is this baby-at power supply safe? (grounds aren't 0) please advise

hello and happy new year!
baby-at (circa 1993) intel 486dx;
brand name: desktop, "Hewitt HR Rand 486 cache "
baby-at psu (115 V / 230 V hewitt-rand switching 200 watts)
PROBLEM: backProbing with digital voltmeters (tried two) gives the following:
P8 6 = -.01 v
P9 1 = -.01 v
P8 5 = 0 v
P9 2 = 0 v
(these are the 4 central black grounds)
looking down on the board with the connectors: the innerMost grounds read -0.01v and the outerMost grounds (next to the above) read 0 volts.
also the hdd power connector's grounds (both) read 0.01 volts. (the power good and the hdd (orange = 5 v and yellow = 12 v)
the pc is giving bizzare and vague errors... the case fan stopped working and i had to unHook the cd-rom to get the hdd detected and bootable (however doesn't always boot) etc..... (the cd-rom had run hot (according to the owner))

just need to know whether the nonZero grounds could indicate a psu problem
i do not know where to find this info
thanks in advance!

 

thank you PeteC for the reply!
the meters are ok and on other psu's the 4 grounds are 0 volts.

the case fan is NOT running (and i think it had -- the cpu fan IS running as well as the psu fan.)

just an upDate on the system: today all it stated was "hdc failure insert floppy into drive a and hit any key" which resulted in "drive not ready insert disk..." etc etc etc.
(the floppy drive has not worked yet -- in dos in windows (but it *IS* in the bios (3.5" 1.44 MB))
so reseated the hd data cable on the controller, the hd data and power cables; removed the modem and removed a connector for the hd led (on the controller card)(red and white twisted)where the white wire was not in contact with the adapter
for some reason it started to work

i'm sure it would like to retire however i would like it to have a running case fan at the very least
for now it's on sabatical

THANKS!
sincerely
Tanya

 

Ok let me get this correct, I may have what you are doing wrong though so tell me if I am.

You are saying the PSU have 0 Volts between your grounds (Negatives) when you place the probes on them, well you wont get any voltage between them full stop, its the laws of electricity.

You also say that you are getting 12v and 5 volts and the power good is ok. That is good news though.

I have had problems with baby-at power supplies before, its the capacitors inside them they tend to fry up, the only way to fix them is to replace them or buy a new power supply, I did the first and the power supply is ok now and its a lot cheaper than buying a new supply if you can find one that is!

I am an electrician by the way

Hope this helps

Cheers
tel33

 

hi
thank you very much for replying!

what i was measuring was the grounds against the psu case
(backProbing -- red on the wire and black on the psu case)
the results were *normal* except for p8 wire 6 and p9 wire 1 (the 2 grounds that are next to each other on the board)
the results were p8 6 vs psu case == -.01 volts
p9 wire 1 (vs. psu case) == -.01 volts
the other 2 grounds were both 0
(p8 5 vs psu case == 0 volts
and p9 wire 2 vs, psu case == 0 volts)
on the hd power connector (again with the pc on)
the 2 black grounds (in the center) (each measured against the psu case) were both -.01 volts

reason i am *concerned* is that afaik small voltage changes etc can really effect pc components so i am not sure whether - .01 volts would be significant in this scenario
thanks again
sincerely
Tanya

 

It should not affect anything at that kind of voltage, you need not worry about any voltage on the ground side between grounds and case. They are all connected.

I still think you have problems due to the capacitors that are failing.

What make is the power supply and what is the wattage output of it?

According to my information this is what you should get on the pins voltage wise.

p8-1 power good 5=v Orange
p8-2 +5v Red
p8-3 +12v Yellow
p8-4 -12v Blue
p8-5 ground (0v) Black
p8-6 ground (0v) Black

p9-1 ground (0v) Black
p9-2 ground (0vBlack
p9-3 -5v White
p9-4 +5v Red
p9-5 +5v Red
p9-6 +5v red

Measure the voltage between p8-1 orange (+5v power good) and p8-5 ground black (0v) when the pc is booting up with all the componants connected, this should stay at around +3v to +6v any lower or much higher than these specs and the psu is most likely faulty.

Once working keep an eye on any voltage that maybe dropping when the pc is operating and the psu is getting warmer, especially on the +12v side as this operates the cd rom and floppy and hard drives so is under more intensive use, if it starts to drop to say +11.0 volts or lower then something is wrong, measure between p8-3 +12v yellow and p8-5 ground black (0).

