Drifting Skipping Axis?

[verhoevc]

This is the first time I've had this issue, but it has me a bit scared now to cut any of the other expensive figures lumber I need to work on. This neck is longer than 24" so it was being milled on a diagonal line (hence the finishing path pattern). However, at the very end of the absolute final path/cut something happened. I don't know what as I wasn't around for it... but it looks like the x-axis lost some steps and so things started cutting too right right. As you can see the roughing pass did just fine. I have also re-tried this file in some scrap and it worked fine, just like all the other times I've run it.
What would cause something like this?
How can I avoid this in the future?
Any help that could put my mind at rest before I strap more expensive guitar wood to the table would be greatly appreciated!
http://i95.photobucket.com/albums/l139/verhoevc/CAD/728DE89C-278F-4632-B24F-B247F5C95192_zpsgnb4qjdo.jpg (http://s95.photobucket.com/user/verhoevc/media/CAD/728DE89C-278F-4632-B24F-B247F5C95192_zpsgnb4qjdo.jpg.html)
Chris

[verhoevc]

Info: 3IPS with a 1/4" ballnose. I believe it's 2IPS plunge. Shopbot Desktop and the path was made in MeshCAM.
Chris

[guitarwes]

Was anything prohibiting the machine from going over on the far side of the Y axis at the end of the cut, like a DC hose or whatever?

[verhoevc]

Hose hangs from the ceiling. When I came back and saw what happened I checked the rails but didn't find anything troubling. I also changed/clean absolutely nothing before running the path again in pine (with success). PS: it's been run in other hardwoods before this issue just fine as well.
Chris

[verhoevc]

This pic may help people understand what happened. Looking more this could be a jump forward in X, or the Y being held back and skipping...?
http://i95.photobucket.com/albums/l139/verhoevc/CAD/69C79D82-0955-4A5B-AF59-EC13DF6EFB0B_zpsp9mb3vuj.jpg (http://s95.photobucket.com/user/verhoevc/media/CAD/69C79D82-0955-4A5B-AF59-EC13DF6EFB0B_zpsp9mb3vuj.jpg.html)
Chris

[Kyle Stapleton]

Where are you? It is starting to get dry in the shop... is you DC hose grounded?

[srwtlc]

You could be loosing steps due to a couple of issues. In the first images, it looks like it happened right at the spot of taking a larger bite into the full height tabs. Depending on your % overlap, along with feedrate, the most stress or force required would be right there. I'd make those tabs much thinner so that they don't interfere with the ballnose path. If you want a better finish, at the cost of more cutting time, drop the overlap % down to about 10%.

[verhoevc]

I'm just north of DC in MD. DC isn't grounded as it's, now now, just my Rigid shopvac hooked up to a plastic hose hung from the ceiling to keep it out of the way. Static could be causing the motor to hiccup?!?!
Chris

[verhoevc]

I initially thought that to SRWTLC but that is a tab left by a previous path from before I flipped the piece over. This path goes up to it, but doesn't cut into it. Also, even if it does hit that tab for some reason, it would have done so at each level of the roughing pass as at the end of each pass there the toolpath runs a perimeter cut.
Also, there's 4 total tabs like this in this toolpath and it's never accidentally bumped one before, especially to this degree.
Chris

[guitarwes]

Could some of that dust skirt have gotten caught under that black hold down clamp and bogged the XY movement on that end causing it to lose XY position? Just throwing out stuff. Could be a static issue also as others have pointed out.

[verhoevc]

I guess maybe but I can't see how it would during that pass and then not do it during an identical pass when I re-ran everything in line?
Chris

[verhoevc]

This video shows me re-running things to check where the dust shoe come in contact with the clamps. At the end the brushes come in minor contact, but barely. This also was the carving path run without the profile pass being done. Aka: this is the equivalent of having full depth tabs at the edges everywhere:
http://i95.photobucket.com/albums/l139/verhoevc/CAD/th_0D7892A8-3FB7-415D-BC46-015807B50A39_zpsqtjn49sb.jpg (http://i95.photobucket.com/albums/l139/verhoevc/CAD/0D7892A8-3FB7-415D-BC46-015807B50A39_zpsqtjn49sb.mp4)

