Why you shouldn't open power supplies (w/video)

[larrymoencurly]

It was always understood to me that its the amps that will kill you, not the voltage.

We've all been misled by action movie torture scenes involving 12V car batteries and jumper cables attached to men's nipples.

A teacher once told me it was safe to stick fingers into a wall outlet that also had a toaster plugged into it because current took the path of least resistance. At another time, that teacher said light took 1/113 of a second to travel, regardless of distance. Master's degree, but not in science, engineering, or math.

 

[Bobnova]

That teacher is dangerous!

 

[Mr.Guvernment]

http://digitallife.today.com/_news/...taking-apart-unplugged-computer?lite=obinsite

 

[(G{in}[AK)TION]]

usually i just open the PSU to replace or remove fan. depending on the PSUs condition.

never got shocked by a PSU. i did mess iwth a wall outlet when i was a kid though. "shocking" experience.

 

[Bobnova]

If you don't take the PCB out and don't touch anything shiny, you'll be fine.
If you take the PCB out it's very hard to avoid touching shiny things, and shiny things may have 380v and plenty of amps (if briefly) for you to enjoy.

 

[(G{in}[AK)TION]]

If you don't take the PCB out and don't touch anything shiny, you'll be fine.
If you take the PCB out it's very hard to avoid touching shiny things, and shiny things may have 380v and plenty of amps (if briefly) for you to enjoy.

i cant resist the shiny things though. they look so beautiful yet so deadly.

#sorryitwasrightthere

 

[Mr.Guvernment]

unplug, press poewr button on computer couple times, then remove.

 

[madhatter256]

CRTs are much more dangerous...

 

[Bobnova]

Much higher voltage, much higher capacitance. The flyback transformer inside them can store quite a charge for a very long time too, as can the CRT itself.

 

[Rey]

For the record, those big 'ol filter caps behind the rectifiers are discharged almost instantly after unplugging the unit, as they are essentially in parallel with the drive circuit.

I've repaired a few switcher supplies in my time, and been zapped by plenty. The best advice I can give to anyone that decides to disassemble one is to work with one hand when possible. Keep the other one at your side or behind your back.

Also, Bobnova, what are you referring to when you speak of APFC caps? I've never known any caps in a power supply to contain high voltage other than the filter caps I mentioned earlier.

 

[Bobnova]

Modern PSUs have an Automatic Power Factor Correction circuit in 'em, it takes the place of the voltage doubler (and switch) in older units.

 

[givmedew]

Ok so how long do I need leave a PSU unplugged before I open it up to modify anything? Also does it discharge quicker if I unplug it then hit the power on button on the computer? That usually causes all the leds and stuff to go dead since I think it just sucked out all the capacitors? The reason I ask is because I want to modify my PSU to have a large top intake fan that would be very low noise and a thermal sensor off the motherboard into the PSU. It's for my wifes computer and it would be a very very silent PC if it wasn't for a PC Power+Cooling 1KW SR that runs the cooling fat at 100% all the time.

Also is it my understanding that if for some reason the fan fails that there is good enough heat protection built in to keep things from destroying themselves?

 

[Rey]

Modern PSUs have an Automatic Power Factor Correction circuit in 'em, it takes the place of the voltage doubler (and switch) in older units.

Ah, fair enough. Tbh I don't think I've opened a PSU that was less than 5 years old.

 

[wagex]

Ok so how long do I need leave a PSU unplugged before I open it up to modify anything? Also does it discharge quicker if I unplug it then hit the power on button on the computer? That usually causes all the leds and stuff to go dead since I think it just sucked out all the capacitors? The reason I ask is because I want to modify my PSU to have a large top intake fan that would be very low noise and a thermal sensor off the motherboard into the PSU. It's for my wifes computer and it would be a very very silent PC if it wasn't for a PC Power+Cooling 1KW SR that runs the cooling fat at 100% all the time.

Also is it my understanding that if for some reason the fan fails that there is good enough heat protection built in to keep things from destroying themselves?

that that 1k psu's fan is going balls out all the time, rma it, not worth the risk/ warranty voiding. plus running a thermal sensor from the mobo the the psu? sounds kind of unsafe.

