Switching power supply issue.

[dbtoutfit]

Hi,

I have a couple of ATX (Switching) power supplies that are the exact make/model that have blown a fuse.

I also have many others that are fine same make/model.

I gave throwing another fuse in one of them a try and it blew. I understand I have a short and that it's common for a diode to go bad and cause this issue.

I have a power supply of the same make and model I have used to compare readings from that is in working condition.

I am not sure which parts I should check and how I should check them but I have checked a few of the black and gray (what I believe is diodes) next to the large capacitors (filter capacitors?) On one of the problematic PSU, 2 of the "diodes" are coming up dead short using my MM set on Ohms. At the same time when I check them if I hold the probes for a few more seconds I begin to get a very low resistance reading but it always starts our dead short on those diodes only. I also understand that the "filter capacitors" could be bad but I am unsure of the best way to check them to know and/or if the filter capacitor is what I think they are (the 2 big capacitors).

I could basically use some help on solving this issue, things to look for, advise, troubleshooting tree etc. As this happens from time to time and simply replacing the unit is not a fix for my issue.

Thank you for any help.

 

[ljcox]

The way to check the filter caps is to set your meter to the Ohms range, connect the black probe to the - wire and the red to the + wire of the cap.

The reading will be low at first, but it will start to increase as the cap charges - if it is healthy. If you are measuring it in-circuit, there may be other components connected across the cap that may affect the reading.

I keep switching the range of my meter up as the cap charges, ie. start with the low Ohms range and switch it progressively higher as the "resistance" goes off scale. It should eventually go off scale on even the highest Ohms range - provided that there is no other path in parallel, eg. a resistor to ground. If so, you will eventually read the resistance of the resistor.

Then I switch the Meter to the DC Voltage range and measure the voltage across the cap. It should read the meter's battery voltage initially (or a bit less) and slowly decrease as the cap discharges through the meter.

 

[dbtoutfit]

Thanks Len will do

How about the Diodes?

**broken link removed**

They look like these.

Thanks,
Eric

 

[Hero999]

The way to check the filter caps is to set your meter to the Ohms range, connect the black probe to the - wire and the red to the + wire of the cap.

The reading will be low at first, but it will start to increase as the cap charges - if it is healthy. If you are measuring it in-circuit, there may be other components connected across the cap that may affect the reading.

I keep switching the range of my meter up as the cap charges, ie. start with the low Ohms range and switch it progressively higher as the "resistance" goes off scale. It should eventually go off scale on even the highest Ohms range - provided that there is no other path in parallel, eg. a resistor to ground. If so, you will eventually read the resistance of the resistor.

Then I switch the Meter to the DC Voltage range and measure the voltage across the cap. It should read the meter's battery voltage initially (or a bit less) and slowly decrease as the cap discharges through the meter.

Will that test show a capacitor that's gone high impedance though?

 

[Nigel Goodwin]

Will that test show a capacitor that's gone high impedance though?

Assuming you mean 'high ESR'?, no it won't, you have to test it at 100KHz.

 

[ljcox]

I agree with Nigel, you need an ESR meter to measure the ESR.

The test I described is simply a rough indication of whater the cap is open.

 

[Hum Bucker]

Before we start, a word of warning.
Switch-mode PSUs - such as these - run directly off the ac mains supply; there is no isolation transformer. Secondly, the filter capacitors (the really big ones) can hold a very large amount of energy at high-voltage - enough to really hurt - possibly even kill you! Therefore, be careful! Always make sure that those caps. are discharged before you get your fingers on any tracks or components.

With that out of the way, it is unlikely that those caps. are defective. The usual mode of failure in ATX PSUs involve short-circuit components: diodes, bridge rectifier, power transistors / FETs. If the main primary switching device is S/C, there's a reasonable chance that the driver cctry. will have U/S components as well. If it uses one of the UCxxxx i.c.s, my experience is that that will have failed also.
Then there is the secondary side of the switching transformer. Check for S/C diodes and filter caps.

