PDA

View Full Version : X is Drifting off Coordinates While Cutting



zeke
06-08-2009, 07:05 PM
I setup 2 Partworks files to cut my Bot-table "through holes" and another with "recessed holes" for the bolt heads. I started out creating both of these in one file to match up the X and Y coordinates. The through holes cut fine on Saturday, yesterday I ran into some other issues (in another post), got beyond that, thought I was golden, not so much. In fact, I pulled up a chair after all the setup work, first w/e cutting and was smiling from ear to ear thinking I had arrived, little did I know my sidekick was up to something
. It's probably me, can't wait to see what is causing this....

I separated out the through holes from the recessed holes in 2 separate SBP files. The x and y coordinates are exactly the same, I verified twice. However as the Bot moves/cuts from hole to hole, it begins to decrement slightly on the X axis. The end result is that on the last few rows, the recessed hole is offset. I included two pictures one at the start and the other at the end row.

Last night I verified that the X and Y at the start of the cut were physically in alignment with the first hole and slightly adjusted SB3 to correspond with my file coordinates. I then executed the same file with the same result, there appears to be a gradual loss in the x coordinate. Today I checked the x and y again back to the first hole and the x is off by about a 1/4 of an inch.

One of the things I am going to do tonight is turn off McAffee to make sure it's not eating too much cpu/memory and slowing down the SB3 processing and try running the same file again. I think I'll monitor the computer processing during the cutting. I never went any higher than 1.25 on the speed and tried some lower settings as well.

Also tonight when I was typing in various coordinates and this is the first time this happened, the stepper motors made two consecutive banging noises, I then hit the stop button.

See the first and last holes below. Anyone run into this problem, any ideas?


3948
3949

zeke
06-08-2009, 09:06 PM
I went ahead and disabled Macaffee, pulled the plug on my garage door opener, checked that all power cables were separated from the data cables all the way back to my laptop. I have full echain on X and Y with dividers so those should be OK. Moved my dust collection system about 6 feet away and I still get the same results.

I confirmed it is only the x that is physically drifting off the logical coordinate, the y is not affected. By the time it gets to 50th hole is out by about a 1/4". When I move the x back to zero it is not in the same physical place it was at the start of the cut it is definitely drifting about 1/4".

Anyone else experience this behavior? I emailed Technical Support my files and pics, I'll contact them tomorrow.

dana_swift
06-09-2009, 09:07 AM
Zeke that is seriously wrong. Your profile does not mention what shopbot you have. If it is a standard, consider slowing your slew and cutting speeds and see if the problem goes away. If so, the steppers are losing counts.

If it is an alpha, something is needing replaced.

TS is definitely the route to go on figuring this one out.

D

scottbot
06-09-2009, 03:18 PM
Zeke,
If I forget to shut off my anti-virus software while the ShopBot is running, it bangs and lurches horribly.
The first time it happened I thought I had broken a gear or something. Now I shut off every other piece of software I can while running the Bot.
I don't know if doing that will help you but it certainly couldn't hurt.

Also, I use a laptop and I have to make sure that I shut off the Wifi and Bluetooth while using the Bot.

Good Luck

Scott

gabepari
06-09-2009, 04:29 PM
I had these exact problems before I grounded everything. Ran a continuous 10g stranded wire from every moving part on the table and grounded it at one end. Problem went away.

zeke
06-09-2009, 06:03 PM
Dana, I have a PRSStandard 48X96, 4HP HSD Spindle new in 4/2009.

TS asked me to increase my speed to 3 IPS from 1.5, I had the same results and ran out of day, it was after 5PM. TS said they would send me another of the same file I sent them tomorrow, seems like I'll get the same results. You guys bring up some good real-world issues that sound very similar to my problem.

It sounds like I could turn a few more pieces of software off. On the grounding, I did perform the ground per the manual.

Gabe could you expand a little more on the grounding steps you took?

Scott, did you use the task manager to identify the processes that were running and turn them off after startup or did you go into the administrative services and turn them off there or other?

Scott and Gabe, did you experience both the banging noise and the drifting off axis?

Thanks for your feedback!

Gary Campbell
06-09-2009, 06:34 PM
Zeke...
If you are having comm problems, this may help: http://shopbotwiki.com/index.php?title=Grounding_your_ShopBot
Gary

zeke
06-09-2009, 07:06 PM
Thanks Gary, that looks like it will definitely help and something I have to do to regardless to eliminate this as a potential problem anytime in the future. I've already got the dust collection done this way, but missing most everything else except for the grounding completed per the manual. Guess I'll be busy for a few days, I'll let you all know how this turns out.... Where do you get your extra t-nuts?

Gary Campbell
06-09-2009, 07:27 PM
Zeke...
I bought some from ShopBot and I bought some weld nuts from McMaster. The ShopBot version that allowed slipping in and "twist lock" were the easiest to use for the ground lugs. The weld nuts require an open end on the extrusion to slip in place.
Gary

zeke
06-09-2009, 09:20 PM
Gary, did you buy the tab or slab base nuts and use bolts from the local hardware store?

http://www.mcmaster.com/#weld-nuts/=28y8ez

Gary Campbell
06-09-2009, 10:01 PM
Zeke..
Tab base, and yes the local hardware store. Our local hardware store carries stainless steel fasteners exclusively.
Gary

scottbot
06-10-2009, 02:54 AM
Hey Zeke,
I just shut off any of the programs that show up in the tray in Windows XP.
I only get the banging. No x drift.

