Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: SMA on December 16, 2011, 04:31:08 PM
-
Hello All,
I have a very basic understanding of the Driver Test but don't have other tests to compare to so it makes it difficult to decide what is good, ok and not so good. Any thoughts on this driver test and if there are some simple ways to improve it.
Thanks
Stefan
-
The spikes at the start, do they happen again and again whilst viewing the graph or are they just a one off?
Hood
-
Hey Hood,
Thanks for the reply. I didn't have a chance today to really check as we are super busy here in the shop. If memeory serves me well I believe they are always there.
I swear my troubles all stem from the Computer I am using. What does their presence indicate if anything?
S
-
If the spikes are always happening every so often then it could cause you issues, they are not huge spikes but could well be enough to upset things. Have you tried doing the optimisation steps outlined on the downloads page?
Hood
-
Hey Hood,
They are there constantly. They get a little bigger and spread out a little further when I use a 100khz setting but still exist when I test at 25. Here are some screen shots that I took during a test to show how they are always there.
-
Ok so I optimized. I didn't make much of a difference. Here is a screen shot of test after optimiztions. I did not complete the one that was described as more for laptops...maybe I should give it a go as well?
-
Well after optimizing there was no real change but then I started going through drivers and updating them. I think the issue may have been from an out of date monitor (?) driver. Now the test looks like better. Wondering if anyone else agrees. It does not resolve my issue of lost steps but I am working at eliminating the potential causes !
There are still some spikes that are momentary and generally throughout the test max vari is around 2.3 or less. The screen shot shows a pretty max spike.
Any thoughts.
s
-
That last one should not be a problem at all, the others may or may not be an issue, probably depends more on your BOB and drives as to whether it would or not.
Hood
-
Ok. The last on is where I am at now. It tests out about the same evertime and as I mentioned this screen shot shows it at it's worst. Is it possible to get better than what I have posted. Is it possible to achieve a dead straight line?
My machine runs very well. Smooth and strong but it loses position after cutting a file on Y. It has been a problem all along. I have a work around but would like to figure what it is that is the issue.
Another small issue is that the computer is losing a couple of minutes of time after being left on for a night. Wonder if this can be an issue.
Learning lots but feeling a little frustrated. Thank you for your comments.
S
-
If its just the Y that is the problem that would suggest a hardware or mechanical issue, have you tried swapping motors and drives around to see if the problem follows.
Hood
-
Yes I have tried swapping motors. It was both Z and Y. I have used coding as a workaround but would like to finally track down the problem.
One thing I did notice is that I have backlash unchecked. Wondering if that is the issue? Talked to the fella who made the machine and he feels it may be a Mach Version issue. He had tested on Version 2 and I am running Version 3.
I have gone over mechanical so many times. No binding, motors sound great and machine is very smooth.
This is a screen shot of my backlash config (or lack of it).
S
-
Leave backlash disabled at the moment as it will just add to the confusion if its not set up right.
You say you swapped motors, did that solve the problem for the Z?
What hardware do you have? Drives etc.
Hood
-
Hey Hood,
Backlash is not setup. The designer of the machine informed me that he had backlash enabled when the machine was tested and is going to send the parameters that were setup at time of testing which, I am told, the machine was running perfectly and was not losing position at all. The big issue I believe is not owning the computer that the machine was setup with orginally.
This has been an ongoing problem. The machine preforms a move called the crossover. After machining one side of the surfboard the machine crosses over to complete the other half. When it did this Z would end up about 2mm lower than it should be cutting one side of the board thinner than the other. Swapping drives did not change a thing. I added a G28.1 command that would of course send z back to home switch and "find itself" and then move down to cut the board. Essentially I coded out of the problem for now.
Y loses itself in the same fashion. About the same amount out as Z was lost. I have added a 28.1 that has reduced the issue to a manageable point but again would like to find what the issue is.
I swapped drives (only x has been consistently spot on) and quadruple checked any kind of binding issues. I did fresh installs of MACH and just recently optimized the system as well as updating drivers (which did improve the driver tests).
There are "sensitivity" settings on the drives which can be increased. This is about the only thing I have not toyed with but will be going there next if all else fails.
I believe the drives are Gecko's. I will investigate the details and post asap.
Thanks once again !
S
-
Ok now you menton surfboards I remember the macine in question :)
Sadly I cant remember all the steps you went through before, I am wondering however what you have the pulse widths set to on the motor tuning page?
Hood
-
It has been a bit of a slog. I keep finding work arounds but I can't help but want to understand what the root of the issue is.
I have the motor dir pulse at 3 and direction at 2. This is the case for all axis.
-
Ok that will likely be fine.
I seem to have it n my head that the axis setup on your machine is not really conventional, could it be that the two axis you have issues with have the most weight on them?
I am wondering if it is your motors are just on the small side for what they are meant to be doing. The builder may well have tested but how good did he test? Do you know if he ran similar jobs to you are doing and was getting the accuracy?
Hood
-
If I remember correctly the builder used some weights to simulate a load on z. X and Y are both horizontal axis. Z must overcome the strongest forces. U could very well be right about undesized motors or gearing of ? These are steppers at 3:1 I believe. I will check in the am as to some more precise information on gearing, motor type, and driver type.