I hope this helps you.
tel33

 

Hi Tanya

thanks for indicating what you were using as reference (ie the case)!

although there are instances where a small difference in voltage can have quite a large impact (like core voltage for CPU), in this case not to worry too much - there's a hefty current flowing through the blacks going back to the PSU - I'd be surprised if you didn't see several hundedths of a volt (negative) at the other (PSU) end of the blacks just due to the resistance of the wire.

also consider - don't know how many digits your DVMs have - but if the "bottom" digit is hundredths not thousandths... 0V004 "rounds down" to 0V00 but 0V006 "rounds up" to 0V01 - making it look worse than it is.

although what does get the alarm bells going a bit is that you are getting the different readings from points so close together physically. Indicates the currents flowing down the pairs of blacks going back to the PSU are out of balance for some reason.

If this is only happening with one particular PSU then Tel's pointer to the elcaps in the PSU has got to be the most likely. They have to take an appalling amount of "ooomph" - and there are a lot of dodgy ones around (this makes me chuckle - was this actually "selling a dummy" perhaps? Industrial sabotage?) Whatever, many reputable manufacturers have fallen foul of the naff ones and are rather embarrassed. Watch out for cans with slight bulges...

(if opening up the PSU take care that all capacitors - particularly the big one at the input! - are discharged before connecting yourself across them. If in any doubt, don't open the PSU.
Tanya - you are probably well aware of this - but BBS is meant to be searchable...)

==

If you can get to the underside of the motherboard - not always easy - check the soldering of PSU connectors (P8, P9). This problem is nowhere near as common as it once was but is really nasty if you hit it.

Look carefully for dry joints (lens, wiggle!) suspicious ring around the pin - but you cannot always see it. If you find one resolder all of them.

(It's if they've wave soldered, and run the board through too fast for the thick metal of these pins to heat adequately.)

Failure of one power supply rail (only) can be totally disastrous, and ruin all your hard work. Tricky if it's due a dry joint on the connector, you can measure at the accessible bit (where you have been) and the voltage is OK - without it actually getting to the board.

This isn't as likely as the PSU elcap thing; and it's often impractical getting to the bottom of the motherboard; but it would fit with your readings, and also

makes me think connector problems. Unless the case fan comes through a thermostat somewhere maybe?

(back in the 1980s I was working for a university which had bought a batch (~25) of VT100 terminals - almost every single one had dry jointed power supply connectors, down to poor wave/flow soldering. A whole range of symptoms (understatement). And it's not a situation which you want to happen: gave rise to ICs being subjected to inputs not between VCC and 0V: very good way to destroy chips)

phew. sorry for writing a book!.

best wishes, HJ.

 

Hi HJ

I would agree on what you have said, seem many psu's with poor soldering and dry joints, plus the tracks raising off the surface due to the heat the psu can generate and breaking.

About the bulging of the capacitors at the tops, that is very much the case of a psu I did recently, it kind off emptied its contents all the place, luckily it did not damage the motherboard when it went.

Still I think that this is the most likely cause of the problems, lets hope eh.

Regards
tel33

 

Hi Tel

yup it's gonna be the elcaps ten to one...

yuck!no doubt, far better to avoid if one can...

Are you old enough to remember "chicken feathers? "

- that DJ thing is so awful if you hit it I thought I'd give it a mention in case - but it's a pig getting to the bottom of the board to check usually...

...and production technology is undoubtedly a lot better nowadays.

If I get a mo I'll try opening one of the blacks that should be in parallel back to the PSU on one of the old 486's (got one that doesn't matter too much - it's about to turn into a 243 soon anyway ) see how much does get dropped in that scenario - guess it would actually a bit more than 0V01?

best wishes, HJ.

 

LOL

They sure can go with a bang also!

Regards

tel33

 

hi tel33 and Hugh Jarss
thanks for replying!
GREATLY appreciated!

briefly (before removing the board which i was likely going to do anyway) a couple of questions:
and answers)
tel33:it is a Hewitt HR rand (pc company also *made* the psu (they are extinct))
switching psu
model no. PW 200APS (so 200 watts)
AC INPUT 115v-230v 5/3 a 50-60 Hz
and guess where it is made:
Taiwan
i think it was made in 1993 (that's when the moBoard was)

i have to retest the volts (afaiRemeber they were within the values that you list)

Hugh Jarss
the dvms were all to 2 (?3) significant digits (xx.xx) i.e. i used the same settings on all of the ones i've tested (4) all except the 1 i post about were @ zero (the grounds)(this is made in Taiwan for hewitt hr rand so i don't know whether it is 1 of the affected one)

-- is there something on discharging capacitors?
and on opening psu's?
(actually would they be discharged by 1 end of something on the capacitor and the other end to a grounded object? or a screwDriver)
i will WAIT for the reply p.t. trying this obviously


i have no qualms re: removing and inspecting the mb (unlike opening the psu)

i really thank you both very much again!
sincerely
Tanya
sorry -- wasn't so brief i guess............

