Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Set Command Variables

#32515

Set Command Variables | 9 February, 2005

Hello there!

I am familiar with the Set Command Variables available on our older UIC through-hole machines, and am investigating the possibility of making these variables available to our operators. Over the past years, our operators have been trained to alter the pattern files in order to overcome mechanical insertion errors. This has caused many headaches and quibbles on several machines (mostly the VCD and Radial inserters) since many of them are run on multiple shifts, by multiple operators, with varying degrees of experience.

My belief is that the Set Command Variables provide a method of overcoming basic errors caused by wear and tear, without "adjusting" individual lines of a pattern file to compensate for something that will ultimatly affect EVERY product and thus, its corresponding pattern file. I believe our maintenance program should be able to correct these mechanical problems and return the variables to their factory defaults. I would like some more input on this, since I am hearing several different views from maintenance, UIC, and other sources who deal with this equipment.

reply »

jdengler

#32557

Set Command Variables | 10 February, 2005

Primus,

Operators should NEVER have the ability to alter patterns. The altering of a known good pattern is a serious event and should be given careful thought before it is done.

I worked at a company that operated UIC equipment 3 shifts a day 5 days a week. When I started we had strict controls on patterns and preventive maintenance. Because of this we had less than 1% average down time on all UIC equipment. Due to restructuring we lost control of the preventive maintenance program. This started to cause a large increase in down time. Management decided to allow "temporary" changes to patterns to reduce this down time. All this did was cause more down time because a change was never temporary.

Under these conditions, when a machine is mis-inserting you don't know if it is the program or the machine. This causes a delay in getting the machine back into production.

Finally we went back to doing preventive maintenance as it should be done. Placed strict controls on Pattern editing and got things back to the way it was. It took some time to undo all of the problems that were hiding in the equipment and patterns due to 2 years of mis-management.

Allowing Operators to use the set commands can be dangerous as I am sure UIC has told you.

Good Luck, Jerry

reply »

#32582

Set Command Variables | 11 February, 2005

Actually, I have never had the opportunity to discuss this with UIC. In fact, our Maintenance Engineer has told me that, according to UIC, the Set Command Variables should never be altered from their default settings except by qualified personel, and only as a temporary fix. Under these circumstances, my company has actually accepted, almost encouraged, the operators to alter the programs that are not running correctly. This is done even when we KNOW that there is a machine error; we simply alter each program individually, as we run it, in order to make it work. This practice used to be okay when there was only one operator, because that operator was (and still is) compotent to know what he was doing. We have since expanded to three shifts, and due to a lack of structured training, each generation of new operators learns less and less about the machine. Now we are at the point when new operators are cursing the machines because the programs are wrong, and cursing the other operators for "sabatoging" the programs when in fact something is wrong with the machine.

So you've answered half my question, that the programs should not be altered under any conditions. This is something I understand, but what about the Set Command Variables? Should these be available (with proper training) to operators who are currently trying to adjust programs?

reply »

jdengler

#32585

Set Command Variables | 11 February, 2005

Simple answer no.

Do the operators currently use set count and set sequence commands? How many mistakes do they make with just these 2 commands?

Then in a few years you will find your operators have turned over again and you will have less knowledgeable operators trying to do things they don't totally understand. It is easy to make a mistake in this area and cause machine damage. Also to do this correctly they would need to bypass interlocks, which also puts them at risk for injuries due to a mistake in the set commands.

Rarely do you improve a machines operation from these commands and you have a high risk of machine damage or operator injury. Because of this I would not allow operators to use the set commands.

reply »

#32586

Set Command Variables | 11 February, 2005

We use these machines to insert swedge pins. When I first started here, there were certian programs that we ran that required us to put in a machine offset of +.120 in the X which is a lot. after learning the machines better, I realized that whoever had written the programs had bad insertion data and just never fixed it. I went into each program and altered the insertions, and step and repeat steps. No problems since then. We do get batches of boards that may be mis-drilled sometimes, .003-.006 so when I set up the machine I load the program and step it through the insertions with the head off, and chek them with a jig we made. I will then put in a machine offset. When the insertion is dead nutz, we run. It is a pain, but sometimes you have to run the boards you are given. Just don't forget to put the machine offsets back to zero before loading another program. We do not have board error correction on this machine. If we did, I would not have to go through the above steps.

reply »

#32590

Set Command Variables | 11 February, 2005

As for letting operators alter a program or machine offsets, I would not advise it.

reply »

#32602

Set Command Variables | 11 February, 2005

The operators use SET SEQ and SET SEG. I know all about getting these wrong, because if you forget the "Q" or the "G" you get SET SE which is SET SENSOR. That REALLY messes up our BEC! But then again, I believe our BECs are all calibrated wrong anyway. For example: A BEC step (we'll call it BZ) is programmed to read the left hole of C4, which is a BC inserted component at +04500 +03200. Since C4 is a 0.200" lead span component, the BZ step should be +04400 +03200, right? Not on our machines... it might be like, +04215 +03503 or some crazy thing like that.

That's why I want to know more about the SET COMMAND VARIABLES. I have used them to make a new program run without altering it from the data. It worked.

As far as BEC we are now trying to use the Dual BEC function (skew correction) since we have as many as 100 window panels... reading EVERY SINGLE UNIT gets rather tedious, especially when we need to use the Board Reject (BR) function as well. What really gets me is that none of our machines, any of them, will run without some type of BEC implimented, and the ONLY reason for that is not because the board itself has errors, but the table isn't straight! And we can't get anyone to admit it or fix it! The bulk of our holes are drilled with a Pluritec Giga 8888, which I must admit is pretty accurate when the operators and QC are paying attention to things like broken bits and runout.

