MPc 4.1 2009, PSU Blows fuses in house

[goodfidelity]

I have one 2009 4.1 MPc.

UPDATED:

Machine keeps blowing fuses upon plugging it into socket for power. Once one fuse is blown (or fuse is blown one time) the machine runs just fine.

I now did have another look at the problem, and seems that the fuse will blow when cable is put in FIRST time. Fuse will blow no matter what socket i use, or what end of the cable is plugged first. That makes me understand this is not a mechanical problem, problem is not caused by temporary short upon forcing the cable in place. If the cable is unplugged for a few minutes the problem can be repeated. If unplugged for less then one minute problem will not repeat. This makes me think it has something to do with charge in capacitors.

First idea was mechanical problem or simple short to ground or something or that nature, from abuse or damage from physical force when cable has been connected. But this does not seem to be problem after looking into it more close.

Since the fuse will only blow upon first connection, and the PSU will run fine after that, i am thinking that it makes a temporary overload when charging the capacitors or something like that. Once they have been charged, it will no longer draw that current that makes the fuse blow. The machine can stay plugged for long time and not blowing fuses even if pressed to the max with stresstest or GPU stress.

Capacitors all have flat heads, i don't see any sign of damaged capacitor or otherwise burned components on the board.

After unplugging cable, it's about 5-10 sec then i hear a CLICK sound. That indicates a relay turning back off i would say.

Is there a possibility that there is something wrong with the relay? Is it common that the relay causes this kind of problems?

Using thermography camera it becomes clear some of the components on the PCB gets hotter than the others, even when PSU is not mounted into the machine. Temperature is about 40 degree C, that is about 20 degrees above room temperature. This is when PSU is idle, sitting on bench, power cable plugged in on AC side but no load on the DC side.

Area that becomes hot is around the heatsink in central part of picture. I am not sure if the heatsink is getting hot from the component attached to it, or if it is from the resistor R403 sitting next to it. (camera has to low resolution to show exact what component is hot). On the image it looks like more parts are hot, but that is just thermal reflection from the shiny large heatsinks.

Anyone with insight on how these PSU work have any good ideas about what part is failing?







OLD POST:



Something makes me think previous owner has damaged the PSU by appying force to cable while attached to machine, and that might have damaged something inside. (connector / plug is bit loose but not much)

Sometimes when i plug the cable, the fuse will blow for this part of the house.

Somethimes when this machine is left plugged in it will blow the fuse for this part of the house.

Fuse is blown direct, even if computer is not turned on.

After fuse is blown, i can easily just put new fuse and run the machine.

For me this sounds like PSU problem.

Anyone has experience with this, or has any pointer as from where i should start looking? Is there any common problems that i can address? Like faulty parts that can be swapped inside the PSU or anything like that?

Problem is replicated in different houses.

Thanks!

 

[absente]

Did you try different sockets in different parts of the house? If you can replicate the problem there, then it's a PSU problem

 

[goodfidelity]

Did you try different sockets in different parts of the house? If you can replicate the problem there, then it's a PSU problem

Yes, i can replicate the problem in different places. Also different houses.

I am pretty sure it is PSU.

I have now taken out the PSU and opened it up, to try to see the problem. Only think that i can see visible is metal parts in the plug has visual wear from cable. And i can also see a crack in the plastic cover around one of the main leads inside the PSU. (cable that runs from plug input socket down to board has a crack in the rubber/plastic shielding around it, but no visual burns from that area of the cable).

 

[absente]

Yes, i can replicate the problem in different places. Also different houses.

I am pretty sure it is PSU.

I have now taken out the PSU and opened it up, to try to see the problem. Only think that i can see visible is metal parts in the plug has visual wear from cable. And i can also see a crack in the plastic cover around one of the main leads inside the PSU. (cable that runs from plug input socket down to board has a crack in the rubber/plastic shielding around it, but no visual burns from that area of the cable).

Safe to say that the PSU is the problem, but opening it up is very dangerous - the capacitors and other parts hold charges that if touched will kill you. This is no joke. If you know what you're doing okay, but personally I would just get a PSU replacement and call it a day.

 

[Slash-2CPU]

I'd replace the PSU, now. Good way to set your house on fire is to keep running that one. If it's blowing fuses, then it is overloading the wiring in your walls when that fuse blows. Think about it.

 

[pl1984]

This is one of those situation where I would say: If you have to ask then better to replace than try and repair. As others have mentioned this part can have severe consequences if one does not know what they're doing. Better to be safe on this one.

 

[goodfidelity]

I dont agree with previous posters. I know how to repair things.
This thread is oriented towards repair of PSU not exchange of PSU.

