Saturn 2 4x4 w/ Nema34 closed-loop - In Canada

[wmgeorge]

The OEM tube is what... 3/16” so 3/8” should hold 100 lbs supported on the tube or shimmed.

[he1957]

Dan,

No video (yet) sorry but as it turns out I had a linear rail block lose its balls (no pun intended).

Spent about 4 hours finding and re-installing all but a few but since it's a warranty replacement (still waiting for it) I bought a set from Amazon and will replace all 4 of them for the Y axis which means I have to take the gantry off again and re-do the procedure so I'll try to video or photo the process. Alas, I have to take off all my top surface and spoil boards to get at the bearing block mounting screws and my gantry weighs about 250 kg so I'm not all that anxious to do it because the machine is working well at the moment.

The alignment is not difficult, just a little fiddly.

1. Start with a straight line drawn with a spring loaded pen/pencil in the spindle on the master rail side on Y (and set X to (close to) zero)
2. Around the center of travel for Y, make a 1000mm set of marks (500 mm each side)
3. Measure and mark your 1000 mm box with a good straight edge rule using the diagonals to ensure it's square
4. Align spindle center (pen) to one line of intersections on X=0 side, move X along line to other side (X+)
5. If gantry is square to master rail, it will follow line and match at the end of line intersections, otherwise diddle the mechanicals until it is
6. Repeat to check
7. Move Y to other perpendicular line and re-check.
8. If there is a discrepancy take the difference of the two to maximise (minimise?) squareness between the allowed distance
9. If there is a discrepancy it shows the master rail is not straight and has a bow of sorts

Any measurement is OK, bigger is better, it's just our old friend Pythagoras for the hypotenuse side.

As mentioned in previous post, if meshing of R+P is not snug and easy for both sizes, power down and move slave side motor. The idea is to find the best full step position for meshing without power. After power on, the meshing may cause a slight movement, so again mechanical adjust if needed.

Then under power move the spindle pen around the box, for the perimeter and the diagonals to show the tracking of the lines.

[Ntl]

Dan,

No video (yet) sorry but as it turns out I had a linear rail block lose its balls (no pun intended).

Spent about 4 hours finding and re-installing all but a few but since it's a warranty replacement (still waiting for it) I bought a set from Amazon and will replace all 4 of them for the Y axis which means I have to take the gantry off again and re-do the procedure so I'll try to video or photo the process. Alas, I have to take off all my top surface and spoil boards to get at the bearing block mounting screws and my gantry weighs about 250 kg so I'm not all that anxious to do it because the machine is working well at the moment.

The alignment is not difficult, just a little fiddly.

1. Start with a straight line drawn with a spring loaded pen/pencil in the spindle on the master rail side on Y (and set X to (close to) zero)
2. Around the center of travel for Y, make a 1000mm set of marks (500 mm each side)
3. Measure and mark your 1000 mm box with a good straight edge rule using the diagonals to ensure it's square
4. Align spindle center (pen) to one line of intersections on X=0 side, move X along line to other side (X+)
5. If gantry is square to master rail, it will follow line and match at the end of line intersections, otherwise diddle the mechanicals until it is
6. Repeat to check
7. Move Y to other perpendicular line and re-check.
8. If there is a discrepancy take the difference of the two to maximise (minimise?) squareness between the allowed distance
9. If there is a discrepancy it shows the master rail is not straight and has a bow of sorts

Any measurement is OK, bigger is better, it's just our old friend Pythagoras for the hypotenuse side.

As mentioned in previous post, if meshing of R+P is not snug and easy for both sizes, power down and move slave side motor. The idea is to find the best full step position for meshing without power. After power on, the meshing may cause a slight movement, so again mechanical adjust if needed.

Then under power move the spindle pen around the box, for the perimeter and the diagonals to show the tracking of the lines.

Thank you very much for taking the time to explain and for the drawing now I understand how you do it. I'm going to try it out and see how close mine is right now..

Thanks again,
Dan

[GME]

The OEM tube is what... 3/16” so 3/8” should hold 100 lbs supported on the tube or shimmed.

