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

 

[itsmoked]

Hi Bob. Naw, just picture lighting. The entire unit looks out-of-the-box new. With the exception of the aforementioned slight darkening underneath the board under the power resistors. There is not even a trace of dust on this 15 year old.

IR; ALLLLLLLLLLlllll the electrolytics have those blobs of what appears to be RTV somewhere on their bottoms. Often between them. It is obviously factory applied in the usual business like manner. Perhaps to mitigate vibration work hardening of the leads holding a big device.



I told them this is just too hard not having documentation nor a working model. They said "I think I know where to get some documentation." They also said there was another machine that was dead NOT because of the supply. And that they could probably get me a working supply.

So that's where we're at on the long path to....

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MacGyverS2000]

Sheesh, if you can get a hold of a working supply, it would be easy peasy to compare major voltage points.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Comcokid]

"There is not even a trace of dust on this 15 year old.".

Suggest you pull some of the electrolytics (both the big and small) and measure them. They're old enough to have dried-out, which happens with time, heat, and use. If they have, you will find the capacitance is 10% or less of what you expect.

 

[MikeHalloran]

Before it became Kaypro, and long before it died, NonLinear Systems sold a magic box that could check the health of most capacitors in situ, among many other things. I think it cost $400, when that was a _lot_ of money. Doesn't anybody make stuff like that anymore?



Mike Halloran
Pembroke Pines, FL, USA

 

[MikeHalloran]

Halfway along the copper-colored sink, near edge:
Is that:
- a water mark?
- a witness mark from a plating electrode?
- bleached from heat?



Mike Halloran
Pembroke Pines, FL, USA

 

[itsmoked]

Mike; Sharp

there.

It looks to be an anodizing drip.. Nothing translates onto the board.

Rather than desoldering the existing caps I soldered on the back, more caps. So if the capacitance was dropping to a point where the zeros and poles were too badly misplaced it would drag them back. Made no difference. I guess I'll wait for them to cough something else up.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

I have been told they will pay several Bills for even an unsuccessful attempt. Yes! Release the hounds! So I am about to commence with a full scale shotgun attack. I will keep you posted.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MacGyverS2000]

Man, wish I could find a company that would pay me money for non-guaranteed results!

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Skogsgurra]

Wow! Thrill and joy are back to Eng-Tips! We will be with you all the way - either way it goes!

Gunnar Englund

http://www.gke.org

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

 

[itsmoked]

Well I've been dragged back onto this baby.
I can now get it to attempt to turn on.

It's 24V supply ramps correctly as does it's 5V.

It's 12V only ramps to 8V and the supply's jungle logic trips the entire supply off in a fault condition that lights the 12V fault light.

This all transpires in about 20ms from live switch-on.

I now can hunt backwards thru the unit to see where the ~8V is coming from.

In the photo below the the red area is the lair of the 12V switcher.

The blue area is the area, including the transistor switch on the heatsink, that cuts the 12V off instantly when the fault is declared.

The Green Area is the actual 12V switcher. This includes the switches on the heatsink. That transformer has around 10 pins on it. (It's rumored to be + and -12V but there are no obvious -12V outputs anywhere.)

My immediate question is why does the 12V cap, (top of green area), have it's positive pin hooked DIRECTLY to one of the pins on the transformer?

If this makes sense to you maybe you can tell me what topology I'm looking at. Then I can continue troubleshooting with one more arrow in my quiver.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MikeHalloran]

A real RS-232 port, common on machine tools, requires -12V at a few milliamperes. They are used to drive attached printers or terminals for doing data dumps. It doesn't have to be a particularly good supply, and may be weird.

<tangent>, not necessarily uh, unconnected.

There may also be a 120V socket in the machine tool, provided for the express purpose of powering said local printer or terminal. That socket's ground and neutral may not be connected where you think they are.

I.e., if you connect a data device to the machine's RS232 port, _and_ power the data device by plugging it into any line socket other than the one provided on that machine tool, ... you will instantly smoke the power supply of the data device, and some very expensive circuit board in the machine tool.

Twice.

I didn't believe anyone could possibly _design_ a machine tool that way.

That's when I started buying opto-isolated RS232 connectors for machine tools. Always.

</tangent>


Mike Halloran
Pembroke Pines, FL, USA

 

[itsmoked]

Excellent point Mike. This power supply actually has the old original RS-232 driver chip in it.

I don't believe there is any printer in the picture. There is a paper tape reader.

Another interesting anomaly: I scope captured the signals seen on the switch hooked to the transformer. There was total random AM hash on it, like the oscillation was uncontrolled and wild.

I doubled the filter cap capacitance - with no change.
Perhaps some small cap in the driving circuitry is bad?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MacGyverS2000]

See if that chip has a charge pump inside... I cannot recall if they made chips without them (requiring external voltage source), but if it has the pump, look elsewhere for an answer.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[itsmoked]

No, its the old original "88" chip. Four drivers and nuthin else. So -12V is required somewhere. Plus since this supply runs a mother board with drive controllers and a dense CPU board something out there probably needs a small amount of -12V anyway. Makes sense.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

This is a cliff hanger! Sounds like your circuit for the presumed 12 V is somewhat unorthodox. An unorthodox circuit is what I would be looking for if several individuals fail after prolonged time. Of course, I would look for traces of high temperature also. But you did that already.

Gunnar Englund

http://www.gke.org

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

 

[itsmoked]

Yes sir. And since there is the zener/logic circuit for disabling the supply on any out-of-range issues there could be an overload that never gets a chance to heat up the board.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

Any news? No gnus are good gnus - or not?

Gunnar Englund

http://www.gke.org

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

 

[itsmoked]

Ah.. I was pulled off onto a job I get no $$ for. (see shooting ones self in the foot)<sigh>.

I was able to get them a backup so the heat is off a bit.
I also ordered every single electrolytic capacitor ~ 30 - and they came in Monday. Hopefully I will be back on it later this week.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

Downdate:

All thirty two electrolytic capacitors were very carefully replaced.

The unit was powered up and now, not even the 5V logic supply functions anymore.

I was also handed 2 more identical <bad> supplies. One has a smoking habit. The other works just like the original except instead of a bad 24V supply it has a bad 12V supply.


If I can ever get the original running again I can at least compare the the alternate good sections between them.

I started drawing a schematic. After about 2 hours I found myself searching for new ways to kill myself.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

KUTGW!

'smoking habit' - good one!

Gunnar Englund

http://www.gke.org

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