Help! My machine is a mess!

The issues I’ve had with my machine have been numerous and have existed since day one. I have contacted support many times but they have stopped responding to my emails so I thought I’d see if anyone here has any insight.

My most consistent failing is an inconsistency in depth control. After loading a job I’ll set zero using my BitZero then start the job. The machine probes the bit using the BitSetter and then commences the cut. Often times it will cut the correct depth but maybe 1/5 times it’ll cut an incorrect depth. Occasionally it’ll run the job ~5mm above the work piece but just as often it’ll plunge into the work piece ruining it. When I recheck the zero settings X and Y are always still zeroed on the correct spot but the Z zero is incorrect. Somewhere between setting Zero and starting the job it forgets Z zero. I have tried to set zero manually, without the BitZero and I’ve tried it without the BitSetter but I get similar results. My solution has been to shut down the machine, remove power close Carbide Motion and leave it for 10 minutes then re-initialize and start over, resetting the process and having continued lack of success with the failure presenting ~1/5 times.

A few other issues with my machine: EMI freezes despite having in place every EMI mitigation that Carbide recommends. Inconsistencies in the X and Y axis also exist. Last week I performed maybe 20 successful jobs but right in the middle of my day 2 jobs in a row misaligned a series of cuts. One misaligned by 3 or 4 mm south so I reset the machine, reloaded the file and the next job did the same thing but this time 3 or 4 mm to the north. I reset again and this time everything was fine and I continued another 10 or so jobs successfully. Everything is clean, appropriately tight and tuned, no physical issues that I can identify.

My machine is a Shapeoko 3. As I read the list of improvements on the Pro and on the 4 I notice that they advertise larger, improved belts, V wheels, homing switches and improved, EMI resistant electronics. It seems to me that all of these are improvements on the components that have left my machine so unreliable. Is it safe to say that Carbide is aware the the 3 had component shortcomings that needed to be improved? I can say that those shortcomings have been so severe that my 3 is nearly useless? If that logic is followed then what is the responsibility of the company to owners of 3s with those shortcomings? What I’m saying is that I think the 3 was a lemon, I think they know it and I think it sucks that those of us who own 3s (or at least my 3) are left in the dark.

Okay, I’m a little sensitive. Rant not withstanding, is anyone else seeing any similar issues? Does anyone have insight on how to rectify the X, Y and but especially Z inconsistencies? I won’t even ask about the EMI because I’ve already invested enough time and energy chasing that down.

Hi @mynamejefff,

You may or may not believe me, but here goes.

There are no intrinsic design flaws in the SO3 that would have been “fixed” in more recent models, I’ve been using a standard size SO3 successfully for 4 years (some folks much longer than that), and hundreds if not thousands of other people too. BUT it’s a CNC kit and everyone’s use is different so statistically speaking there are a variety of hurdles that can present themselves. I’ll try and give you a few tips that helped people seeing similar things in the past on the forum.

The Z inconsistencies can be due to:

  • zeroing somewhere that is not consistent with where you defined the zero point in the design file. Often times when people see the machine cutting air, it is because they set the zero reference to “stock bottom” in CC but actually zeroed on the top of the stock. It’s an easy mistake, I’ve been there.
  • slippage of the endmill inside the collet due to inadequate tightening force, or dust in the collet, or a plain defective collet.
  • use too aggressive plunge rate and losing Z steps, which messes up with the Z0 reference.
  • using the BitSetter but not quite following the recommended process, in particular swapping the tool without using the “Load New Tool” button or when prompted, results in a wrong Z offset then being calculated when starting the job
  • on belt-driven Z axis, loose setscrews on the Z motor pulley
  • a problem in the Z wiring to the controller. The Z motor may be getting spurious pulses. Quite the unicorn, but theoretically possible.

