Y axis lost position

[bthoron]

Our PRS is intermitently, but regularly jumping out of position by about 1/2 in the Y. Specifically, it runs for a while (20min) and then jogs (I think) about 1/2" less than it should in the Y direction. It will run more of the tool path and then do it again. This has happened on two different files and during an air-cut.

What steps should I take to figure out the problem?

[Brady Watson]

1) Check for loose Y pinion. Remove motor, & crank down hex grub screws til you bend the wrench...this is why you were given 3 of them.

2) Check your unit values.

3) Check pinion to rack mesh. Make sure it isn't loose.

-B

[scott_smith]

Spindle or router? Could it be electrical noise from the spindle wires to close to the Y-motor wires? I have a PC router but have heard of lots of noise problems with spindles. Just a thought.

[bthoron]

Thanks Brady and Scott, but I haven't licked it yet.

It happens in a jog move, smooth operation followed by a shudder and change of motor noise. the shudder is like the motor goes negative in the Y a short distance and then continues. but arrives 1/2 in short.

Any body in San diego have time for a cutting job?

[Brady Watson]

Ben,
At what speed does this occur?

Have you tried running the SpeedTest.exe in the SB3/Diagnostics folder? (Close SB3 before running). This will test communication speeds. There is a noticable change in motor 'note' when running at high speeds if you don't have good communication.

-B

[harryball]

I run at 12 ips and my speed test shows 75%... if ANYTHING causes an interrupt it skips. I must eject the network card to run any of my groove files. The first time it happened I was at a loss to figure out why, I'd been running all day when it hiccuped. It took a few days before I noticed the network icon changed when it occured. With the network card ejected, problem solved.

Just sharing experience that might help you track it down.

Robert

[bthoron]

Speeds are low- my jog speed is 6ips.

Robert- did it always happen in the same place in the file or was it random. I can repeat the problem in the same line in the file every time.
I'll try the speed test file next.

[harryball]

It was random. I jog at 20 ips.

If you aircut the file does it happen?
If you comment out the line and run the aircut does it still happen?
If you change the jog to a move does it still happen?

Robert

[bthoron]

Yes it happens when I air cut,
Haven't tried the commenting
Haven't tried changing it to a move.

What we did first was to change the Y and Z motors, when that didn't work we also switched the driver cards in the box (a suggestion from Chris at SB). This seemed to work and we have run one sheet. It was too late to run more.

We ran the speed test and it returned 73.4%.
The fact that the fault mostly happened at the same line may be a red herring, you know how these things go, but it was dependedable enough to be a good test point.

Another thing we noticed is that the z motor mounted to the Y car was quieter, less prone to that buzzy moan sound when ramping. (another red herring?)

Anyway thanks- we'll see what happens tomorrow.
Ben

[bthoron]

We still have a problem.
We managed about 9 hours of cutting with 2 stops for loss of y axis.
SB suggested that we install a usb 2.0 hub to increase data transmission. So we did, they also suggested turning the drive box on its side to minimize overheating, and we did. But none of these measures have helped in the long run. Could heat be causing a problem? The shop is 75deg the motors are running about 100 and the drives are in the 96 deg range.

We're getting annoyed and stressed out watching for the machine to screw up.
The latest thing we tried is to eliminate all Jog moves and make them just moves- sure slows down the file. We ran three sheets that way. But its a bummer of a workaround.

Has anybody had these problems before?
Ben

[Brady Watson]

You are absolutely 100% positive that the Y pinion is tight? I thought I was positively 100% sure of this a few times before to find I was mistaken. It's worth another look. Drop the Y motor, and crank those grub screws down. If they have been losened (which it sounds like they have been by swapping it out for the Z), then you'll want to add fresh loctite to the screws.

The chances that there is something wrong with the motor or driver are extremely slim, especially since you said that you completely swapped out the Y for the Z motor and driver. Temps you list are well within the operating limits of the electronics. There is a very small chance that you have a defective control board, and this could be tested by plugging the driver into the channel 4 and assigning it to the Y axis in SB3. I haven't heard of anything like this happening on an Alpha, so I'm just offering it as another thing to try.