 

[givmedew]

that that 1k psu's fan is going balls out all the time, rma it, not worth the risk/ warranty voiding. plus running a thermal sensor from the mobo the the psu? sounds kind of unsafe.

I think that running Balls Out is the design spec. The quote it as being a server PSU in a normal PSUs body.

The fan is only hooked up using 2 wires. A 12v + and - the fan that it came with is a 3 fin delta. I have put a 4 fin delta with similar CFM specs. When I run it at 80% duty it is actually tolerable and seems to be pushing more air than the older fan at 100%

The thermometer I am speaking of it capable of sending data to the motherboard to be directly used in the motherboards bios fan controller. I can tie it directly to the fan I put in the PSU and let the motherboard control the fan based off the internal temps of the PSU.

Just wondering if this is all a bad idea? I feel comfortable working with a PSU however I have never modified one like this.

Also wondering if having no flow at all is ok if very little or low.

I also have another idea where I could put a small 80mm fan in place that has a thermal sensor built into it.. and but even at full load that little fan blows a tiny amount of air but I have a 140mm fan that I could cut out a spot in the top of ple

 

[Bobnova]

I don't think I would be comfortable changing the fan design that much, the location of the components and the heatsinks are rather specifically set up for that fan location and speed.
Given that you aren't using anything close to 1200w you probably could swap a slower fan in, but you'd be risking some out of the way component now getting little to no flow and overheating.
Changing to a top mount, large, fan I can't recommend without looking more closely at the design. There has to be somewhere for the air to go after hitting the heatsinks, and the current design is set up to wash over everything, something a 120mm top mount fan doesn't really do in some cases.
I'll spend some time hunting for a review of the unit at some point today so I can give you a better opinion.

 

[givmedew]

I don't think I would be comfortable changing the fan design that much, the location of the components and the heatsinks are rather specifically set up for that fan location and speed.
Given that you aren't using anything close to 1200w you probably could swap a slower fan in, but you'd be risking some out of the way component now getting little to no flow and overheating.
Changing to a top mount, large, fan I can't recommend without looking more closely at the design. There has to be somewhere for the air to go after hitting the heatsinks, and the current design is set up to wash over everything, something a 120mm top mount fan doesn't really do in some cases.
I'll spend some time hunting for a review of the unit at some point today so I can give you a better opinion.

Bob, did you ever get around to finding that out so you could get back to me or did my frivolous attack stop you from doing that? I totally understand if it did but after our other conversation I hope you'd be in the mood to help me out on this one.

I definitely do not want to burn this PSU up but at the same time I can't have it sounding like a jet engine in a computer that would otherwise be silent because of the water cooling.

The fan exhausts out. The "cool" air comes from inside the case through the mesh back side of the PSU. The PSU is mounted at the bottom and it is very cool down there, no HD's just a radiator several inches away with the air being sucked through it from outside the case. The rad doesn't actually have to do much cooling and the fans always run on low so the air going into the PSU should be good.

Like I said I did take out the old delta fan. Even PC P+C started putting 4 fin delta's in shortly after this one was made. The new fan is actually capable of more CFMs and at 100% will still be very loud.

My hope is that I can safely run around 40-60% and have it raise up if it starts to get warm inside the PSU.

Also where would be a good place to locate the temperature sensor?

I am going to edit this post and include some pictures but I need to hop over to my macbook for that since I do not have an SD reader for this PC.

edit: Here is the shot of my PC the bottom most fan plug is also where I plug the temperature gauge in I need to put the temp sensor somewhere that it can at least acknowledge a change in temp. Even if the temp is way off so long as it can see the temp rising I can use it to run the duty load on the fan.

 

[Bobnova]

I'd forgotten about it. Looked now, I wouldn't replace that fan. It's an exceedingly cramped unit, the fan has to run full blast just to get air through it.

 

[givmedew]

I was editing just you posted...

My bad... do you think the 4 blade fan is ok in there? At 80% it is a lot quieter and seems to actually flow more air. The last ones of these PSUs had a 4 blade delta instead of a 3 blade.

 

[Bobnova]

As long as the airflow coming out the other end is about the same, it should be fine.