The EMI gremlins are often related to sub-optimal grounding of the machine and dust collection, but since you went through fixes with support already I won’t mention the usual suspects again, however you may be interested in this thread: Grounding your Shapeoko

The X and Y offsets, in my experience, relate to the following root causes in 99% of the cases:

  • lost steps due to too aggressive feedrate or the router/gantry hitting a hard limit or the endmill hitting a clamp
  • loose set screws on a motor pulley. If you haven’t already, you should definitely inspect EACH pulley and make sure one set screw is aligned to the flat on the motor shaft, and both set screws are tight.
    Your description of things working “for a while” and then getting X/Y offsets makes me think the pulley set screws hypothesis is likely (the setscrews could be tight enough than it works most of the time, but slips when there is extra force due to the toolpath or material being cut)

If you are willing to work with the community (along with support depending on where you are with them), we can probably figure out what happens and get your machine to a reliable working condition. Call me biased but my SO3 has been a workhorse, so there is no reason why yours shouldn’t be the same, but it’s likely that it will take some effort to go and inspect it mechanically, and possibly review the kind of setup/workflow you are using.

11 Likes

You perhaps should search this forum. There have been a lot of issues relating to Z height issues resulting in plunging too deep into the stock or cutting air, for both the SO3 and SO4.

This is the thread I started, when this one got closed down.

Good luck!

1 Like

I appreciate the detail in your answer, thanks for taking the time.

Fair enough, maybe the 3 is not a lemon, maybe mine is or maybe it’s me. I can admit my own shortcomings as it has been a steep learning curve and I’ve made mistakes. With that said I’ll address the suggestions you’ve made and I think we can agree that there seems to be a deeper issue.

In terms of the Z zero:

  • zeroing somewhere that is not consistent with where you defined the zero point in the design file. Often times when people see the machine cutting air, it is because they set the zero reference to “stock bottom” in CC but actually zeroed on the top of the stock. It’s an easy mistake, I’ve been there.

I almost always use the same zero points, SW and Stock top. But to put a finer point on it, why would I see the Z zero issue while using the same file 5 times but only see it once?

  • slippage of the endmill inside the collet due to inadequate tightening force, or dust in the collet, or a plain defective collet.

I have seen this issue before so I make sure that the collet is clean, tight and effective as part of my workflow. I even started measuring the depth of the end mill by hand before a cut in order to detect any slippage.

  • use too aggressive plunge rate and losing Z steps, which messes up with the Z0 reference.

Entirely possible as I am no expert on feeds and speeds. My only push back would be the same question about using like files. If my file has too aggressive a plunge rate wouldn’t I see issues every time? It’s also worth mentioning that support has seen the files I use, they have made some recommendations which I’ve implemented but to no avail.

  • using the BitSetter but not quite following the recommended process, in particular swapping the tool without using the “Load New Tool” button or when prompted, results in a wrong Z offset then being calculated when starting the job

That makes sense but I don’t think it applies as I don’t have that particular bad habit, I believe I use it as described and, most importantly, I use it the same way for each job yet get inconsistent results.

  • on belt-driven Z axis, loose setscrews on the Z motor pulley

I’ve been over my machine with a fine tooth comb while on the phone with support. Every time I begin using my machine I start by cleaning it and double checking all screws, belts, V wheels, etc.

  • a problem in the Z wiring to the controller. The Z motor may be getting spurious pulses. Quite the unicorn, but theoretically possible.

Interesting. Haha, not sure how to investigate that.

In terms of the XY:

  • lost steps due to too aggressive feedrate or the router/gantry hitting a hard limit or the endmill hitting a clamp

Same answer as above for the federates. I have made the second mistake though. If I err and lose a step due to hitting a clamp or any other cause I shut down and re-initialize.

  • loose set screws on a motor pulley. If you haven’t already, you should definitely inspect EACH pulley and make sure one set screw is aligned to the flat on the motor shaft, and both set screws are tight.

Done.

Again, I appreciate your thoughtful response and I don’t want to sound exceedingly negative. After having worked with Support for nearly a year on these issues and having tried all of those solutions I still get these anomalies so exceedingly negative is what I’ve become. I’m happy to look inward but if my workflow is consistent, the file I’m running is the same but my results vary I have to point to the machine. While looking at the machine, I have tuned it up as instructed (set screws etc.) but the issues persist so I question inherent flaws, design, components, etc. Perhaps I’m wrong, I only have my machine to base that opinion on, I sure wish I could describe mine as a workhorse but so far…nah.

