Skew and runout

[letterman7]

Hi folks,
Not a regular on the boards here, but ran into an interesting issue today. My PRT96 has been cutting relatively well for years (mainly, after Brady stopped by to help me tune it up a little), but it was never "just right". There was always just a touch of non-repeatability that bugs me - no two identically cut parts were exact. Very close, but maybe 1 or 2 mm off in one direction. Anyway, the past week I was cutting a file and one of the X motors hesitated and jumped a tooth or two. It happens - it's an older computer that gets buggy in certain weather, but usually both X motors will hesitate and screw up my cut (yes, I'm planning to upgrade very soon). Anyway, I reset and re-squared the Y carriage as close as I could get, measuring from the rail ends to the ends of the X rails, then following it up with a known good T square. All seemed good, and I continued on cutting my file.
Fast forward to today. I decided to surface my tabletop just because it was getting a little beat up. Today I just used my keyboard keypad to run the router via the user input rather than running a file to keep any errors out of the equation, so I ran the 2" end mill around the outside of the table first, then worked my was back and forth across the X axis. As I neared the far side of the table and had that strip of uncut material left, I thought it would be interesting to see how accurate it's actually cutting... and I was in for a rude surprise. The variation is about 1/8" across the length of the table - far too much for me (sample measurement shown):

21104

So the question comes into play: what do I need to tweak to get it to track straight? I thought it might be the motor that jumped and that the Y carriage isn't exactly perfect, so I unplugged and moved the carriage back by a tooth or two. I didn't want to go too far as I fear it might bind. Continuing the surfacing with the relocated pinion, it didn't seem to make any difference, so I took the strain back off that particular motor and re-squared the Y axis again. I'm a little stumped. I flunked geometry.. that's why I'm in the sign business!

Anyone have suggestions?
Thanks,
Rick

[Brady Watson]

Gary's tutorial on squaring the PRS gantry is good: http://www.shopbotblog.com/index.php/2008/12/tuning-up-the-prs-gantry/

There is no way your tool should be out that much. I am not sure how you measured the deviation, but it may be a result of just winging it with the keypad instead of running a proper flattening routine.

The quickest check of squareness is to chuck a small bit into the router (or plotter pen) and do a 30x40" rectangle by scoring it into the table about .01" or so. Then measure the diagonals. If it is 50", then you are good. If both diagonals measure something else, 'read it' and it will tell you which way the gantry is out.

Your machine also should in NO WAY be 'grinding gears'. I am wondering if the sound you hear is one of the motors stalling due to slow COM speeds or mechanical resistance. It is also possible that one of your motor drivers is 'tired' causing one side to lag behind. Please check that BOTH of your X motors are completely engaged into the rack. Drop each motor via 4pcs 3/16" hex bolts, then slide them all the way up while you wiggle to make sure they are FULLY engaged into the rack. As time goes by and the pinions get worn, they slowly disengage from ideal mesh because of wear.

Not for nothing, but it might be a good time to have someone come out and inspect/repair your tool so that you can get back to the business of making money.

-B

[letterman7]

Both motors are fully engaged, Brady. It's been over the past year that the computer itself has been hanging and losing connection to the 'Bot (that nice yellow rectangle comes up on the screen). This last time it was in motion and happened to stall the one motor - I saw it jump the tooth.

Winging it on the keypad might be the reason for the runout. I did plan to run a test pattern in the next few days as time allows - I'll know for sure then.

I do plan to upgrade to the new(er) computer and drive units in the next month. It's due, even though the machine sits idle more than it works. I'd rather have the knowledge of an upgraded machine waiting for me than one that doesn't have the resolution I'm looking for. After the upgrade, if there are still issues going on, I'll give you a holler and we'll set up another "fix-it" day...

[bleeth]

A couple thoughts:

"Fully Engaged" Make sure your tensioners are not too loose or too tight. Either could cause x axis tracking issues. Brady is, of course, dead right that a deviation like that should not be at all. And squaring wouldn't cause it unless it were so far out that the rollers were not staying centered on the rail somewhere. Check things out this way:
1. with control box off and motors unplugged slowly move gantry all the way up and back keeping an eye on all 4 rollers.
2. Look for wear on the rails. The angle steel will wear over time and needs to be dressed back to a nice v with a file or diamond rasp.
3. Also look for resin build-up on the rails and in the rollers. This can cause some resolution issues but shouldn't affect true tracking.
4. On the computer side: Dump your temp files and make sure start-uip is empty. If you have been using the control computer for other things it may have gotten a bunch of things running in the background that affect comm. I strongly advocate a dedicated computer stripped ala Black Viper.

[letterman7]

Thanks Dave. I'm pretty obsessive about keeping the rollers clean. I cut quite a bit of pvc and hdu, both high-static items. The rails are pretty good. Might could do a quick run with the rasp, but otherwise ok. The computer is a dedicated unit, nothing on there but the SB program and a copy of VCarve Pro. I'm fairly certain the lost comm issue is from the SB box and not the computer, but then, the computer itself is old(er) so it, too, may be having some issues.

[Mayo]

You mentioned static - make sure your ground wire hasn't loosened up.

Make sure there's no build up of dust in your control box and in your computer. If there is, shut them down before blowing them out. I found out the hard way that fine dust is highly flammable and the power supply can ignite it!

Also make sure Windows isn't trying to update something even if the computer is not connected online.

Double check all the cable connections - motors and serial port and power, etc.

With everything off and power unplugged, you can make sure all chips on the PRT board are pressed into their sockets fully.

And the last thing (or maybe it should be the first thing) to check is your unit values.

[letterman7]

Unit values are fine. Dust in both computers is a definite. Taking them apart and blowing the particles out is on the schedule for the next week!

[bobmoore]

Rick;
One thing to check that I don't see mentioned is your wago connectors. I have had a couple issues lately and both times it was corrosion (like a car battery between the wire and connector) at the connectors. I am buying and replacing all of them and trimming the wires back and restripping them on my PRT-9648. Bob

[David Iannone]

Many years ago I took the wango connectors off and soldered them together and never had a problem since.

Just about a month ago I was able to upgrade my 2001 PRT to the 4g board upgrade. What a difference. Easy board swap upgrade and the motors just sound so smooth. And of course the increase in speed is awesome. Cleaner cuts, ramping is best I have seen my machine do. I never thought I needed the extra speed, but once you have it just for the jog speed in between tool up and down makes a Hugh difference. I also put new pin ions on, as mine were slightly worn too.