Troubleshooting a switcher.

[itsmoked]

I have been handed a PS out of an old but valuable machine tool. "Can you fix it?"

Argh!


It's about a 100~200 watt unit. It outputs 5V, +/-15V, and 24VDC.

I have NO schematics and only a vague field wiring diagram.
Extensive web searches for a schematic have only resulted in wasted time.

I did a very careful visual and see nothing other than the usual slight board discoloration under a couple of the square ceramic power resistors.

The fuses have never blown.

I hooked up the required 220V to it, and while cringing appropriately behind my eye protection, threw the switch.

A relay clicks and one of the LEDs lit declaring "source". A voltage check shows 5V is present on a bevy of logic chips.

But! I see no other voltages and if you listen very carefully you hear a faint chirp-chirp-chirp with about a 1.5 second period.

So my question is: Any suggestions as to what should I focus on when the oscillation doesn't seem to get off the starting blocks?



Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

Hello Smoked! It hasn't smoked yet? Has it?

It might be that there's nothing wrong with that PS. Some won't work if not loaded at least five or ten percent of rated load. Try put som load on it and see if it makes up its mind. The problem could be in the external wiring in the machine. No connection - no output.

Or scrap it and buy a modern unit. Smaller, more reliable and definitely less costly.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[MacGyverS2000]

That sounds almost like a standard PC supply. Gunnar is right, it needs to have at least a light load to regulate properly, and this may be what the chirp is... the units will turn power on, look for a proper current draw, then turn off for a second or two if none exists. The PSs for home walkways lights work the same way.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Comcokid]

Non-working SMPS, but no burns.

The power resistors you see with a little brown around them are probably snubbers. They always heat.

Chirp-chirp-chirp. SMPS when unhappy will click, buzz lightly, or chirp. These sound come from the magnetics. Could be a overcurrent/softstart cycle re-occuring. This could also occur if the minimum load requirement is not being met.

Generally, as a SMPS ages, the electrolytics slowly dry-out and loose capacitance. They may look perfectly fine, or may be slightly bulged, have the bottom rubber seal slightly bulged out, or show a faint sign of leakage. This can happen to both power electrolyics, and small electrolitics used for signal conditioning

Optoisolators can degrade with a lot of age, CTR increases and the control loop may "open" resulting in the output going (usually) higher in voltage which may cause over-voltage protection circuits to activate, making the supply then shut-down in overcurrent. A properly designed SMPS may pull little input current when it's in overcurrent mode.

Also check the thermal pads under power devices. If it's the old zinc-oxide or silicon based thermal transfer, these can dry-out. Clean it completley and re-grease, or use a more modern type of thermal pad. But, your problem doesn't sound like this. A thermal problem usually shows up under load.

Also, some older SMPS use various inrush current protection schemes so the input electrolytics don't spike the line when power is applied. These inrush circuit sometimes eventually fail making it so there no or inadequate power on the primary circuit side.

 

[itsmoked]

Greetings Skogs!
Smoked? Negative! Nice and clean. All the smoke is still inside.

Not loaded enough? Dang, since all the modern supplies don't care anymore I forget that this 20 year old might. I shall try that.

Scrap it? That was my first thought but subsequent inspection showed that it has an 80 pin IDC connector and about (25) 74TTL chips on it. It plugs directly onto a mother board.

Very weird bird. If you can find a used one they're $1,400. They have six of their 12 machines down with bad supplies. They're afraid to turn the other six off for fear of them not restarting.



Dan; Yeah! Easy to test that theory. I'll try it today.


Com; Electrolytics Caps... I always suspect them babies. There are 2 large line-voltage units and about 6 low voltage. The low voltage have the classic flat aluminum circle with the three lines meeting at the center - zero bulge. Now the high voltage ones? A different story. They have a solid black disk in lue of the aforementioned aluminum. This black disk is seriously domed. Perhaps 3 or 4 mm / 1/7". Both identical. No other bulge evident. I'm not sure if that is normal or not.

There is at least one opto I can see. Excellent point.

Thermal pad comments - Agreed.

The one relay that clicks - could that be part of an anti-surge start up? If it is I should expect a brief delay after power is applied wouldn't I? I haven't actually paid enough attention to notice a delay.


I suppose we should think about 6 of these supplies, with most, probably having the same issue.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

Update.

Adding loads didn't do the trick.

The relay seems to pull in the instant power is supplied and then drops out as the "source" LED fades out.

So I went for replacement of the opto. It's a Sharp PC817.

