Need Help!- TC4 Spindle Parameters

[emvali]

During the last month or so I noticed the spindle speed is different from the one set. The machine is a TC4 with DX32 control and was always very precise, within +/-1-2 RPM.

I managed to observe up to 10 RPM difference - usually lower than the setting- and this number is unstable.
The same a few lines below, where the deviation percentage is shown, numbers play constantly between 7-12% and sometimes I am able to catch even 17% jumps.

This is with no load and no tool installed.

I read some threads mentioning the spindle gain parameters adjustment, specifically param. 71 which on my machine is set at 3.778076.
As I understand, there should be another parameter which is usually set at 250 and when changed can stabilize the speed within 1-2 RPM.
Which parameter is that?

By the way, I played with parameter 71 and managed to bring the speed to within 2 RPM of the set speed, but the percentage line still oscillates.

If this helps, I had a problem some time ago with rigid tapping, solved by changing the C11-1 parameter on the Yaskawa controller.

I also made a S0M3 test, the spindle drifts very slowly but no eror is generated.

[machintek]

Yes, in theory, PAR 71 is for adjusting the output of the BMDC so the Yaskawa spindle drive will do the correct RPM. But adjusting this will also have an effect on the spindle (S0M03) drift which must be kept below 2 RPM of the tool change will fail as the spindle drifts out of orient position. This is adjusted in the Yaskawa spindle drive.
The other parameter you mentioned is for tightness of the spindle when doing a accel or decel. I usually go to the maintenance page and do a rigid tap and watch the S axis position error. I adjust that parameter to achieve a following error less than one degree.
If you are having a stability issue, it may be that there is a component failing, a bad connection, encoder feedback issues, etc. There are a lot of small electronic pieces working in unison to achieve the correct spindle speed.

George

[emvali]

Thanks George,

I don't believe things are just happening, speaking logically.
This spindle difference between the set speed and actual speed started a while ago, but I ignored it. Then, I got the "auxfuncs: spindle time-out" error during rigid tap (regular tapping was ok). This is when I thought I solved the problem by adjusting the C11-1 from 21 original to something like 16 if I remember right, any value between 10-18 worked, I mean the rigid tapping was at least executed.

What happened next is the machine started to break tools - not immediately, but I had a part with some 30 threads and I worked very slowly on them (aluminum part) and after 18 cycles it broke the tool. This is something that never happened.

This is when I started to investigate the issue and went back to the spindle speed problem, I think I bypassed one problem but I left the original one unsolved.

I played with the par. 71, which was 3.778076, I changed it only a little, 3.783 gave me speed between +/- 2 RPM of the set value at all speed ranges.
After a while, I observed the computer "corrected" this value to 3.782something, but the percentage line still shows high oscillations.

OK, so my question is, how do I initiate the rigid tap from the maintenance page and which parameters should be adjusted in order to keep the spindle error at 1%.

Thanks

[machintek]

If you write a program with a rigid tap cycle in a loop, you can call it up and run it from the maintenance page. Look carefully. You can run a program, jog, MDI from that page and watch how the machine is positioning. You can also watch any switch (input) or output on that page. Run a SCANDISK, DEFRAG, and the B3T BMDC test.

But, you have other issues going on. Spindle encoder failures are common for those that are not Heidenhain. It is belt driven so it can be that the teeth are gone and the belt is slipping. Rigid tapping evokes a situation where the spindle becomes the S axis and is mathematically tied to the Z axis, otherwise you will pull threads or break taps. If the encoder starts to fail, you will have the same situation. BUT, again, the entire machine needs to work correctly to make a good part.

George

[emvali]

Got it, thanks.

The encoder is Heidenhain, and I will take the cover off tomorrow and see the belt's condition.

I tested it today and the spindle error is -1.092 to +0.941, jumps between these 2 values, I mean steady jumps, one time I see one reading and second time the other, every time these 2 numbers, I mean it is not running between them, but jumps from one to the other.

I made the test with no tool in the spindle, means no load and I used parameters long enough to observe the error.

The S0M3 oscillates from 1 to 2, from time to time I manage to spot a 3.

So far it looks ok, I guess I will have to run a part and see how it does.

However, as you are so helpful, I have another thing that bothers me with this machine:

I work on many parts with octagonal shapes, contours and pockets.
While everything is ok with rectangular and circular motion, a change of 30-60 deg causes a knok from the table.

It sounds like if I move on X for example and take a corner which is smaller than 90 deg and then move on Y (no matter if the first is X or Y), during the corner, when both X and Y are moving together it makes a knock like something is displaced and when changing to one axis at the second corner it makes a second knok like whatever moved goes back to place. The knock is heard at the change point and is a single one when both X and Y start working together and the second time when one of them stops. Maybe I imagine, but it sounds like the first one is different from the second, it is like they come in pairs, if I am not looking at the machine's movement and only listen to the knocks I hear them in pairs.... again, it might be my immagination.

I tried to hear where this knock comes from, I am almost sure I hear it from the X gear, on the right side of the table.
I disassembled sevral times all the covers of both X and Y, looked for a loose bolt or anything that moves and nothing. The knock is loud enough to have the impression something is moving, I runned the machine also with no Z and makes the same.

It quite drives me crazy, if you have any ideas maybe I will manage to keep a little hair on the head.

[machintek]

The DX32 control was not very good at 3D molds and such. If the motion angles were too severe, it would bang around or sound that way but in reality, it was not hitting anything. A few years back there was a class action lawsuit against Bridgeport because of this defect in their software but Bridgeport went bankrupt and the law suit was dropped. With a high degree of probability that is what you are experiencing. If straight line moves and circular interpolation is quiet, than I believe you do not have a mechanical issue.

George

[emvali]

I had no idea the machine has this problem.

The big question is, if this knock will eventually cause a mechanical problem or is just there and I should live with it.

I noticed that if I place a short brake at the point where the turns begin the knock dissappears in most cases. Maybe I can solve part of the problem by software.