Xnaron Printable Magnetic u-joint option for Rostock Max

[inventabuild]

I found threaded ball bearings at a few places:

http://www.painfulpleasures.com/xcart/c ... php?cat=64

http://www.thechaingang.com/threaded-su ... -ball.html


and something called a threaded ball stud here (not sure of the precision quality of these balls):

http://www.lucidhardware.com/index.php?cPath=70_73_56


and something called polished steel ball knobs here (although these look to be rather large balls):

http://www.jergensinc.com/site/sfg_hand ... oup_no=690


Just do a google search for threaded ball bearings and other search results will pop up too.

[mrblade]

I rested the ball in a small socket and balanced the screw on top. I have used pliers on mine to tighten them..

I meant drilling deepening of the screw head.

[mrblade]

I found threaded ball bearings at a few places:

http://www.painfulpleasures.com/xcart/c ... php?cat=64
http://www.thechaingang.com/threaded-su ... -ball.html
and something called a threaded ball stud here (not sure of the precision quality of these balls):
http://www.lucidhardware.com/index.php?cPath=70_73_56

and something called polished steel ball knobs here (although these look to be rather large balls):
http://www.jergensinc.com/site/sfg_hand ... oup_no=690

All balls in your links are made from nonmagnetic metal.
Only last link of magnetic metal. But they are very big.

[xnaron]

I heard that someone had a ball come off a screw. Make sure you soak the balls and screws in acetone to remove any grease or other contaminants. Make sure you mix the jb weld thoroughly. Cure it under a desk lamp so they get warm. I dure mine for at least 8 hours. After a day or so I can't knock the ball off the screw.

[Flateric]

I used JB weld stick on my 2 spare carriages, acetone, scuffed the bearings slightly with 400 grit and applied heat. I also have yet to have any issue at all with them coming loose. Even at excessive speeds/accels with high jerk settings.

[Broose]

I heard that someone had a ball come off a screw. Make sure you soak the balls and screws in acetone to remove any grease or other contaminants. Make sure you mix the jb weld thoroughly. Cure it under a desk lamp so they get warm. I dure mine for at least 8 hours. After a day or so I can't knock the ball off the screw.

We are having the same discussion over on this thread: http://forum.seemecnc.com/viewtopic.php?f=41&t=1919

I think scuffing the bearings with sandpaper in the area where they adhere as Flateric mentioned is also an important part of the prep process so that the JB weld can get somewhat of a mechanical connection with the ball.

[barnett]

Hey I have some questions about Xnaron's magnet joint mod & firmware adjustments.

• When I trim the rods to 240mm, that measurement is before the boom_ends are attached, yes? If I understand correctly, the goal is for the center of the one ball to the center of the other ball to be 269mm or else I need to edit #define DELTA_DIAGONAL_ROD in the firmware.
• Similarly, if I use Xnaron's cheapskate_ballends.stl, then the bearings are supposed to line up with where stock ujoints were so that no firmware edits are needed, yes? I'm looking at DELTA_CARRIAGE_OFFSET in this delta geometry diagram.
• And lastly, if I use Xnaron's effector.stl, then I need to edit END_EFFECTOR_HORIZONTAL_OFFSET to 36.5. If I try some other effector, then I should use END_EFFECTOR_HORIZONTAL_OFFSET to tune the firmware to the new effector.

I've got everything ready except the effector. I hope to try installing it next weekend.

Thanks!
Barnett

[jetpad]

Hmmmm. My DELTA_DIAGONAL_ROD is set to 267.5 but I don't have anything that I trust to measure the actual length. What should it be set to for the standard Rostock Max rods?

[barnett]

Yeah my firmware also says 267.5 and I can see a prior value of 269 was commented out. When I measure my stock arms, I'm getting 269 to 270.

I'm hoping someone who know's what's going on can make this less mysterious... otherwise I'm just blindly entering values when I'm told.

[Polygonhell]

Yeah my firmware also says 267.5 and I can see a prior value of 269 was commented out. When I measure my stock arms, I'm getting 269 to 270.

I'm hoping someone who know's what's going on can make this less mysterious... otherwise I'm just blindly entering values when I'm told.