I think the thought that the was caused by static is the winner as once that was mentioned I did see that my dust collection hose's connection to the shopvac literally rested on the x-axis stepper motor. Given that the picture appears to be a jump forward on the x instead of something being held back, I'd say this is the most probable answer now.
Re-posting the OP pic since the link got broken and I can't edit the post for some reason.
http://i95.photobucket.com/albums/l139/verhoevc/CAD/728DE89C-278F-4632-B24F-B247F5C95192_zpsb7biowg3.jpg (http://s95.photobucket.com/user/verhoevc/media/CAD/728DE89C-278F-4632-B24F-B247F5C95192_zpsb7biowg3.jpg.html)
Chris

[scottp55]

Bare copper wire(stranded works best it seems) INSIDE the DC hose, and dead end it maybe on reinforcing wire of the hose itself. Don't let it touch any metal on the dustfoot.
I've got a steel can DustDeputy, so I just ran the copper to the base canister and screwed it, then on bottom of canister I put a mish mash of heavy bare iron and cutoffs of grounding rod on wire screwed to canister base so it drags on bare concrete floor.
Probably better to ground it to the Fein motor, but it works.
Every year at this time people who haven't grounded the hose have problems.
Hiccups and worse.
scott

[verhoevc]

I'm gunna try and get to this tonight, but had a question.... if my dust collection right now (at least where the CNC is) is just a shopvac I'm gunna have a hard time grounding. Can I ground to the metal of the Shopbot itself? This would be handy as it would allow me to run wire from the dust shoe, up the hose, ground it to the Shopbot frame there, and then have another leg of wire go down the hose towards the Shopbot. This would be handy as it'd still allow me to disconnect the shopvac easily for use elsewhere.
Thoughts?
Chris

[scottp55]

NOT to the Desktop!!!
Should be totally isolated.
Bare concrete floor?
Easy way to run wire ,if clear hose, is to wrap wire around a small cloth that will fit inside hose. Turn on vac and feed it into hose until almost to vac,then fish it out.
IF bare concrete floor, then maybe poke hole in hose before the adapter to shopvac(seal with silicone after),and run to mish mash of conductive metals(maybe one loop around outside of hose to prevent pulling on it?). OR if Desktop is in a fixed position is there any way to maybe drill small hole(or insert wire brushed TapCon) in concrete and put wire in wedging it with a nail or similar? Any bare metal legs nearby in direct contact with floor.
Some body who understands VOLT Gremlins and ground loops PLEASE chime in!!! :)
Electrical idiot trying to give advice:(
scott

[verhoevc]

OK. Well I ran some copper wire through the pipes. Attached them to the ground of a plug. And have kept the hose away from any servos. Seems like so far this is a win! Thanks for the help you guys! I now have successful cuts under my belt in a cold environment.

Exactly what I did in case it helps someone in the future:
1- took copper wire and ran it from the dust foot to the junction with my shop vac hose
2- attached the halfway point of that wire to a new copper wire that I kept the plastic coating on and ran to the ground of a plug
3- sent the other half of the wire from step 1 down the hose to the shopvac
This lets me discoonect my shopvac while still ensuring copper flows down that hose as well.
Best,
Chris

[verhoevc]

OK, now I'm pissed! A $5-$10 neck blank loss is fine. But this was a $50 African Blackwood rosewood fretboard this time!
I thought this was addressed. As you can see above, I have grounded everything and the extraction hose no longer runs anywhere near the machine's motors.
However, once again, it decided to only screw up when I finally trust it enough again to go do something else while it runs. For some reason it decided to lose Y-steps again while running a 3D path again made in MeshCAM (exact same scenario as previous failure above... ironically on the replacement neck for the previous failure).
As you can see, this wasn't a single "jump" in position, it continued to wander more and more as it went along:
http://i95.photobucket.com/albums/l139/verhoevc/CAD/DEAC1CAD-FB6A-44A0-A43F-311356D217A8_zpsztriocxw.jpg (http://s95.photobucket.com/user/verhoevc/media/CAD/DEAC1CAD-FB6A-44A0-A43F-311356D217A8_zpsztriocxw.jpg.html)
However, this was a file I had previously run successfully! So first thing I did is re-zero'd the axes (yes, the Y was incorrect at the end of the paths) and ran the exact same file again. As you can see here, it finished all the way to the end of the fretboard without issue this time:
http://i95.photobucket.com/albums/l139/verhoevc/CAD/D8F4F7E7-7EDF-49A6-A575-9558D5006E5E_zps3wem9y2o.jpg (http://s95.photobucket.com/user/verhoevc/media/CAD/D8F4F7E7-7EDF-49A6-A575-9558D5006E5E_zps3wem9y2o.jpg.html)
(Granted, it looks like I needed to set the Z a little lower so the radius'd fretboard goes all the way across the board, not missing the center as seen here).

