My meanwell blew up

My meanwell went bang after charging for a few minutes last week. It has recently had the tppacks current controller added, but that seemed to be working OK.

It was definately stable on 6A current (at 52V) when it was turned on. It's not obvious what went wrong yet - the only obviously dead component is the fuse. What is the correct replacement (it says 4A 220V on the back - was it supplied with a 4A fuse?). The dead one is too blackened to read.

Putting in a replacement 3.15A I have to hand nothing bad happens, but neither do I get any volts on the output. I'll have to do some more investigations.

can you measure the high voltage on the input bridge in front of the inductor?

The original fuse should have it's ratings engraved or stamped into the metal end caps. (assuming a glass tube fuse).

Theoretically, 250V/4A would be the same power rating as 120V/8A, but fuses are typically rated just for the current flow thru them, not power. So a 4A fuse is probably what it takes.

That's 4A *input*, so the 6A output at whatever DC voltage doesn't have any (direct) relation to the 4A input current.

An actual "bang" sound is usually either a cap or a FET, and almost always leaves a lot of visible (and olfactory) evidence behind. Should be easy to find that part of the problem, though finding the root cause might be a little harder.

One common cause for power supply failure (when modded like many MW are) is the power ratings for them get exceeded. 52V at 6A is gonna be about 416W output power, plus whatever gets wasted inside the MW as heat/etc. Unless that's about a 420-450W MW, it's fairly likely something overheated from the excess power dissipation, and that's what blew up. (or shorted internally and then caused something else connected to it to blow up).

I could only find one end cap - the other had gone flying. And that one wasn't legible. The bang in this case could have been the fuse, as it was shattered and blackened. Nothing else shows obvious signs of harm.

This one is a 48V nominal 350W, and at 52V 6A it's running at 312W+losses (where do you get 416W from?) What I suspect is that current regulation went screwy and then indeed something overheated. Fortunately I have a spare so it's not too hard to compare.

Presumably it's OK to run them out of the box so I can get at things so long as no real power is dissipated? (when they are out of the box the FETs/power diode are not heat-sunk). I just need to check they are not getting too hot? I hate fixing switchers - I've had very limited success with it in the past.

when you had limited success in the past, did you measure the voltages in the front end or back end or just replace parts that you guessed were broken? if you can put in a new fuse and see if the current will pass into the bridge then you can remove the ICL from possible failures.

wookey said:

I could only find one end cap - the other had gone flying. And that one wasn't legible. The bang in this case could have been the fuse, as it was shattered and blackened. Nothing else shows obvious signs of harm.

Click to expand...

That's gonna be some serious overload to cause a fuse to explode. Iv'e seen them blackened but even in the worst case nothing more than a small crack in the center of the glass (and that was with a lighting strike at the office the equipment was at).

To make it explode (with sufficient force to throw and endcap out of the power supply case, presumably thru a vent hole unless you run it without a cover), the fuse strip or wire inside would have to be heated so hot so fast that it turns into superheated high-pressure gas instantly, rather than melting somewhere. I am not sure what kind of overload could happen to cause that that doesn't involve either power surge from outside or other downstream components.

Reagrding the end cap that's missing, is it possible that it or a piece of it is underneath a component somewhere, shorting it out?

This one is a 48V nominal 350W, and at 52V 6A it's running at 312W+losses (where do you get 416W from?)

Click to expand...

wrong button on the calculator I guess. :lol: Sorry!

Fortunately due to a simple topology, they usually roast the high-side push-pull transistors (at the end of the case) and usually take out the bridge rectifier as well. Check the high wattage resistors driving the transistors as well, they go open circuit sometimes too when this happens. I've fixed a few with the same problems - the transistors are as common as muck, they are 2625's (IIRC) and are usually found in computer power supplies (as is a suitable bridge rectifier).

Also - these came out with a batch of common-mode chokes that were dodgy (one lead isn't soldered well) - check this too.

Cheers for the clues. A new fuse just goes pop instantly so somethig is definately poorly. This schematic from http://endless-sphere.com/forums/viewtopic.php?f=14&t=20118 is really handy: http://www.hordsoffun.com/ebike/pics/Meanwell_S-350-27.png

Bridge seems OK, and resistors, but both big transistors have gone short-circuit. The parts are 2SC3320, which remarkably digi-key don't stock (nor RS, CPC, Mouser or Farnell). In fact they don't seem to have any 15A, 400V, TO-3P-3 parts in stock. There is a fairchild KSC2751RTU which looks about right (although Dc gain is twice as much as the 3320 - which I assume doesn't mater in this application).