 

Hi there

Ok so its a small psu and very old at that!

Best if you do check the voltage check again when the pc has been running a while, thats when the caps start to go.

I think its the psu going, check as HJ says all the connections in it and the motherboard power connections for dry joints etc.

The capacitors will be discharged naturally over a short time.

The best way is to unplug the pc from the mains and turn it on for a short while this usually drains out what power is left in any caps, if you are still worried then short the pins out on them quickly, you should be ok though.

Wear rubber gloves if you are still worried when handling the psu and stand on a rubber mat. (I am being serious )

To be honest I think you will have to find a replacement psu unless you can change the caps that are gone, if you can change all of them to be sure you have got the faulty/failing ones.

Let us know how it goes on the checks

Just found this online, not sure how old it is but it might be of interest to you.

http://www.cfusion.com/cgi-bin/getPart_2.cgi?strSearchFragment=PW 200APS

Regards
tel33

 

hello,
i want to THANK you very much for the help, information, advice and the link looks very useful (for a new psu)

i'll follow the instructions re: the gloves and mat even (and letting it run w/o the cord, etc)
(but first will let it run for a while and reMeasure ALL the voltages)

i have another electrical question i guess i should post it as a new topic (voltage regulator and psu)
i hope you all see it

MANY THANKS!
sincerely
Tanya

 

Tanya
a very big sorry for not replying sooner - and now I find I've been suffering from foot-in-mouth disease too

when I said about BBS being searchable it wasn't meant to be sending you on a trawl searching for postings about discharging elcaps - the "you probably already know this" was in case you thought I was belittling your commonsense or skills...

...but I felt I had to put some kind of caveat in case some twerp who doesn't know one end of a soldering iron from another and doesn't have the wit to take heed of the plain "High Voltage - Do Not Open" warnings, searches BBS, runs across this thread, decides to open a PSU when it's turned on and ends up splatted all over the ceiling as a result. Although it's arguable that the world would be a significantly better place with fewer such I could understand BBS feeling I'd let the side down somewhat - and then there would be the cost of redecorating the ceiling to consider, let alone the awful smell of roast idiot hanging around for days afterwards...

attempting to atone for my gauche ambiguity earlier - an addition to Tel's excellent "rubber mat" advice: something I picked up from a TV repair man: when dealing with high (or unknown) voltage, use one hand only - keep your other hand firmly in your pocket.

reason: current flowing "arm to arm" or "arm to leg" goes across the heart - which is what you are really trying to avoid. "In one finger and out another" (same hand) might hurt like the blazes but isn't in the same league.

Discharging Cs generally - usually it won't hurt to short them, but preferably between their two terminals (rather than terminal to ground/chassis). Keeping the discharge current localised is a very good idea - particularly until a fault gets localised...

There's an excellent tool to discharge Cs - a cement coated wirewound resisitor; not only is it a bit more gentle but the cement body makes a convenient insulated handle. (value largely unimportant 10R to 1k ideal). Dead monitors are good hunting grounds for these.

The can of an electrolytic capacitor is typically connected to its -ve terminal; I mention this because it sometimes matters, like when the capacitor negative doesn't connect to ground...

In the PSU I have open in front of me there are two elcaps as "reservoir" (220microFarad @ 200V) fed reasonably directly from line (mains) power - they are arranged in series. I can see bare metal tops to their cans (occasionally the whole of the can is insulated - but commonly you will see bare metal at the top), So when power is on (or if the bleed resistor's absent or failed) there's enough voltage between these two adjacent exposed metal surfaces to hurt... Easy to check with a meter first!

best wishes, HJ.

 

hello Hugh Jarss,
didn't answer before now b/c i was laughing about the clean up of and the redecorating... <lol>

i haven't done anything w/ that pc or its psu (trying to fix the 1 w/ the HOT VRM and HEATSINK....)

i was aware of the 2-arm rule (was warned when i wanted to repair my microWave oven (which i did since it was <only> a blown fuse -- luckily -- i suppose))


also, i cannot post my question re: moBo capacitors here without ruining the "flow" of this thread
...so i'll just mention a couple of things
when i let it run and then test the voltage, it's been <spontaneously> reBooting <lol>

also, reTesting the mains on the board, this time it did not involve p9 and p8 but the 2 grounds from p9 but still the hd connector (4 wire) grounds were ALL -.01volts.
(with the autoRanging dvm, it would not read steady on the board grounds fwiw)

the reason i haven't attempted to fix it (if i did find the caps blown are they savageable? or would i need to find some?) is b/c i have total 5 baby-at psus here i cannot imagine that they are all toast.

do you know whether there's a photo of an open psu?


so now i am going to post in the "voltage regulator and psu" thread -- i hope you have a chance to read it
THANKS
sincerely
Tanya (electricianWannaBe)

 

hmmm - something "funny" going on here - it feels like those PSU rails may well not be smooth ie have an AC voltage component superimposed - might be what's upsetting your DVM. Not necessarily enough to matter though.

