Gremlins

[garyb]

For 2 years my prt has been flawless. As maintenance time came around we decided to move the shopbot to improve shop production, at which time we rotated 90 degrees & moved 10' to the North, did the usual leveling realigning, adjusting etc - everything back working normal.

Also decided to upgrade the SB software at the same time. Installed the sb233sx and ran load 46 set the VU and VR values.

Now for the Gremlins!

First thing I noticed was the XYZ read out no longer matched the given move distance eg MZ,1.0 would result in a move of .999 or 1.001 same with X and Y moves

2. While using the K command and in the middle of moving the tool head to a new zero point the software decided to cancel the "K" command on its own.

3. Now for the biggest problem, with the SB sitting idle and the sofware running I would hear a clunk from the drive motors at odd intervals, on closer inspection I found the SB moving on its own down the "X" axis and sometimes down the "X & Y" axis. This resulted in the 0,0, point moving by the same distance of movement, but it does not show up in the "X Y" read out boxes.

After going through all the normal trouble shooting I found this Phenomenon only seemed to happen when the dust collector to the main shop was running, this is on a 220 dedicated circuit and the SB in on a separate 110 dedicated circuit.

This dust collector is not connected nor does it touch the SB in any part and in moving the SB increased the distance from the dust collector.

So I'm open to any suggestions for a fix
thanks gary

[paul]

I TOO LOADED UP THE NEW SOFTWARE AND FOUND IT READ .999 INSTEAD OF 1 INCH. VERY STRANGE. HAS ANY ONE NOTICED WHEN YOU LOAD UP THE SOFTWARE ON WINDOWS MILLENIUM THE MOTORS RUN MUCH SMOOTHER??? I DID!!! I HAD MY JOG AT 3 FOR XY AS I DID WITH VERSION 2.29 BUT WITH 2.33 RUNNING IN MILLIENIUM IT WAS MUCH SMOOTHER ( I HAD WINDOWS 95 BEFORE). PROBLEM IS AT 3 JOG THE MOTORS WERE JUMPING STEPS!!! I LOWERED THE SPEED TO 2 FOR JOG AND IT HELPED BUT I DIDNT LIKE THE POSSIBLE PROBLEMS I WAS SEEING THAT I NEVER HAD BEFORE IN WNDWS 95 WITH THE OLDER SOFTWARE! ONE THING IS FOR SURE MY MOTORS WERE ALMOST SILENT WITH THE MILLENIUM AND NEW sb SOFTWARE BUT THE STEPS WERE OFF! I WENT BACK TO 95 AND VER 2.29. WISH I COULD HAVE STAYED AT 2.33 SINCE THE MOTORS SEEMED TO RUN SO MUCH BETTER. WHAT IS GOING ON?? ALSO THE JOG AT 3 IN 2.33 SEEMED FASTER THAN THE JOG AT 3 ON THE OLDER SOFTWARE? AND YES I DOUBLE CHECKED MY UNIT VALUES ON BOTH SOFTWARES. gremlins maybe? confusion yes!

[bjwat@comcast.net]

Gary,
The first 2 places that I would look would be:

Make sure that you have a good solid ground on the Bot.

Secondly, check to make sure that your unit values for each axis are correct after upgrading to the new software.

Also check all your connections from the control box (make sure that has a good ground too) to the steppers. Maybe something came 1/2 way unplugged or got chaffed in the moving process.

Hope that helps,
Brady

[Joe Crumley]

Gary,

There is a chance that the unit value's could have changed but can't imagine why.

I would be more inclined to think there is a static grounding problem. Two weeks ago I had a very simular problem due to my vacuum hose collecting static electricity.

Here is what I did.

Drilled a whole in the concrete floor and drove copper rod down two feet in the ground. Lowes has 6' copper rod for $7.00

Sodered a lead from the SB Table, Control Box, and the Vacuum. Also, I put a wire inside the flexable vacuum hose from the router area down to the vacuum itself and grounded it.

This took me approx a half day and solved the problem. Cost me $300 in down time but will never have to do again.

By extending the legnth of your vacuum hose, you
also increased the static problem.

I suggest you also look on SB Forum page for "KEYWORD SEARCH" and see how many people have had this problem.