If you decide to wade into a SMPSU with test equipment - whilst mains power is applied, please run the PSU from an isolating transformer, for your own personal safety. If you intend to probe around with a 'scope - then a isol. transformer is a MUST - otherwise you'll short-out the a.c. supply, damage the PSU - and probably the 'scope - with a big bang!

In the days on the original IBM PC, the 'big silver box' always had a warning notice on it saying 'do not remove lid' (or something like that). This notice wasn't there for fun.

Be careful - and good luck.

Hum Bucker.

 

[Hero999]

Or you can just disconnect your scope's earth connection but that's just retarded.

 

[mneary]

Or you can just disconnect your scope's earth connection but that's just retarded.

uh, a shortcut to the grave.

 

[ljcox]

Thanks Len will do

How about the Diodes?

They look like these.

Thanks,
Eric

Eric,
I don't understand the question.

Are you asking if those diodes will do as a substitute for the blown ones?

If so, I would need to see the data sheets of both the originals and the substitutes.

 

[Hum Bucker]

Or you can just disconnect your scope's earth connection but that's just retarded.

Yes - obviously you can do this - but this is a bad idea for two reasons:
1. You have now lost your 0v reference for making measurements. Any display on the scope will be heavily modulated by a.c. The display will be useless as far as gaining meaningful information is concerned.
3. Between the -ve. side of the main filter caps. and true earth - the 'scope case is at true earth - there is a large a.c. voltage from the mains supply. Should you connect yourself between this -ve. rail from these filter caps. and any true earth - you'll get a potentially life-threatening shock.

When working on ANY item of electronics that doesn't give isolation from the raw a.c. mains supply, not using an isolation transformer is an absolute no - no . . . unless you want to meet the Good Almighty.

Hum Bucker

 

[Hero999]

I agree, I've heard it being mentioned on this forum before and it's a very stupid thing to do.

A good way of making an isolation transformer is to connect two transformers back to back but be aware that the output voltage will be less than the input voltage due to the losses in the transformers.

 

[Nigel Goodwin]

OK, I'll mention it YET again - disconnecting the earth on a scope is an EXTREMELY common practice in the service industry.

For many purposes it's FAR safer than an earthed scope, personally I wouldn't ever use an earthed scope, I don't consider them safe enough.

But as always, you need to know EXACTLY what you are doing, and WHY.

I wouldn't advise anyone to do so, if they understand the reasons they already know if they need to or not.

As a matter of interest, all benches at work are run from isolation transformers - but that's still no reason to earth a scope!.

Claiming it's 'stupid' or 'retarded' is just showing you have no idea or understanding of the reason for earthing or not earthing test equipment!.

But if you don't know why, then DON'T DO IT!.

 

[Hum Bucker]

OK, I'll mention it YET again - disconnecting the earth on a scope is an EXTREMELY common practice in the service industry.

For many purposes it's FAR safer than an earthed scope, personally I wouldn't ever use an earthed scope, I don't consider them safe enough.

"Extremely common" it may be - probably due to a lack of suitable isolation transformers - but just because it is "extremely common" does NOT make it a safe practice! The "Service Industry" is notorious for its cost-cutting measures - usually to the possible detriment of the staff. Plus, you'll still get the display problem that I referred to earlier.

Using an isolation transformer and an earthed scope is the ONLY safe and sensible option in the scenario under discussion. An unearthed 'scope is a safety hazard under ANY circumstances.

Hum Bucker.

 

[Nigel Goodwin]

"Extremely common" it may be - probably due to a lack of suitable isolation transformers - but just because it is "extremely common" does NOT make it a safe practice! The "Service Industry" is notorious for its cost-cutting measures - usually to the possible detriment of the staff. Plus, you'll still get the display problem that I referred to earlier.

No, there's no display problems at all.

