Testing a power supply

[godspeed]

is their any way to test a power supply to see if its working correctly?

[AruisDante]

Yes. Make sure you at least have a couple of fans attached to it for load (PSU's can be damaged if you start them up with no load) and then jump the green wire on the 24pin connector to any of the black wires.

[Deluded]

That only sees if the PSU even works.

Making sure that it delivers clean power under sustained draw is another matter.

For that, I would plug it in a computer, firing up LinX and furmark so both CPU and GPU are drawing power, and then measure the voltage with a DMM.

[AruisDante]

Well yes, obviously. I was just assuming he meant to see if it is what is broken in a non-functional system. Measuring actual PSU performance is a totally different beast.

Fun related story, I was working on a robot once that had a piece of control equipment that would restart sporadically when the robot's motor controllers were power cycled. It was powered off of two 12V lead acid car batteries. We literally spent days trying to track down the problem. Using a very nice oscilloscope we eventually discovered that the motor controllers had a flaw that would essentially cause them to short to ground during start up, which was causing a 12 nanosecond voltage droop on the 24V line. But that was all it took to f the whole system up. The solution? A bank of six freaking enormous capacitors. They held enough charge to keep the LED lights in the E-stop button lit for over 5 hours.

[godspeed]

thanks guys

[BigE4u]

You definately need a DMM(DigitalMultiMeter) when checking your 12v/5v/3.3v rails and when shorting the green and black ground wires on the power connector to power on, its best to connect 12v fan or similar device to put a small load on the PSU. You can also buy a power supply tester and use as shown below...

Link - http://www.youtube.com/watch?v=cVc9f...eature=related

You can pick up a pretty decent DMM at your local walmart(automotive section) or at your local radioshack, a good $20-$30 auto-ranging type will suffice.

[901-Memphis]

Originally Posted by BigE4u

You definately need a DMM(DigitalMultiMeter) when checking your 12v/5v/3.3v rails and when shorting the green and black ground wires on the power connector to power on, its best to connect 12v fan or similar device to put a small load on the PSU. You can also buy a power supply tester and use as shown below...

Link - http://www.youtube.com/watch?v=cVc9f...eature=related

You can pick up a pretty decent DMM at your local walmart(automotive section) or at your local radioshack, a good $20-$30 auto-ranging type will suffice.

What he said except almost nothing uses the 3.3v line anymore, the 12v/5v are what you need to test and both can be tested from a single molex.

[larrymoencurly]

Originally Posted by 901-Memphis

What he said except almost nothing uses the 3.3v line anymore, the 12v/5v are what you need to test and both can be tested from a single molex.

Can modern motherboards work with just +5V and +12V rails (assuming the Power_Good signal would still go high)? Because a budget Socket 775 DDR2 mobo (not exactly the state of the art) drew about 0.4A @ +3.3V while MemTest86 ran and built-in video was used.

[901-Memphis]

Originally Posted by larrymoencurly

Can modern motherboards work with just +5V and +12V rails (assu;ming the Power_Good signal would still go high)? Because a budget Socket 775 DDR2 mobo (not exactly the state of the art) drew about 0.4A @ +3.3V while MemTest86 ran and built-in video was used.

I'd guess so since i believed DDR1 was the big user of this line. 0.4a is a pretty small draw but something had to be using that, maybe a motherboard signal or some weird part?

[AruisDante]

The CMOS battery is also charged off the 3.3V line usually if I remember. 3.3V is standard CMOS logic level operating voltage, so I imagine there are a few low power chips on the mobo (like the BIOS and such) that run directly off it without the need for voltage regulators to shift down the higher voltages.

[901-Memphis]

The CMOS battery isn't a rechargeable Lithuim Ion battery, its a primary lithium battery, so it can't be recharged. But still i am sure something runs of the 3.3v for some sort of signal.

[AruisDante]

Originally Posted by 901-Memphis

The CMOS battery isn't a rechargeable Lithuim Ion battery, its a primary lithium battery, so it can't be recharged. But still i am sure something runs of the 3.3v for some sort of signal.

Either way, it's a 3.3V battery, meaning whatever it's supposed to power when the PSU is off is likely powered (and in a higher power state) of the 3.3V line when the PSU is on

I also think it depends on the board as to whether they use a rechargeable or not, but don't quote me on that one.

[larrymoencurly]

I think laptop computers have rechargeable lithium CMOS batteries, at least some made by Compal did (disk covered with shrink wrap, 2 wires going to a connector), and I am referring to a small coil-size battery, not the large main battery. But I've never seen a desktop motherboard with a rechargeable lithium on it.

I had another desktop motherboard, probably the AM2 type, that consumed about 0.9 amp @ +3.3V

[BigE4u]

Nope... most 12yr old laptops to current ones use replaceable non-rechargeable 3.3v lithium cmos battery, some are shrink wrapped with wires protruding from it that can be disconnected from the main board and others snap into a battery holder like ones seen in desktops.

[Sheep221]

CMOS battery is not and never was rechargeable, it just has long time durability because it powers only cmos system - Bios ROM and time crystal. Rest of the system is powered from PSU.

As well, I'm not sure if the PSU when running empty needs to have load connected to prevent any damage, it does have own fan which is drawing power either. I don't recommend running empty PSU just like that for extended period of time. That could put some good wear on the capacitors inside.

[larrymoencurly]

Originally Posted by BigE4u

Nope... most 12yr old laptops to current ones use replaceable non-rechargeable 3.3v lithium cmos battery, some are shrink wrapped with wires protruding from it that can be disconnected from the main board and others snap into a battery holder like ones seen in desktops.

Originally Posted by Sheep221

CMOS battery is not and never was rechargeable,...

The maintenance manual for my Compal-made Toshiba laptop: http://tim.id.au/laptops/toshiba/satellite%20a200.pdf

says the clock battery needs ~24 hours to recharge:

-----------------------------------------------------------------------------------------------------------



-----------------------------------------------------------------------------------------------------------


The schematic shows a charging circuit for the RTC (real time clock) battery, with a signal named +CHGRTC, apparently standing for "charge real time clock", and it's not from the regular power supply or big battery because there's also a signal, +RTCVCC, for that:

-----------------------------------------------------------------------------------------------------------



-----------------------------------------------------------------------------------------------------------


The battery specified in the Toshiba manual is type ML1220/B, which Panasonic describes as rechargeable: http://industrial.panasonic.com/www-...09+ML1220+7+WW

-----------------------------------------------------------------------------------------------------------

[godspeed]

brought me a psu tester for 12 bucks, real nice and yes it was the psu, thanks again guys because the dvom meter was reading like it was ok.