The quest to fix a broken argon switchmode PSU

[Things]

OK well, a while ago I "bought" a free argon laser from Dave. He offered them without power supplies or any information. Thats fine, I got one anyway.

I originally intended to build a simple linear PSU, however my mains voltage being 240V, I can't just rectify mains and chuck it into a passbank, I'd require about $200 worth of transformers!

I saw that timelablasers had a complete, but broken argon laser for sale, so I asked him if he'd be willing to sell me the PSU. And he did

First obvious things were a FET that was cracked, and a burnt up resistor. He had already replaced a fuse next to the blown FET.

I bought some new FET's and replaced the cracked one, hid behind my kitchen window, and powered the PSU up. It didn't explode, but it didn't do much either.

There is a 200 ohm resistor on the neutral input line. Beside it is a relay. I have come to the conclusion that this resistor is to limit inrush current, and is bypassed by the relay once the PSU has powered up. However, this relay is clearly not doing anything, and the resistor just gets hotter and hotter. No lights or anything show on the board either, just 1 hot resistor.

So, I unplugged it, and being a switchmode PSU, I left it for a while, and tested the tank caps. They had no charge. This could be because they were already drained by other components, or they didn't get any power at all.

The only components between the mains input and those capacitors is the rectifier, and 2 big inductors. I have removed all the boards from the PSU, but have no idea where to start testing for dead components.

I would check the rectifier first, but can I do that while it is still on the board, and with a regular DMM? If the rectifier was dead, that'd explain nothing getting power, but it wouldn't really explain a dead FET, unless the rectifier was killed after the FET (Tube wires shorted or something maybe?) I don't know.

Here are some pics:

Considering the component count on this beast, I have absolutely no idea where to start. The fact nothing seems to get power points to a PSU issue though, so I guess I can eliminate pretty much all the logic circuitry.

Any help or pointers would be great If you need any more information, please do ask, I'd like to give this thing a 2nd life instead of having it sitting out in the garage for the rest of it's days

Cheers,
Dan

 

[LSRFAQ]

Congrats, you got the one that in 15 years of trying, I have never gotten a schematic for.

What usually fails in a ion switcher:

#1 FETS
#2 Whatever drives the fets GATE
#3 Its a buck switcher, so the fast commutating diode(s)
#4 Any caps on the anode bus.
#5 transorbs are evil and kill fets
#6, interlocks.
#7, stupid little auxillary psus that power the switcher.
#8 bridge rectifiers
#9 any small fuses
#10 innerconnects and sockets on the board.
#11 tiny board traces.



FETS usually die shorted if old age, or partially fail from gate "shoot through" taking the gate driver out. Therefore the FETS need tested under load with a bench supply. Any caps with 100-200V on them need tested with a bench supply.

You have about half a microfarad of gate capacitance there, so make sure the BEEFY gate drivers are fixed and working. That gate needs to transit from on to off in less then 200 nanoseconds.

1.7 max kilowatts is a lot of energy, and it likes to go places.

A friend hosts the Omni 150 schema here for me, so you can get a idea of the buck regulator concept, even if you don't have the same thing:

Omnichrome 150R Power Supply Check Out


4 FETS tends to imply a 240V european psu or a high current one

Find the pins on the PWM chip and opamps and probe around them based on the data sheets, do NOT assume the whole unit has a common ground all around the board, usually there is a hot side and a cold side. The cold side talks to the remote interface and is usually optoisolated from the hot side.


Good luck, Steve

 

[james1095]

Switchmode power supplies can be tricky. You have to be sure you get all the bad parts, or the whole thing can blow again the instant you power it up.

Often times when the switching device (FET in this case) fail, they short, and the rectified DC from the front end appears on the gate which can cook whatever is driving it. You also sometimes find that the power source to the controller chip is a simple resistor and zener shunt from the front end. Start by looking up the datasheet for the control IC and measure for voltage on the power pins. If that's there, look on the output that drives the FET for a square wave of several kHz.

Check carefully for open resistors, often a resistor will read below its rated value in circuit but it won't read higher unless it's bad. Diodes can usually be tested in circuit, as well as transistors using the diode test function on most DMMs. Normally the drop will be about .5V, or about half that for a Schottky diode. If it's not in that range, the suspect part should be removed from circuit for testing.

 

[daguin]

I have a 200V Spectra-Physics 263-A PSU here that you can have for shipping cost. Unfortunately, shipping cost is ~$90

Peace,
dave

 

[Elektrotechniker]

I have a 200V Spectra-Physics 263-A PSU here that you can have for shipping cost. Unfortunately, shipping cost is ~$90

Peace,
dave

It's better than shipping it toGrece :crackup:

 

[daguin]

It's better than shipping it toGrece :crackup:

A 14 pound package (like this one) to Greece (Priority) is only ~$70

This Spectra-Physics PSU is not as heavy as a JDS Uniphase PSU

Peace,
dave

 

[jayrob]

Dave, if Things doesn't want it, I sure would be interested in it if it could power my Argon that I got from you...

 

[daguin]

Dave, if Things doesn't want it, I sure would be interested in it if it could power my Argon that I got from you...

The challenge for you is that it is 200V. Do you have a 220V line available?

If so, I have a JDS Uniphase 220V PSU as well that you could have for the cost of shipping. It is heavier than the Spectra-Physics PSU (24 pounds vs. 14 pounds) so it costs a lot more to ship internationally. However, domestically, it would only be ~$30 to ship it to you.

With either of these, you will have to trace back the pinout (or get it from a schematic) and make some sort of "adapter" or just wire it straight to the laser.

@ Things -- If you are willing to take a chance, I also have a Spectra-Physics 263-C PSU (100V-240V) that I cannot get to start a laser. It powers up, but won't start the laser. It would have to be "repaired", but it is small enough to fit into a large, flat-rate box ($55)

Peace,
dave

 

[jayrob]

Hmmm... Out in my garage I have 220, but I'll probably just keep sitting on the Argon and hoping to see something easier come along...

Thanks though Dave...

 

[james1095]

The challenge for you is that it is 200V. Do you have a 220V line available?

If so, I have a JDS Uniphase 220V PSU as well that you could have for the cost of shipping. It is heavier than the Spectra-Physics PSU (24 pounds vs. 14 pounds) so it costs a lot more to ship internationally. However, domestically, it would only be ~$30 to ship it to you.

With either of these, you will have to trace back the pinout (or get it from a schematic) and make some sort of "adapter" or just wire it straight to the laser.

@ Things -- If you are willing to take a chance, I also have a Spectra-Physics 263-C PSU (100V-240V) that I cannot get to start a laser. It powers up, but won't start the laser. It would have to be "repaired", but it is small enough to fit into a large, flat-rate box ($55)

Peace,
dave

200V is easy, although a bit odd. Is it intended for Japan? A surplus 24-48V transformer can be used to buck 240V down to something close enough to 200, although the secondary needs to be capable of carrying as much current as the load requires.

Virtually any house in North America will have 240V available, clothes dryer, electric stove, etc. In a pinch you can even make an adapter with a pair of plugs that you plug into two receptacles that happen to be wired to opposing sides of the panel, then between the two hots you have 240V. Some apartment and condo buildings have a 3 phase feed and 208V instead of 240.