[sunchiro]

Hello all. Have our Bridgeport R2C3 with a TNC145 making chips. Got everything working including using a Windows XP computer to store the programs and then transmit them back to the controller. Now have a new problem.

Last week was first one in limited production with it. Broke a couple of tools and figured it was just our bad programming, but now that we are doing larger runs of 20+ parts each it appears that the Z axis is not returning back to zero, rather it is loosing a bit in the negative direction on each pull up. After about 3 or 4 parts the lost distance is enough to require the knee to be dropped to compensate if you want to keep running. You can keep this up till you run out of z travel. We are using it like a large automated drill press so it's not critical, but have some parts to make where it will be (stepped milling).

Machine has the SEM servos and I don't think they loose steps. How do I trouble shoot this? Is it the servo, the encoder or the board or something else I should check?

As always, I really appreciate the help I have gotten off of this forum. Dennis

[gus]

Since you mention no error, the ony thing I cna think of is the pulley on the z servo being loose.

Try this

LBL1

Z I-1.000 F500

Z I+1.000 F500

LBL0

CALL LBL1 REP 50


Tell me if the numbers change on the display

tell me if the location changes as in the toolholder vs the vise/table

[sunchiro]

Thanks Gus!

Ok, ran the program you provided over several different iterations.

Ran it as incremental and then redid it as absolute.
Ran it with the spindle on and ran it with the spindle off.

Over 50 reps the z dropped negative by approximately 3/8 of an inch as measured over the several trials with my pocket ruler. None of the iterations above made any difference in the results.

The control showed the Z axis readout at the end of the operation was still at zero (the control did NOT know that the spindle had fallen) hence why it just happily kept running till a tool broke.

What do you suggest we do next? Is it a loose pulley like you suggest?

Dennis

[sunchiro]

I thought I would add this.

On the Z axis it will trip the +Z limit switch sometimes, but not always, while finding the reference points on start up. I mentioned this before and thought that might just have a bad parameter in there for the Z+

Mentioning it again, as after running this sample program to test Z I could not get the Z all the way back up to change the tool. The control showed a +Z limit switch error, but the spindle was down a good 3 inches. Nothing I did would allow the the spindle to go up past the point where it was, although I could easily move it within the range of where it was and lower negative no problem.

I cut the power and cycled it. It showed itself as past the Z limit on startup. I held down the power and cylcled it past the limit and then it referenced itself no problem and worked normally.

[machintek]

The encoder is on the top of the motor.
Thus there is a mechanical issue between the encoder and the quill.
I wold be focused at the mechanics between the motor and quill because that is where you are working against the forces of gravity. I would check the belt first. Both for tightness and quality of the teeth. A machine with a TNC145 is probably 30 years old. And as Gus said (thank you!) it could be the pulley on the motor as this has a taper lock on it.
Doubtful if it is a loose encoder disk but cannot rule it out.

George

[sunchiro]

After Machintek's comment we checked the belt and it didn't look too bad.

So we decided to check if the belt and pulley were working correctly as it was suggested the pulley might be loose.

What we did was make a mark on the belt and a mark on the pulley so they lined up.

Then we made a mark on the Spindle that was the reference for where the pulley and belt mark lined up.

Then we cycled it up and down 50 times and had that 3/8 inch negative drop of the spindle while the control said it was at zero.

Also, the two lines on the belt and pulley did not line up.

But, when we put it in manual and brought the spindle back up to where it should be as determined by the mark we had made on it, the marks on the belt and pulley lined up again as well.

I take that to mean that the pulley and the belt and the spindle are all moving together as they should. No loose pulley and no problem with the belt. Rather the encoder is wonky and wrong in it's position of the spindle.

I'm just winging it here with my son so we very much appreciate all of your input as to what we should try next.

[machintek]

Do the same thing but now place a mark on the pulley AND Z motor shaft and see if that moves. So close!

