replacing FET's with contactor

[flippy]

i need some advice on how to do this.

my BMS is officially rated for 50A continuous (100 peak) but i blew the ass out of 8 of the 20 FET's after just 300km of driving with my controller at 60A peak (15A cruising).

you can see part of the blown asses here:

now i want to replace the FET part (that got mighty hot anyway) with a regular "proper" contactor.

in order to keep driving i have bypassed the FET part so the BMS is reduced to cell balancing duty.
but that means i lost my overvoltage protection during charging.

how do i get the control signal of the controller's FET's to power a decent sized contactor? the actual signal going to the FET's is obviously not powerful enough for this task.

on a related note: where do you get decent contactor's that can handle 100A without melting?

 

[dmwahl]

Gigavac makes some nice contactors and they'll do credit card orders for individuals if you call them. I've used the P105 (or P115) which is rated for 50A continuous and will do 100A for a few minutes without overheating. I paid around $50 each for the P105 contactors. That said, I'd suggest you replace the FETs with new ones first. There's no telling if the ones on there are real, or if they were damaged already and that's why they blew up. They should easily be good for your current levels assuming you're respecting the voltage limit.

If you do use a contactor, you'll need to devise a way to shut it off when you're not riding, as it will draw a few watts constantly when on. You'll also need to protect whatever is driving the contactor from the voltage spike when it opens. None of that is particularly complicated, but will require some engineering work.

http://www.gigavac.com/catalog/power-products/minitactor
http://www.gigavac.com/sites/default/files/catalog/spec_sheet/p105.pdf

 

[Punx0r]

Slightly ironically, what you probably want for driving a contactor is a FET... You might even achieve this by removing the 8 blown FETs on your BMS and see if the other 2 still work.

 

[flippy]

i am well within the voltage rating.
this is the BMS in question:

https://bmsbattery.com/bmspcm/713-10s-13s-30a-50a-lipo-battery-bms-system-bms-pcm.html

i prefer to replace them all with a contactor (how would i hook that thing up)? or a need advice on what fets i can replace them all with that are highly overrated for this task. i do not want to burn out this thing again.

i am also thinking about replacing it entirely with a better BMS.

any recommendations?

 

[Punx0r]

The simplest, smallest and most effective solution would be to replace it with a better BMS. The FET gate drivers probably won't be sufficient to switch a large contactor themselves.

 

[flippy]

The simplest, smallest and most effective solution would be to replace it with a better BMS. The FET gate drivers probably won't be sufficient to switch a large contactor themselves.

yes, that is my issue. i can probably boost the gate signal with another FET or transistor but then i would need a separate 12v line to power the contactor (or get a 48v one) wich complicates matters a bit more.


what "better" bms would be advised? i personally hate to use FET's for inductive loads as they burn up a shitload of power with loads like this. a contactor does not have this issue.

 

[parabellum]

I believe, what Punx0r means, 1 of the fets (which is still working) can switch your relay, from main pack voltage. Fet will only need to handle relay coil current, few dozens mA usually, being kind of driver

 

[flippy]

yes, i can try that.

still need to find a decent contactor. i hope i can find a supplier in europe. postage and taxes from america triples the price...

 

[Punx0r]

Sorry, I've insufficient experience with different BMS to recommend one.

What system voltage are you running? The common little automotive relays are available up to 80A (12V coil, 30V contact rating) and cheap. You might be able to push this to 48V or perhaps run two in series. Your load at power on (relay make) is basically just charging the controller caps. Breaking under load (fault condition or over-discharge) is probably what will test it. Normally running shouldn't be a problem, as the already closed contacts are just carrying current.

For what it's worth, the motor is an inductive load, but I don't believe the battery side of the system sees this - it's just keeping the controller caps topped up.

 

[friendly1uk]

Hold out for some fet advice. You want low resistance ones (rds) so not much heat is generated.
That board design looks like the fitted fets could of done double with adequate cooling. Some thermally conductive epoxy could hold a heatsink on.

Why did it blow though. It sound like you were barely using it. Were the bits fake or have you smothered it so it cooked?

New fets and perhaps some cooling method. Anything. I big metal plate over the lot would help, but it's better to draw the heat out via the metalwork.

 

[flippy]

The fets WERE cooled with a heatsink. You can see this on the link to the site given above.
But the sink is so poorly made it only contacts the metal lip, not the black epoxy of the fet.

Ps: automotive relay's will melt in this application. They are not rated for continuous high amps and high volts. The sheer wattage will melt them.