Tanya, do you have an analogue multimeter (cheap and cheerful will do)?

going over to have a look at your other (VRM) thread, will come back to this one after

best wishes, HJ

(PS microwave ovens are geblinkin' lethal inside - when I had to open ours to try to fix the fan(!) I looked first - of course - looks like 2.2kV, peak rectified, voltage doubler configuration. Which gives about 6 kV (peak, unsmoothed) on the magnetron anode & heatsink if I got it right - large exposed metal area - yikes!)

 

thanks for answering,
Hugh Jarss...........

i do have an analogue however, not sure what to set it on especially for the negative ##'s

it is a "Protek A802" and has no user's manual
i imagine i could "calibrate" it (extrapolate) by measuring some batteries...

i've heard they are the most dangerous appliance...

thanks very much again for the reply
(i'll try the analogue tomorrow when i figure out how to use it)

sincerely
Tanya

 

Hang on! - I want you to use the analogue meter to explore the behaviour of your digital ones

but the analogue meter will also turn out to be your friend too... they spot quick changes which digital ones can miss. A digital meter averages out over its sampling period. You can see a pointer "kick" when a digital meter would miss it.

don't worry about measuring -ve rails - just put the red from the meter to ground, and use the black to measure with. Easy to tell if you are the wrong way round! If you have the meter set to a voltage range you are highly unlikely to do any damage. Start with a high range and work down if in doubt.

(Ohms ranges are different for general "prodding around with" - you can get current out of the meter prods, and with an analogue meter it's backwards, as you are about to find out!)

try this: set your analogue meter to an ohms range (the lower the better); set your digital meter for a DC voltage measurement; connect the two leads from the one meter to the other (red to red, black to black) you are expecting -1.5V on the digital meter here...

now, the interesting part. Set your digital meter to AC voltage instead, (everything else the same), and let me know what you get, pls...

while you are at it, one last check. Analogue meter to its highest ohms range, digital meter set for DC volts, same connection, what does the digital meter read pls.?

best wishes, HJ.

(what's a-caring: some digital meters will block DC when you set them to an AC range - others won't - particularly, "true RMS" reading meters won't. Trying to suss out which type you have.)
(edit! the -1.5V might well turn out to be -3V if your analogue meter has two cells)

 

hello Hugh Jarss,
thanks again for answering!
i ran the test(s) and this is what i got:
(the analogue does not go to zero (when off or on) it goes to 1/10 and it is fairly new)

with the analogue set to 1X ohms:
autoranging DVM:
dc: -3.220
ac: 19.7 - 19.9 (wasn't steady)
manual DVM: (20v) dc: -3.18
ac: (20v): 0.00

analogue set to X1k ohms:
autoranging DVM
dc: -3.192
ac: 19.5 (!)
manual DVM: (20v) dc: -3.02
ac: (20v): 0.00

the autoranging wavers a lot -- i like the manual better.
is this what was expected?

thanks so much again for replying...
i really appreciate it
sincerely
Tanya

 

Hi Tanya

thanks for the results - good!

deal with zeroing the analogue meter first:

look for a adjustment scroo above where the meter needle pivot point is - this is a mechanical adjustment to bring the pointer to the zero mark (no electricity involved)

(if you have a mirror scale get the pointer superimposed over its mirror image - then you know you are looking straight on to the meter's face)

do this adjustment first... when done you are OK to measure voltage or current (AC or DC)...

...but if you want to measure resistance, you need to set full scale correctly as well, as follows:

Calibrating for resistance measurements: before you connect to whatever you are going to test, set the meter to the most appropriate ohms range (guess!) and short the two prods together.

with the two prods shorted, adjust the preset called "Ohms Adjust" (or the like) to obtain full scale on the meter - calibrating for zero resistance between the prods.

then make your measurement!

note: the zero ohms adjustment for full scale will probably require a different setting on the different ohms ranges - this is normal. Also, this adjustment will alter as the batteries in the meter age - so do it immediately before you measure.

Hint: don't leave an analogue meter set to ohms when you aren't using it - avoid flattening the batteries.

==

best way to check the accuracy (after setting up as above!)is to measure some DC somewhere with the analogue and the manual digital meters at the same time, in parallel.

9V battery is good for this!

Even cheapo analogue meters are not too bad usually - you'll probably be pleasantly surprised. The AC ranges are worse than the DC ranges.

more soon, a bit busy here at the moment!

best wishes, HJ.