RS have the STMicro 2STC5242 which is 230V 15A. I don't think 230V is enough headroom is it? Is the DC voltage higher or lower than the AC input after rectification? I can't remember.

Voltage after rectification is higher. On either voltage (240 or 120) - you will see 1.414 X V = 340/2 (170V) per capacitor (and each transistor in the push-pull pair) on 120V its essentially running as a voltage doubler (this is what the 220/110V switch is for - and why you DON'T put it on 110 for 220V!!!)

Also - bridge rectifier on these are weird. They test OK, but put mains voltages on them and they short. I had one OK at 12V but went shorted at 36V - very very weird. Replace to make sure, or you may need to replace your underwear when the fuse exits stage left again. You don't want a brownout event!

OK, so can someone tell me what a suitable replacement part is? (or where to get the 2SC3320).
2SC3320 datasheet: http://www.svntc.com/TPDF/2333.pdf

Digikey have the KSC2751RTU http://media.digikey.com/pdf/Data%20Sheets/Fairchild%20PDFs/KSC2751pdf.pdf
which is almost identical, but also obsolete.

I can get them from ebay for 6.20 euro each, which seems extortionate.

RS have the BUH150 which seems similar: http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=5449494
but it's in a TO-220 case. Pin ordering is the same, so it can probably be bodged OK.

Aha. STW204 is a TO-247 (very close in size to TO-3PN package of 2SC3320 with matching pinout and spacing).
500V, 20A, 125W. And actually available at GBP 1.90 each:
ttp://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=497-8794-5-ND
Higher hFE at 20, rather than 10, and faster switching. I reckon that'll be OK. Anyone disagree?

Yep should be OK - I found a few of these in use in older ATX power supplies (non-P4) I think they are an older make.

maybe i missed something, but i did not see any pictures. how do you know that these transistors are dead? what function do they perform and did you look on ebay for replacement transistors? they can be really cheap especially when shipping is included, plus you can buy the 10 pack for the price of one from digikey usually.

I think they are dead because they measure short circuit in both directions between the pins (with the diode-test mode of a meter). (actually of the 4 possible diode junctions you should see in 2 transistors, 1 of them still looks like a diode to my meter, so only 3 of them actualy melted

Function: They are the main power transistors in the meanwell switcher.

And yes, I looked on ebay which, as quoted below, did have them, but only at great expense (GBP 5 each, plus £2 delivery).

Digikey don't have them listed at all, in 1s, 10s or 1000s, and neither do any of Farnell, Mouser, RS, or CPC

I've ordered the STW204's now so we'll see how that goes.

these are the transistors on the far right side of kingfishes diagram? or on the left side? he has them labeled so that would help locate them for me in the circuit. i have never had to figure out if a transistor is dead, but that makes sense. usually they are smoked so it is a nobrainer and i can manage nobrainers.

when i used ebay to look for transistors i found a lot chinese ebay shops which had pages of transistors offered, and shipping from china is cheaper than digikey for sure. i never buy from digikey myself because they charge so much more for parts and shipping, i use either mouser or newark which is farnell in britain.

OK, parts arrived and I got round to doing some soldering. Replaced both mosfets with the STW204's I found at digikey eventually, and it worked fine. It's run at 340W for a hour or so with no trouble, and no excessive heat, so looks like those are a perfectly acceptable replacement. The bridge and resistors both seem to have survived.

Hey, is this power supply still running?

Also is it a genuine meanwell or a knockoff?

This is an old thread but I have about 10 MW knock off from egay that are blown. I just pulled the transistors out of the back and they are all blown. I would like to find a upgrade to replace them with.

hello,

i have got a blown meanwell clone s350-60. replaced both 2sc3320 transistors and also original bridge rectifier with Diotec KBU12K. rectifier was also blown.
now the power supply seem to work properly, i can regulate voltage normally with poti. when i try to charge my 57 volt battery block i only got about 0,5 a current. checked almost all resistors and diodes. everything seems to be ok. has anyone an idea what can cause this problem? r33 is original with about 380 ohms.

thanks in advance, harald.