Bill,

Are you proposing steel or aluminum? Aluminum needs to be thicker, both because of less rigidity and thread holding. Aluminum has roughly 1/3 the rigidity of steel and is obviously much softer. Also, the tubing is advertised at 4mm, or right at 5/32", 1/32" thinner than you guessed.

There is also the question of thread depth. Folks have different rules of thumb on thread depth, but the published literature indicates minimum depths of 1 times the nominal thread diameter for steel and 2 times the nominal thread diameter for aluminum. That works out to 8mm for steel (the rails use 8mm cap screws) and 16mm for aluminum. For steel: 8mm is close to 5/16". If you are going to mill a reference edge of 1/8", you are at 7/16" thickness, or in standard sizes, 1/2" thick. I recognize that the steel tubing is thinner than it should be to hold 8mm screws, and it would be no better for attaching the mounting plate. Sadly, You have to work with what you encounter on the ground, and applying conventional standards isn't always possible. On the plus side, you can get away with a lot more with steel than aluminum. Still, the tubing thickness is far away from being ideal. It's probably less than marginal, even though it has been working in practice.

Aluminum: 16mm converts to .630", or just over 5/8" If you round up, you get a 3/4" thickness for aluminum. If you mill an edge to butt the rail against, your are a little under the 2D rule. The weakest link is having to attach the plate to the tubing, which, as said, is too thin.

Maybe the best of all worlds would be to stay on the thinner side on top, and insert a steel bar on the inside. The top piece would have holes in it, and the inside bar would be threaded. Either steel or aluminum would work, but if I was doing it, I would probably go with aluminum on top and steel inside. That's just because aluminum is easier to work with.

A thinner top plate would work best with some sort of rigid foundation underneath to bridge gaps between shims - assuming that shimming of some sort even becomes necessary. I like a bedding (no sag) epoxy. FWIW, one metal mill seller suggested using epoxy under the column, when tramming adjustment was needed. Shim it to position, and fill the gap with epoxy. You've made it clear in other threads that you're not a proponent, and believe that shims alone should be sufficient. Let's just agree to disagree on the point and let Leon decide what works for him. No skin off my nose either way.

I believe grub screws would be easier to use than shims. I've used the technique before, and while fussy, it worked when shimming wouldn't. It partly depends on variability. If you determine that the amount of shimming is entirely consistent over the length involved, then shims may well be easier. However, if there is variability, grub screws may be a better solution. Unfortunately there is extra work drilling an tapping a lot of additional holes for the screws. Seems like everything involves some sort of trade off.

Gary

[wmgeorge]

Gary, Steel of course and the 3/8 inch is thicker than it has now for thread engagement. You would not place shims just under the bolt locations but along the surface. The short distance between the shimmed and not shimmed area would not be great, and remember you are dealing with a 100 lbs.

Ideally if the height did not need adjusted no shims required and in that case the inside bar as Crash proposed would be a much better solution.

I spent my entire working career solving other peoples problems and sometimes working with engineers showing them why their design needed changed. For this I got paid very well.

Granted most of the work was either electrical or commercial HVAC related but with some mechanical tossed in at times.

[Ntl]

HE1957

I used the square method and I was off by 0.03 from my original squareing. Made the adjustment to the gantry on the slave side re-ran the square and it is as close to perfect as I can measure with the tools I have. Now I just have to remember to make sure it's lined up with my marks every time I start up the machine. I can't wait for the update for my controller to be released that has auto square. It's will be one less thing to remember to do.

Thanks again for the info and help.

Dan

[he1957]

Dan,

I'm impressed! -- Well done! With 0.03 (mm?) now I'm envious! How did you measure to that degree of accuracy?

If the mechanics are aligned then start-up should be fine. The auto-squaring using the sensors for Y1 and Y2 homing can be setup using feeler gauges and you should be about as square as the tolerances allow. Problem here might be the sensors sensitivity/proximity will have some variation. If your software has diagnostic/LED sensor indicators that might make it a bit easier.

Now for an easy and inexpensive way to get the spindle trammed, open for suggestions...

Perhaps this should go into a new thread to avoid hi-jacking?


HarryE.

:-)

[wmgeorge]

You would think finding a true .500 dia round steel bar would be easy, but not in my shop! Finally found one and MIG welded a flat CR flat steel bar about 2 inches long at 90 deg and drilled a mounting hole for one of my dial indicators. Went to the center of my table and had to shim a little in the Y direction bot the X was fine. You can purchase one like Gary GME has to use, but I have so much stuff laying around I hated to buy even more.

