Mazak CNC power supply repair

[HackedFridgeMagnet]

Hi
I'm a bit out of my depth here, no schematics, no datasheets (Yet).

I posted this on CNC zone but didn't get much response. So I will try here. I have no experience on these machines.

 My employer has a Mazak CNC lathe. The other day it stopped working.

 On startup the Power supply MDS-A-CV-260 shows and Alarm number 'A' for less than a second which means 'Contactor melt'. Then after that it shows 'n' which means 'Power module overheat'. Though it hasn't time to really heat up. Obviously nothing much happens after this.

 The main contactor seems fine and is showing open circuit when it should. So I think the contactor melt is a 'furphy' (slang for misleading statement)

 I took it apart and the 3 boards show no signs of damage. (edit see below)

 Can anyone suggest any fault finding tests I can do?

 Does anyone have a schematic, or can offer any other ideas?

 Ideally I would like to be sure the Power supply is at fault before it gets sent away.
 I am in Australia, can anyone suggest a good repair place? Not necessarily in AU.

I had another look and one small electrolytic is fried. I will replace and try. 
I want Datasheets on these proprietory components, if anyone has info.
PM200DKA060
PM100EHS060
Meldas DK495 (sub board)

Any help appreciated.

[HackedFridgeMagnet]

Some photos of the offending unit.
Sorry about the image quality.

I found a burned out cap C21 from the board with the busbars, which had caused damage to its PCB (shorted ~30v rail) so cut away the offending copper. And put in only matching cap I could find. Lol 'Suntan' brand apparently low Esr  and 105 degrees.

I will put it back together tomorrow and turn it on.

I am still not really sure about the about outputs of this PSU. It had two busbars marked 'P' and 'N'  so I jumped to the conclusion that it is AC. Wrong. (Phase and Neutral ? no. Maybe try Plus and Minus???!!)
It must be DC but I am not sure if it is Regulated, Unregulated or Variable voltage output. I am guessing the later but the lack of documentation is frustrating.
http://www.uscomponent.com/buy/Mitsubishi/PM200DKA060/ is as close to a datasheet as I can find. If you click on the picture you can see the little gate drive boards have a minimal schematic.

[HackedFridgeMagnet]

It lives! Moving, Cutting no errors.

Everyone is pretty happy about saving the 3.5k, I am surprised it was just cap and no other damaged components.

Still no schematic. Or useful info.

When I wired it back into mains I had connected phase to Earth until one of the fitters pointed it out. Its amazing how easy it is to make these small mistakes.


 

[EPTech]

Hi there,

I own a Leadwell VC25 milling machine which has a similar DC supply for it's spindle. The servo drives each have their own supplies. It is a Meldas M50 system (1995).

Usually these things have a SCR (Silicone controlled rectifier) bridge to produce a variable DC bus. The SCR's are the big (often white) power modules mounted on the heat sink with the aluminum bars screwed directly on to them. You have to look at it more as a power regulator than a voltage regulator. The capacitor bank served as a buffer to even out peaks in power demand. When the load draws current out of the bank, it's voltage will drop. The circuit senses this and sends more current to the bank and load by opening up the SCR's at a wider angle. a certain power but you have to look at it more as a power regulation than actual voltage regulation. When the load brakes, it sends back current to the buffer and the voltage will rise. The circuit also senses this and will in it's turn "waste" power in a braking resistor or send it back on the net.

If a capacitor has failed, I would replace them all. It is kinda costly but it will make sure the supply will last you many more years. Also replace the smaller caps on the switch mode supply that supplies the control circuit. It is likely located on the board opposite of the power  bank. It can even be a multi stage supply with a buck converter first to take the edge of the 800VDC folowed by the actual supply. No real need to change all capacitors on the board but concentrate on the ones that experience a higher frequency ripple.

Also be careful please! The capacitor buffer may hold many hundreds of joules of energy, enough to start several harts but to stop yours instantly. It  is not  the voltage that kills you, it is the potential energy
The buffer needs several minutes to bleed out. Check that the P-N terminals are below 30VDC before handeling. I use an external bleeding resistor of 1k 500W and when I am at approximately 0V, I short out the terminals. Why? Because the voltage in the buffer can ramp up again after you take away the bleeding resistor, especially in systems without internal bleeding resistors.

Nice fix by the way.

[HackedFridgeMagnet]

Thanks for answering! I really appreciate it as I have been scratching around for the littlest dribs and drabs of information.

We have ordered replacement caps for the caps that sit next to all the little driver boards (x7).  One of these was the problem, some of the others are showing possible leakage.  just 22uF 35V.

I wasn't going to replace the bank of big caps though, these looked ok, but I will mention your suggestion to the guys in the workshop.
Ok I will bleed the bank out before doing anything.

Nice to hear a description of what this supply does.
The big Semiconductors are black and marked on the 3 top connections  (from memory C1, E1C2, E2) so I guess its a totem pole igbts or transistors.

[EPTech]

Hi there,

I thought the bad cap was one of the buffer elco's. If not, I guess there is no immediate need to replace them all. If it were on of them, I would replace them all.

But like you said, replace corresponding caps in mirrored circuits. Maybe those caps are of a bad batch or lesser brand.

If the power modules are marked like that, they are probably transistors, more likely IGBT's. Be careful! If you ever disconnect the board from the power modules, I recommend you connect the small terminals together with a piece of wire or solder. Just wrap it around there. An accidental discharge (ESD) may destroy the gate (base) of an IGBT. But if you leave the modules alone after disconnecting them, you should be fine.