I look at my 25 page un-organized list of stock ICs. There is listed a PS2401. It's pretty much the same part by NEC. I go thru all three large boxes - many hundreds of packages and tubes looking for the PS2401.. I get down the absolute last thing in the last box, a piece of weird foam the size of a postage stamp. I pick it up and turn it over. One part... A one year newer PC817!! Figgers. But wait! This has to be a good omen.

I replace the one in the supply and.... Nothing. Same exact dweep, dweep, dweep.



Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

One thought on the relay: some older units have a very low power auxiliary supply which is there to allow the unit to 'wake up' and kick the main converter into life. An avenue to explore perhaps.


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

If we learn from our mistakes I'm getting a great education!

 

[Skogsgurra]

"Six of their machines down with bad power supplies"

Sounds like the swine flu! Epidemic disease among power supplies.

Such things usually are because of external factors. If marginally designed and grid voltage getting low (more load on grid?), it *could* be the reason. Have you tested with higher AC input? I would also test with lower AC. You never know...

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[BobM3]

Could there be a needed logic state on the pins connecting to the motherboard?

The moderator (Milicron) on the following forum rebuilds machine tools. I know he's troubleshot machine tool boards and power suppies:

http://www.practicalmachinist.com/vb/forumdisplay.php/cnc-machining-21.html

 

[itsmoked]

Scotty; Agreed. There is a pre-supply that is running the logic. Hence that 5V. I see a separate bridge rectifier which I think is running that.

Gunnar; I think these have died over about 2 years. Which seems to not point at line voltage. I did try uping it and lowering it though. No joy.

Bob; I am definitely fearing that. I would not even try if it wasn't for the dweep, dweep, dweep, which means the supply is trying to run.

For your entertainment here's a picture of this pig.
You can see it isn't anything typical... Its a CPS-10NB.

it has LG(as in cr*p cell phones) on it and LCK ELECTRONICS, and its in a KIA Heavy Industries machine tool running the Yash-something drives.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

Enjoyed the entertainment immensly! Porn for EEs.

You mentioned that the electrolytics were somewhat swollen. And they are, indeed.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[itsmoked]

But they're so uniformly swollen voluptuous...

Are you saying I should replace those??

I would just expect them all not to be swollen exactly the same if there was something nefarious afoot.

The bottom two are 200V and the next ones up are 63V so they aren't even the same model.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MacGyverS2000]

I'm no expert in repair, but I can't say I've ever seen an electrolytic with a rounded top unless it was bad (even in the top-down pic they look swollen). Take that for what it's worth.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Skogsgurra]

No news are good news? Is there any more entertainment from Flamin Systems?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ScottyUK]

Well, we're late for the 4th of July but the UK has much the same kind of show on 5th of November and that's only 5 weeks away...

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

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

No I am dithering over telling them "go fish" or doing more on it.

If you press on the 'bulged cap' tops there is no resistance. The little domes un-dome.

It is really rough trying to fix something this complex that you have NO schematic for. Nor even a functioning example to at least make comparisons against.

As mentioned I don't even know if they come ON without some external signal. I could replace everything on the board and discover that I wasn't supplying ENABLE.

If this was my supply and was preventing my money machine from spewing money I would sit down and draw a full blown schematic. Trouble is, that would take probably (3) eight hour days. Something I know they don't want to pay for.

If I had a full schematic I could actually do several things. I could layout a motherboard with the crazy logic on it,(one PLD), buy an OTS supply, and drop it on.

I could probably sell a hundred since I could come in way below the $1,400 going price.

I could also mark every device on the supply. Send it in to a repair place. Pay the $800 repair charge and see which parts get replaced. Then do the the other 5 supplies the same way. LOL

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

Definitely not entertaining. For anyone.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[MacGyverS2000]

Keith,

If the caps were suddenly cooked at some point (rather than slowly going bad), they may have bulged due to gas expansion. Once the pressure is gone, the can is still puffy...

I would be tempted to pull one out and see what a cap meter has to say.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[BobM3]

I'm just a novice on this stuff and am interested as much as anything else in learning how you guys toubleshoot these things. So please excuse my questions if they are too basic. Looks dark in an area around what look to be two fuses on the lower right hand side of the picture. Is that an optical illusion? Overheating?

Also looks like some components in that area that would make up the input rectification and filtering for the supply. Have you checked the voltage of the two big caps?

 

[IRstuff]

Is the goopy stuff around some of the caps epoxy?

TTFN

FAQ731-376