Electronics can be dangerous, and so can driving a car. I put myself into more danger each day driving to job then repairing a PSU. This thread is not about risk, its about how to repair a faulty part or what to look for.

As most problems with the Mac Pro, someone has doe it before and knows how to fix it. Its usually a matter of time before someone with the proper knowlegde replies.

This problem could be as simple as a bad relay or common connector problem, maybe a short between main and ground. I just want to know what to look for, and if anyone has been through this before.

thanks

 

[hfg]

Although this does sound like a power supply issue ... it might be worth it to try a different AC cable just in case there is an internal cable short when it is bent a certain way.

 

[goodfidelity]

Although this does sound like a power supply issue ... it might be worth it to try a different AC cable just in case there is an internal cable short when it is bent a certain way.

That is interesting, i will double check that.

However i can see from the pins in the PSU, that i looks like there is some damage from sparks on two of the three pins. I cant see if this is normal wear or from jumping electricity.

Ill measure up the cord with multimeter and check the connector.

Thanks!

 

[Slash-2CPU]

This is all assuming you can find the individual PSU part to repair the old PSU. You’re very likely to spend more time and money trying to fix the old PSU than purchasing a good, working used PSU.

 

[goodfidelity]

This is all assuming you can find the individual PSU part to repair the old PSU. You’re very likely to spend more time and money trying to fix the old PSU than purchasing a good, working used PSU.

Yeah, you must be right.

Instead of posting threads on the internet asking questions how to fix things, and getting good ideas from people who actually knows something i should just drop it straight and don't even try. So when a GPU burns out i should just buy a new, not try to bake it in the oven like i have successfully done with 4 or so GPUs. and this i learned from the internet, and posting on forums. Instead i should have bought new, paid more and waited days for the delivery when i have had the problem sorted out in less then 2 hours myself for free.

And when the processorboard of the MPc failed from cracked rivets on the north bridge i should not bother asking people on the forums if they have had the same problem and try to fix it, instead i should just order new CPU board straight away and not bother asking people who know a thing or two. Well, wrong again because i did google it and ask questions and fixed it fast and cheap at in a matter of less then one hour and at the price of a bag of peanuts.

When the RAM failed, i should have just bought new. Not bother trying to find out what was wrong or fix it, and not asking on the forums what could be wrong. Well wrong because getting the advice to clean the connectors and re-seat the ram saved me buying new ram and was way faster then ordering new modules.

When the mac-pro suddenly stopped booting and i suspected the EFI chip i should just have bought a new mainboard and not try to fix the old one. Well, wrong again. After alot of tinkering the computer came alive again, and there was no explanation to why. Maybe a bad connector somewhere. I should have gone with your advice rather then asking around the forum right? That would have saved me what?

And now i have a machine with a broken PSU. Should i follow your advice or should i do what i always did?
Well. Ill try to ask if someone who actually has the knowledge, and has seen this problem before, knows how to fix it. There is a chance that someone has seen the exact same problem before me and knows that this is a blown fuse, broken rectifier, bad relay or defect grounding/isolation from some cable. I dont know, that is why i ask. Chances are that this problem is just as simple to fix as all the other problems i hav had with other machines.

Instead of using your time, and my time, and posts on the board arguing and trying to make people buy new things before they try to fix it you could try to steer away from posting in threads when you cannot contribute in a positive direction. Your posts might actually waste someones money, and you are for sure wasting other peoples time posting them, not to mention yours. Avoiding those posts would make the board/forum a more resourceful place and save us alot of time.

I think i proved my point here. No harm intended. Now lets solve the problem in the PSU if its possible.

Thanks.

 

[Slash-2CPU]

Yeah, you must be right.

Instead of posting threads on the internet asking questions how to fix things, and getting good ideas from people who actually knows something i should just drop it straight and don't even try. So when a GPU burns out i should just buy a new, not try to bake it in the oven like i have successfully done with 4 or so GPUs. and this i learned from the internet, and posting on forums. Instead i should have bought new, payed more and waited days for the delivery when i have had the problem sorted out in less then 2 hours myself for free.

And when the processorboard of the MPc failed from cracked rivets on the north bridge i should not bother asking people on the forums if they have had the same problem and try to fix it, instead i should just order new CPU board straight away and not bother asking people who know a thing or two. Well, wrong again because i did google it and ask questions and fixed it fast and cheap at in a matter of less then one hour and at the price of a bag of peanuts.

When the RAM failed, i should have just bought new. Not bother trying to find out what was wrong or fix it, and not asking on the forums what could be wrong. Well wrong because getting the advice to clean the connectors and re-seat the ram saved me buying new ram and was way faster then ordering new modules.