Did you happen to change any of the red dials on the front of the Alpha drivers by any chance? You might want to look to see if all of them are the same across the board. One of the dials is a torque setting.

-B

[bthoron]

We will re-examine the pinion - We "bent a wrench" the other day when we swapped motors.

Unfortunately, it's a Standard, not an Alpha. should I move the driver to Channel 4 still?

We've been running for 5 hours without a crash after we set up the files with no Jog moves.

Thanks for your advice.

Ben

[Gary Campbell]

Ben...
I experienced many of the same problems you explained. The main difference was that my loss of table coordinates were not confined to the Y axis. X and Z were affected also. After many hours to the patient crew at SB Tech Support, here is what solved the problem, or at least cut it down to less than 1 event per week of cutting.

Ground the machine as follows:
1) Install a ground lug that is attached to the ground of your electrical panel. (my electrician did mine) (not to an earth ground or rod)
2) Using #12 or larger stranded wire go from the lug to the aluminum rails (both) of the machine.
3) Using the same attach to the X car extrusion.
4) Using the same Attach to the Y Car (scrape to bare)
5) Using the same Attach to the Z Extrusion
6) While you are there, it wouldn't hurt to attach to the ground wire of the dust foot.
7) Using the same, Attach to the other end of the dust collection hose wire.
8) Using the same, Attach to the control box chassis and VFD Chassis (if using spindle)
Using the same, attach to chassis of SB computer and ground lug of UPS if using.
9) Use a dedicated line for computer with isolated ground.
10) inside the control box, slide a small stranded wire into the serial to usb connector and run this to the ground bus inside the control box
You are now officially grounded.
Use a many strand, high quality, low impedance wire. Use Crimp connectors on the ends and star washers for good contact.
I am not sure if separate home runs or a daisy chain is better, but I daisy chained my setup. I also dipped all the connections in the tinning pot to try and delay the effects that our salt air has on open connections.
Good Luck, Gary

[bthoron]

Wow, Um... OK...
We'll try that

Thanks

[dollarbill165]

Ben,
I've been having the same problem with my new benchtop. I thought I had mine grounded, but by the looks of Gary's advice... I've got some more work to do.

The only alibi I have is my laptop... I'm going to shut off my internet connection, automatic updates, screen saver, and uncoil my excess motor cable wire. I'll let you know if it works for me.

[fleinbach]

Ben,

The one thing I noticed here is there has been no mention of whether your dust collection hose is grounded. Although proper grounding is important grounding of the vacuum hose is the most important and overlooked of them all. If you do not have a wire wound hose you need to run a bare stranded copper wire through the hose and attach to either a ground rod or your power panels ground.

When I first got my Shopbot I had simular problems for nearly 2 months before properly grounding my vacuum hose. All those problems dissappeared once it was done. The strange thing is I should have known better since I have worked with electronics all my life but I was so excited about using the Shopbot I ignored the obvoius.

[bthoron]

The Hose was installed with a ground wire. I'm pretty sure it still is connected properly, but we'll check.
After changing the Jogs to moves we've been OK for the last day.

[richards]

Ben,
Have you tried slowing down the jogs just a little? If the machine if holding position properly while doing moves but loosing position while doing jogs, then it seems that either the jog ramp rate is too high or that the jog speed to too high.

[bthoron]

I had the Jog speed on a PRS Standard set to 6 ips. Is this really too fast? Besides which, We had no trouble like this prior to last Friday.

But what should the jog ramp rate be set to?
Ben

[richards]

Ben,
Sorry about the delay in responding. I was 100 miles from the shop when you posted.

I don't know how fast is too fast for the PRS machines. Six ips doesn't seem too fast to me, but I really don't know. Your jog ramp settings are probably factory default, which should work perfectly fine. To check, type VR and then let the mouse hover for a few seconds near each setting. The range, as well as the factory defaults, will pop up.

