Problem with old Makino EC7050 wire EDM breaking wire at higher currents

[JST]

One of our suppliers has this machine. They have already had the Makino folks look at it, with nothing past a "gee that's kinda dirty, you better clean it" to show for 3 days work. The Makino guys apparently admit to not knowing much about the machine, and I don't know if the Makino tech even ran through the calibrations.

With nothing to show for that, the owner asked us, being the electronics folks, to take a look at it, so I went over there today.

Essentially, when doing the same sorts of thing in the same sorts of material, using the same settings that used to work, they have had to drastically reduce current to avoid breakage. They used to be able to run at 10 to 15A, now they can't go above 3 with confidence.

Normally, the various things they are doing differently from recommendations would be suspect. But they used to be able to use those settings and get good performance, but now they cannot. Therefore, I am inclined to not put a lot of weight on the differences in what they are doing from what the manual says.

So, what I observed is that, yes, when they get up to about 5A pk current, set at 100V , they got wire breakage when cutting HSS for form tools (they mostly use it for in-house mold and tool work).

Sparking seemed pretty even through a 10mm part (tool blank), standard white sparking that was all in the cut. I did notice that there were intermittent bursts of red sparks, that were more similar to grinding sparks, and which appeared to exit the cut. At that time the cut was only about 50 to 80 thou into the part.

Not knowing so much about wire EDM I don't know if that is normal, I thought maybe not.

They had the wire speed set what appeared to be pretty high, with 20+ inches per minute wire speed. That's higher than the speeds recommended in the manual, and I would think should if anything, be easier on the wire in terms of abrasion/wear. The wire coming out at 3-4A current appeared abraded, but not too far reduced in diameter. I didn't mic it, but the original size was 0.25mm (0.01) and it wasn't a lot smaller by visual check. However, it was not *evenly* reduced, but seemed more rough than I expected.

We tried a current calibration per the manual, but didn't get too far, I suspect the manual doesn't quite apply to this machine. Measuring on the control input (TP310 and 311) while cutting gave approximately 1 mV per amp of current per the setting (about 2.3 to 3.3 mV at 3A set current), but seemed a lot more variable than I expected.

The machine was clean, with no evidence of bad tension, etc. Guides, contact blocks and ceramic wheels had been recently replaced, and were pretty clean. Dancer arm was steady. The mechanical parts appeared to be well maintained, and looked like the book.

Water nozzles were farther away than may be best, but apparently they have run that way for a long time and have not had problems.

The wire broke three times in about 0.1 inch of cut, always at 5A setting. It broke once in the middle of the part, once near the top, and once near the bottom (so it was not consistent in location). The one case that I captured the wire and looked at it, I saw that the wire was very inconsistently abraded. The broken area was thin, with thicker areas nearby, and other areas nearly as thin visible in a space of maybe 100mm (4 inches).

What I came up with is that the current is uneven , possibly from a defect in the control circuits. A noisy part or bad connection in the current measurement circuit could do that, or a similar problem in the current control circuits might also.

Any other ideas or suggestions?

 

[emcmike]

Not being familiar with the machine. What your saying about the wire breaking (thinning out) sounds like a frequency control. If it uses that as a control , I would check the capacitors. Under daily use most manufactures say about 5years before wear down. I've seen a lot that go 10 or better. In high frequency mode the amps are less and as the caps lose there working value the frequency gets lower and amps go up. I believe that's what your seeing at the breaker , not what the control panel is saying.
look to see if the caps have bled oil. Another thing is the cooling jacket clean and has good flow. The die-electric jacket for the wire feed leaking power, might need that meg tested.

 

[JST]

Not being familiar with the machine. What your saying about the wire breaking (thinning out) sounds like a frequency control. If it uses that as a control , I would check the capacitors. Under daily use most manufactures say about 5years before wear down. I've seen a lot that go 10 or better. In high frequency mode the amps are less and as the caps lose there working value the frequency gets lower and amps go up. I believe that's what your seeing at the breaker , not what the control panel is saying.
look to see if the caps have bled oil. Another thing is the cooling jacket clean and has good flow. The die-electric jacket for the wire feed leaking power, might need that meg tested.

Discharge caps were tested, came up at normal. This one has a bank of film type caps, not oil or electrolytic. The film caps have a very long life, and anyhow they checked out at capacity. There is a smoothing cap before that which is electrolytic, but it came up OK also, and ripple was not excessive, supplying another clue that it is OK.

Anyhow, I'm not seeing it at the breaker but in the return wire for the spark, which is where the machine senses the current for the table drive servo speed. I was reading the sensor voltage, same as for the cal check.

What I see is about the right current, but with a lot more variance than I expected. The sparking is visually and audibly quite even, but the current is not.

As for the dielectric, the water shows low conductivity (but not zero) while the spark is powered, but the machine is not back in the cut yet. Only a little current showed up in that condition, and it was nearly rock steady. As the wire got into the cut, it started sparking a bit and picked up when it got to the "working face".

The wire is obviously going to thin, just due to the normal action, I know that. This was intermittent very much thinner spots in a wire that was otherwise pretty normal for what I have seen... With the magnifying glass, the thin spots didn't look like something from particles causing excessive sparking at one place. Seemed like what I would expect for spots that had sparked at a higher power than that for the rest of the wire length.

But, again, I am not that familiar with the process to say if it is normal or not. They don't generally look much at the used wire, so they could not give much help about what it looks like normally, or after a break. I was only able to save both ends of the break one time, so I just had the one example.