Like I said, I use isolation transformers at work on all benches (and isolated areial sockets), but I wouldn't consider earthing the scope, it adds an extra unwarranted danger for no useful purpose.

Using an isolation transformer and an earthed scope is the ONLY safe and sensible option in the scenario under discussion. An unearthed 'scope is a safety hazard under ANY circumstances.

In your personal opinion! - if that's what you think, you obviously don't have the slightest idea why appliances are earthed, and what the advantages and disadvantages of earthing are!.

Like I said before (and always say) if you don't understand the reasons for and against, then you should earth everything - it might not be any safer, but it will probably give you that impression!.

 

[Hum Bucker]

You simply don't understand the whole concept of "earthing", do you?

By placing the electrical circuitry inside a metal box, which is connected to earth, and with one side on the a.c. supply joined to earth (at the sub-station), an electrical failure that results in the metal case becoming 'live' blows the supply fuse - thus saving the user from receiving an electrical shock. Therefore, to ALL test equipment should be earthed at all times in the simple interest of user safety.


You don't seem to understand the basics of SMPSUs, either.

If I wish to examine the waveform across the filter capacitors in a switch-mode PSU, where these caps. are fed from a bridge rectifier DIRECT from the a.c. mains, I MUST, repeat MUST connect the earth lead of my scope to the -ve of the caps.- otherwise, there is no reference for the 'scope to work to. To do this, a mains isolation transformer is ESSENTIAL between the a.c. supply and the PSU - otherwise there will be a dead short across the bridge.

There is no debate to be had on ANY of these aspects: they are simple, provable facts. Only the ignorant and mis-informed would disagree.

No more need be said. Period.

Hum Bucker.

 

[Nigel Goodwin]

You simply don't understand the whole concept of "earthing", do you?

Obviously far more than you do?.

By placing the electrical circuitry inside a metal box, which is connected to earth, and with one side on the a.c. supply joined to earth (at the sub-station), an electrical failure that results in the metal case becoming 'live' blows the supply fuse - thus saving the user from receiving an electrical shock. Therefore, to ALL test equipment should be earthed at all times in the simple interest of user safety.

That's the view of someone who doesn't understand the principles involved, or the alternatives available - a service department ISN'T a domstic environment, and forcing domestic standards on it makes it LESS safe, not more!. Hand on earthed scope, other hand on live wire - oops! - full mains straight across the heart. Maintaining an earth free service environment is considerably safer than an earthed one.

As you're blindly rattling off domestic earthing standards you're obviously far too young to remember the huge controversy over earthed and earth free environments?.

I'll fully agree it makes great sense in a domestic environment, particularly when you have washing machines, dishwashers, steam irons - anything that has water anywhere near electricity. But a service environment is a completely different affair!.

I might also mention I've been on training courses at huge international electronics companies, and in their training rooms there are no isolation transformers, and all scopes have their earth leads removed. I can't remember the last time I ever used an earthed scope?, at technical college all scopes were unearthed - again, for greater safety.

Mind you, at technical college they also had live bare brass 440V three phase terminals on the walls! - that was in the heavy machines lab, which we had a number of lessons in (by complete mistake as it happened).

You don't seem to understand the basics of SMPSUs, either.

If I wish to examine the waveform across the filter capacitors in a switch-mode PSU, where these caps. are fed from a bridge rectifier DIRECT from the a.c. mains, I MUST, repeat MUST connect the earth lead of my scope to the -ve of the caps.- otherwise, there is no reference for the 'scope to work to. To do this, a mains isolation transformer is ESSENTIAL between the a.c. supply and the PSU - otherwise there will be a dead short across the bridge.

Yes, you connect the chassis of the scope to the -ve of the caps, that's blatently obvious - and I do so on at least a weekly basis, and have done since SMPSU's first appeared in the Thorn 3000 series TV's. That's why you don't have an earthed scope in service departments - it's then perfectly safe if you use an isolation transformer or not.