Joe

Good luck

[rgbrown@itexas.net]

Static discharges of electricity in dust systems can start fires and knock the bearings out of dust collectors too.

Ron

[garyb]

Well guys here's the update, went to the shop today and spend 6 1/2 hours putting the SB through all kinds of tests in conjuction with the other equipment trying to reproduce the same effects I had previously encounted.
I got very minimal reaction today.

However I did manage to have a number of gliches with the SB223sx that I never had with SB229sx.
1.The read out not matching the move distance as state above continues and is inconsistant.
2.During a move X axis operation I hit the space bar for an emergency stop, this stopped the software from responding, no "resume" no "quit" no nothing. Had to do a end task and reboot to clear and be able to restart the SB program.
3.Another thing I noticed is that at start up the SB program is inconsistant with running the timing check and its very noticable in the ramping when using the arrows under the "K" command.

Now before everyone gets too excited, the computer connected to the SB is a celron 600 with Windows ME and has run sb221 (I think was the ver. I started on) to sb229 with no problems at all. Installing sb233 last week was the first software gliches I've had.

As per my post above I set the unit and ramp values after installing and they remain as set,
Joe what you did with the grounding rod I also did when I installed the SB and all connections have checked out OK.
The Gremlin side of this is that the interference when it happened came from a dust collector that is in no shape or form connected to the SB, the only difference is the rotation of the SB, before, the SB sat parallel to the dust collector and main duct pipe, it is now perpendicular where both "X" rail stepper motors face the collector and main duct, so if this makes a difference I don't know.
The ducting and collector are all metal and are grounded at every machine including the building so I don't believe this to be a static problem.
I believe this to be an frequency problem & turning on the dust collector triggers a reaction to the stepper motors, if thats possible but then I'm just guessing.
As far as SB233, I think I will try a reload if all else fails there's still 229 !

[bjwat@comcast.net]

Gary,
Listening to your prognosis so far, I am inclined to point the finger at Millenium...It is the absolute worst version of Windows. If you can go back to Win98 2nd Ed, you'd be much better off. (I know...there goes another day of loading up software) The other thing that I would try is, completely unloading the SB software. Make sure that you restart the machine to refresh the registry, and reload the SB software again. You have more than enough procesor to run the Bot by far.

If I was having this problem, I would wipe the machine, put on 98 2nd Edition and reload the SB software with ZERO additional software. I find that my machine ran best with the least amount of software on it. I do CAD/Part Wizard stuff on another computer.

Just a suggestion...I hope you get it worked out.

-Brady

[Ted Hall, ShopBot Tools]

I'm going to jump in on this one ... and hopefully won't add too much confusion. There are at least 3 issues in this thread that I'll comment on:

1. Gremlin Movements (Gary)
2. Location Display and other issues is new software (all)
3. Smoothness & WinME; Timing check (Gary, Paul)
4. Grounding (all)

1. GREMLIN MOVEMENTS. I believe, Gary, that your control box is picking up some electrical noise generated by the dust collector. It is good to keep these things on separate circuits ... but the main power line is only one route for noise to enter the control box. The second is via the ground line. The third is induced noise picked up by other wires on the tool (long antennas). The motor wires and especially the wires for limit switches and zero plate are good candidates for noise entry.
The first way to attack ground noise (after grounding considerations, see below) is putting the control box on a UPS or line conditioner (preferably one that 'conditions' the ground as well). Cleaning up the source with a filter or capacitor may also work.
The new alignment of the tool and wires, operating as an antenna, might be another entry point. A little experimentation might solve the problem quickly ... especially since it seems obvious when you turn the dust collector on.
In your particular case, Gary, we have found that versions of the control box shipped before 4/1/01 (ones with large black or gold heat sinks inside) is prone to pick up noise. We do have a filtering solution that we can add. If you don't get a quick improvement from above, let us know and we can get you a modified control board that may deal with it.

