I will break down and go through everything this week. My parts should be in by Thursday and I will report back my results once I get things replaced.
I’m excited that I might get my machine back to true working order. Let’s hope my motivation to do it stays high enough. Dealing with this issue has made it less fun than normal to be around the CNC. I have hope now though.
1 Like
I received my parts pack today. I replaced belts and gears so far. It seems that the only gear that was worn well down was the right y axis one. Not sure what would cause that on one and not the other. Pressure from the tensioner?
One of the shoulder bolts that came in was defective. I reused the shoulder bolt from the left y for it. The allen key hole want machined correctly and I couldn’t tighten it without slipping. I’m not sure the difference between the new versus the old type. The older has a black oxide coating.
I ran a quick test after replacing to compare the circles it would make against the circles from before. Definitely a difference. However, I probably need to adjust the homing sensors because I adjusted them to get the machine sways when not writing correctly previously.
I want to understand why the one gear was so worn compared to the rest. I would expect the same amount in both sides. The x is different, I would not expect the same necessarily.
The upper left 2 circles are the test after changing. I can visibly understand the difference between the two. That’s great, now I need to fine tune the setup to make sure I can get accurate holes again.
I have no proof of this, just speculation, but I think the machine has a slight twist in it. And it doesn’t have to be a huge twist either! Just a tiny amount will cause problems.
When I design moving assemblies that require high speed and vibration free movement, key #1 is that floating bearing system.
It allows those imperfections to exist without causing wear and vibration on adjacent parts. Since our machines are missing floating bearings for whatever reason, all I can do is speculate based my design experience.
And yes, blaming imperfection is easy because we know that it is always there but in this case, we are witnessing the symptoms. We just don’t have the scientific observation to back it up, yet.
At least you know the issue and what to look for. That is half the battle. Do a walk around and a checklist like a pilot. Before the machines moves, inspect it and use that checklist so you don’t miss things.
I’ll have more time this weekend to dig in. I can see the idea of a twist causing that issue. With such a large machine, it’s harder to diagnose fully because I can’t even think of a good way to find a twist in it.
Next I’m going to square the gantry and try and see what happens over time. Watch the movements throughout jobs, at the gear itself to see if I can feel it out.
I line with the recent comments, I thought I’d share a story to support your theory.
When I finished assembly on my machine 3 years ago, in spite of being very diligent, I found my gantry to be excessively out of square for my standard (around .060”) The machine would square itself up successfully during the homing operation, but when the motors weren’t energized, taking my hand and applying force to the gantry, I wasn’t happy with the inherent torque being exerted to square the gantry. The solution for me was to shim. This has worked very well over the years and was a solution supported by AVID at the time. I wonder now if you don’t have a similar dynamic (as Subnoize suggested ie twist). I have no doubt the load on one of my Y axis drives would have caused uneven and excessive gear wear had I not corrected this.
Just my .02. Hope it helps!
1 Like
Good suggestion. How did you determine square? Did you push it up against the end stops and measure?
I had the same question as MichaelDot ^^^.
I did not foresee this as an issue. Very good to keep a look out for.
Correct! I moved the Y axis to the very front position BY HAND. I first removed the rubber stops and replaced them with a standard bolt which I threaded into the bracket holding the proximity switches until they mic’d to be protruding equally with .001” to the very front of the machine and measured off the front proximity switch ‘ears’. Visually,m, the gap on both sides was obvious even without the dial indicator. To verify, I powered the Motors and homed the machine, letting the process square the machine. I then remeasure off the bolts I put in place. The gap went from +.060” on one end prior to squaring with motors de-energized, to .0005” after homing. The amount of force required to physically correct for the .060” was not insignificant, though I never measured it. It would have put a significant amount of preload on the gear on one end which would have caused accelerated wear.
The wear that this would have caused should be obvious. The type of preload on one end would cause flank wear on one side of the tooth. This would be visible by comparing the centerline of the crest (top of the gear tooth) between the worn segment and the unworn segment. Flank wear on one side would cause a shift in the crest centerline. Also, the width of the top of the crest would have narrowed.
Your description of the wear is exactly what I see.
I have the old style proximity switches but can do the same thing. I made a spacer printed with resin to stop the switches at an exactly equal distance instead of attempting to measure that. With the older ones in the stop itself it’s hard to measure without getting caught on something. Physical spacers really helped with that.
I’m going to try your suggestion tonight and see what I find. When I originally out everything together, the end covers on the x on one side were machined wrong. The front stuck out some and if I had tightened the rails down they would have had to bend to make up for that. I then milled the spaces out so they would sit with interference. I assume a bracket could have also been machined out of tolerance and caused this problem I have now noticed. I guess the extrusion could be an issue. Compounding variables throughout the build as well.
I do have to shim my mount for the tramming already. Aluminum foil ftw.
We may be onto something! One of my end covers on the end of the gantry was machined slightly oversize (stood proud of the extrusion) on the opposite end (back side) so it was making contact with the bracket prior to the extrusion getting snugged up to the bracket. In my previous picture, you can kind of see that this is the case on the shimmed edge as well. My end caps did not interfere with the guide rails but I am beginning to remember that in order for them not to, I had to offset the endcap before snugging you the bolts or my rails would not have sat flush to the beam on the ends. Very interesting that you would mention that! Shimming was just easier for me at the time or I would have milled the edge down. It wasn’t the entirety of the .060” I measured, but it was at least .040” so it was the major compounding factor.
Good to know that I’m not the only one and that other should look for this.
I got my 90mm partially threaded bolts in today and they are threaded more than enough to engage well. They are stainless steel and have flat heads with Allen holes. They seem more than enough robust and cannot rub on the brackets any longer.
Just following up on the bracket update from the earlier discussions.
1 Like
Thats how its supposed to work. Being out of square 100 thousandths before homing on a 4+ ft gantry is nothing, especially on these bolted together aluminum extrusions - they aren’t very rigid (the two motors working in unison is what makes it rigid when in use). It takes very little torque to square them up in the homing process. If you spend a lot of time shimming them to get them square when the motors are not powered, they aren’t going to stay square anyway. The first time you hit a clamp or something on one side of the gantry and stop it you are going to knock it back out again, or even pushing the gantry around unpowered will do it as well.
If one gear is significantly worn vs. the other side, I would suspect belt tension differences, or maybe you do more machining on that side of the gantry so it takes more of the load, or some other mechanical issue.
1 Like
I literally watched your backlash and stiffening videos today. I’ve seen them before but wanted to go back just in case things were in there that I may get use it of.
I have reduced the tension on my belts and it does seem that the right y was stretched more than all the rest I replaced.
Not a single one of you checked the square or dimensions on that 8020 before you assembled it.
I can tell because none of you have said what I am about to say 
There is nothing precise about 8020. Not a single piece of it was cut square, cut to length or was straight to start with.
So, expect your machine to wear in odd places and at odd rates. Its a “feature” of the 8020.
Now, quit staring at your crooked nose in the mirror and go cut something 
1 Like
I think that’s true. I just want to cut something well right now. I had it work before, I know I can again. Almost there. Picked up a 60 inch caliper today front work and going to get this working well by the end of the weekend. Lots of projects coming up and they have to be right.
…and I thought I was crazy.
Borrowed it. Didn’t buy.
MITUTOYO Long Range Digital Caliper: 0 in to 60 in/0 to 1,500 mm Range, IP66, ±0.004 in Accuracy - 4CGN4|552-315-10 - Grainger
1 Like
Just want to get an overall feeling for this. Did everyone here tune their steps per our just stick with the defaults from avid?
Defaults and things just…work
2 Likes