[Alex_Cole]

Ok, lets see if you guys can help me out here. I am working on a servo system for my mill (dyna 2400 benchtop). I have setup my servo drives and tested them the other night and all was working fine. Now I started working on adding my limit switches.

I created a board that has 5v coming in then it is split off to 3 seperate 10k resistors. at the end of each resistor it splits with one wire going to my parallell port for the input and the other goes to a terminal where the switch is connected. The other end of the switch goes to the - on the 5v supply.

I only hooked up one limit switch and started to test my machine.

I jogged the X and y axis and everything worked fine. Then I tried to jog the Z axis and it moves fine going down but it seems like it is stuttering when trying to move up. This was working fine the night before? When it is moving up it seems like it is moving faster than when it moves down and it only stutters when moving up? I tried to compensate for the weight of the head my pushing up to assist the weight and it didn't help.

Things I have tried.

I tried slowing the jog rate to 10 ipm. same result
I tried to help by pushing up on head same result
I made sure there was no mechanical obstruction including making sure the system was well oiled. same result
I tried lowering the acceleration in mach 3 (was at 11) I tried it at 6) same result

does this sound like a servo tuning issue?

Servo drivers: The servo drives are an Hbridge design using pic micro controllers. These drives are running a router machine fine. These drives were running fine just the other night so I know that they work with this setup. The drives were tuned with these motors buy the guy that built them.

I guess I am not sure if I am struggling with a noise issue or motor tuning or what. Does anyone have any ideas. If you need more info let me know and I will provide it.

The guy I have been working with on this has been great but I am starting to feel bad about pestering him all the time with my little issues that I cannot seem to figure out. I am new to the electronics of this stuff but I am trying to figure it out.

Thanks

AC

[Eurisko]

It sounds like the problem started when you wired up the limit switch...

If that's the case, make sure that the wiring to the parallel port is correct, that the proper pin is connected.

I'd also use 4.7k pullup resistors and make sure that the 5V ground is properly grounded to the port. A floating ground can cause all kinds of weird problems.

[Alex_Cole]

I do not know much about electronics at this level but isn't the 10k resistor I have between the 5v and the port doing the same thing or is the 4.7k going to do somthing else?


AC

[Eurisko]

The 4.7k pullup simply provides a slightly lower resistance, and ensures that the pin will be at a high logic level. It is pretty much a standard value for TTL interfacing. Probably won't matter much, but better to play it by the book.

Try undoing the changes you made (disconnect the limit switches and undo ANY port changes). Test the axes again to see if they misbehave.

[Alex_Cole]

ok so I removed power to the board for the limit switches and stuff and tested again. The results this time are better but not quite there. I now can run XY and Z smoothly. So I decided to do a test cut. For my test cut I mounted a piece of 6061 aluminum to the table of the mill and put a
.005 dia engraving tool into the spindle. I wrote a program to cut a word. I set it up to cut .015 deep into the face of the aluminum. I have run cnc machines before and I run the Haas at work all the time. I know how to read gcode very well and am sure the program is good. I setup the program to run and it started out well. It cut the 1st letter fine but when it re-positioned to cut the 2nd letter the z didn't come down all the way to where it was before when it cut the 1st letter. it was a small distance above the part. It ran the path to cut this letter and then retracted in z to position for the next letter. This time it was a little higher yet. It continued to do this moving up in z each time until the program was done. By the time the program was done the z was about .150 above the part.

Now this is a little perplexing to me as this is a closed loop system using encoders on servos. Now, the z axis never faulted out at any time and I even tried running the program slower even though I was only running at about 10ipm anyhow. I got the same result at a slower feed. I am thinking that the rapid retract move in z going up is where I am having the issue.

Does this sound like I may have some noise in my encoder lines that is messing up my z position? if so could it be normal that it is in 1 direction?

Any ideas on this one would be greatly appreciated. I talked with my buddy about this who has been helping me with the electronics and he is trying to come up with somthing also, we just havn't had the time to get together and trouble shoot.

Thanks again

AC