Magnetic vs. ball-joint rod ends [poll]

[626Pilot]

It looks like SeeMeCNC has capitulated to the awesomeness of ball joints, and produced new arms with ball-cup joints. Like the Trick Laser arms, these work by allowing a ball joint to swivel inside a constraining cavity. SeeMe uses all plastic, whereas Trick Laser uses metal balls in plastic holders.

This type of joint is very good, but using plastic on plastic - or metal on plastic - guarantees that the joints will wear down over time, eventually requiring replacement of the whole set of arms. Even with PTFE lubrication, this happened to me. It might take five hundred hours, or two thousand hours, but sooner or later it will happen. I believe using rubber bands/"zero lash straps" exacerbates this problem by putting more friction on the joints. Those will keep the joints lash-free, but not positionally accurate, until they're totally shot.

To be sure, if you're printing statues and other things that don't have to be dimensionally accurate, this won't matter until the joints are REALLY worn out. However, I've been printing stuff with screw holes that have to match up with circuit boards, and the amount of play in the joints was enough to make the screw holes not line up. This was with carefully-calibrated filament (measured it in ten places and took the average, then printed a hollow single-loop cube and dialed down the filament flow multiplier until it was within .02mm of perfect). I also put the boards on my scanner and texture-mapped the images to mockups of the boards so that I could be sure my model was correct, and everything lined up like it was supposed to. The printer was obviously the problem.

I think joint wear is also why several auto-calibration systems (mine, dc42's, and maybe OpenDACT) will compute an arm length that's one or more millimeters off of what it's really supposed to be. With 269mm Trick Laser arms, I was always getting arm lengths somewhere around 271mm. Yesterday, I installed brand new 300mm Trick Laser arms, and dc42's calibration decided the arm length was 299.91mm one pass and 300.11 another. Well, what do you know! Far less error.

I thought of building some rod ends with a combination of ball bearings, to create a typical U-joint, but that seemed overly complicated and too bulky to be practical. The other options (as far as I know) are to use magnetic arms, or alternate rod ends. Mag arms are a neat concept, and one I'm looking into, but I also thought it'd be worth it to look into precision rod ends. The nice thing with rod ends is that you don't have to weld a steel ball to a screw - it can be awfully hard to get the ball centered on the screw. The BerryBot platform, designed for mag arms, can be bought with the steel balls pre-welded using some fancy proprietary process that centers them perfectly, but holy crap, $195? There's gotta be something better. I was able to find another shop selling mag arms and carriages, but they were all sold out.

This guy made mag arms using 6mm OD arrow shaft, 6mm ID/8mm OD PTFE tubing, and 6mm OD neodymium magnets. He uses the PTFE tubing to hold the magnet to the end of the arrow shaft, and cuts it so that it just barely holds the Nd magnet from touching the steel ball. The ball rides on a ring of PTFE, keeping it perfectly centered. I like this solution a lot, and certainly PTFE is strong and low-friction, but I wondered if that wouldn't wear down over time just like regular ball joints. The inventor of the BerryBot posted a very informative thread about his mag arms, and he too uses a PTFE liner between the ball and magnet.

It was then that I looked at McMaster's selection of rod ends. They have quite a few configurations, many of which are all-metal. The ball rides in the socket on top of a liner, either PTFE or PTFE-impregnated brass, and requires no external lubrication. Many only have 25-30 degrees of swivel, which I don't think is quite enough. A few have 50+ degrees, which looks more promising. The ones with metal liners seem like the way to go. The trouble is that they're expensive, often costing ten bucks for one rod end!

I also found some cheaper all-metal rod ends on Amazon - four for $10. Reasonable! They don't specify how much swivel, but judging by how far the ball projects from its yoke, it looks comparable to a Traxxas rod end. I ordered four of them, and will be checking them out when they arrive.

What kind of joints do you use? If you ever looked into using ball-joint rod ends (other than Traxxas' plastic ones), what did you find?

[mhackney]

Check out dc42's Igus ball ends he used on his large delta. I've used Igus material in the past and it is interesting low wear stuff.

Timely post. I'm building a tricked out Max Metal delta and just started thinking about the arms and what to use. I use mag ball joints and traxxis arms on my 6 or 7 deltas today. I've never had an issue with the mag balls coming apart even on my Kraken equipped machine. The traxxis ends are all over the map when you buy a pack. I bought 36 and got 12 really tight ones, 12 so so and 12 throw aways. David mentions this in his post above too.

The original R Max universal joint was not a bad approach, it just needed higher precision joints and better adjustability. There are commercial U joints out there but I haven't really investigated them yet.

