[Cerbera]

Hi all,

I have a 1996 Yang Eagle/SML-12 with a Mitsubishi Meldas 50L control I am trying to get operating again. Long story short, it's been in storage for a while (>3 years) and I am only now trying to get it operational.

In typical fashion, the control battery died, so I replaced it and attempted to reinitialize the machine so I could get to re-inputting the parameters. CS2 to 6, wait for 7-segment to display 60, off, CS2 back to 0, power NC, set type and input power, power cycle NC, etc. However, the instructions say at this point the LCD should come back and start to display things in Japanese, so I'll need to change the language setting. Instead, I get a blank screen with only the cursor flashing.

I've tried a few things, like instead of returning CS2 to 0 after re-initialization, move it to 2 (manual says this pulls initial settings from EEPROM?) - no difference. Tried inputting other settings during that initial prompt immediately after reinitialization for the control type and input power, but there was no difference. I noticed immediately after reinitialization and before power cycling the NC, it does throw an alarm of S01 20 X, while the corresponding X servo drive status display flashes F1 20. I checked over the wiring to the encoder, but figure this is a bit of a red herring until I get the control parameters input in. As a test, I disconnected the servo drives from the main control and it has the same behavior of giving a blank screen after reinitialization, just throwing a Y3 01 -- error on the 7-segment instead over nothing being connected to it.

I called Mitsubishi Electric about this and one of the gentlemen there said the control likely needs to be serviced. Normally, whenever the NC is powered and not throwing an alarm like it does after reinitialization, it displays 0.4. on the 7-segment. The rep said this is a code indicating the control has had something interrupt the boot sequence and is hung up on it. I tried finding some sort of literature about this myself on the internet, but couldn't, so I'm assuming their techs have some more detailed information I'm not privy to. He also elaborated that the alarms were likely red herrings, but apparently there was a previous case file about a similar control needing capacitors replaced due to age-related failures, causing miscommunication with the servo drives. He suggested it was possible our control unit had a similar problem and to send it in.

I was a bit short on cash, having just paid for the rigging for the machine to bring it to its new location, so I decided to look into the control unit myself. Sure enough, I found a leaking capacitor. I believe it was the C17 one from my notes. I had an electrical engineer friend inspect the board, and his assessment was the C17 capacitor was the only one that leaked and there was no damage to the PCB. There was a neighboring -12V linear regulator that had some burnt residue on the pins, but this was likely electrolytic fluid dribbling onto it from that C17 capacitor and burning up when it shorted. I contracted his help in having the linear regulator and all but the largest capacitors replaced on the board as insurance and put the control back in the machine. Given the failed capacitor, we figured this was a slam dunk repair and it should resolve any issues the control had, but essentially, nothing changed. Still a blank screen after reinitialization and initial settings input. Now, we still do have those largest capacitors we can replace (the big 200v, 270uF snap in caps), but we skipped those over given how many smaller capacitors hadn't leaked yet (according to EE friend, the larger the capacitor, the less likely it is to burst, and he had also done a connectivity check as a rudimentary health check), so I'm doubting it will make a difference.

It's increasingly looking like I will have to send the control to Mitsubishi, but I wanted to consult with you gentlemen if any of you have experienced a similar issue. I've done a lot of fruitless searching to see if anyone else has gotten hung up in the same spot I have, but haven't come across anything yet.

Thanks,
Luke

[MrMetric]

Unfortunately I don't have a 50 series control, only an M3, so I'm not sure I can help very much. However, I can tell you that leaking caps is a problem on the M3 machines. The interesting thing, though, is that this seems to be more frequent on the amplifiers than on the control boards. To a degree this makes sense simply because there are more capacitors on the drives too, though. Even so, I'm a little surprised how well the control board caps look versus the drive amps.

So, a couple of things here.... Have you checked the voltages on the control power supply for both level and noise? Your EE friend should be able to help you out with the noise side of things as that will require an oscilloscope.

Next, I'd try to narrow the problem down. Bear in mind that I don't know your machine, and I have never even seen a 50 series control. This is my VERY pointed statement that you assume ALL liability for both the machine any your body, including death. I am an amateur, not a certified Mitsubishi technician or even a professional CNC service person. Working on these machine can absolutely be dangerous and can kill you! So, liability and responsibility statement done....

Narrowing down... On an M3, and I *think* the rest of the Meldas line, the drives are complete separate devices. To avoid interaction, you can consider removing the drives out of the equation entirely. Most have circuit breaker that you can switch off and you can disconnect the data cable from the control cage. It is likely daisy chained to all the amplifiers and spindle drive. The idea here is to eliminate stuff that you don't need and that *might* be causing a problem. You are essentially isolating the card cage. Now, clearly the machine won't work this way, and it will likely throw lots of errors, but the important thing is whether or not you have a difference in behavior. If your display suddenly pops up, then you know the issue is downstream of the control cage. I strongly suspect that will *not* happen and you'll still have the dead display, but it is a diagnostic step.

The drives themselves should require positive control from NC unit control cage to actually energize, but it is still best to be as "safe" as possible. That is why I'd flip the breakers to the drives, if you have that option. I'd also make sure that the estop mushroom is locked in the E-stop state. Usually, but not always, this is a direct connect into the drive itself so that it cannot energize. The M3 drives actually won't even charge the high voltage DC bus without communication with the NC unit. Mitsubishi 'grew' their designs from a software perspective, and I suspect they did the same from a hardware/safety perspective too (but, as always... your responsibility to verify, be safe, and accept all liability/responsibility).