If you have a dial indicator, such as the one shown in this link (http://www.talkshopbot.com/forum/show.cgi?tpc=7&post=40868#POST40868), you could clamp the indicator to the table at the end of a Y-axis jog, so that some part of the router or spindle pushes against the plunger on the indicator, and then repeatedly jog the Y-axis from 0,0 to the point near the indicator. Each time you jog, the indicator should indicate the same length of movement (within +/- 0.015 or so). In any case, you should not see an additive change. If you do get an additive change, then you're either losing or gaining steps - probably losing steps. Lost steps are normally caused by going too fast or by binding. If you slow the jog down by about 25% and still get lost steps, then drop the motor(s) away from the rail(s) and move things around manually. You should be able to easily feel any binding.

(A basic dial indicator, without mounting accessories, can be purchased for about $20. Every one that I've purchased has had an 'ear' with a mounting hole on the back of the indicator so that you can easily bolt the indicator to a piece of wood.)

[bthoron]

Mike- The loss of position is sudden and quite specific. In the middle of a jog move (positive move in both Y and Y, the Y car jerks and resumes travel. Same behavior each time it happens. Very nearly the same loss of position ~.5". We have a dial indicator and will experiment with it. Other than the jerk which obviously results in a loss of position, we're not seeing any other position loss at this time.

[cnc_works]

Ben, I recently experienced somewhat similar symptoms that turned out to be worn pinions jamming on the rack.

Donn

[richards]

Also, check your Windows computer. I had almost exactly the same problem, but random axes, caused by a wireless network card causing the computer to be erratic whenever communications were lost. The result was always a 'stutter' and then some relatively small unexpected move - just enough to ruin the part. Removing the wireless card solved the problem.

[djharris]

My brother, Mike, works with you, Ben. I don't have a Shopbot but he was telling me what a "wonderful" week you guys have had and it got me to thinking.

My 2 cents: I didn't see any mention of switching out the wires themselves to the motors. The grounding suggestion seems to be the most promising, but a bad control wire or connection at the motor could also be the culprit. I'm also curious if a channel change makes any difference.

[gabepari]

Ben, I'm sure this won't help you, but I have been experiencing a loss of ~.5" when I run my spoiler surfacing routine. Never happens with any other files or jogs. Just when I surface the table, yet no jerking of any kind. When it returns to 0,0 after surfacing the table, it is -.5" off in the Y direction. Very weird.

It's not mechanical, I'm VERY SURE of that. Since it ONLY happens during surfacing, I haven't spent a bunch of time tracking it down, but I suspect it's the new control software, since that is when it started.

Good luck, and if you need any help with production, we are in Ontario, Very Sunny, Southern California. Not real close to you, but in a pinch it could work.

Gabe Pari
www.socalteardrops.com (http://www.socalteardrops.com)

[gabepari]

Ben, did you get this figured out yet. I just started getting intermittent ~.5" loss of position in the Y again. But, now it is happening after a file ends. Never happens in the middle of a file. After it's done cutting, it returns to what it thinks is 0,0 and it is between 3/8" and 1/2" off in the Y. The X not affected, and there is NO dramatic jump or stutter, just doesn't stop in the correct spot.

Again, I will stress that I am positive that it is not mechanical. I am in the process of regrounding the table (kind of did it half-assed the first time). We'll see if this helps. If not, I'm going to revert back to an older control software build, as I never had the problem before 3.5.3.

Thanks,

Gabe

[gpari]

OK, so we grounded the machine correctly, I may post some photos next week if I get time. Guess what, ALL of the loss of position problems have gone away. The machine is very stable now.

If you haven't grounded every single moving part on the bot to the table and then to an appropriate ground, do it now, it's worth it


Gabe

[Gary Campbell]

More info and pictures posted on the ShopBot Wiki
http://shopbotwiki.com/index.php?title=Grounding_your_ShopBot