Hypex SMPS400A400 broken

kjb · 2018-09-07 9:38 am

Hi
I have a Hypex smps 400A400 that is dead, If I measure the main outputs, the negative rail is shorted to gnd. I think it become that way when I accidentally shorted the output with a probe when measuring output. I'm not sure tho. Cause i remember it like it worked for a while after that, but I could be wrong. This was I while ago.
I thought the smps where protected against shorts anyway so.

Because I bought it second hand, I get no warranty. Hypex wants €80 + shipping both ways =~ €130 to fix it. More than a new unit costs.

I cant see any damaged components at all. And no small piece of metal junk shorting components can bee seen ether.
I know its a long shot but does anybody know if it can be fixed by replacing some components? Feels like a waist to throw away an almost new smps.

JonSnell Electronic · 2018-09-07 9:48 am

If the output is grounded that is because one diode is short circuit.
On the heat sink, surrounded by capacitors, there are four TO92p diode packs. One or two may be faulty and the power supply will be in trip mode. Remove and replace the faulty diode with EXACTLY the same item.

kjb · 2018-09-07 9:54 am

Wow quick answer! Thanks! Great to know.
Hopefully its possible to see the part-number.
Is it possible to measure them before I desolder or do I have to desolder all before I check them?

JonSnell Electronic · 2018-09-07 10:20 am

There are three pins, the centre is the joint between the two diodes. Yes, with your DVM on diode test, check centre (pin2) to pin1 and then pin3. They should read like a diode one way and open circuit the other. The part number is on the devices. RS component, Mouser and Farnell etc will have them in stock.
Hopefully for a few pennies your power supply will be up and running again.

kjb · 2018-09-07 10:42 am

Thanks! Great info, will check them this weekend.

kjb · 2018-09-07 5:16 pm

The part has this number. Trying to see it without desoldering. Its not easy to see but I think I got it right.

LT 3228
LVP 640p

I cant seem to find it when google it.

MAAC0 · 2018-09-07 5:34 pm

LVP640P 640P LT TO-220 MOS 5PCS/LOT Semiconductor Chip IC CCFL Backlight,LED Backlight Kits,TV Parts,PC Parts,Projector Parts,HIFi Parts LVP640P 640P LT TO-220 MOS

There are some other diodes like these starting with MBR, Could be a IRF640 also.
Are You sure those are on the outputs connected to the elkos. Usually they are larger like these

iCCFL.COM Electronic Parts , Repair Parts.

kjb · 2018-09-07 5:58 pm

Those are the ones closest to the output.

kjb · 2018-09-07 7:21 pm

Great to find a supplier. Tho 1.5 euro turns into ~ 20 euro before i have them. 1.5 + Shipping + vat and the 7 euro handling fee for international orders we have in sweden. Still much cheaper than a new psu.

JonSnell Electronic · 2018-09-07 9:03 pm

Could be the IRF1640G may be the other. If it is short circuit, that will be the one to change. I think L3 supplies the -ve voltage but not sure as I haven't seen one for a while.
You are looking for a short circuit diode pack, there are four components on the one heat sink. A job to see without removing them or the capacitors in the way.
Had a good look but haven't got one to check. Sorry.

MAAC0 · 2018-09-08 12:35 am

Check also the smaller ones between the capacitors.
I was referring these, datasheet MBR10200
but I'm sure there are others with 3 legs, because I have some laying around somewhere.

kjb · 2018-09-08 7:30 pm

As a status update, I ordered the lvp 640p. But it will take a long time before they arrives. Maybe I could prepare and desolder the one/or two that could be faulty.

MAAC0: I can only see 3 pin smd components around the same area. Around the capacitors and the 4 lvp 640p on the heatsink. And they are on the bottom side. No more to220 components besides on the HV section of the board. On the lower left side of the bottom layer.

IVX · 2018-09-12 7:21 am

LVP640P : 18 A, 200 V, 0.18 ohm, N-CHANNEL, Si, POWER, MOSFET, TO-220AB. s: Polarity: N-Channel ; MOSFET Operating Mode: Enhancement ; V(BR)DSS: 200 volts ; rDS(on): 0.1800 ohms ; Package Type: TO-220
So it looks like chinese copy of irf640(I very surprised if somebody still using that).

Cresnet · 2018-09-12 9:05 am

@kjb

Testing those Mosfets on the PCB is not the best way to determine if they are faulty or not.
You must take them out, then test them.

Good luck

IVX · 2018-09-12 9:28 am

Cresnet, no problem to test soldered mosfets yet

kjb · 2018-09-16 9:21 am

I desoldered the suspected mosfet and indeed it is shorted when measure between the pins. Measuring between negative and gnd on the supply shows no short. So I wait for the new to arrive from china and then se if a new one fixes the problem.

IVX · 2018-09-16 11:34 am

usually mosfets dying together with drivers, good idea to check this out.

kjb · 2018-09-16 3:27 pm

Ok! So wich one should i check? I dont know if diode tester works but one of the legs, center left side to top right, of t41 could not be read even if I flipped the probes (pos-neg, neg-pos). But it seems a bit far away from the broken T27. The one that is desoldered. All other seems ok. Gets a value on the diode test.

Cresnet · 2018-09-16 8:23 pm

kjb said:
Ok! So wich one should i check? I dont know if diode tester works but one of the legs, center left side to top right, of t41 could not be read even if I flipped the probes (pos-neg, neg-pos). But it seems a bit far away from the broken T27. The one that is desoldered. All other seems ok. Gets a value on the diode test.

Check transistors T16 - T18, and the diodes surrounding them.

the area where T41 is located has nothing to do with the synchronous rectifier driver stage.

Hope that helps

kjb · 2018-09-16 9:01 pm

Thanks. Checked and all seems fine. No shorts at least.

Hypex SMPS400A400 broken

Member

Member

Member

Member

Member

Member

Member

Member

Attachments

Member

Member

Member

Member

Attachments

Member

Banned

Member

Member

Attachments

Member

Member

Attachments

Banned

Member