2 Likes

About the Z issue:

I did not get from your first post that the “1/5 times” related to using the same gcode file repeatedly. Indeed this rules out a mistake in consistency between the design file and zeroing point.

So you also ruled out endmill slippage, one less root cause to consider.

Understood. I will just mention that it’s easy to do things slightly differently (or more likely, from a different starting situation?) and not notice it.
One thing you said in your first post is critical in the investigation:

Are you 100% positive that you still get Z inconsistencies 1/5times when you disable the BitSetter from CM settings ? I believe you of course, it’s just very important to double-check this as this steers the investigation in a very different direction. If you can, try working without the BitSetter for a little while, to confirm.

Now if you do get those issues with BitSetter disabled, did you double-check Z0 just before starting a job ? With BitSetter disabled, CM does not alter Z0 at any time other than when using the BitZero (or setting Z0 manually)

It’s not easy indeed. A quick check can be to jog Z up and down, and while it is moving, gently wiggle the Z motor wires, see if the movement gets jerky/erratic. If there is an intermittent break/poor signal, it should show. And if not, you’ll know you should look elsewhere.

Now for the X/Y:

  • at the risk of sounding redundant, I like to double-check whether my pulley slip by drawing a line across the pulley and shaft (pic is from a Pro, but you get the idea):

And then WHEN you notice an X/Y shift in the cut, go and check that the line is still straight.

  • the other cause I could think of is again an intermittent wiring issue. X or Y wires may have been pinched, or may have been weak from the start (just brainstorming here…), and intermittently transmit signals incorrectly.

No worries, I get how annoying this situation must be. Slow and steady, we’ll probably figure out what’s going on.

5 Likes

Ive used my Shapeoko 3 xxl almost everyday for 30 months now. With No Major issues to Report.

Try checking your z manually before starting job and after finishing.

Rapid to xy. Rapid z+6
Move over work piece.

Hit 3 and jog down 6 steps. If right on material all is good.

Do the same after the job is complete.
See if you see a variance.

That would be the simplest way to verify if your loosing steps due to mechanical issue during job since you’ve ruled out EMI
If your having belt issues maybe look into Z plus? Plunge rates directly effect Z belt slippage.

Stuff gets loose over time, I try and do a machine check at least every two weeks. I have come across issues before they really became a problem. Loosening vwheels, belt stretch, nothing else really. Maybe a few brushes for router here and there.

5 Likes

One other quick thing you can do to check for endmill slippage is to mark the endmill at the point where it enters the collet with a sharpie.

5 Likes

I’m not 100% sure to be honest. I started working with the BitSetter disabled and the first time I saw the issue I ruled it out. Perhaps I’ll run without it a little more to really accumulate some data.

Interesting, I’ll give that a try!

I like the sharpie idea on the pully and as Ed.E mentioned for the end mill slip. Great ideas!

Good food for thought here but I have a few questions for ya.

Do you mean to do this before starting the job? So the workflow would be to set zero using the BitZero, hit start job, the bit would be probed using BitSetter then you’re prompted to start the router and start the job. Are you suggesting that I select “remain paused” and jog back to the zeros to see if they’re still accurate? Can you even jog the machine once you’ve selected Start Job?

I can’t say that I have ruled out EMI. I’m still getting occasional EMI freezes so perhaps EMI is affecting depth control glitches as well?

I have thought about this but the petty part of my brain struggles to throw more money at this damn thing. At some point my wife is gonna kill me. lol. But perhaps that’s what I need to consider. I will point out though that the root of my complaint is that I believe there are component weaknesses with the 3. I don’t think I should have to replace these lousy components out of my pocket because carbide made a machine with inadequate components.

Please let us know about all these specifics at support@carbide3d.com and we will do our best to work this out with you.

As other folks have noted, the SO3 line was available for a long while, and has worked well for many folks — Serial #6 is my main machine and works reliably and well so long as I check things per:

and use feeds and speeds which are suited to it and the materials being cut.

