Losing z zero

[phil@philsfoils.com]

I was cutting a rudder on my 'bot - did one side yesterday, and flipped it and did the other side this morning. Big problem - it was cutting a couple mm deeper than it should have been. When finished, I rechecked z-zero at the top of my flip-jig, and the bit was a good 1/8" deeper than it should have been.

This is not the first time that this has happened.

I'm not always consistent in what order I power up my equipment - could the "errant movement" problem in the Shopbot Msg board have been responsible? If not, what? I NEED to figure this problem out if I'm to have the accuracy that is crucial to my application.

All my tool paths and feed speeds are generated from VisualMill. I cut softwood with x,y move speeds of 2"/sec and z move speed of 1"/sec.

Thanks
Phil
www.philsfoils.com

[gerald_d]

We also get this problem twice a year or so. Very irritating and very mysterious. We used to get it much more often when we had bad electrical grounding but we sorted that out. Last week we had it again - blew out dust from all the electrics and connectors and are holding thumbs till next time.

But it really is not good enough to know that this problem does occur and there is no definitive reason or cure.

Some more threads relating to this:
http://www.talkshopbot.com/forum/messages/26/1258.html
http://www.talkshopbot.com/forum/messages/26/104.html
http://www.talkshopbot.com/forum/messages/26/1446.html

My latest theory relates to damp dust build-up, but I am clutching at straws here......

[toys]

If I were to guess. . . I would say that it relates to the computer and timing causing missed steps. My bet would be that the PRT controler in conjunction with version 3 of the software which moves the timing to the controller will fix random mysterious problems like this. Just a hunch based on my former work in the US Navy on a large data collection system. Timing is critical and sometimes unpredictable when dealing with interfacing electronic equipment with computers.

[elcruisr]

A last consideration is to start with the simplest things like, plunging too fast? Dull cutter? A cutter not designed to really do plunging moves? I've had these bite me a few times when I was first running my 'bot. Followed by grounding issues. A quick test might be to air cut the files and see if it occurs without cutter resistance.

Eric

[phil@philsfoils.com]

for what its worth, I cut a duplicate rudder a day earlier, using the same cutting files, and that one went fine. Cut this one (the problem one) with a brand new bit too.


These files can take 90 minutes or more to run (per side of the foil), and sometimes I'll get a ridge across the object where the z has changed slightly (and enough to be annoying). Yesterday's had no ridges, but was simply off by 2mm or more in the z, making me think that it lost its zero between sessions, since I carved one side of the rudder one day, shut down, and did the other side the next day.

I really have to sort out ALL z-zero issues - losing z-zero is just not acceptible in this work. If anyone can give me tips, I appreciate it.

[Ryan Patterson]

Did you re zero the next day. If you shut the control box off or turned the moters off gravity will pull the z down.

[davidallen]

I'd be more concerned about the 'ridge' effect. does this occur during an xy translation or between xy moves (during a z movement)?

if the later, you're driving the z too fast. either slow it down or reduce the distance moved.

if the former, you're probably getting a spurious pulse on the z control lines (grounding issue?).

da

[elcruisr]

For what it's worth we went through a bunch of z issues like this when we first set up our machine. After a long series of phone calls it turned out to be several issues and not one. The biggest problem was a small defect on the driver board itself. After solving that we had a small problem with grounding. The static buildup in the dust extraction hose was causing very small additional z moves. Solved that one with a ground wire through the hose. Also the motor control wires are nice long antennas that can pick up interference from other souces in and around the shop as well as from each other. We had to pay careful attention to wire routing and interference issues. There have also been issues with power supply line "noise" with some other machines. If it's not an obvious problem, I'd call technical support at the factory. They've dealt with many of these situations and often will find one the little thing we're missing.

Eric

[gerald_d]

Was hoping that tech support read the forum and post their expert advice here. Calling somebody 10000 miles away who has a very strange accent won't help me - plus, the toll-free number doesn't work from here.

[phil@philsfoils.com]

re: static. From the beginning I've had a bare copper line through my dust hose and attached to the z-axis. So grounding shouldn't be a problem. If its picking up spurious signals somewhere maybe sheilded cables would have been the way to go. Real pain though, if that's the case.

re: plunging - just trying to think about this - if its plunging too hard and skipping steps, then its going to think the bit is lower than it actually is, and will later be cutting too high, not too low, right? I'm using the Makita RF1101, not the big Porter-Cable, and it never sounds like its bogging down or being driven too hard.

Regardless, I take my passes in parallel 1/16" apart as the router works its way up the rudder, usually cutting cedar or spruce. Thats not a heavy load on the router, and as I say, the toolpaths, including ramps & plunges, are all calculated in VisualMill.

Re: ridges - I take passes in parallel working up the rudder - it sets X and then adjusts Y & Z constantly in a vertical arc (all as a set of M3 instructions in very short steps). Fairly regularily I'll get a ridge in the work, presumably because for some reason Z has lost its zero by a small amount.

[artisan]

Phil....is your Z a Rack & Pinion or Ball Screw? I too have had this problem and it was simply build-up on my ball screw causing the motor to stall momentarily....but losing some steps. It's one of the reasons Shopbot switched to R & P for the Z......D

[phil@philsfoils.com]

Mine's a rack & pinion - bought in November last year.