This is so frustrating! Both times this has happened it's been when I haven't been around to watch it. Having a CNC I can't trust to do what it's supposed to do without me watching it really defeats the purpose of having a CNC! Now I'm scared to put anything on it!
It feels like I have Schrodinger's CNC: it has both destroyed and not destroyed my work until I check on it.
Please help!
Chris

[Ajcoholic]

OK, now I'm pissed! A $5-$10 neck blank loss is fine. But this was a $50 African Blackwood rosewood fretboard this time!
I thought this was addressed. As you can see above, I have grounded everything and the extraction hose no longer runs anywhere near the machine's motors.
However, once again, it decided to only screw up when I finally trust it enough again to go do something else while it runs. For some reason it decided to lose Y-steps again while running a 3D path again made in MeshCAM (exact same scenario as previous failure above... ironically on the replacement neck for the previous failure).
As you can see, this wasn't a single "jump" in position, it continued to wander more and more as it went along:
http://i95.photobucket.com/albums/l139/verhoevc/CAD/DEAC1CAD-FB6A-44A0-A43F-311356D217A8_zpsztriocxw.jpg (http://s95.photobucket.com/user/verhoevc/media/CAD/DEAC1CAD-FB6A-44A0-A43F-311356D217A8_zpsztriocxw.jpg.html)
However, this was a file I had previously run successfully! So first thing I did is re-zero'd the axes (yes, the Y was incorrect at the end of the paths) and ran the exact same file again. As you can see here, it finished all the way to the end of the fretboard without issue this time:
http://i95.photobucket.com/albums/l139/verhoevc/CAD/D8F4F7E7-7EDF-49A6-A575-9558D5006E5E_zps3wem9y2o.jpg (http://s95.photobucket.com/user/verhoevc/media/CAD/D8F4F7E7-7EDF-49A6-A575-9558D5006E5E_zps3wem9y2o.jpg.html)
(Granted, it looks like I needed to set the Z a little lower so the radius'd fretboard goes all the way across the board, not missing the center as seen here).

This is so frustrating! Both times this has happened it's been when I haven't been around to watch it. Having a CNC I can't trust to do what it's supposed to do without me watching it really defeats the purpose of having a CNC! Now I'm scared to put anything on it!
It feels like I have Schrodinger's CNC: it has both destroyed and not destroyed my work until I check on it.
Please help!
Chris

Try running a file without your shop vac on. Just get it out of the way. I think it's a static issue. You'll know if you eliminate that possibility.

Unless you are using an expensive vacuum like a festool, with an anti static hose, vacuuming chops causes a lot of static buildup and discharge.

[verhoevc]

I finally broke down and called the Shopbot folks. They're thinking it's likely a bearing issue. I one-by-one took the linear bearings off the gantry to test them. I'll upload the videos hopefully tomorrow. See what ya'll think. I think they're right and that at certain, random, points they are binding for some reason and causing these issues.
Hopefully replacement linear bearings aren't pricey...
Chris

[verhoevc]