Scott

dana_swift
06-10-2009, 10:13 AM
If it is a grounding problem it would be likely to influence Y and Z also.

Your problem seems to be confined to X. Grounds problems are not "usually" particular.

My suggestion is to look for a bad connector in the X wiring. That could be at the driver, the driver to the control PCB, or at the stepper motor.

You can exchange the ribbon cables from your X driver to the PCB inside the box with the one for Y. If the problem moves to the Y axis you have found your problem.

You can also exchange the X and Y driver modules. Again if the problem moves, you have found your problem.

There is (probably) a connector in the wire from the X stepper to the box. Try wiggling it while the machine is running, if there is a loose connection in there all kinds of weird things will show up while you are wiggling it. I located a problem on my bot that way..

My suspicion is you will end up with a great grounding system and still have your problem. Since I can't check your system myself, I am just passing along my suspicions after being a practicing electrical engineer for over 30 years. Yet sometimes it is the grounding system. If you have an oscilloscope you can find out for sure which one it is (assuming you know how to use one).

The biggest problem I have found with electrical systems over the years is "connectoritis". Connectors are the least reliable part of many electrical systems, especially where there is motion or vibration. A CNC machine clearly has lots of that. My suspicion goes there..

After a long search I often find a bad crimp, loose contact, nicked conductors at the point where the wire is stripped, etc. Fixing that fixes the problem forever.

Hope that helps

D

zeke
06-10-2009, 11:47 PM
Here's an update.

Item 1 - Banging Steppers.... It dawned on me that I had been using the machine for a few weeks with no banging and I had not made any significant changes to the machine. I wasn't doing any cutting yet, but I was still moving it around quite a bit. The banging noise started a few days ago and only happened a few times. So what changed, we have had a ton of rain every day several times per day for weeks, translating to allot of moisture in the air and most likely on my equipment in the garage. Seems like moisture would be a great conductor. However the last two days have been dry, so I thought I'd crank it back up just to see if the banging noise occurred again. I ran the machine for over an hour tonight manually moving, stopping, cutting, no noise. I still need to tackle the optimal grounding scenario in the very near future to be prudent and avoid this particular problem.

Item 2 - Drifting off X axis while cutting.... I was eager to figure out what the other "drifting" problem was and get beyond that. I think I may be close to an answer and I think I can use a work-around in the software to not have to use the same "pocket toolpath" using clear pocket for cutting the 50 1 1/8" in diameter and .438 recessed circular holes. I worked with TS and they tried to diagnose the issue without success. Although we did determine I didn't have the latest version of SB3 so I upgraded. I tried cutting with the same file and had the same exact problematic results.

I was thinking if I could cut the same recessed holes using another method and received the same results then it might point more towards my machine. If I had different results possibly a software bug, file corruption, etc.

One of the things I noticed before trying a different cut method was that while I thought the x was drifting progressively with each row of the 50 holes. The shift off axis occurred on a specific row with consistency, the last three rows. Seems software related versus the Bot. I validated all the coordinates in the two different Partworks / SBP files, both the cut through holes and the recessed holes and they were all exactly the same x and y coordinates, yet the recessed hole file cutting caused a drift off the x axis.

As a test, I used the exact same recessed file and changed only the tool path to a profile toolpath, ran the file and it cut the holes as expected without issue. Hoorah! At the end of the cut, the x and y came back to 0 as it should have with the profile toolpath. This may be a software bug under certain conditions. To confirm that, i think someone else or TS would need to create the same type file with the same setup and coordinates. Then physically make the cuts and get the same erroneous results. Although TS said they are going to be very thin over the next few days so I don't see that happening. TS did look at the file in preview earlier today before my test scenario and said it looked fine, so I think this would have to be a cut scenario to unveil the issue. I'd be happy to pass on all the file information if someone wants to give it a go.

The concern I have for all of us is if this is a bug, it may only present itself in certain software setup conditions and could destroy some good material. I've got some type 1 PVC I need to cut and I guess I need to make my cuts on mdf or something less expensive to ensure there are no issues before cutting into the PVC. While on the subject, anyone have a suggestion for a good Type1 PVC cutting bit?

I'm going to pass on my findings to TS.

dana_swift
06-11-2009, 02:41 PM
If your problem is software, that will impact all of us. "a software bug under certain conditions" is just a "bug". If it happened all the time, the folks who put out the software involved would have applied a little cyber-raid to the cyber-arachnid.

If it is in the software any bot running the same version of software as you will probably exhibit the problem. TS should be very interested in that. They may be swamped, but your problem may represent more of the swamp than just you.

I wish you a speedy resolution of the problem. Then post the lessons learned back here for all to benefit.

Suggestion to the moderator- why not a section on the forum of "Solved problems" starting with original symptoms, and eventual solution. Including the diagnostic method used. That way anyone with a symptom can just look up "things that worked", or "what does it mean when..". Perhaps that is already on the wiki, a case of "RTFM", or "RTFW" (wiki) in this case.

D