2 Likes

I have been using my Shapeoko 3XL since early this year and have had issues. Z-0 issues were the most common for me as well and I built an extra step into my workflow as a check. After every bit change I would check zero again. How I did that was to wait for the machine to move the bit back to centre of the front of the machine. Then I place a straight edge across my piece and run it to where the bit was sitting at the front of the machine. I would then put a ruler at the base of the bit and run it down to the bottom of the straight edge. I would then compare that number with the number showing as the current Z location on Carbide Motion. If the numbers matched then I know the bit change went well and the cut would proceed as expected. If the numbers did not match I would stop the cut, reset Z-0 and usually just export a new nc file that starts at the point where the last run left off by disabling the tool paths that have already run. I still have not figured out why it was occurring but by checking after every change I know the Z issues does not come up. While I find the extra step a little tedious It has saved numerous pieces from Z damage. The funny thing is that since I started the process I have only ever had it be off probably 2-3 times whereas before I would notice it quite often with air carving plunges that wrecked the piece. I feel that building in that step may have made me stop whatever mistake I was doing that caused the issues. Even if it does not fix it, this will stop you from wrecking pieces.

With regards to slight issues with X-Y I have only had a few but they were always related to belt slippage. Once I accidentally had something in the path and that caused a slight slippage. Another was when the piece had a knot that slowed the machine unexpectedly. Another was related to over tightening the bit. I actually pulled the X over slightly while tightening. I felt it happen but thought it slipped back. Once the carving started I knew I was wrong though. I still have that cribbage board sitting near the work area as a reminder. The error is very slight but enough that I am not willing to sell it.

6 Likes

I will try that extra step, thanks for the tip. I am having these Z issues even when a bit change hasn’t occurred but still a good call.

I will say, however, that I don’t think it should be up to us as operators to add these creative steps to our workflow in order to have the machine operate as advertised. A good machine shouldn’t need these additional steps designed to prevent malfunction.

I too added a step to my workflow in an attempt to avoid the malfunction. What I started doing was setting everything up as normal, starting the job but removing the workpiece and letting the machine do an air cut to see if it was behaving. If it was I stopped the job and reset to let it run with the actual work piece. A time waster and a nuisance, to be sure. Ultimately it proved ineffective as the issue seems to occur at random and on several occasions it ran a successful air job but then malfunctioned during the real job. I stopped doing that extra step.

I would use the BitZero to set z, then manually jog over work piece and verify.
Just to check that off the list.

Run your job and proceed. Then once job is complete.
Rapid back to x,y and +6 z
Jog over material and step down z 6mm 1mm at a time.
Verify if Z is still correct.
If you are in the air once at current z then you have lost steps, more than likely EMI
If you are touching before current z then you are more than likely skipping ribs on belt- Check Z belt for damage, Make sure z belt is properly tensioned.
(someone want to chime in here my brain hurts right now trying to get orders done/ I could of had some Dyslexia happen just now lol)

I feel you on the z plus or hdm but if you are having issues directly related to skipping belts then this will directly take that problem out of the equation.

As for EMI, having router properly grounded and cable routed away from machine cables will help. I had EMI disconnect issues when I first got the machine… Guilty of routing my router power cable into drag chain. So when I would turn router up past 3 I would have issues almost every time.

I put a EMI jacket over router power cable and re routed up and away from machine and on a separate dedicated circuit. If running dust collection do the same for it. And be sure to run a Ground Cable. Like actually use a grounding rod and wire ground from vac. line to it.

2 Likes

Oh thanks! I understand now. I will try that but I suspect that the Z reset occurs at the BitSetter or through a software glitch some time after setting Zero. So I suspect it will look good right after setting zero but may still reset before running the job. I will double check that though. Good tip.

I was guilty of the same thing with the router cable. Now I have the cord suspended above the machine and away so it’s never anywhere near any other components. I haven’t tried the EMI jacket yet though, maybe that’s next. The EMI issues appeared to have gone away until last week when I started getting freezes again.