Here's the videos of messing with the linear bearings independent of the gantry:
http://rs95.pbsrc.com/albums/l139/verhoevc/CAD/63C964FE-9D23-422A-A010-2D3D23482C26_zpsfogfmcxm.mp4?w=160&h=160&fit=clip (http://i95.photobucket.com/albums/l139/verhoevc/CAD/63C964FE-9D23-422A-A010-2D3D23482C26_zpsfogfmcxm.mp4)
http://rs95.pbsrc.com/albums/l139/verhoevc/CAD/BC2036EF-3544-4D89-A7AD-69CD67A7DA61_zpssnbyphvw.mp4?w=160&h=160&fit=clip (http://i95.photobucket.com/albums/l139/verhoevc/CAD/BC2036EF-3544-4D89-A7AD-69CD67A7DA61_zpssnbyphvw.mp4)
http://rs95.pbsrc.com/albums/l139/verhoevc/CAD/60B0E32F-C0AC-4884-A72E-D8EA993F0681_zpsagdmmfve.mp4?w=160&h=160&fit=clip (http://i95.photobucket.com/albums/l139/verhoevc/CAD/60B0E32F-C0AC-4884-A72E-D8EA993F0681_zpsagdmmfve.mp4)
This one is easily the worst bearing as you can hear me say in the video. I also later pulled this bearing off the rail and it is missing a ball:
http://rs95.pbsrc.com/albums/l139/verhoevc/CAD/20EE0EFC-D00B-4A4D-9CB3-6C9A66620CAA_zpsqut2vynj.mp4?w=160&h=160&fit=clip (http://i95.photobucket.com/albums/l139/verhoevc/CAD/20EE0EFC-D00B-4A4D-9CB3-6C9A66620CAA_zpsqut2vynj.mp4)
Chris

[verhoevc]

Lastly, here is me moving the gantry pushing on the left side (side with the worst bearing imo):
http://rs95.pbsrc.com/albums/l139/verhoevc/CAD/B6138D74-2901-4025-8BA9-13DCAC595082_zpsyodu668r.mp4?w=160&h=160&fit=clip (http://i95.photobucket.com/albums/l139/verhoevc/CAD/B6138D74-2901-4025-8BA9-13DCAC595082_zpsyodu668r.mp4)
Towards the end of the video you can hear it bind several times as it runs along.
Sadly, I've also pushed the gantry around without that bad bearing attached and I still get these binding issues. However, perhaps that could be a symptom of simply not having equal load support without that 4th bearing attached? Who knows...
Chris

[scottp55]

I wouldn't push the gantry around like that. Motor plus electricity= movement...motor plus movement= electricity.....you can fry components on your board unless motor wires have been disconnected:(

[verhoevc]

That's what I would have figured too... I never had until the Shopbot folks told me to...
Chris

[jerry_stanek]

Be very careful doing that with your steppers plugged in

[verhoevc]

I had that discussion with them and the rates at which I'm doing these tests damage is not likely.
That said: hopefully this has reached resolution! Apparently the linear bearing sided in my year's Desktop production were succeptible to degradation from two very specific dust types: pine and corian. Almost ever single test I did while learning was done in cheap, Home Depot pine... sigh.
Good news: since these bearings are no longer used they have a few spare that they were happpy to send me and help me with some other stuff to keep ALL dust off my rails.

They've been super helpful! Will update once I've made the swaps.
Chris

[verhoevc]

Ok, so I replaced all the bearings on either side of my Desktop but am just now getting around to actually testing them. I'm still not convinced I've gotten everything covered though. I wanted to include a video of some strange. Poses the machine makes sometimes when moving on a diagonal path. It is still doing it after the bearing swap so I figured I'd see if people had any opinions on it:
http://rs95.pbsrc.com/albums/l139/verhoevc/58546FCE-D19E-4F3C-8D51-5DE570B9A912_zpsw234cbdo.mp4?w=160&h=160&fit=clip (http://i95.photobucket.com/albums/l139/verhoevc/58546FCE-D19E-4F3C-8D51-5DE570B9A912_zpsw234cbdo.mp4)
Thoughts?
Chris

[scottp55]

What's the Move Speed on that roughing move?
I've had similar sound/vibration on radio button moves in that direction(I forget the MS) and changing MS down .2IPS made it smooth again.
Maybe run the aircut, and using Shift(holding it down) and the < > keys to see if a few .1 IPS changes makes it run smooth, and then mod the the file MS and spindle speed?
Every machine has slightly different spots with bad resonance, and seems you found one.
scott

[verhoevc]

Changing the speed did not have an impact on the vibration on the two legs where it does it... at least not in a positive sense. If anything things got worse as the feed was slowed.
I can do another video if it helps? But I don't know how much will be learned from seeing it do the same, just at a different pitch, from the last video.
Chris

[srwtlc]

Have you checked into adjusting the pots on the drivers? You may be able to dial out some of that resonance.