There is no debate to be had on ANY of these aspects: they are simple, provable facts. Only the ignorant and mis-informed would disagree.

Plenty of debate, only the ignorant and mis-informed would think otherwise!.

 

[Hero999]

As you're blindly rattling off domestic earthing standards you're obviously far too young to remember the huge controversy over earthed and earth free environments?.

I'm only 25 and I've heard about it and to me it's common sense why the earthed, mainly because if it isn't earthed it can float at whatever voltage it wants to.

• If a an earth fault occurs on any live conductor, the other lives will float at 400V.
• If you're powering a device that uses an autotransformer to get a high voltage (e.g a neon sign transformer which runs at 20kV) and the live output developed an earth fault the neutral will float at 20kV which would damage the insulation in the distribution system and cause a fire.
• There's no earth path for high voltages induced from lightning strikes.

However, on a controlled environment like a test bench, it's normally only single phase and you know what you're connecting to it so the above problems don't occur.

Hand on earthed scope, other hand on live wire - oops! - full mains straight across the heart.

Not if you're using an isolation transformer as there's no return path to earth.

Yes, you connect the chassis of the scope to the -ve of the caps, that's blatantly obvious - and I do so on at least a weekly basis, and have done since SMPSU's first appeared in the Thorn 3000 series TV's. That's why you don't have an earthed scope in service departments - it's then perfectly safe if you use an isolation transformer or not.

No, it isn't safe if you don't have an isolation transformer. The -Ve pulses up an down at mains potential with respect to earth so as soon as you touch that scope you're a goner. You're better off blowing a fuse or rectifier then learning from it and buying an isolation transformer.

The only time I would agree with disconnecting a safety earth connection is when you're testing something that's got its 0V rail bonded to protective earth (PCs and some audio amplifiers). This is because failing to do so will result in a troublesome ground loop. This isn't dangerous at all as you're still earthing the appliance by connecting it to the scope's earth bonded chassis.

I might also mention I've been on training courses at huge international electronics companies, and in their training rooms there are no isolation transformers, and all scopes have their earth leads removed. I can't remember the last time I ever used an earthed scope?, at technical college all scopes were unearthed - again, for greater safety.

Mind you, at technical college they also had live bare brass 440V three phase terminals on the walls! - that was in the heavy machines lab, which we had a number of lessons in (by complete mistake as it happened).

When was this?

In the 1960s may be, but not in 2008 and just because you've seen it done before it doesn't mean it's safe.

 

[dbtoutfit]

Maybe I should go to another forum; I did not mean to get something going. My bad.

I simply read an article on common reasons these power supplies blow fuses.

It described that it's highly typical for the diodes next to the filter cap/s to dead short for no reason. This seems to be true in my case but I want to make sure other components down the line after the diodes were not affected.

It was described that it is possible that the filter cap/s could go bad.

I am sure that this is typical but I thought I would just mention that there is a capacitor that crosses beside with the fuse. After installing a few fuses as I was trouble shooting I notice this capacitor would finally "pop" so to speak. Blow a little ceramic pc off it's self. I would assume this wouldn’t help as it is probably a result in the dead short.

Eric

 

[ljcox]

It described that it's highly typical for the diodes next to the filter cap/s to dead short for no reason. This seems to be true in my case but I want to make sure other components down the line after the diodes were not affected. You will have to measure them to know. Some components can be measured in-circuit if there are no parallel paths.

It was described that it is possible that the filter cap/s could go bad.

I am sure that this is typical but I thought I would just mention that there is a capacitor that crosses beside with the fuse. After installing a few fuses as I was trouble shooting I notice this capacitor would finally "pop" so to speak. Blow a little ceramic pc off it's self. I don't understand this, what does pc mean? I would assume this wouldn’t help as it is probably a result in the dead short.

Eric

Eric,
Some times you have to unsolder one end of a component in order to test it.