When the mac-pro suddenly stopped booting and i suspected the EFI chip i should just have bought a new mainboard and not try to fix the old one. Well, wrong again. After alot of tinkering the computer came alive again, and there was no explanation to why. Maybe a bad connector somewhere. I should have gone with your advice rather then asking around the forum right? That would have saved me what?

And now i have a machine with a broken PSU. Should i follow your advice or should i do what i always did?
Well. Ill try to ask if someone who actually has the knowledge, and has seen this problem before, knows how to fix it. There is a chance that someone has seen the exact same problem before me and knows that this is a blown fuse, broken rectifier, bad relay or defect grounding/isolation from some cable. I dont know, that is why i ask. Chances are that this problem is just as simple to fix as all the other problems i hav had with other machines.

Instead of using your time, and my time, and posts on the board arguing and trying to make people buy new things before they try to fix it you could try to steer away from posting in threads when you cannot contribute in a positive direction. Your posts might actually waste someones money, and you are for sure wasting other peoples time posting them, not to mention yours. Avoiding those posts would make the board/forum a more resourceful place and save us alot of time.

I think i proved my point here that I’m a complete asshat and an inter web tough guy. No harm intended. Now lets solve the problem in the PSU if its possible.

Thanks.

Here’s how you fix it for free. Remove the PSU and disassemble it. Desolder the power cord socket. Cut the end off the power cord and strip the wires back. Solder the wires into the pin holes where the socket was. If it all works ok, then you know it was the socket. Be careful to check with a meter that removing the socket didn’t interfere with case ground. You don’t want to grab a hot chassis.

You can then acquire a dead PSU and swap the socket with yours.

It's in fact a terrible and dangerous idea to open a PSU. You can easily end up shocking and/or killing anyone who touches the case of the Mac Pro if you mis-wire or improperly reassemble the PSU. If you think the risk is worth the $50 or so you'd save, then go ahead and give Mr Darwin his chance to prove his law.

 

[Slash-2CPU]

Now that you say it only blows fuses on initial plug in, that's a completely different story. Bad main filter capacitor and/or bad rectifier. 100% guaranteed. Rectifier can be tested by desoldering and testing for continuity in the wrong direction. Swap it out and you'll be like-new.

Still, please be aware that mains power is an animal that wants to kill you. Always check with meter before touching anything, especially the housing.

 

[goodfidelity]

Now that you say it only blows fuses on initial plug in, that's a completely different story. Bad main filter capacitor and/or bad rectifier. 100% guaranteed. Rectifier can be tested by desoldering and testing for continuity in the wrong direction. Swap it out and you'll be like-new.

Still, please be aware that mains power is an animal that wants to kill you. Always check with meter before touching anything, especially the housing.

Thanks for helping out.

With the rectifier you mean the larger 4 leg diode bridge that sits on the heat-sinks together with the coils?

https://www.wontop.com/uploadfiles/56/sort_excel/pdf/gbj15.pdf

There is two of those.

I measured across the legs in both directions according to this video:

I got results between 400-600 as he mentions. As you say, the component might have to be desoldered before tested. Especially since they are located two pcs in paralelle. I am not sure if i measure the voltage drop on something else on the PCB.

Another thing i noticed was that the area that was getting hot, had a K3564 MOSFET mounted to the sink.
https://toshiba.semicon-storage.com/info/docget.jsp?did=856&prodName=2SK3564

And when using this guide to test it i cannot really seem to get the same results:

I am not sure if that component is good or bad. The mosfet has a capacitor part of it, that should "charge" and make it stay open or closed. I was thinking that component might give similar problem as it can change the way it works regarding to how it was previously charged. The MOSFET does not short in any direction, but i cant get the "latch" to work.

The capacitors i tested by using a multimeter, just to see that they charge properly and block current. I did not unsolder them and test them with a proper capacitor tester as i don't have one. I will try to get component tester so that i can test components. All the capacitors tested Ok this way. I Followed the guide of this video to test capacitors:

Attached are 3 images. One that might be the rectifier, one of the MOSFET that i removed, and one of the "hot" location on the PCB where the MOSFET was located. (Se Thermal image in first post)

In the hot area, there is also one diode that i tested, and one large resistor.

I am not sure the rectifier needs to be removed to be tested, as the video shows testing on the PCB. But i am open to the fact that other components might give me a fake measurement.

Right now i want to find out how to properly diagnose the MOSFET so that i can solder it back or get my hands on new one. Reason i want to check this first is because of these component emitting heat when there is no load on the PSU. I understand the problem might be elsewhere, but showing up at this spot as heat. However i am used to failing components emitting heat. Heat might also be from diode or resistor in same area.