[verhoevc]

No, I'm fairly new to this whole thing so a lot of this stuff, like adjusting pot values?, is Greek to me.
The reason I thought this might be an issue isn't because it's ever failed while making that noise... but instead both times it has failed has been when I've been milling something along a center-line that's a diagonal between the lower left and upper right of the table. Moving along this line is when it sometimes makes that noise and vibration... moving along the X-axis for a part along that line on a finishing pass is when it has failed both times.
Chris

[scottp55]

About minute 5.
https://www.youtube.com/watch?v=3iwKFDER494

Have you checked tightness of X rail bolts, and X stepper?
Both of those loosened up on me after about 18 months.

[verhoevc]

Yup. Everything has been tightened, nothing was what I would have called loose.
I may get in there and try and adjust those pots. But before I start messing with electrical boards I wanted to ask: if slowing/raising the feed rate at the time that it is loud and vibrate-y didn't affect it, do we still think adjusting that pot will have a positive impact?

I wanted to throw something else out there. Could I be perhaps asking too much out of my machine? I know during the second failure I was running a 2.6ips,2ips,16k feed/speed. That's just under (.004875) the chipload range (it's a .005-.007) that my Onsrud chart says I should be for that 1/4" bit in a 2-flute. I used to run at slower feed/speed but while research these calculations I read somewhere that folks were doing all their hardwoods at 16k RPM, which makes my 2.6ips make some sense.
This may have also been my feed/speed I was running the other time the failure happened as well, I just can't remember; but it's likely.
I've recently been reading more about feed/speed recommendations for hardwoods and I see a lot around the 1.5/.7-1/12k (chipload or .00375... so even lower out of the Onsrud range) suggestion. This is a pretty big difference over what I was running. Granted, the second failure that I know was at the higher feeds/speeds was a finishing path through a mere .050" of wood... hardly what I would have called taxing... but I thought it was worth throwing out there.*
Chris
*I thought it might be worth throwing out there because when I first got the machine and Iw as just playing around with it not cutting anything I set the IPS higher than the Desktop can handle and that made it definitely lose track of where it was supposed to be!

[scottp55]

I wanted to watch all the vids and reread everything and think about it.
Never having cut long diagonals,or that big a 3D, I still might try 2,2 and even 1.5,1.5 for Move Speeds. 2.6,2 is faster than I've ever 3D'd.
A friend with a Desktop in OZ cuts 2,2 all the time, and I tried it, but on small stuff it was a little too jerky for my taste.
I wish one of the Desktop Luthiers would chime in:(
Chiploads for the Desktop are usually only a very rough guideline I've found.
No way on the finish toolpath it will ever get up to 2.6, but it will try.
Try some scrap at 1.5,1.5 and then 2,2?
Have you read any of Brady's VR threads or blogs?
Here is one;
http://www.talkshopbot.com/forum/showthread.php?21218-Public-Service-Announcement-Make-More-in-Less-Time-via-Tuning&highlight=public+service
Wish I could help more:(
scott

[dmidkiff]

Thanks Scott. This was very timely as I just tried the smallest V-carve for me. Very jerky. Will try some of these settings and watch the machine for smoother cuts. Merry Christmas.

[scottp55]

Merry Christmas all:)
Glad the shop is up and running David:)
Yep, no sense telling the machine to go faster than the cut allows.
Quite frequently on Teeny fonts and such, I'm down in the sub .5IPS range in dense hardwoods because of tearout.
I'm still on Desktop default VR settings because I'm mainly one of's now....I'm chicken:)
scott

[verhoevc]

I don't want to claim victory just yet... but I successfully milled out a new maple neck using 1.5,.7,12k. I still don't see how when the machine says it'll cut up to 4ips that doing a finish path through a minuscule thickness of rosewood would give it pause... but it's looking like it did.
I'll be doing more and more tests on cheaper woods before pulling out the fancy stuff again, but hopefully this is the last of my woes?
Best,
Chris

[Gary Campbell]

Chris...
I will take a stab at your question: "I still don't see how when the machine says it'll cut up to 4ips that doing a finish path through a minuscule thickness of rosewood would give it pause..."

It will. Give it a vector that is 4" long.

The advertised speeds for most any mechanical product are the maximums. Like the speedometer in your car. You will seldom, if ever, run at the maximum attainable speed, but we as humans often make our decisions based on those maximums. What you need to realize as you go forward with this craft is that when doing 3D work you may seldom have a vector more than a few hundredths of an inch long, many just a few thousandths.