Another common operator mistake when using the SET SEG command: even the UIC manual warns that it is VERY EASY to put the machine out of sequence if this is done incorrectly. And no matter how you try to train them, they never understand exactly what they are supposed to do to set a segment repair. But I can't see that as the operators fault, because nobody wants to try to use the set commands properly; the subject is taboo, and so the operators never learn what's really going on.

Oh, sorry, this was in response to the last thread, hope nobody gets confused... :) my bad

reply »

#32608

Set Command Variables | 12 February, 2005

I have seen BEC holes drilled wrong also. On a couple of products We run, We insert a Pem Nut. The insertion hole is too large for BEC to read. So We have the board house drill a small hole for each array, circuit. If these smaller holes are drilled a little off, or I think that the machine is drifting a little bit, I will modify the X-Y position of the actual BEC step and check it with the jig. This way it will make the correction by using basically a BEC offset instead of having to alter a lot of different steps and save You from a bunch of # crunching and data entry. Hope this helps a little.

reply »

#32610

Set Command Variables | 12 February, 2005

We don't have any automated machines (UIC) that insert anything into holes larger than our terminals, so about 0.050" is as big a hole we ever have to use before we've plugged them with something. Since we also do a fair amount of Man-U-Sert (yes, you read that correctly, we still use them!!!) there's almost ALWAYS a nice little pinhole to use.

We also use two Zeirick terminal inserters... the BEC for that is called Dipsy... she runs our DIP inserter too!

reply »

#32613

Set Command Variables | 12 February, 2005

I think you may be assuming that our operators aren't already doing that. We need to be able to bypass the interlocks... how do you think our operators make line adjustments? Step the machine, drop the tooling, and eyeball it. That's it. The operators FULLY understand the dangers of the machine: a large rotating bevel-edged metal disk capable of crossing 24" range at a high rate of speed is a dangerous looking thing! My feeling is that if the operators are already making adjustments, then we should be able to use the set command variables INSTEAD of altering the programs to achive the same results. The holes in the board haven't moved (though it is hard to convince some people of that) so their coordinates shouldn't. If a product was running at 99.44% accuracy Monday and is down to like, 74.69% accuracy by Friday, and the PCBs are from the same lot number, then the Machine has changed. Chances are the table has drifted due to the encoder belts stretching, or some kind of parasitic resistance in the servo cards, or some crazy thing like that. But the boards are the same, so the program should be the same. The machine needs to run, but maintenance can't just shut it down to recalibrate it for a few thousandths of an inch every week. So the operator looks at the components and determines that they need to move X+4 Y-3. Then the operator uses the SET ERROR to see if the BEC is still reading +00000 +00000. If yes, then the SET MANUAL-OFFSET.X.Y needs to be adjusted. If no, is the SET E reading -00004 +00003, or close to it? If yes, then the BEC tower has drifted. Adjust the SET SENSOR.X.Y That's how it should run, I think... does anybody else do it like this? I know we don't, but should we?

reply »

jdengler

#32633

Set Command Variables | 14 February, 2005

Yes maintenance should shutdown the machine and align it once a week.

reply »

jdengler

#32634

Set Command Variables | 14 February, 2005

If you already have trouble with two set commands, you will get many more when you add 3 more.

Maintenance MUST learn to align the machines correctly. The BEC scenerio you describe proves that the alignment is wrong. The table can be adjusted for square so you would not need a skew BEC reeding.

Send Maintenance to UIC for training or have UIC come in. Maintenance is the problem, this is where you need to fix it. Allowing operators to alter set commands will cause more problems not less.

reply »

#32671

Set Command Variables | 16 February, 2005

I would agree with you there, but I am only a shift supervisor. I am do not actually work in the maintenance department, although I do perform many repairs to the machines. So I've learned quite a bit on my own, and it was only recently that I was shown the programming guide for our Radial 5, when I learned that many of the theories I had devised were actually true.

It seems to me that MAINTENANCE should be responsible for any changes to the SET COMMAND VARIABLES, and a few senior operators should also be trained to use these when it is possible to prevent downtime (such as a slipped encoder belt or a misalignment of the head because it was removed to replace a broken tool, etc.) Then again, these "operators" would probably need to actually follow the maintenance program as strictly as any other mechanics.

As far as maintenance goes, I believe that our company is doing everything they can to improve their current program, and have recently decided to take an even larger step in the amount of prevenetive maintenance. For example, we will probably be replacing more components due to wear, rather than waiting for things to get REALLY bad. We will probably be performing more alignments, and I must suggest that they have their "meter box" checked for accuracy. I believe it's the culprit in our BEC problems.

As far as table skew goes, there are two possible causes for that which I am aware of. There are probably more adjustments possible; it would require a UIC tech to actually get it right I believe. I have also talked with our production engineer (my boss) about becoming more involved with the engineering process and how it relates to the auto-insertion department. I am hoping that this will give me the leverage to begin restructuring the operations in this department.

And everything you all have suggested has been very helpful. I look forward to continuing discussions... we will be purchasing more equipment so I may start a few more threads. Thanks again to everyone for the insight!

reply »

#32673

Set Command Variables | 16 February, 2005

Align the machine... are we talking about the table, the BEC, the clinch, the insertion tooling, EVERYTHING? That sounds right to me too, but I don't know if we have the capacity to do that at this very moment.

reply »

reflow oven profiler

Thermal warpage test services