2. LOCATION DISPLAY. All recent versions of our software (from well before 229) display the actual location of the tool. Because your particular tool is stepping at something like .0017" per step, there are times when the location will not appear as the exact distance the tool was sent too because you are getting the rounded value of the closest possible step. This should never appear as any greater discrepancy than +/- .001. What surprises me is that you seem to be seing something different now than before using 2.29, which I would have thought worked exactly the same. The problem is particularly obvious after an XY diagonal where getting the closest XY step location may create the .001 discrepancy in some cases and not others.
I have not heard (or experienced) problems with the Keyboard control in this version, but we will make a point of checking it out. I'm wondering if it is possible that your table base coordinates have gotten very large (as would be the case if you're not using this system) and that might create a problem with the keyboard control. Maybe send in your PROBLEM.LOG file if you get a chance and we can check this.

3. SMOOTHNESS & ME; TIMING CHECK. I should say first, that in terms of actual functionality, there is not a lot of change in the last few versions of the software. Of course, the little fixes that we make can always lead to problems we have not expected ... but, in general, I don't think you should see much operational difference.
However, we are always trying to improve the ramping because this has such an influence on cutting, particularly complex files. I imagine that the differences that Paul is refering to are related to smoother ramping. (Alternatively, there could be slight differences in speed computation that might cause motors to resonate less at a particular speed and thus sound smoother.)
Ramping is heavily dependent on timing in your specific computer. Thus, it is the most easily disrupted by things Windows is doing and Windows may make it difficult to determine the exact speed of the computer. We have attempted to improve both of these things in each version (note that the new Windows version puts all timing outboard ... thus these problems go away). Two additional problems we've noticed measuring timing are that: 1) Windows ME seems more likely than 95/98 to get bad timing measurements and thus cause bad ramping; 2) the faster a computer, strangely, the harder it is to get a good timing measurement in Windows. Following this message, I will put up an explanation of how to hard-code your timing measurement. This can improve the situation considerably ...

4. GROUNDING. On grounding, you want to have a good situation. Grounding is important for: a) safety; b) static and noise surpression; c) common ground for some operations such as Z-Zero checks. The best all around approach is to first make sure that you have a good building/system ground. All grounds should ultimately tie into this system ... branching out from a common point. Grounds related to your ShopBot, for example, should all be tied to the nearest power box/conduit ... which itself is, hopefully, well connected to the system ground.

[Ted Hall, ShopBot Tools]

HARD CODING TIMING

Timing primarly influences how ramping is done. If we don't get a good timing measurement, then the tool may either ramp very slowly or too abruptly. The measurement is typically only a problem with a fast computer in Windows, or a computer that is running a number of programs. Assuming you are running in Windows, here's how to deal with it (the new ShopBot software will not have this issue):

Use the [UP] Command to re-sample the computer speed half a dozen times. Write the values down and look through them. Throw out the extremes, and take the average, then decrease the number by about 10%. Let's say you got about 500 ... so we'll call it 450. OK ... let's put this number to use.

Find the file called SbCut.bat in the Sb200 folder. Open it with a Text Editor like Notepad. The last line of the file currently looks like this:

sb.exe %1,,,,

Change it so it looks like this:

sb.exe %1,,,{your number}, e.g. sb.exe %1,,,450,


This procedure will hardcode the timing check on your computer so that windows does not get any strange timing values. Funny timing values are usually the cause of uneven ramping. This problem does not occur in DOS.

[garyb]

Thanks to everyone's comments. We are in the middle of field install so don't expect to get back to trouble shooting till the end of the week.
As far as sb223 I either have a bad install (my fault) or Windows threw another curve ball as I have now noticed another oddity not related so I believe its time to clean house and reinstall.

Ted, as far as the display read out, previously if I did a jog or move on the X Y or Z axis the read out equalled the move, but in honesty those moves were limited to mz, 1.0 mx,8.0 and the likes just to move the bit out of the way or maybe jx, 65.0 to park the carriage at the end of the table. I do a lot of movement just using the "K" command therefore would not notice a difference in read out. But when cutting I keep a close eye on the Z axis depth and they have been spot on, I mostly noticed the saftey pull up is the reading that is off, not that its a big deal as the measurement is so minute but its one of those oddities that just became noticeable.

The odd part of this is that the coordinates x,96 y, 48 were no more than 3' from the edge of the dust collector for 2 years with no problem. We pick the table up, rotate and move to where x,0 y,48 is 8' from the collector with no changes to grounding, wiring etc and the wrath of Gremlins desended upon us.
Has definitely been interesting these past few days.