Toolpath parameters at near max speeds cause over acceleration and deceleration, which in most cases cause a loss of position or lost steps. The machine will keep attempting to reach the speeds you set, until it fails, as you have found out.

[verhoevc]

So this appears to be back... whether I just got lucky for awhile, or it was happening in smaller amounts to a point I couldn't tell as readily, or I just don't use my machine that often (probably cause I have to deal with stuff like this all the time!)... I don't know.
But here's the newest data: It seems to ONLY happen when I'm working in 3D! It also never happens on the roughing paths, just the finishing paths! I contacted the MeshCAM guys and got this answer:

"It's tough for me to say what's going on in the part since I don't know what it should look like or what the orientation is but I'm familiar with the skipping steps problem.
First, when the job is done, jog the cutter back to 0,0,0 and see if the zero point moved. If it did, you're skipping steps.
This is almost always due to:
- ESD/Static electricity if you're in a dry environment.
- Aggressive acceleration and speed limits set in the router itself.
- A bad motor or drive board.
- Mechanical binding or jamming on the linear motion hardware
MeshCAM may show this behavior more because the toolpaths raster back and forth more than 2D CAM programs so there's more opportunity for the problem to occur."

Apologies for anyone going back and reading it, for unexplained reasons over the last couple years there have been a bunch of words randomly repeated, etc. in the posts. Through this thread I've already addressed the static issue. We've talked about my feeds/speeds; although I still think 2.25 xy/2z IPM doesn't sound crazy. Not to mention it works MOST times just fine. I'd have also thought if this was the issue I'd see it on both the roughing and the finishing paths? Lastly, we also addressed mechanical jamming as Shopbot gave me a bunch of new bearings... sadly they were all for the Y gantry. Maybe I need to call them up and ask for the X axis bearings too?

Please help, I'm back to being afraid of my bot :(
Chris

[scottp55]

Sorry to hear Chris...I thought we were done.
One thing if you could check, that we didn't address.
Type UT in the command line of SB3, and read the UD instructions.
Then type MX,4 an doing nothing else, type UD and let us know the Packet_ET number please.

It's been almost 2 years...have you blown computer out? USB ports? Checked for any wires that may be crimped/broken? checked connections?
I'm fishing a little here you can tell:)

Does it do it on an air cut?
For kicks and giggles try 1.5,1.5 IPM again?

Check the Packet ET first though. That's Comms.
scott

Oh....Has anything been added to your computer just before this started again?

[verhoevc]

Packet_ET - 16.5 (ms;good<20)
No idea what that means though hahaha.

It has indeed been awhile but I've been doing mostly 2D work with my machine and therefore I guess I just wasn't seeing it happening. I was lucking out or something.
Connections seem to be good. I have not blown my computer out, but I guess I could give that a try.

Yes it does it with air cuts. Over the last few days I've run a long (35 minute) 3D path over and over. First round it failed after the second run of the path. I came back and the bit had lost about 1/32" in the X and lost about 1/4" in the Y.
Next round I ran it probably half a dozen times (same path) before it failed. This time only lost in the Y, but it was closer to 3/8". I was also watching the machine this time and saw the "random slow downs" you can see in the video below. I have had this issue before but for some reason didn't think to connect the two things. Basically the machine will just random lower the feed rate to super slow. I can then stop the path and resume at which point it speeds up to normal again. I don't know if this is the CAUSE of the lost steps but seems interesting. Another thing to note is that it seems the steps are lost regardless of whether I stop/resume or just let it finish at the slower feed rate. I have seen this happen NUMEROUS times in 3D paths. I have seen it happen in a 2.5D path as well, but it's very rare.

https://www.youtube.com/watch?v=sl7WQZ77aKY

Same computer, nothing added. Doesn't even connect to the internet and is used for 3 things: notepad (notes on path orders for complex projects), Vectric, and SB3.
Best,
Chris

[srwtlc]

On the slow down issue, what version of SB3 are you running? I had that same slow down issue with 3D cuts on my PRS Alpha until one of the latest versions. I'm running 3.8.48 now and haven't had the slowdown issue. You can update to that one or one of the next two version (3.8.48 or 3.8.50).

As to the lost steps, you should look at adjusting the ramp settings for running 3D files. The default settings can cause lost steps upon direction changes if they're too abrupt. You can save your defaults and then set up for 3D and save those. That way, you can return to them as needed. Post a clear shot of your current ramp settings (from the VR screen). Use the snipping tool in Windows and choose a rectangular snip and select just the fill-in sheet screen. That way it will be large enough to see here.