Those red grinder-type sparks that looked like some bigger particles getting blasted off seemed to occur much more with higher currents, more than doubling in frequency from 3 to 5 amps. There were several seconds between those at 3A, but about 1 per second at 5 amps, (until the break, which took under one minute.) While I cannot prove it, they seemed to happen at about the rate that would account for the thin spots.. I suppose they could be from already removed particles getting zapped a second time, dunno. Kinda hoping someone here knows.

Is that red sparking a clue, or a "red herring"?

 

[emcmike]

A drop or low spike in the frequency control would make the red park ,unbalancing the voltage control circuit. But mainly want to look at the frequency driver. Intermittent Igbt if it has one or output mofset regulator.

 

[JST]

"Frequency driver"?

This is a pretty simple type, it's 20 years or so old, which is why Makino techs know nothing much about it. Voltages were looked at and are stable, so the variable is the so-called "on" and "off" time, along with voltage and peak current. This seems to be mostly dependent on charge resistors and capacitor size, which are both selectable.

The current regulation seems to be largely the movement of the table, which is normal, it drives the table speed to regulate the spark current. Narrowing the gap increases current, and widening it decreases.

But it seems as if something is messed up, since the current varies more than expected.

As for the red sparks, a DROP would lead to a small spark, which might be red, but these red sparks are larger "explosions" that shoot particles out of the water streams. Mostly you do not see anything but white sparks in the gap, you see nothing in terms of particles.

These red sparks seem to be hot particles leaving the area more like grinding sparks shooting off the wheel/work. It looks as if a much larger energy was suddenly applied, and a bigger piece was blasted off the work, shooting hot/molten pieces of it out. That would seem to line up with the current variation, and the intermittent thin spots in the used wire.

I'm suggesting an electronic cause, but that may be a case of fitting the solution to what I know, instead of letting the symptoms lead me to the cause.

I'd like to solve their problem, since Makino has essentially cut them loose.

 

[emcmike]

Check the boot straps (grounds) , what ever you call them on the table. These on older machines tend to get build up under the bolts , clean them really good ! Check all other grounds too , even the control board. After years of service the connections get like diodes and won't allow current to pass at full value. Make sure that movable table is good and clean on the electrical connections.
remember 20% of the metal your removing is just getting vaporized and gets into everything, particles much smaller then surface grinder.

I'm still working on one for a friend, not a Makino.

 

[greggv]

Not familiar with Makino machines at all, but when I had the same problem on our Sodick, it was the power cables, and the ground cables. Replacing those solved the problem. If your cables are old (3= YRS), I'd replace them.

 

[JST]

Thanks.

They did replace the ribbon type cables that go from the cap bank to the wire head and work which is I think what you mean. They say recently, last year or so as I understand it. So that should be good. Those suckers are expensive, too.

 

[emcmike]

Sorry for the multi post. At lunch I had a server problem and dupe the post.

 

[JST]

No problem.

As for all the grounds.... well, it would be cheaper to buy another machine.... there are a ton of screws. Most of the important wires have been dealt with, though, and the wire handling and contacting parts renewed.

We are next going to see about swapping the current programming boards around, in hopes of identifying a bad one. They seem to go in order, and if we can get one of the three swapped to the high end, evenif we do not fix it, that would help considerably. We are allso going to scope the current signal, since on my previous visit we didn't get a chance to do that.

The signal goes through one op-amp and a 4066 analog switch, then gets directly converted to digital then and there. Not much to go wrong, but a good look won't hurt.

The manager over there had little patience for troubleshooting, but the boss is much more interested in fixing it. He'll be there when I/we go back.

 

[JST]

No more good ideas?

We haven't got over there yet, so any more good leads on possible causes are appreciated.

I'm not as excited about swapping boards around, they are a bit buried, and I want to do more investigating first. My boss is one who tends to jump to a conclusion, and he's the one excited about swapping boards. The rest of us find that a good survey and investigation first is usually better.....

With old electronics, I much prefer to let anything that doesn't have to be disturbed lay as it is. I don't want to make things worse.

 

[rwbaldwin]

The only time I've seen symptoms like that, it turned out to be a holed dielectric filter. The little explosions and fluctuating current were due to tiny metal particles in the flow.

 

[JST]

The only time I've seen symptoms like that, it turned out to be a holed dielectric filter. The little explosions and fluctuating current were due to tiny metal particles in the flow.

That's what I suspected in terms of the red sparks. Would that be likely to also burn through the wire? Seems like it could.

The sparks are not the top-level problem, although they may be a clue. The real problem is wire breakage. If they could avoid that, they would be pretty happy, I think.

Any reason why it would be more of a problem with higher current? Maybe because it is actually not washing the particles out well enough as they are created?

Seems like it might also drive the current control to do a lot of things to try to compensate for the current peaks. Or not, the control systems seem to be pretty sluggish, and there is no visible change in the spark pattern or sound other than some red sparks. Seems to go good, and then the wire breaks. Good even sound of frying bacon, no changes that I picked up.

 

[rwbaldwin]

It also can cause havoc with the window discriminator/comparator part of the control circuit and this can lead to eratic motion of the workpiece.

 

[JST]

We'll ask about the last filter change..... I thought the problem had persisted through a change of the water purity stuff, but I got that second-hand.

I can see that if the wire gets shorted to the work, with the old style R-C current control, the current might get pretty high while it is shorted, and that could thin the wire until it breaks.. That's pretty much what it looks like, intermittent hot spots.

I don't think the controls are running the wire into the work, or we'd see it in the spark, but we might not notice a short-time interruption when a particle or several short the wire to the work and get blown away. That could be either the filter OR bad water flow, I'd suppose.

Thanks