George

[sunchiro]

George,

We had already headed out of the shop for the night when you messaged. We'll mark the pully and shaft first thing tomorrow and report back.

By doing this what would be testing for?

[machintek]

To see if pulley is slipping on the motor shaft.
This is what Gus first mentioned.

George

[sunchiro]

Ok, did what George suggested.

You can see the marks. There was no question when done that the shaft was slipping inside the pulley, by the time we were done doing 50 reps it was off by several full revolutions.

So we took the belt off the pulley and it fell right off the shaft. Evidently the belt was holding it on.

There are three allen set screws and the pulley is in several parts. Kind of like a collet. We put it back on but for the life of us can not figure out how to tighten it. Or at least what we have tried has not worked.

Instructions on how to tighten this pulley back on its shaft anyone? Any type of calibration required after it is back on. As always thanks!

[John Bennett]

Looking at picture of pulley. You reinstall the three set screws into the plate that has 6 holes (three of which are threaded for the screws) but do not tighten. Place pulley back onto shaft with belt to make sure it is on the right distance. Now start to tighten the three screws (in a tightening pattern you would use when putting on a car tire). These screws press against the collet and tightens it. Side note, the small hole on plate, has a very small set screw which is used to lock the plate , which is threaded onto the collet body. Since this has not been altered, at this point ,you do not have to touch.

[John Bennett]

Pictures might help

[sunchiro]

Thanks for the pictures and instructions John,

It's back on the machine and we have tested it. No more negative Z drop. Thank you ALL!!!!

Thank you, Thank you. Thank you.

[gus]

glad you got it running just make sure to reset your z limit parameters. If you set them just before they touch the switch you might end up with a hair under 5 inches z travel. This machine is designed to never ever touch the actual limit switches, properly set up it never will

[sunchiro]

Gus,

Actually setting the software limits is something I want to look into as I would like to start it up without it ever going over limit. What I have found is the machine actually has two limit switches. On the z axis the first is mounted vertical and when the quill comes up it touches and the control says limit +Z limit switch. Then there appears to be another higher up mounted horizontal which is more like the button limit switch on our other mill that actually has bolt that comes in contact with the button that can be adjusted. What I notice is that when we trigger the first limit and get the message on the control we are still about 1/8 inch from the bolt depressing the button on the second switch. Is that second switch the actual hard emergency limit switch?

[John Bennett]

This is for the V2Xt with Dx32 control, but should I think be close, look at 10-13 for reference.

[gus]

This is not how this machine works.

There should be 3 switches on the z. One is a trip dog and that is what is used when homing. There are then a z+ and z- limit switch. Your machine should never ever ever ever touch those switches. You program the z+ and z- parameters so that you get the z+ limit error if you even try to go to a position that is past the software limit[IE z+5000]

when you fire the machine u p and it is not on the trip dog, it should travel up until it hits the trip dog, changing the display to non boxed or non bright or whatever, you hit the button again and it will reverse direction until it sees the encoder marker pulse, and the z axis display goes away and you then home the y or x[I changed the order of homing since as mine came it would try to clean off the coolant drains when it homed] In metric mode zero the z where it homes, move the axis down with the cover open so you can see just before it touches the lower limit switch[or just before it actually trips the limit] and the metric number on the display 'should' be the value for your z- parameter. The z+ parameter is actually above the trip dog[kinda has to be] and is again a small amount below where the limit switch trips.

You will now get limit switch errors without actually hitting a switch, which is as it should be, unlike stupid stupid machines that actually hit a switch to know that they are in the wrong place.

[gus]

just to add, if you hit the limit while setting it up just move it off by turning the pulley then clear the error and hit the enable button, no harm no foul, but it should never hit them while normally running

[machintek]

If I could make a suggestion: I also use a long socket head cap screw to align the other holes. Why? Glad you asked! To remove the pulley, you have to back off the set screws and thread in the socket head cap screws and tighten them to pop the pulley apart (by pulling it together).

George