I sound like a broken record now but don’t you agree that a machine should simply work as advertised out the box? I can own my mistakes like the router cable but should we be expected to go out and buy EMI jackets, USB isolators etc? It seems to me that by doing such things we are simply covering up the underlying design flaws that seem to exist. Again, broken record time, but it’s interesting to me how many people tell me that their machines run great but then tell me about all of the bandaid fixes and goofy extra workflow steps they take to avoid malfunction.

Quick question What Version of CC and CM are you currently running?

As for the EMI issues, there are many possible causes. I believe the team has worked on isolating and improving these issues on the newer machines. But they are in a constant evolutionary state. That’s one of the things that appealed to me. They are never happy and always making improvements.

A machine can work fine in one shop and have issues in another directly correlating to incoming power variances. I’ve had issues with this in the past with completely other products from Companies selling Millions of $$ product every year. Not really something they can test, best thing is to have an voltage booster in this situation.

You can see the price tag on something like this and see why having this for every machine would directly boost the costs.

I have built Plasma CNC Tables in the past and certain control boards would just go haywire and have so much EMI. Having to upgrade these boards to ones that cost nearly $800 per unit from a swiss manufacturer to not deal with the added EMI of plasma torch. Which was working well for a spindle attached but once torch turned on a no go. It did directly effect the product pricing and sales. That is nearly half the price of these machines. So while they have had their hurdles and continue to overcome, for the price point I feel they are top notch and with a great customer service presence and a community here with the knowledge to help you with anything you could imagine.

I myself hope to have a fully industrial machine in the next few years. Spending 10-20x more than the cost of a Shapeoko I would totally not expect to have any issues.

3 Likes

I really hate to see anyone not liking their machine and will do what I can to help get alleviate that pain. Met a few people over the years about to toss in the towel with their machines and either traveled to them or came to me. We were in close proximity of course. I’m currently traveling across the country with my machine in a 32’ camper and if my path takes me close I’d be willing to swing by and do some troubleshooting.

4 Likes

I’m running CM build 537 and CC build 530.

I applaud them for their work as well, always striving for improvement is a great virtue. It’s a little hard to digest owning an older machine (just over 2 years old) and hearing about how they’ve solved the problems of old. Great for new owners, I guess, but what about the rest of us?

truth! I had mine set up in the workshop but the EMI was so bad I had to move it into the house. That, in addition to all of the other mitigators seemed to work until just recently.

Indeed!

In my case the “overcome” remains to be seen. You’re right about this community, great resource! I have varied thoughts about customer service though. The individuals I’ve dealt with have been amazing! Solutions, once identified are shipped overnight! Some good stuff there. I’ve also had email chains go silent for weeks though. Ignored emails and great customer service are not compatible thoughts.

Did you by chance add another appliance to the circuit you run your machine on? If just recently started having issues again.

Maybe try running without the Bitsetter? See if issues still arise.

Those are indeed the current “Stable Versions”
Did you have issues after Updating Software?

Maybe revert back to old Software and Compare Results.

I try to stay with the Beta myself. And if issues arise revert to old and report issues.

Current CM Beta is at version 540
https://carbide3d.com/carbidemotion/beta

Current CC Beta is at version 620

2 Likes

No changes there, same setup. I suspect the low humidity with the approaching winter may be a factor. My EMI issues were terrible last winter and only went away in the spring. I was troubleshooting all winter with support and when the issues resolved I worried that was a factor but we agreed to re-evaluate this winter. Now I fear that my machine may be a summer only tool haha

I disabled BitSetter for a time and ran some successful jobs. I saw the Z depth issue once or twice with the BitSetter disabled so I eliminated that as a cause. I think I’ll run with it disabled a little more to accumulate some better insight though.

Good idea. I’ll see what I can do.

Sounds like you may have solved your own problem then…

If your summer humidity levels average 45-55% and winter drops down to 25-35% then try getting a digital humidifier and set within the summer parameters to see if this directly correlates to the issue.

I would totally start there. If that was the issue last winter/spring

2 Likes