If you need more info on saving ramp values for later use, just ask.

[verhoevc]

Yeah, I'm a little behind hahaha. I'm running 3.8.14 which is just what came on the original install when I bought the machine (eg: I've never updated it ops). Firmware version is 220.
I just installed 3.8.50 so let's see how things go!

Here's the screenshot of my ramp settings. I have literally never been to this screen before, so I'm sure this is what they've always been. This shot was taken when I was still running version 3.8.14
32084

Best,
Chris

[scottp55]

Yeah Chris, Go to 3.8.48 like I'm running also. I remember a LOT of issues with 3.8.14 which my machine had.
https://support.shopbottools.com/support/control-software
You've been fairly lucky as there was a MASSIVE Pause/Resume glitch right around that time if you spacebar Paused during a Z move.
I also think there's something funky sounding at the reverse direction moment.
Can't remember at the moment if you HAVE to UI (Utilities tab Install control firmware), but I always do anyways.
Think capital S Scott nailed it:)
scott

[srwtlc]

Ok, first off, get up and running with the new version. Be sure to perform the firmware update if it didn't automatically do it. Go to Utilities/Update Control Box firmware and follow the prompts. Some of those early 3.8 versions were really testy!

On the ramps, read up on it here (http://Ok, first off, get up and running with the new version. Be sure to perform the firmware update if it didn't automatically do it. Go to Utilities/Update Control Box firmware and follow the prompts. Some of those early 3.8 versions were really testy! On the ramps, read up on it here.). Try 150 for the 3D ramp threshold, .08 for the minimum distance to check, and 35-40% on the slow corner speed (these would be for 3D work, leave the .08 as is, but put the others back for normal work) Your feed rates will differ from those in the article, but I'd try the 1.5 to 2.0 range for both. The slower the z, the more the xy have to slow down for the z.

[scottp55]

Chris,
I was also thinking of other things that weren't mentioned.
Before the machine arrived I talked to everybody from Ted and down at Shopbot, and some other Desktoppers asking stuff to do before it arrived.
We put a separate 20A breaker in the panels/subpanels that is used ONLY for the Shopbot in all 3 Desktop locations.
We located that breaker away from the other large draw sources.

I'm on country power with lots of "Blips" and a few second power losses, also had power drops when 220A compressor/hot tub/dryer/furnace kicked on etc.
Decided to add a CyberPower Uninterrupted power source to all 3 machines(more as a line conditioner, but saved my butt several times on the short blackouts).
A lot of the little glitches and gremlins disappeared after that.

Have any problems showed up when you noticed something kicked on?
Are all wires coiled neatly in the same place close together....WRONG! Crosstalk can cause all kinds of problems, usually intermittent.
Spread them out a bit, and keep large power cords away from the USB cord.

Scott's probably got the major problem sussed, but thought I'd mention these as may be why I have had so few problems for years.
scott

[verhoevc]

Been MAD busy. But I’m most way through reading the suggested materials and making changes. Hoping to find time to run a bunch of tests soon too if the day-job will chill out. Will keep you all posted. Thank you thank you!
Chris

[verhoevc]

@srwtlc the ramps learning link doesn't seem to be working?
Best,
Chris

[scottp55]

Good reading here Chris;
http://www.talkshopbot.com/forum/showthread.php?21218-Public-Service-Announcement-Make-More-in-Less-Time-via-Tuning&highlight=corner+speed

[verhoevc]

So far with the new software and firmware upgrades and VR settings my tests are going well (even at my old feed rates!... no point in changing everything at once or else it'll be hard to tell what helped and what didn't).
More to come!
Chris

[verhoevc]

So I had a LOT of success with these new ramp settings in 3D! I did a HUGE several-hour run with no issues. Along with a couple other smaller 3D runs. All went off without a hitch.
Need to remember to switch back to the 2D settings afterwards though as I forgot to and got... less than desirable results with that. Not the end of the world, nothing that won't sand out, but hoped for better.
I did still have one instance in a 2D path where it stopped and paused for a good while. I stopped the machine and set everything back up. Who knows, maybe after it started moving again everything would have been fine... but given my past experiences I wasn't really inclined to find out.
Still not comfortable leaving the machine to run on it's own though... which kind of defeats the purpose of a CNC!
Chris

[Brady Watson]

Please run the US command in SB3 so that it creates a snapshot of your settings and creates an .SBD file. Give it a meaningful name and save it to your desktop or you might have trouble finding where you saved it. ZIP it up or rename the .sbd to .txt and post here. - OR, just open it up, copy the text and paste into the body of a post/message.

