Powering up an older KM3

[rbalister]

Folks,
I got a KM3 from work a few months ago and finally got around to trying to power it up. I wanted to find out what works and what doesn't before I went too far.

So far I've separated the control power from the spindle power so I can power up the electronics separately. I'm feeding it 120VAC and the results are:

1. Solenoid activates on the control cabinet interlock to lock the door.
2. Only three LEDs turn on inside the control cabinet. They're +/-12V and XFMR, all on the Fault/Logic board. No other lights are on, anywhere.
3. Nothing comes on at the console. No screens, LEDs, not even the internal fan.
4. One of the card cage fans and the power amp cooling fan doesn't come on. It looks like they're gummed up by coolant mist mixed with dust.

The controller is an Ultimax II. The machine has 10,000 hours on it and has never cut metal. It's has only cut glass, granite, and plastic - it was in an optics shop. They did some work on the spindle bearings a few years back. Other than that, I don't have much history on it.

From what I've read on the forum so far, things that need to be done for starters are:

1. Clean all of the crud off of the boards. There's plenty.
2. Clean and reseat the connectors.
3. Reseat all of the ICs on the PCBs.
4. Replace the Dallas chips.

I've got new memory chips and batteries but I'd like to wait until I've verified there's no critical parameters left in memory to recover, however, I may have to replace them to even get this running.

Any other suggestions as to what else to do with this machine? I'd like to get it running as opposed to scrapping it.

rayb

[Jackal66]

Ray,

There is a guy named "bloke" on here that works for Hurco. He can tell you anything about this machine.
Also" Buddy" at AMTS is a former employee of Hurco. He designs a lot of his own parts to repair & replace obsolete Hurco items.

Another is Fasto, Rich Carlstedt, Parkerbender, etc.

A lot of help on this site.

JAckal

[underthetire]

Bloke is your guy, but your gonna need to check the power supply mounted probably to the side of the card rack. should be 5.1 volts, 12 volts and 24 volts. Hurco will also tell you to check the AC ripple on the DC bus, so put your meeter on AC and check the dc bus, should be very low MV numbers.

[bloke]

Good evening, gents.
First thing I notice from your post is that you seem to have 115vac present as the door interlock works.
Next, check the power supply (gold or black mesh box about 12" x 5" x 3") has 115vac at the input side (marked L, N and GND) measure on AC scale between L & N.
If the voltage is present, Measure the output voltages on the DC scale between the following wires on the power supply:
grey to yellow 5V
grey to brown +12V
grey to orange -12V
grey to blue 24V

If some / all of these are not present, then it looks like new power supply time.

Another thing to check would be to make sure there is no swarf on any of the boards (including the console boards) that could be pulling a particular voltage down.

[rbalister]

Well bloke, it does appear to be the DC switcher. I saw it in the side of the cabinet last time I was working on it and meant to check it but ran out of time. I went to measure the voltage on the terminal block and before I even measured anything I noticed inside the cover a short silvered glass tube with silver end caps (AKA blown fuse). Well, there's the problem. Pulled out the supply and replaced the 4A fuse with a 3A (closest I had on hand) and tried again. Sure enough, it blew as well. The supply is rated to draw about 2.5A so the 3A should have been enough for a test.

So now I'm on the hunt for a new supply or repair this one. I didn't see anything obviously charred or burnt inside. It could be anything from a shorted bridge on the AC input to bad transformers. Whatever it is, it almost looks like a dead short across the AC input.

Do you know if these supplied care if they're run with no load? Some switchers don't like that and it blows the regulators.

[bloke]

That'll need a load applied to make it run correctly. Depending on what you want to spend on the repair, you could replace the 5v, 12v and -12v voltages with a PC power supply and then just add a seperate 24v supply for the 24v side of things. Cheaper and easy to fit.

[rbalister]

Well, between travelling for work trips and real work around the house I had some time to pull the supply and check it out. It turns out there were two shorted diodes in the input bridge. Newark was the first place I found that you didn't have to order 2500 of them. Expensive items at $0.20 apiece. Fourtunately they were on sale for $0.12 each. I got 12 and replaced all four and that fixed the supply. Reinstalled it into the control cabinet and now I have both screens lighting up with meaningful images.

