There is a new unit available with added remote voltage sense.
https://de.aliexpress.com/item/1005001458325206.html?spm=a2g0o.productlist.0.0.3822ccf1zniWmj&algo_pvid=219f12ff-32b4-4855-bf41-41d5690f1ee2&algo_expid=219f12ff-32b4-4855-bf41-41d5690f1ee2-7&btsid=0bb0623116012052587916854ed059&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_
I had a similar one and it let the magic smoke out on the third use
I have one and can report mixed feelings about it. The software is useless. The current is noticeably jittery at low values in CC mode. However, it is sometimes handy to have around. The photo shows the external temperature sensor being used to monitor the heatsink of the switching FET of the power supply.
I have one and can report mixed feelings about it. The software is useless. The current is noticeably jittery at low values in CC mode. However, it is sometimes handy to have around. The photo shows the external temperature sensor being used to monitor the heatsink of the switching FET of the power supply.
I have one and can report mixed feelings about it. The software is useless. The current is noticeably jittery at low values in CC mode. However, it is sometimes handy to have around. The photo shows the external temperature sensor being used to monitor the heatsink of the switching FET of the power supply.
Couple of quick questions, if you could:
1) Once it hits the low-voltage-cutoff and turns off.... does it stay off (at least until user intervention)? Earlier versions apparently hit the cutoff, turn off, but then turn back on as soon as the voltage drifts back up above the cutoff (which every battery known to mankind will do)
2) Is there any output on this thing that can be used to switch a relay ? I want a battery tester that'll discharge a battery down to my specified cutoff, then switch in a charger or my choosing. I know this is alot to ask....
Anyone have the DC Load shown in the pic?
SO I bought the version that doesn't come with a heatsink - you must add your own. 1000W, don't ya know. https://www.aliexpress.com/item/1005001345772275.html
1) The included 12v PSU is garbage. It cuts out intermittently, and you will waste an hour or 4. Roundbin it immediately.
2) The FET will promptly explode. The problem outlined by Syonyk (https://syonyk.blogspot.com/2018/06/the-atorch-purple-fan-mosfet-destroyer.html) still exists. I think the problem is an opamp stage driving the fet: it is bode unstable, and the oscillations exceed the ±20v Vgs spec. My fet failed short circuit - which will cause big problems if connected straight to a Lipo battery (it's funny that the brand name is "ATORCH" - this should be taken literally).
Replaced FET with FDH44N50. The specs to look for in this application: 1) lowest possible °C/Watt on Junction-to-case, and, 2) 175°C max junction temp (most big fets are 150).
I put a 16v Zener from gate to source to clamp the oscillations. There's probably an easier/better fix by changing something in the opamps' (there are multiple) feedback loop to improve stability, but I was too lazy.
3) Heatsink: my heatsink box yielded a nice copper sink with blower (not fan), 12v @ .8A, harvested from some old compaq P4. I drilled/tapped a 6-32 thread in the copper, and a screw holds it nice and hard against the fet.
4) Do not use the included white plastic standoffs - they interfere with your heatsink. Use rubber feet.
5) One other problem: there is a TO252 diode on the PCB, with a .7v forward drop. At the ATORCH's rated 20A... that's 14watts into the PCB, which may already be stinking hot from the fet. If you run high power at high amperage... I can foresee melted solder on the TO252.
Anyway, it works pretty good now. Reasonable accuracy (at higher voltages/currents at least; I didn't test a low current. You might have problems if you want to characterize an LR44 or somesuch).
I put 250W through it for a couple of hours, draining a 48V ebike battery, with no issues.
I put 500W through it for maybe 10 seconds, before dialing it back cuz I'm chicken. I wouldn't trust it more than maybe 300W.
Overall, not bad. For the money, it is alot of watts.
I found lower Rds mosfets than the factory one. This would drop the heat on the mosfet considerably
I have an old square CPU heat sink. I think it might work in place of the factory one and reach over to the diode too.
I found lower Rds mosfets than the factory one. This would drop the heat on the mosfet considerably
I have an old square CPU heat sink. I think it might work in place of the factory one and reach over to the diode too.
rdsOn doesn't matter at all in this application - the fet is operating in its linear region (ie. the fet is never fully "on", as in rds"on").
The fet specs you want are:
- lowest possible junction-to-case °C/W
- 175°C operating temperature (most big mosfets are 150°C)
- highest power handling
- VGS of 20v or more (cuz of the dumb ringing on the gate driver amp, that blows up the fet. That's what the zener is for. A better fix would be to figure out how to stabilize the opamp that drives the gate).
The best I could find is FDH44N50
The diode and your fet are likely two different heights, making both sharing the same heatsink problematic.
I think the diode is there strictly for reverse polarity protection. The problem is it is not schottky, so it will smolder at least 0.7W per amp you shove through it. But schottky's generally have poorer reverse-voltage standoff.
The 2 things I see in favor of the IXTH130N20T are the package watts (830w vs 750w) and the amperage (130a vs 40a).
Did I miss anything?
The 2 things I see in favor of the IXTH130N20T are the package watts (830w vs 750w) and the amperage (130a vs 40a).
Did I miss anything?
The 2 things I see in favor of the IXTH130N20T are the package watts (830w vs 750w) and the amperage (130a vs 40a).
Did I miss anything?
You should add thermal resistance to your list as well - the IXTH130N20T is a bit better.
40a is more than the diode will handle, unless you have good heatsinking on it, and/or swap in a schottky, and/or jumper it.
830w vs 750w - I doubt you can get this without chilled water cooling or LN2 or something. IIR the chinglish manual that came with mine (hey, at least it had a manual) spoke of paralleling fets. I'm not sure if this would work (hinges on whether the temperature coefficient is positive or negative, and I can't remember what fets are *EDIT: fets have positive tempco: resistance increases with temperature, so paralleling identical partnumbers should be ok), and there is the mechanical hassle of a bunch of thick wires and fets and big heatsinks all hanging loose.
Anyhoo... my record is 300W sustained. Let us know what you achieve.
I have a couple of the bare-board version coming. Is the original FET a IRFP260N of dubious origin, either reclaimed or counterfeit, as some of the review comments suggest?
Also, as I can't see them from the pics... what are the shunt resistors on this board? Will I need to plan on rebuilding the stack to be able to pull 20A through them without fireworks?
mnem
*taking a punt*