You should also reseat each of the boards a few times. This could be an issue of corrosion/crap on the connector pins. By reinstalling the boards, you are effectively wiping the pins and that may improve the connections.

I'd also look at the display head carefully. If this is like an M3, the display is actually 120VAC and obtains the video signals from the main NC card cage. However, there are still a few boards out in the head that provide an interface point between all of the buttons and the NC card cage. This makes it so that a gazillion cables don't have to be wired through the long tube that separates these pieces; instead, everything can be condensed into a communications stream, thereby reducing to only 3 or so wires.... There should be some blinking lights, or similar, on this coalescing board to signify communication with the NC card cage. If you don't see that, then that too is a diagnostic step. The fact that you *are* seeing a cursor tends to indicate that you have power and whatever is providing the video signal is, in fact, providing something (a cursor might not be a character, but it is still a video signal).

Finally, on the M3 there is a row of DIP switches that have to be configured correctly. If those are not set properly then you'll definitely see zip on the screen. I know it sounds stupid, but you should confirm what the appropriate settings are and make sure that everything is correct. It is better to say "that is dumb" and check it than spend 5 days searching for the unknown (and probably breaking something else in the process) only to say, "what an idiot I am for not checking those stupid switches first!"

Oh... Mitsubishi.... They are mixed bag. I've talked to people there before that have *excellent* knowledge of their controls and are super helpful. I've also talked with a few that have not really impressed me very much. The latter are kind of the "you turned it on and it didn't work? That is bad... you need to have a technician come out and look at it" type. Based on your comments, I tend to think that you probably had one of the good guys. Even so, take what they say with at least a little bit of questioning. Don't get me wrong.... I respect the MEAU support guys. Dealing with these older machines is not easy, and they've generally been very friendly and knowledgeable. But I'd still not immediately assume that they are always right.

Best of luck!

[meric]

Hello Cerbera,

I have a similar problem with Meldas60. There is only the cursor is flashing. did you find a solution with your machine? If yes would you please share with me.

Thank you

Hi all,

I have a 1996 Yang Eagle/SML-12 with a Mitsubishi Meldas 50L control I am trying to get operating again. Long story short, it's been in storage for a while (>3 years) and I am only now trying to get it operational.

In typical fashion, the control battery died, so I replaced it and attempted to reinitialize the machine so I could get to re-inputting the parameters. CS2 to 6, wait for 7-segment to display 60, off, CS2 back to 0, power NC, set type and input power, power cycle NC, etc. However, the instructions say at this point the LCD should come back and start to display things in Japanese, so I'll need to change the language setting. Instead, I get a blank screen with only the cursor flashing.

I've tried a few things, like instead of returning CS2 to 0 after re-initialization, move it to 2 (manual says this pulls initial settings from EEPROM?) - no difference. Tried inputting other settings during that initial prompt immediately after reinitialization for the control type and input power, but there was no difference. I noticed immediately after reinitialization and before power cycling the NC, it does throw an alarm of S01 20 X, while the corresponding X servo drive status display flashes F1 20. I checked over the wiring to the encoder, but figure this is a bit of a red herring until I get the control parameters input in. As a test, I disconnected the servo drives from the main control and it has the same behavior of giving a blank screen after reinitialization, just throwing a Y3 01 -- error on the 7-segment instead over nothing being connected to it.

I called Mitsubishi Electric about this and one of the gentlemen there said the control likely needs to be serviced. Normally, whenever the NC is powered and not throwing an alarm like it does after reinitialization, it displays 0.4. on the 7-segment. The rep said this is a code indicating the control has had something interrupt the boot sequence and is hung up on it. I tried finding some sort of literature about this myself on the internet, but couldn't, so I'm assuming their techs have some more detailed information I'm not privy to. He also elaborated that the alarms were likely red herrings, but apparently there was a previous case file about a similar control needing capacitors replaced due to age-related failures, causing miscommunication with the servo drives. He suggested it was possible our control unit had a similar problem and to send it in.

I was a bit short on cash, having just paid for the rigging for the machine to bring it to its new location, so I decided to look into the control unit myself. Sure enough, I found a leaking capacitor. I believe it was the C17 one from my notes. I had an electrical engineer friend inspect the board, and his assessment was the C17 capacitor was the only one that leaked and there was no damage to the PCB. There was a neighboring -12V linear regulator that had some burnt residue on the pins, but this was likely electrolytic fluid dribbling onto it from that C17 capacitor and burning up when it shorted. I contracted his help in having the linear regulator and all but the largest capacitors replaced on the board as insurance and put the control back in the machine. Given the failed capacitor, we figured this was a slam dunk repair and it should resolve any issues the control had, but essentially, nothing changed. Still a blank screen after reinitialization and initial settings input. Now, we still do have those largest capacitors we can replace (the big 200v, 270uF snap in caps), but we skipped those over given how many smaller capacitors hadn't leaked yet (according to EE friend, the larger the capacitor, the less likely it is to burst, and he had also done a connectivity check as a rudimentary health check), so I'm doubting it will make a difference.

It's increasingly looking like I will have to send the control to Mitsubishi, but I wanted to consult with you gentlemen if any of you have experienced a similar issue. I've done a lot of fruitless searching to see if anyone else has gotten hung up in the same spot I have, but haven't come across anything yet.

Thanks,
Luke