0.1uF across the zener in any case.

The zener is a 1n5247b what about running 2 in parallel and a .1 cap across them. I have a order on its way from digikey but it will be here ~3 Friday and at 8 am I need to get this thing to the university for a open house.

 

[fechter]

The zener is a 1n5247b what about running 2 in parallel and a .1 cap across them. I have a order on its way from digikey but it will be here ~3 Friday and at 8 am I need to get this thing to the university for a open house.

I don't think running two in parallel will do much (but it wouldn't hurt). You could test the ones you have to see if there are any with higher or lower zener voltages.

For that application, you could use an even lower voltage one, like 5.1v. You can usually find those at a Radio Shack store. Two 5.1v in series would be in a perfect range.

The cap will help no matter what. It will increase the precharge time a bit.

 

[Arlo1]

The zener is a 1n5247b what about running 2 in parallel and a .1 cap across them. I have a order on its way from digikey but it will be here ~3 Friday and at 8 am I need to get this thing to the university for a open house.

I don't think running two in parallel will do much (but it wouldn't hurt). You could test the ones you have to see if there are any with higher or lower zener voltages.

For that application, you could use an even lower voltage one, like 5.1v. You can usually find those at a Radio Shack store. Two 5.1v in series would be in a perfect range.

The cap will help no matter what. It will increase the precharge time a bit.

Oh thanks. I never thought of the 2 in series. I have some 5.1v Zeners.

 

[Arlo1]

Ok I was wrong they were 5.6v zeners 5w. I used them anyway 2 in series and left the others there as well and it still failed. I'm not sure what's going on. This all started after I had the one issue when I mentioned turning on the switch after unhooking the power supply. Now I'm trying to use the actual battery to precharge which does not sag like the power supply and it has never worked with the battery maybe a resistor in series with q1?. Otherwise maybe something else is wrecked or I have a wire crossed.

 

[Arlo1]

Think I figured it out. Had a diode from the gate shoted to the cap imput to the drain on q1 and I have the resistor in the wrong spot... lol. It must have worked with the resistor from the cap from the gate on q1 to the - imput lol.

 

[fechter]

Ah, well, that could explain it then. If built to the schematic, it should be pretty bulletproof.

 

[Arlo1]

Ah, well, that could explain it then. If built to the schematic, it should be pretty bulletproof.

Nope I have another problem... Now maybe one of the components are wrong. But when I hook the battery wire up it arcs then voltage on the CA goes to about 70( pack voltage is 84) and then it will drop to ~60v on the CA and fluctuate around 60-65v on the CA when I turn the switch on it does what it should and the contactor engages but when I turn it off it only goes back down to ~60-65v the 12v supply is shutting off so Im kinda stumped. I pulled the batteries out for now and will show the bike as is then I will try to finish this weekend. Tomorrow I have some more zeners some 15v and 12v 1w zeners and a bunch of 1uf caps coming. So I will try rebuilding the whole thing from scratch.

 

[izeman]

I would advise this as well. Build from scratch. The circuit does work very well. Once done correctly no more sparks again. I sometime take my battery inside to do a charge under close observation and then need to plug it to the controller again. No spark, and now the battery stays connected all the time. no drain at all, and starting the bike is a flip of a small switch.

 

[Arlo1]

I would advise this as well. Build from scratch. The circuit does work very well. Once done correctly no more sparks again. I sometime take my battery inside to do a charge under close observation and then need to plug it to the controller again. No spark, and now the battery stays connected all the time. no drain at all, and starting the bike is a flip of a small switch.

Are you using a contactor or just the fets?

 

[izeman]

i'm using the circuit with a switch. layout is somewhere on around page 5??
4x 4110 for 80v and 4kW. they don't even get warm.

 

[Arlo1]

i'm using the circuit with a switch. layout is somewhere on around page 5??
4x 4110 for 80v and 4kW. they don't even get warm.

I'm using the one from the top of this page its a little different

 

[Arlo1]

Fetcher you have D2 as a 1n914 what is this for? I looked and digikey says small signal? Is a 1N4001 going to work there?

 

[fechter]

Fetcher you have D2 as a 1n914 what is this for? I looked and digikey says small signal? Is a 1N4001 going to work there?

Yes, that will work. Almost any diode can go there.
The diode give a .7v drop that ensures the precharge is completed before the relay turns on.

 

[Arlo1]

Ok thanks fetcher. I made a new Proto board and laid it out cleaner then before usually that helps me alone. I should be able to test it tomorrow night or Sunday.