It appears to me that you have some settings that are not working for what you are doing. This would be especially true if you have been messing with VR commands ~ it's VERY easy to get them to a point where the tool does weird things.

It's also important to note that you are running the tool in all 3 axes when rastering, which uses up a lot of COM. 3,2 might be a bit too aggressive - but lets look at your settings first.

-B

[tlempicke]

I have the same desktop and have dealt with a similar problem. If the machine is making a "groaning" noise when the Y axis moves you have fouled the linear bearings on the left and right side of the machine. I see from your earlier posts that you have already been into this, but let me add some extra light to the situation.
I can see that you are running speeds that produce chips rather than sawdust, a very good thing. But keep in mind that there is also some very fine wood dust being produced. It lands on the rails of the Y axis and will eventually work its way into the linear bearings and cause the balls inside to jam and stick.

There is no real adjustment for squaring the Y axis, but you can help it a little bit by loosening the screws on each side and holding it as square to the X axis as you can while you tighten the screws.

Another good thing to do. At the front of the machine loosen the screws that hold the end of the ball screw for the Y axis. Run the axis all the way forward and tighten the screws. Now loosen the screws on the motor at the rear of the machine. Run the axis all the way to the rear and tighten these screws.

I also removed the bearings, one at a time, and cleaned them in mineral spirits and lubricated with WD-40. I have two sets of bearings now so that I can just change them out if need be. Finally I made a couple of "Elephant Ears" that magnetically attach to the cheeks of the gantry and keep dust from falling on the rails of the Y axis. As part of the daily shut down I wipe the rails down and use a little 3 in 1 oil on them. The real answer is two motors driving two jackscrews for that axis but the foregoing has kept my desktop running smoothly for a couple of years now.

[Brady Watson]

For sure the Y axis bearings are very susceptible to getting jammed up from chip/dust infiltration. Take care when servicing them. Without a complimentary section of rail to slide them onto when removing - the balls will fall out. SB sells replacement bearing blocks. Be sure to note what brand they are - THK or PMI when you order...they are not cheap. The placement of the rails on the DT was a poor design decision. The added protection of the DT Max is way better.

As Tom pointed out - a chip guard of sorts will help keep them clean. I did the same thing with some 6" tall (just enough to clear the bottom of the gantry beam) coroplast mounted on some angle iron. It makes a huge difference. As also noted, after or during - it's a good idea to clean those rails. You can vacuum them off or blow them off, but avoid blowing chips into the bearing blocks.

I saw your VR settings posted above, but since then you commented about altering your VR settings. Please post them up again or copy/paste the content of your SBD file AFTER you run US.

[CNCDave]

@tlempicke: Ugh.... I'm getting the "moan and groan" again on my DT (THK bearings). Last time this manifested itself as jittery behavior while doing small-ish circles (under an inch in diameter -- same conditions repeating). Problem was solved by replacing one of the bearings on my gantry.

I ass-u-me-d that once the bearings "failed" they were done for (though at the replacement cost, I kept the old one for posterity). It sounds like you have had success with cleaning and re-lubing the bearings, and in fact, have a good maintenance plan for keeping them in good shape. Any side-effects to the mineral spirits and WD40?

Also, I make it a habit of vacuuming and brushing all dust from the rails (both x and y) after most tool changes, but liked the idea of your elephant ears. Do you have a picture of your elephant ears?

[tlempicke]

Sorry to take so long getting these to you. Here are four pix. The first shows the "elephant ears" in action, the fourth shows the magnets that hold them to the supports for the gantry.http://www.talkshopbot.com/forum/attachment.php?attachmentid=32555&stc=1http://www.talkshopbot.com/forum/attachment.php?attachmentid=32553&stc=1http://www.talkshopbot.com/forum/attachment.php?attachmentid=32554&stc=1http://www.talkshopbot.com/forum/attachment.php?attachmentid=32556&stc=1