EV Global charger smoked, MOSFET replacement questions

I'm using original EV Global charger to charge the 12s4p LiPo pack that I stuffed into a factory box. Turned it up to 50.2v and 3A (by adjusting VR1 VR2) - was running fine for a while, then I removed the cover (which hosted a fan) and turned the VR2 all they way up - got almost 3.3A out of it. The FET did not last long - puffed the magic smoke - not surprised, as it is rated for 50w dissipation (surprised it lasted this long). The FET sits on a L-shape heatsink.

Questions:

Originally the MOSFET is a K2645 N-Channel MOSFET, rated for 600v 10A, 50W dissipation. Can I stuff something more powerful in place of it (I was thinking NTE2970 - 500v 22A, 278W power dissipation, or Irfp240 Power Mosfet N Channel -200v 30a 150W)?

What should be my main points of comparison?

Obviously I'm more concerned with amp rating than max voltage (being a ~50v charger application), should I even bother finding a >100v equiv considering, is there more to it than just comparing max values and physical dimensions?

not sure what you tweaked since there is no picture of the charger even. what is the IC current controller chip and what is the spec for the gate drive output of the chip? if you have the 3842 then there is a large power resistor on the source leg that is used for the current feedback detection and you can add something in parallel to it in order to increase switched current levels.

Sorry, this is a pic of a charger from the other thread. You must be talking about the pink 1/4w resistor. If I adjust the current handling levels, don't I need to replace the FET still to match the power levels?

This unit charges three 12V 12AH SLA batteries.

This charger seems to be a be a two stage charger.

VR1 adjusts the maximum charge rate, and should be set to 3 amps when the
charger red LED is on.

VR2 seems to set the level at which the full current charge ends, the red LED
changes to green, and the fan turns off. and the voltage then remains level at
the voltage determined by VR3.

....

I assume that I should limit the total charge time to half a day or so to avoid
overcharging, since this unit does not seem to drop the voltage for a float
charge.

Click to expand...

Is blue disk a polyfuze, if so maybe it is not the FET that released the smoke? How do I test?
The chip is:
Lm324n
w9965002

another one: sd3843a

My original question still stands, which values should I be comparing MOSFETs by for this application?

what is a polyfuse and where is it located in the circuit?

lm324 is an opamp, still no info about your IC current controller.

You should use a charger for lithium, that's a lead charger.

You don't mention using a bms, just lipo. Gonna burn the house down.

DOGMAN,

I respect you for your valuable input on the topic of EVG and such, but it is a CC/CV charger with three adjustable ranges, coming in a neat "always there" package, how is it not suitable for lithium?

because the lifepo4 batteries have to have the balance maintained as they age and the internal resistance and capacity change with age and abuse.

this requires using a BMS to allocate the needed charge to the cells that need it and that has to be done by the balancing charge at final voltage.

so that charger is useless to maintain a lifepo4 battery pack with a BMS that balances the pack.

Understood, and I'm balancing them once a month (used to be once a week). The cells are lipo.
We digressed, this little charger can run a 10s pack (lowest settings) or a 12s (highest), got me going many many times, really want to fix it, any ideas on the MOSFET replacement strategy?

The charger needs to go to 0 amps at the end of charge, not a trickle charge like lead uses. That's why it can overcharge your lico.

This is pretty basic, Its NOT SAFE to use a lead charger to bulk charge RC lipo packs. Sure, you can get away with unsafe practices for a while. But you won't see me tell you bulk charge RC lico with a lead charger.

I don't care one bit if you burn your bike, just don't park it where you will burn the house. That can still happen even when the bike is not being charged.

You have the packs in the factory battery box, which means it's very easy to remove the box. Then just get a decent charger for lithium if you want to bulk charge. I recommend the three voltage choices one from EM3ev. Safest is to use the 4.1v setting.

But if you had lifepo4, with a bms, you can get away with using the lead charger. But it will likely undercharge, and the bms will not do much to balance the pack. It can work fine, with a newer pack.

So I replaced the MOSFET with the FCPF22N60 and it does not appear to turn on. The charger turns on, but does not push any current. I've compared the Vgs and the values seem to interlace (not identical though). This is the datasheet for the original one, btw.

The circuit of the charger is very similar to the below schematic with some minor differences. Can somebody explain the "two transformer approach"? Also if I wanted to drive a PC fan (variable speed with hall sensor) where would I tap in?

One thing that I've noticed, the new MOSFET "bled off" the static charge from the gate fairly quickly, performing this test:

TEST #4 Check Drain-Source junction for proper conduction.
make shure that you have not touched the Gate lead since you turned the device "on". probe with the RED (+) on the Drain and the BLACK (-) on the Source. the expected reading is a short circuit. then probe with the polarity reversed with the RED (+) on the Source and the BLACK (-) on the Drain. this should also show a short circuit. if either of these shows an open circuit than the Drain-Source is blown open and the device needs to be replaced.

on most devices, once the Gate is charged, it will remain charges for several minutes. long enough to do the tests. you just have to be careful not to disturb the gate lead and accidentally discharge it.

Click to expand...

The DMM just flashed a value indicating there was a short circuit, then immediately showed an open circuit, without me removing the static charge.
Is this characteristic of a faulty batch (I tested several of them, all do the same) or just the feature of a device?