I am impressed with Dan's work also.

[GME]

You would think finding a true .500 dia round steel bar would be easy, but not in my shop! Finally found one and MIG welded a flat CR flat steel bar about 2 inches long at 90 deg and drilled a mounting hole for one of my dial indicators. Went to the center of my table and had to shim a little in the Y direction bot the X was fine. You can purchase one like Gary GME has to use, but I have so much stuff laying around I hated to buy even more.

Here's a link to the tramming gauge I use: https://www.edgetechnologyproducts.c...10-000-09-000/

Gary

[Ntl]

Dan,

I'm impressed! -- Well done! With 0.03 (mm?) now I'm envious! How did you measure to that degree of accuracy?

If the mechanics are aligned then start-up should be fine. The auto-squaring using the sensors for Y1 and Y2 homing can be setup using feeler gauges and you should be about as square as the tolerances allow. Problem here might be the sensors sensitivity/proximity will have some variation. If your software has diagnostic/LED sensor indicators that might make it a bit easier.

Now for an easy and inexpensive way to get the spindle trammed, open for suggestions...

Perhaps this should go into a new thread to avoid hi-jacking?


HarryE.

:-)

Agreed on starting a new thread. And no on the mm we still are Imperial over here. The 0.03" is what I was off originally from using the laser level. Now it's so close that I don't really have a accurate way to measure it, I honestly don't know think it's getting any better than I have it since it's rack and pinion and there is backlash.

If you look at my thread you will see the dial indicator mount I made that I used for tramming my spindle. It worked really well and only cost like $8 in material and hardware.

Dan

[wackocrash5150]

Ok, so!.... I've finally had a bit of time to see what's what on this thing. I've been fixing and prepping and riding some snowmobiles here for myself and the kids the past bit. Weather can be odd here when it comes to good snow/sled conditions so you have to take it when you can get it. Since I last posted, a friend of mine who's a machinist, lent me his machinists level. This thing is considerably better than a carpenter's lever let me tell you, although smaller. Having said that, it turns out my horizontal leveling along the x axis wasn't too bad. Y axis leave a bit to be desired but I can always fix. The level results and photos left me with some cause for concern. The X axis rails seem parallel and fine. The Y and A axis are slightly off over the length of the machine, but what more concerns me is that they are tilted in towards each other; enough to pin the level in opposite directions. I will post the pics and where I took the measurement.

These photos show:

1) Front table/stand surface
2) Machine overall view
3) Front middle of spoilboard
4) Overall view of gantry and general level positioning
5) Top of gantry
6) Top of upper x-axis rail
7) Top of lower x-axis rail

[wackocrash5150]

These photos show:

1) A-axis rail towards front
2) Y-axis rail towards front
3) Y-axis rail towards front
4) A-axis rail towards front
5) A-axis rail towards rear
6) Y-axis rail towards rear

[GME]

Y and A axis are slightly off over the length of the machine, but what more concerns me is that they are tilted in towards each other; enough to pin the level in opposite directions. I will post the pics and where I took the measurement.

These photos show:

1) Front table/stand surface
2) Machine overall view
3) Front middle of spoilboard
4) Overall view of gantry and general level positioning
5) Top of gantry
6) Top of upper x-axis rail
7) Top of lower x-axis rail

I am so sorry to hear the bad news. Major bummer.

The inward tilt you describe explains the shims between the gantry mount and interface plate. The frame on the one I sent back also had an inward tilt. I took a different approach to measuring, but I had the gantry off and it was easy. Also, I had the frame in a vertical orientation on a dolly, which didn't lend itself to using my machinist level.

If you want to better quantify how much the frame is tilted inward, remove the gantry, put a straight edge (a long, but accurate level will do) across the aluminum interface plates and measure the gaps between the straight edge and interface plates. Feeler gauges work great for this. The measurements on mine were different on each side, meaning one side was rotated inward more than the other. Of course, that's not the end of the inquiry. Are the frame rails rotated inward the same amount at the front of the machine as at the back? There is no reason to assume they are.