If you are using my firmware I changed the DELTA_DIAGONAL_ROD because my printer was printing slightly undersized, and adjustments to the rod length effectively change the X/Y scale.
I set it based on a 100x100 test print, and it was pretty close at 267.5, Having printed some reasonably large parts with tolerance requirements recently I'd say it's probably very slightly oversized right now, and I keep meaning to change it back to 268.

There are only two values that really matter DELTA_DIAGONAL_ROD and DELTA_RADIUS, all the other values are just there to make measuring and setting DELTA_RADIUS easier.

[barnett]

As usual, thanks for a clear answer. I'll try to make the arms 240mm (w/o caps) and then see how it goes.

My effector won't be exactly the same width as the old one, so I might need to make an adjustment there. Do you suggest tweaking the delta radius adjustment for that?

It sounds like after installing the new arms & effector I'll have two things to check:
- whether the nozzle is higher/lower in the middle vs at the towers (adjust radius)
- whether prints come out too big or too small (adjust diagonal rod)

[bplemmons]

I have issues with the mag arms coming off on very large prints, has anyone else had this issue?

Bill

[xnaron]

I have issues with the mag arms coming off on very large prints, has anyone else had this issue?

Bill

This will happen if you don't get all of the arms the exact same length. Build a jig to make them like this http://www.flickr.com/photos/13723140@N04/8771948345/ and http://www.flickr.com/photos/13723140@N04/8771958335/

You also need to have all the firmware values correct as documented in the first post of this thread.

[Eaglezsoar]

I have issues with the mag arms coming off on very large prints, has anyone else had this issue?

Bill

This will happen if you don't get all of the arms the exact same length. Build a jig to make them like this http://www.flickr.com/photos/13723140@N04/8771948345/ and http://www.flickr.com/photos/13723140@N04/8771958335/

You also need to have all the firmware values correct as documented in the first post of this thread.

Where on the arm do you measure to obtain the value of 269mm? Is it at the end of the balls, the middle of the balls etc?

[Broose]

The middle of the balls. My ball bearings were a hair under .375 dia, so the measurement to the ends of the balls was 278.5. I put balls on the ends of the rods and held them in a pair of 12 inch calipers with the position locked at 278.5mm. I ended up using epoxy instead of CA so I could have a little more working time to get them set just right.

[Eaglezsoar]

The middle of the balls. My ball bearings were a hair under .375 dia, so the measurement to the ends of the balls was 278.5. I put balls on the ends of the rods and held them in a pair of 12 inch calipers with the position locked at 278.5mm. I ended up using epoxy instead of CA so I could have a little more working time to get them set just right.

I can agree with the 278.5 but that figure does not allow for the space taken up by the thickness of the end of the ball ends.
We start with a 9.525mm ball then the thickness of the ball end then 9.525mm for the magnet length then 240mm for the rod then 9.525mm for the magnet length then the thickness of the ball end then 9.525mm for the ball.
The length of everything WITHOUT the thickness of the ball ends is 278.1 or 278.05 for your smaller ball. Am I missing something? It would seem the rod should be shorter than 240mm to compensate for the thickness of the ball ends. The ball ends sit between the magnet and the ball and do have a thickness that is not shown in the calculations.

[jetpad]

You can use one of your existing stock arms to measure the length then it'll match your current firmware settings. The center holes of the stock arm (where it attaches to the machined piece of aluminum) should be right in the center of the two ball bearings.

[Flateric]

I think a good tip for everyone to further enhance your experience would be to do as is done in the machinist world.

Label where and the orientation of each piece and that way the same arm can go back into the same spot with the same orientation each time if removed.

Very subtle variances in lengths, angles, placement of balls can of course affect the calibration of your specific setup on your sepecific machine. So if you have it all nicely setup just right and have likely spent a good deal of time getting it that way. You really owe it to yourself to take a sec and write 1-6 on each ball on the carriage and 1-6 on each ball position on the uprights. Also number the arms and label which side is up.

That way when you remove the carriage which is so nice to do now to change the hotend or adjust a nozzle you can get the whole machine back exactly to where your hard earned calibration brought it before removal (accidental or otherwise) of arms or carriage.

