Help fixing a power converter

[Boston]

its been a while since I did much with circuit boards and I recently burned this one out



there is no visible damage except for one of the tower looking things that had some thermal grease on it sinking to the body of the device. One of the leads is melted off but nothing in the board itself looks like bad, solder looks pretty good at the connection not that looks can tell me much but still, looks good ;-)



you can see the burned wire on the left side of the right hand heat sink node, ( two big white things sticking up in the middle of the picture with a whole in it at the top )

Its a 240 watt 12 to 24 volt converter. I'm running a 240 watt motor with it and there's a $500 + price difference between this one and the next size up. So here's the plan, I want to eek a few more watts out of this thing, maybe 2 or 3 amps worth, lets call it 300 watts instead of the listed and clearly at its limit 240 watts

Question is
what do I change and is it even possible to eek out a few more watts without altering everything, I'm thinking that what blew was a resistor and all I need to do is throw two bigger ones in and see what shakes loose. I"m hoping a few minor changes and I could probably get this thing up to a reasonable margin of safety so it doesn't burn out again.

What do you say folks, any of our more electrically minded folks have a few jewels of wisdom for my attempt to save a few bucks

cheers
B

[Mark Cat]

Do you have a schematic?

The towering inferno device is a TO-220 package Trans-blister. Can you read the part number on it and/or its neighbor.

Without a schematic, my guess is that the design will not have the margin you suggest, either in power gain. or even if we achieved greater output, a way to manage heat. We can, however, select a comparable power device that will run cooler. If it is a symmetric output stage, we would match the two devices.

Mark

[Boston]

No schematic, sorry, I knew that would be the first question someone would ask. Its an Orion step up DC DC converter if that helps any. Also says Victron energy/blue power on it. I'm not seeing anything written on it but I'm still looking, this time with a jewelers eye piece

what are my chances that piece is the only fault in the system, or does its failure imply other failures

I could easily replace the damaged component now that you have identified it ( thanks by the way ) and just out of curiosity where do you find numbers on those things ? But I'd love to replace it with something that will either run cooler or something that has a higher amp rating. If not, I"m ok with it only delivering 10 amps at 24 volts but it would be nice if it didn't burn out again, the motor apparently draws more than ten amps at 24 volts under load. Maybe something that limits outgoing power to 9.5 amps so its never operating outside its range. I realize the easy way out once its fixed is a simple fusible link but then I'd be changing fuses all the time. I need to limit the draw or limit the available power such that the converter is not being taxed beyond its capacity.

thanks for the help
B

OK got it
MBR2545CT
1017A

the one thats cooked also has a crack in the dark part of the element and a fried wire

going over it with the jewelers glass I'm just not finding anything else that shows any damage, I also found the part on line, so progress. Thanks Mark

[DaveJ]

You could put a higher amp rating device in, but not sure if would be that simple, i'm guessing its not a complex cct with no current/heat limiting built in (otherwise it would not have destroyed itself) so the other components in line will have be able to handle the extra current.

http://www.datasheetcatalog.com/data...BR2545CT.shtml is a page that links to the data sheets for the device. Look like that are all limited to 30amps, see if you can find a schottky rectifier that has a higher current rating.

[Boston]

ok so it looks like the thing that burned out is a 45 volt 30 amp schottky rectifier MBR2545CT
I haven't figured out what the numbers under those mean but I'm working on it
1017A
is still a bit of a mystery but I'm working on it, I'll be accepting hints for at least another few minutes so feel free to dive in any time

I found a 45 volt 40 amp L-series schottky rectifier ( all of 65 cents each ) that could do the trick if it fits in the circuit board but I've yet to dig up a schematic, designation MBR40L45CTG
I'm hoping that by upping the amperage a little I get a beefier unit and by keeping the voltage the same keep the resistance the same.

found both here
http://www.onsemi.com/PowerSolutions/parametrics.do

[Boston]

Hi Dave
looks like we were thinking along the same lines
and posting on top of one another as well

thanks for the info
found that slightly larger rectifier
might work and I've go nothing to loose, things broke already.

I just noticed something
the back of those two components are slightly different from one another
I can only read the back of the undamaged component, the other ones labeling is melted off
could be different manufactures of the same component but could be slightly differing components as well
hmmmmmmm

working on it

[Mark Cat]

Boston,

So this is not the Trans-blister, but a Fry-ttky dual Recti-FIRE.

Can you give me the number off of the plastic controller integrated circuit. Its the little black device with 8 tiny legs or pins. Once I know this IC number I can then recreate the design based on the photograph. According to the part number you provided, the converter topology is such that it is using Schottky Barrier Rectification instead of a synchronous rectifiers for the output stage.

Unfortunately, power dissipation for this class of device from ON SEMI is tied for the most part, to the forward voltage drop and its recovery characteristics per the existing TO-220 package in dissipating heat.

