Cloned Y losing Steps....calling all CNC masters:)

Greetings,

I appreciate your help...

I have a Workbee that is driven by lead screws. I'm having issues with the cloned side of the Y axis. I can home the machine, and both sides will home properly. I get the same measurement from the end plates to the gantry for clone and master sides of machine (26.5 mm). Then I'll send the gantry a little over halfway across the machine (700 mm travel), and take the same measurement. At this point the cloned side has lost 1mm. (726.5 vs 725.5) If I send the gantry to the far side of the machine (1200 mm travel), Ive lost 2 mm on the cloned side. Im losing .00167 mm/mm on the cloned side. When I send the gantry back to the home position, both sides come back to the starting position, and are once again equal distance from the end plates (26.5 mm), so the irregularity is happening both directions.

Here is what Ive done so far: ( not necessarily in this order)
1. Checked thrust bearings, lock nut, etc. for play.
2. Replaced nut blocks on both master and clone side of machine.
3. Swapped motors to see if problem would migrate to other side.
4. Swapped drivers to see if problem would migrate to other side.
5. Swapped controller input to driver to see if problem would migrate to other side.
5. Replaced motor plug.
6. Increased current on driver.
7. confirmed eccentrics are properly adjusted. Wheels are all snug, but movable with one finger.



Unfortunately, none of the above steps have resolved the issue. The error remains consistently on clone side, and consistently losing .00167 mm/mm despite numerous adjustments and tests.

 

Thank you Alex! I needed that push. The machine is a OpenBuilds Workbee. It was originally belt drive, but I converted it to leads a couple of years ago. I was under the impression it was working well. I've been working on numerous improvements and recently had to tear it down to move it on to a new stand. I thought this was a ware issue and any ware would have occurred at the Delrin nut blocks that I had already replaced, so I started looking at everything else. What I didn't take into account is likely the manufacturing tolerance of the leads. When you look at the leads, it's impossible to see the subtle variation in pitch. So before taking your advice and swapping the leads, I checked the resolution of each lead in the Y and X axis. Y = 197.799 Steps/mm, Clone 198.081 steps/ mm, and X= 197.552 steps/ mm. I swapped the X and Clone leads trying to get 2 leads with the closest resolution, and retested. The result was an improvement. Now clone gains 1 mm/ 1200 mm. instead of losing 2mm / 1200 mm. When I reconfigured X with the previous clone resolution it was spot on again.

The two takeaways are:
1. This issue has likely been there since I originally converted to leads but I didn't catch it.
2. I need two leads with the same resolution for my Y axis.

( Considering I bought the previous leads all together from Openbuilds and ended up with three slightly different pitches might make getting two matching leads tough!)

I was previously homing the machine in the front left corner. I also cut most of my parts in the front left corner, so the effect of the leads were reduced to a fraction of a mm. As a result of my recent remodel, I changed the control schema, and was implementing a self squaring gantry. I started homing in the rear right corner because I could nest the limit switches against the machine and keep the cutter over the table. This made the irregularity more evident because I would home the machine, and then move to the front of the table to work using a machine offset. Obviously now the further I get from machine home the more accumulated error is visible.

Any suggestions on where I can get Tr8*8-2p (4 starts) leads with tighter tolerances?

 

Hey Peter, Let me start by saying I have a lot of love for my Workbee. It's been a great machine and I'm not intentionally besmirching OpenBuilds. You guys are the eight hundred pound gorilla that makes all this possible. Respectfully, nothing is loose. If you would like to discuss my installation steps, Im happy to do so. I generally tighten the set screw on the coupler first to let the screw bite the shaft. Then reduce the circumference of the coupler with the larger set screw. The ends of shots are marred

 

Sorry accidentally hit the post button trying to correct typo, as I was saying the ends of the lead are marred from the coupler. I use a tensioning nut on the other side for preload. When I installed the nut blocks, I only install them from the opposite end of the coupler to avoid scoring them. I leave them loose until I get the lead installed, so the lead can easily pass between them. I only tighten them after getting preload on the lead. Everything is tight. Obviously, I am able to move these leads around and dial in the corresponding resolution and get the correct movement. I would think this would be fairly difficult if my setup was loose. I will say that my Z lead does exactly match one of the other leads, but unfortunately way too small to help me with my clone Y.