[artisan]

If your router "stalls" momentarily going "UP" (which is when it usually happens), then it will reset Zero as a lower point...depending on the severity of the stall. It may overcome the build-up and work well from that point forward....but your zero has already been reset by then....juat a thought....D

[tlempicke]

Another point

If you are using Parts Wizard you must always work in either metric or inches. If one of the dimensions is set to metric and others are set to inches it will do what you have described.

For instance your tool settings - your project size, etc.

[gerald_d]

Our recent z-glitch came while surfacing the table. After about 20 minutes of rock-steady holding, the z-axis decides to go about 2mm deeper. Happened at a direction change point and it was one instant move (as if there was a step in the file).

- We were not running dust extraction (no static electricity)
- The z-height had been steady (non-motorised) for at least 20 minutes.
- There was no file with a wrong step.
- All the mechanics (set-screws) etc. were 100%
- All the ground wires tight and making good connection

All we could do was blow dust out of everything (as we do every week in any case) and hold thumbs. But our thumbs are turning blue - and we KNOW that it will happen again......

[toys]

Gerald, you should email tech support directly. It seems like they only look at this forum ocassionally. If you can get in touch with Ted Hall directly he has a lot of good ideas when it comes to troubleshooting.

[gerald_d]

E-mail sent - will keep you informed.

[Ted Hall, ShopBot Tools]

Well guys, we had been watching this thread all day ... and you seemed to have had such good suggestions for each other that we didn't really see a reason to jump in. But here are some thoughts.

First, I don't think this applies to most people above, but in the case of a PR tool with ball screw having a Z problem -- the issue is most likely to be mechanical, usually crud in the ball nut that just every once in a while, produces a tight spot during movement. It is also very easy to try and go too fast with the ball-screw Z. Slowing down will even help get by with a dirty ball nut. (The PR Control Box is virtually immune to electrical noise and this is practically never an issue with this vintage tool.)

On Phil's problem at the top, there are a couple of possible concerns with turning the control box off and not re-zeroing the next day. Part of the problem is that the Z does not have anything holding it rigid when the motor goes off, any slight bump (or an effect of gravity) and it will not return to where it was left. The other issue is that when the Control Box restarts, the microsteppers return to the top of their cycle, which can cause as much as 1/2 of a full step of movement. Normally, this should not be more than .005, but you would not want this to happen if you can avoid it. Thus, it's best to either leave the box on overnight, or re-zero all the axes after you have turned the box on and off.

Always check mechanical issues first: slipping bit, damaged collet, slipping router, loose pinion, pinion not tightly engaging the rack, broken spring (or too much or too little load for the spring). You'd be surprised how often a bit that was not intendended for plunging is the culprit is losing Z location. [In that regard, as great an idea as they sound, we have lost a little faith in the constant force springs that pull up some Z's. They can occasionally bind, it turns out. Normally there is plenty of power to deal with this, but on a fast pull-up, binding might contribute to causing a problem. Try and keep a little film of oil on the springs (we've switched back to just using standard coil springs).]

Then consider speeds. Are they set too near the stall point, for either plunging under load, or pulling up quickly. Be particularly careful that the jog speed is not too high for pull-ups ... especially if the problem is that the Z seems to go lower.

One of the problems with the faster drivers in the PRT tools is that they can be sensitive to electrical noise. I don't believe it actually happens at all frequently or on many tools, but it is a possibility. We have seen it with routers that have worn brushes, and occasionally from other equipment on the electrical circuit. Static from the dust collection system is a possibility, but I have not seen it. Good grounding (connection of system components back to the system ground), router on separate circuit, and a UPS on the Control Box has taken care of any problems that I know about.

I don't have a good explanation for Gerald's problem ... maybe noise. But I'm guessing that the big bit that is usually used for surfacing may be contributing??

FYI, in some applications where speed has been a primary issue for the Z movement, users have added a prox switch to the Z axis so that the location of the Z could automatically be confirmed at critical locations in a file.

[gerald_d]

Ted, re "the big bite for surfacing", why does the tool tend to dig down rather than letting go? Ny impression is that a stepper motor, that is forced to loose steps because of binding and crud, will make a discernible noise. But we have heard no strange noises when this happened, and again; it digs down instead of releasing. (For the resurfacing, we obviously heard the big bit digging deeper, but the other times we heard nothing liked a missed step)

Any comment on my theory that excessive dust in the open Wago connectors can cause leakage currents? (Surfacing the table while the dust collection was broken, was very dusty).

[toys]

Gerald, This might be too basic but was the bit tight enough? I've had bits slip down when I thought they were tight enough. My collect was no good and was accually worn enough not to grip properly. I could tighten the collect very very tight and the bit would still slip. This was using one of those reducers the reducer was baddly worn. Maybe a simular thing could happen with the regular collet. What size and kind router are you using?

[gerald_d]

David, we use a Metabo die-grinder with a 1/4" or 6mm collet. The collects are good and hold well. We had to ground the alu. nose section of the grinder as well (to get the zero plate to work). A loose ground connection to this alu. nose can cause huge problems with z-drift.


3754

[gerald_d]

3755....Rough spec. (http://www.coastaltool.com/cgi-bin/SoftCart.exe/a/meta/ge700.htm?E%2Bcoastest)