You commented that the Y axis linear rails "are slightly off over the length of the machine?" Just curious, but what does "slightly off" mean? Remember, the out-of-parallel tolerance for linear rails is only at or about .001" (.0254mm). "Slightly off" is a relative term. In terms of a .001" tolerance, slightly off might mean .0015", but I get the impression you are talking about significantly more than .0005".

You now know that inserting a bar inside the frame will not solve your problem, and the fix, if you choose to go that route, will be far more extensive.

There is another guy on the forum who has experienced problems with his machine. Might be worth a read. He found some issues I didn't look for, but would have, had I decided to keep the machine. Here's a link: https://www.cnczone.com/forums/finel...066-posts.html.

Gary

[wackocrash5150]

You commented that the Y axis linear rails "are slightly off over the length of the machine?" Just curious, but what does "slightly off" mean? Remember, the out-of-parallel tolerance for linear rails is only at or about .001" (.0254mm). "Slightly off" is a relative term. In terms of a .001" tolerance, slightly off might mean .0015", but I get the impression you are talking about significantly more than .0005".

I can't put a number on the "slightly off" comment but if you look at the 1st and 2nd photos of the 2nd batch, those are running the level the same direction as the rail. If I understand correctly, those 2 readings should not necessarily be level, but pretty darn close to one another. Now this machinists level is pretty sensitive in comparison with my other experiences of consumer levels. We took a reading on a flat desk, and then my friend stuck a sheet of paper under one side at the very end, and it showed a readable change. Pretty sensitive if you ask me ... lol.

I was thinking about pillaging my wife's sewing basket for some thread and making the X across the opposite ends of the rails. I have some magnets that should hold the thread to the rail.

[GME]

I can't put a number on the "slightly off" comment but if you look at the 1st and 2nd photos of the 2nd batch, those are running the level the same direction as the rail. If I understand correctly, those 2 readings should not necessarily be level, but pretty darn close to one another. Now this machinists level is pretty sensitive in comparison with my other experiences of consumer levels. We took a reading on a flat desk, and then my friend stuck a sheet of paper under one side at the very end, and it showed a readable change. Pretty sensitive if you ask me ... lol.

I was thinking about pillaging my wife's sewing basket for some thread and making the X across the opposite ends of the rails. I have some magnets that should hold the thread to the rail.

Sorry. I was thinking "slightly off" referred to linear rail parallelism. I see from the photos you referenced that wasn't the case. I also didn't review the entire thread. Going back through it, I see that you already found the linear rails to be out of parallel by a little over 1/16", and I commented on it.

Yes, machinist levels are very sensitive. Even dust on a surface can affect a reading. I don't use mine often, but when I need precision, it's great to have.

Using thread to do a cross check is a good idea. You won't be able to make a very accurate measurement, given the nature of thread, but it will tell you whether your frame is out of plane. Hint: Be sure to take one of the threads and put it under the other, then put it over the other. You are looking for a gap. The threads may touch in one orientation, but have a gap in the other.

So much for the claim of being welded in a jig and stress relieved.

Gary

[wackocrash5150]

Hint: Be sure to take one of the threads and put it under the other, then put it over the other. You are looking for a gap. The threads may touch in one orientation, but have a gap in the other.

Good call. I'll make sure I do that and will post results. Should be either tonight of Wednesday evening.

[Ntl]

Maybe use some fishing line or if you Have a mig welder use some wire that way you don't have to worry about it bowing. Just a thought if the thread doesn't work out.

[PLJack]

I'll just chime in real quick.
I would not recommend removing the gantry if you do not have to. It's not necessary to "X" the rails with string if you move the X axis to the side. There is room under the gantry.

Jack.

[wackocrash5150]

I'll just chime in real quick.
I would not recommend removing the gantry if you do not have to. It's not necessary to "X" the rails with string if you move the X axis to the side. There is room under the gantry.

Jack.

Roger on that one Jack.

[difalkner]

So much for the claim of being welded in a jig and stress relieved.

Gary

I was thinking about this yesterday and one thing that came to mind is that they all could very well be welded in a jig and held to the tolerance of the jig/fixture. But if the fixture isn't very accurate or has changed then the resulting frames will be out, as well.

David