 

[friendly1uk]

I see it now, yes.
Looking from here, it's one side that cooked the most, giving clues towards a thermal issue. I imagine they didn't use any paste and them tabs don't look flat from here.

 

[Punx0r]

Ps: automotive relay's will melt in this application. They are not rated for continuous high amps and high volts. The sheer wattage will melt them.

The voltage is irrelevant to the carrying capacity of a relay - only the current. I just checked the datasheet for an 80A relay and it's actually "only" 70A continuously rated. They're not ideal, but they're also $5 parts rather than $150.

 

[flippy]

i contacted bmsbattery twice so far and i am getting the usual silent treatment.

i need to mod this bms so i can use it again. i found some FETs with much lower RDSon: IRFB7730
the original IRFB3307 has a RDS on of 5~6.3 wich is bastically sub-par and also the partial cause of the heat buildup.
the IRFB7730 has a RDSon of 2.2~2.6 wich is clearly much better and has higher specs overall.

i can order 20 of them on mouser for 45 euro's so price wise it is a better choice then getting a contactor and supporting stuff and it should be fairly impossible to break.

amp rating of a relay is not the issue. running high voltage over a low voltage contact that has not been treated to deal with the sparks (12v contacts are almost never coated) will cause the relay to stick/melt the contacts. making it impossible to release.

and running a 80 amp relay at 40 amps continous for example and them give it some city traffic regen and accelerations will make short work of a cheap automotive relay.

 

[parabellum]

You could take motorcycle start relay, feeding lower coil voltage, for continuous duty. Good for several 100s A continuous. This does not solve arking at higher V tough.
If your BMS uses Fet's as shunt, then with irfb 7730 it will think your current is ~3 times lower then it actually is and act accordingly. Dunono, maybe a good thing.

 

[flippy]

It has dedicated shunt wires.

Even so: i limit current on the controller and the pack itself is capable of 90A continous and 250A peak within spec of the cells. 13S30P made with samsung INR18650-29E is quite a capable little battery. Unregulated it would be capable of 1kA at least before burning out the cell interconnects.

In short: i dont give a flying fudge about the current limit cutoff of the bms. I need it to keep the cell voltages in a safe range. I use a 100A dc mcb to protect the bike from burning up.

 

[friendly1uk]

Half the rds has to be a good thing.

There have been threads in the past regarding spark arrest and pre-charge revolving around the fact people connect and disconnect the battery from the controller a lot and it erodes the connectors. They offer a lot of fet choices to hold a lot of power on these threads. I think your using the same part numbers, so a search might find more threads with alternatives.

 

[flippy]

Yes, i have been trying to find some more info here. But information is all over the place and not very coherent or complete.

Actually comparing them is a pain to say the least. I found the 7730 by filtering on the international rectifier site by sheer luck.

Ps: love your subscript about bmsbattery.... I do know the feeling as well.

 

[DVDRW]

yes, i can try that.
still need to find a decent contactor. i hope i can find a supplier in europe. postage and taxes from america triples the price...

http://www.ev-power.eu/DC-Contactors/
Latching contactors are interesting tough.
zero power draw wheh activated.

 

[friendly1uk]

A search for your part number produced

Also now have some IRFB3077 equipped Infineons (even lower resistance than the IRFB4110) as well as the IRFB4110 and some lower cost IRFB3307 Infineons

The 3077 has an rds of 3.3 and the often talked of 4110 has 4.5 with some 3.7 variants. The 4110 is 100v and the 3077 75v like your 3307 with it's 6.3

Is see talk of the 3077 being replaced with the 7730 as an upgrade here https://endless-sphere.com/forums/viewtopic.php?f=30&t=73907&p=1114934&hilit=IRFB7730&sid=776bd3f8e9d5ada664bb1fc6a482c352#p1114934

It all looks good to me, though I have not looked at pin-outs or the drive they might require. Is it a bestec? he may advise when he see's you blew it.

 

[flippy]

Is it a bestec? he may advise when he see's you blew it.

no, it's a bmsbattery. if i look at your subscript i think you already know them.

i think i am going to order a bunch of 7730's.

right now i reduced the output of the BMS to charge and balance duty. right now a put a line directly from the battery terminal to the controller with only a 100A MCB between it.

when i get the new fets and ran some tests i will remove the MCB and put the BMS back into full operation.

 

[friendly1uk]

Oh yes, I know bms battery. Like a bad woman I keep being drawn back for more, because they are cheap. Than realising it's not what I really wanted.

Good luck with your project. I think you have made the right decision. Mechanical switching would be large and risk welding. Then you expect everything is fine because the bms will protect you, and mess up your pack.

This will be a lot better if they are direct replacements