[Broose]

I can agree with the 278.5 but that figure does not allow for the space taken up by the thickness of the end of the ball ends.
We start with a 9.525mm ball then the thickness of the ball end then 9.525mm for the magnet length then 240mm for the rod then 9.525mm for the magnet length then the thickness of the ball end then 9.525mm for the ball.
The length of everything WITHOUT the thickness of the ball ends is 278.1 or 278.05 for your smaller ball. Am I missing something? It would seem the rod should be shorter than 240mm to compensate for the thickness of the ball ends. The ball ends sit between the magnet and the ball and do have a thickness that is not shown in the calculations.

It is conceivable that there is only supposed to be .2mm between the magnet and the ball. The closer the better - without touching. I found that my arms ended up being a little longer than 278.5, but I thought it was because the magnets weren't bottoming out properly in the holes. I used natural ABS and I could see the stress marks through the outside at the bottom of the hole where I pushed the magnet in. Its no big deal if they end up being a little longer as long as they are all exactly the same length since you should calibrate DELTA_DIAGONAL_ROD afterwards anyway. You can calibrate as Polygonhell says by printing a good size part and measuring the actual vs. programmed dimension. That also takes into account part shrinkage for the filament material used. I just used some 12 inch calipers and measured the distance from the nozzle to the Z tower with two different Y locations. If the actual dimension is too large, then increase DELTA_DIAGONAL_ROD. If it is too small, then decrease it, reload firmware, and test again. After you have that calibrated, I would recheck DELTA_RADIUS to see if you have to adjust the offset (See calibration step 2 here: http://www.robertshady.com/content/begi ... eemecnccom).

[Eaglezsoar]

I can agree with the 278.5 but that figure does not allow for the space taken up by the thickness of the end of the ball ends.
We start with a 9.525mm ball then the thickness of the ball end then 9.525mm for the magnet length then 240mm for the rod then 9.525mm for the magnet length then the thickness of the ball end then 9.525mm for the ball.
The length of everything WITHOUT the thickness of the ball ends is 278.1 or 278.05 for your smaller ball. Am I missing something? It would seem the rod should be shorter than 240mm to compensate for the thickness of the ball ends. The ball ends sit between the magnet and the ball and do have a thickness that is not shown in the calculations.

It is conceivable that there is only supposed to be .2mm between the magnet and the ball. The closer the better - without touching. I found that my arms ended up being a little longer than 278.5, but I thought it was because the magnets weren't bottoming out properly in the holes. I used natural ABS and I could see the stress marks through the outside at the bottom of the hole where I pushed the magnet in. Its no big deal if they end up being a little longer as long as they are all exactly the same length since you should calibrate DELTA_DIAGONAL_ROD afterwards anyway. You can calibrate as Polygonhell says by printing a good size part and measuring the actual vs. programmed dimension. That also takes into account part shrinkage for the filament material used. I just used some 12 inch calipers and measured the distance from the nozzle to the Z tower with two different Y locations. If the actual dimension is too large, then increase DELTA_DIAGONAL_ROD. If it is too small, then decrease it, reload firmware, and test again. After you have that calibrated, I would recheck DELTA_RADIUS to see if you have to adjust the offset (See calibration step 2 here: http://www.robertshady.com/content/begi ... eemecnccom).

I had forgotten that link and I think it is great, thanks for reminding me.
So basically as long as they are very close to the 278.1 and ALL the same exact length, I should be fine. Thanks for explaining it so I could comprehend. I will build a jig that will guarantee that they are the same length.

[Broose]

So basically as long as they are very close to the 278.1 and ALL the same exact length, I should be fine. Thanks for explaining it so I could comprehend. I will build a jig that will guarantee that they are the same length.

278.5 to the ends of the balls. Here are a couple of photos

[Eaglezsoar]

A Starrett no less, I am impressed!

[xnaron]

How many people have done this conversion now?

[Eaglezsoar]

How many people have done this conversion now?

If you create a new topic, one of the options should be to setup a poll.
You can ask users via the poll it they have converted to the magnetic arms.

[Cleveralias]

How many people have done this conversion now?

I have the work completed but haven't put it in service yet as I'm also changing over to the E3D hot end and EZ Struder at the same time and am just now re-running wiring.

I planned to post some pictures and give you a big THANK YOU for your excellent work on this once up and running. There is of course zero slop and I've had a lot of fun working on this upgrade. I really appreciate you taking the time to develop and share this excellent mod! I'll follow up with results once the machine is back in action.