 

So what your proposing is a 8mm Stainless steel lead screw is stretching uniformly across its length before the threads in a much softer brass tensioning nut fatigue?

 

Im not trying to be argumentative, but Ive gotten the impression your interest is not sincere. The only reason OpenBuilds was mentioned was because I was asked. In my mind two leads cut at the same time have a higher probability of having the same thread profile. Tools ware out, and a lead cut at the beginning of a tools life is not likely to have the same profile as a lead cut at the end. I was simply pointing out I had purchased all the leads at the same time increasing the chances the leads came from the same lot. This forum ceases to be helpful when we can't deal with each other honestly. The tensil strength of 304 stainless is 73,200 psi. I can think of several fail points that are far more likely to occur before the elongation of the lead. Additionally, I would think in order to apply torsional forces necessary to distort the threads, one end would need to be captive, and if were both being honest maybe that can happen in a collision; However the distortion would likely reside between the point where the lead was captive and the motor. Not the whole lead. Ive broken a fair amount of bolts in my day and they all failed at their weakest point. I never extracted one that had a distorted thread profile across its length. I think occam's razor applies, The lead is the strongest component in the mix, and the most likely explanation is any distortion in pitch was because it was milled that way.
If you genuinely believe I stretched a the lead by over tightening, then accept my apology. At this point Im more concerned with getting two leads with matching resolutions. If you can help me do it, that would be appreciated.

 

Hi David,

Touché, but to be fare my lead manufacturing example was more of a metaphor for explaining this is a tolerance issue than an opinion on making leads. We are in complete agreement on needing to calibrate for lead imperfection. Clearly there is a big difference between "spot on" and "repeatable but not accurate." To answer your question, the overall length of the lead screw is the same; However the length the "lead" is unfortunately variable. The lead being the overall length of movement achieved with a full rotation. The effect of which happens evenly across the length of the lead screw. The truth is you can't see the difference between leads with the naked eye, but you've clearly observed it and calibrated for it. In my case I had one lead screw losing approximately 2mm per 1200mm traveled or .00167 mm/mm. (When compared to the master drive.) On a single rotation (8mm Lead) you're looking for a .01333 mm difference in lead. That's less than the thickness of a human hair. You're not going to see that looking at two leads side by side; However, that error is multiplied by every rotation. It take 150 rotations to cross my table resulting in a 2 mm loss at the far end.

The problem I'm having is how to calibrate the clone in a dual drive system. You can calibrate the master drive and get it "spot on"; However, subtle variation in lead causes the clone to be retarded/ or advanced. To put it another way, I can home the machine and ensure that both sides of the gantry are equal distance from a reference point. As I move the gantry away from home, the master drive side will be where it's expected to be regardless of its distance from home because it's calibrated. Conversely, clone side is just along for the ride. It duplicates the master motors steps. This would be great if the leads were identical. However, if the leads are subtly different, the further it gets away from home, the greater the accumulated error on the clone side. Then as the gantry is moves back toward home the accumulated error will be reduced until it's completely gone and your machine is once again on home.

It's for this reason it went largely unnoticed. I was doing most of my cutting in the 25% of machine I homed in. The machine always zeroed on both sides of the gantry and because I wasn't getting too far away from home the accumulated error was small. Ignorant bliss. Recently I made changes and started to home in the back of the table and offset to the front. That made the error far more visible.

Now that Im aware of it, I'd like to correct the issue, but that requires getting leads that ideally match, or are close enough that I can live with the error.

So why did Openbuilds "customize" and go to a 7.8mm diameter lead?

 

Well meshing them together at this point just confirms what I already know; However, might be a good preliminary step for trying to match 2 new lead screws. Im currently using a E5X so the clone is jumpered. I get that you made the shaft a little smaller to make it easier to get into the nut blocks but, does that mean you're using the same nut blocks, tensioning nuts, and bearings as a standard Tr8*8(P2)? Do you have a known tolerance on your leads?