The bad news is that I get a servo fault on all axes and the trackball doesn't move the crosshair. I tried the manual, power, reset, calibrate sequence but that didn't do much other than activate some relays. I couldn't get to the cabinet quick enough to see if any lights changed state. I still have no LEDs on the servo amp cards, however, on the dual axis cards I get 2 LEDs on the front board and one on the rear board. The trackball I suspect is dirty and gummed up from old coolant. The jog wheel turned the pot and loosened the jam nut instead of turning the shaft - same problem. Along with the diodes I got some PCB cleaner to wash down the boards and cage. That's next when I get another spare afternoon.

At least I got a version number on the software now; v8.50.

[rbalister]

After spending some time for in the smoke of the New Mexico high desert for work I finally got a chance to do some more checkout.

Some things I've learned:
1. Trying to cheat and send 120VAC directly to the DC supply bypasses the transformer for the servo bus voltage. Result, servo faults. Now I know that PBU on the Randtronics fault card means Power Bus Up.

2. When trying to clean your trackball, don't take off the black housing holding the encoder wheels. The LEDs and sensors have their silicon dies mounted directly to the PCB and are wire bonded (very delicate) to the copper traces. Result, a paperweight that's impossible to repair and expensive to replace.

3. Alternatives to the Hurco trackball exists. Some require more modifications than others. I've been able to locate one that looks like a reasonably close fit that uses a laser trackball as opposed to the mechanical encoders. Price is about $200US. It's made by NSI (P/N X50-70023D).

I've cleaned all of the boards, card cage, and fans. Turns out there are two bad fans, one on the card cage and one on the door. The Dallas chips have been removed and replaced with standard static RAM and lithium batteries have been soldered in place. The power has been changed to use 240VAC and the transformer wiring has been put back to original configuration. Now I get the console to come up and power to the servos. Resetting the servos and calibrating I get movement on X, Y and Z initially. The Z and X axes move and find home. The Y axis moves and then stops. Hmm, the Y is showing a Y+ and Y- limit. That can't be real. Taking the front panel off of the Y and moving the belt gets it off of the Y- limit, but the Y+ is still there. One more thing to fix now.

With the servos powered I hear what sounds like oscillations from the servo motors. Is this normal or a sign that the serovs need to be tune?

It looks like I might be getting close to something that works. What's left now is to fix the limit switch and get a VFD to drive the spindle.

[fasto]

Sounds like you're making good progress. I found the Y dual limit switch was totally full of rust, corrosion, and old coolant residues on my machine; it's also mighty expensive to replace. The plungers were very sticky which I cleaned. The internal switches were regular Micro switches and I replaced one of those instead o fthe whole switch assembly. If the switch is attached to part of the machine that moves you might check the cable, too.

With the servos powered I hear what sounds like oscillations from the servo motors. Is this normal or a sign that the servos(sic) need to be tune?

You have the Randtronics servos? Before I knew that my machine had Fanuc servos (an adventure all their own) I downloaded a document called "Rand Servo Setup" that described how to tune Randtronics servos. I think I got this from Hurco USA's support site or else it was attached to a post on this site.

[rbalister]

Thanks for the suggestion on the limit switch, however, mine are inductive switches not mechanical. Doing a quick check showed I had a bad limit switch. The original was made by Visolux. Visolux was bought by Pepperl+Fuchs. The original part number is no longer valid. The direct replacement is 101XK5-1 from P+F and is in current production. After replacing that I had full control of X, Y, and Z. All three move to their + and - limits and the machine will calibrate. Success!

I finally got the VFD (Teco JNEV-203-H1). I mounted it on one of the DIN rails in the magnetics cabinet. After rewiring the FWD and REV contactors to provide direction commands to the VFD it was time to power it up.

Once the VFD was set for 60Hz and everything was powered up, I had control of the spindle as well. It looks like the only thing left is the trackball. I have an old Logitech serial mouse that I'll rewire and use that in place of the trackball for now (remove the trackball and put a mouse pad over the hole).

I was able to run a short test that traced out a circle and it appeared to do just that.

The machine is now working. Now it's time to get some tooling!