Based on my search for comparable pin outs from ON SEMI, device power dissipation is not readily reduced. Some devices have lower forward voltage but at the expense of other characteristics that the controller IC may find problematic.

Would you pictures of the heatsink and mounting hardware for the TO-220 to the heatsink?

Also, please provide pictures of the enclosure.

Is the location of the converter in a hot environment?

Mark

[Boston]

C=RD
3578
RM

it also has a symbol ahead of the C looks like a sideways Z with flourishes on each leg.

The thing is small, about 3/16 square with eight legs, black, legs are about 1/8 long each but vary in length according to where they attach in the board. looks like a memory chip.

The heat sink is nothing more than the body of the device and it was attached with two clips, not very substantial at all. Some thermal grease was involved but I'd just checked the thing by placing my hand on it just before it quit. Wasn't even remotely hot to the touch. It was mounted in the shade under the motor on the fuel processing tank, top left hand side, you can see the wires going into it but not the actual unit very well. Although the tank and its attachments are painted flat black to help keep the fuel warm enough to process the fuel itself prevents the tank from every getting to hot to the touch. I'd be fine with carving out a spot for a more effective heat sink or just bolting one through to the rectifier, which is what I'd assume the holes in that thing are for, a heat sink.



I'll post picts of the body of the device in a few minutes
thanks again

B

[Boston]

not sure if it was the camera not wanting to focus this morning or if it was me but couldn't get a very good focus on it for some reason, you can see the scorch mark just above the thermal grease. I think its upside down by the way, so the scorch is actually down where the body of the rectifier was below the contact area on the sink. You can just make out the groove where the board slides into is on the top instead of the bottom of the housing.

I'm noticing that this board seems to have some extra spaces on it for additional caps and stuff, could I modify the board with a few new components and increase the units capacity that way, the wound devices seem pretty substantial for such a small load. Could be just a mater of adding another parallel circuit or something.

it might also be kinda fun to Frankenstein something together from the existing design so if you want to map something out that would be great.

whatever will keep the issue from popping up again

[Mark Cat]

Boston,

It looks like the IC is the popular LM3578 used for boost control 12 to 24 volts DC. I have attached an app note on the device. The design uses a power switching transistor to increase the power output.

The TO-220 devices, Schottky and Transistor, must be electrically insulated from the heat sink. This is done using mica+screw insulators with thermal grease or dry pads with screw insulator. In this design they have replaced the screw hardware with spring loaded clamps. The extruded aluminum case heat sink approach is a good choice for this size design. However, sometimes the clamps have problems. Also, an inside clamp like you have is not as reliable as a ridge clamp, in a design where the top of the housing is removable.

It is possible the Schottky failed due to poor heat transfer to the case, due to clamp aging, too little specified clamp force, or assembly error.

Many times, for a difficult assembly (inside clamp), the assembly workers will abuse the clamp so they can put it together. Not realizing that they have just killed the clamp force.

So lets take this a step at a time:

Step 1:

Do not use the existing clamps for the power devices. Remove them and then use Mica + screw mounting kit for two TO-220 + thermal grease (on device and inside of enclosure, use just a slight film of the grease). TO-220 Mounting kits are available from Digikey.

Position the board in the housing and determine where to drill a hole for the screw(s). Remove the board, drill, and deburr holes inside and out, you may have to use a larger drill on the outside of the case to deal with the fins for hardware clearance. When tightening the hardware best to use a mini torque wrench as per the mounting kit instructions.

Replace Schottky with a new identical part.

Put into service.

Step 2:

If the Schottky fails again. Then we should increase the package size of both devices, Transistor and Schottky to TO-247. Use the same device family for both, but larger package (power capability).

Mark

[Boston]

cool
it will take me a few days to get the bits and pieces and then a couple more to find the time to put it together so once I get something done I'll post a picture

thanks
B

[Boston]

Well that was more than a couple of days. What can I say I'm a procrastinator. Been working on the truck and a doing a historic restoration for a really difficult customer most of the summer.

Ok so I'm inclined to just upgrade this thing right off the bat rather than try it again at max capacity. Basically I'm running a 24V 10A motor off a converter with no Watts to spare. I think I'd rather have a few extra electrons in the safety column than none at all.

I am a bit fuzzy on this statement.

If the Schottky fails again. Then we should increase the package size of both devices, Transistor and Schottky to TO-247. Use the same device family for both, but larger package (power capability).

I kinda thought I'd just be replacing both Schottky diodes, or are the diodes and the transistors integrated, cause it kinda looks like they are. IF not I'm going to need a hint, please. ;-) otherwise I'm just making an expensive fire cracker.

?

thanks again for all the help
B

I did find a 45V 40amp TO-247 here, http://actionelectronicswholesale.co...ODE-P9828.aspx I could always buy 50 of them and put the rest up on eeek bay. From these guys for about the same price http://www.onsemi.com/PowerSolutions...do?id=MBR40L45