I saw those ally rod ends and they caught my attention too. They'd be worth a try for kossel but too small for a Rostock I think.

[Jimustanguitar]

Anybody remember the original u-joints that required all of the sanding to tolerance correctly? Man was I glad when they came out with the clip style arms.

[mhackney]

Yes I do. Those were the ones I was referring to! Once they were dialed in they worked great but it took perseverance and skill to get them "perfect".

[Polygonhell]

Anybody remember the original u-joints that required all of the sanding to tolerance correctly? Man was I glad when they came out with the clip style arms.

I had them on my machine until a couple of months ago.
The were fine, just more messing around than most people were willing to deal with.
My biggest issue with the original arms was their propensity to be less rigid than ideal in some directions, the trick-laser arms have obviously less effector slop if you do the grab and wiggle test.

[Generic Default]

I recommend nylon or acetal rather than PTFE; PTFE has a very low modulus and creeps easily. I used PTFE cups on an older arm setup. Eventually the ball and magnet will deform the PTFE enough to make contact with each other, at which point you lose centering and the low friction.

Friction of a ~5mm cup radius at a distance of ~270mm is negligible. As long as you avoid stick-slip, you're good to go.

[Tincho85]

Check out this guys delta printer http://scheuten.me/?cat=2
It's in German, but google translate does an acceptable job.


He is using lubricated felt pads as spacers.

[img]http://i0.wp.com/scheuten.me/wp-content ... =660%2C495[/img]

[bot]

I've got some magnet arms. The biggest drawback is that they are magnets. Tools are always attracted to them. On my next delta I will seriously consider real cup and ball alternatives from industrial suppliers.

[U.S. Water Rockets]

I have the second generation SeeMeCNC universal joints on my RmaxV2 but they have thousands of hours of wear on them and the joints got sloppy. I made tensioners by cutting a bike inner tube across one side to make a rubber ring. That made it much better. I ordered the Trick Laser Arms and was waiting for a down period to install them and then the new carriages came out and I was putting off installing the Trick Laser arms until I figured out the best way to mount them. Then the new SeeMeCNC arms with the ball socket joints came out and with the special price I figured I could get them as a backup, since my stock arms were getting loose and I started having problems with the axles sliding out over long prints, even with the clips installed. I was going to try to replace them this week, but the weather has been so nice I was spending time outside.

Anyhow, I was doing a 20 hour print and went outside when it was 50 percent complete (looked great) and when I came in the room stunk of ABS and I knew something was wrong. When I checked the printer I found a rats nest of extrusion and the hot end flopping around on one arm. Two of the axels had walked out and the hot end fell down.

So, it looks like I will be trying out a new set of arms tomorrow.

One question, is it bad or good to spray the new SeeMeCNC Ball Joints with PTFE dry lube? Will this help them last longer?

[geneb]

The ball arms are made from Acetal, not PTFE. No lubrication at all is needed.

g.

[mhackney]

Some comments...

Those felt bad will introduce variability. Felt compresses and wears very quickly.

"Tools are always attracted to them." Wow, I want those magnets! I've not had this problem on 2 deltas with mag ball arms running for about a year now. I'm interested in what kind of tools!

Delrin aka Acetal is the bees knees for low wear low friction components in this application. I've run a CNC mini mill with 1000s of hours with Delrin gibs and other wear parts and they are going strong. This is a high force repetitive motion application. Our printers are low force repetitive motion so I anticipate even less wear.

There is a 1.75mm printable acetal filament now. One of my options for the new metal max is mag arms with printed acetal ends. I could machine the ends too.

PTFE creeps like a snake. Delrin does not. Teflon is PTFE. Nylon (most varieties) is hydroscopic and changes dimension with humidity changes.

I use a LOT of Delrin in my machined fly fishing reels. It is a near perfect material in my application. It is common n black and white so it looks good too. A friend brought a 2" diameter Delrin rod in BLUE to my shop last weekend, it was cool!

[bot]

I've got the magnet arms from Werner Berry. Shopspt.com. I can't count how many times a screw driver, set of flush cutters, or pliers has slammed into the things. They also tend to get dirty and build up grime very quickly under the connection point. Berry sells teflon pads to go between the rods and magnets -- maybe not such a good idea.

They are very precise and perform amazingly... but is it worth it? Hmmm. I'm going to have to say no. I'd rather have some well-built non magnetic ball joints of industrial quality.

[mhackney]

Odd. The mag balls with sockets that you can print don't have any of those problems. I can't imagine why Berry's would. Maybe others have had issues with the printed types but I sure haven't. I can't imagine pliers sticking to them!

[bot]

These magnets are fairly large, mostly exposed and very strong. I don't see how it's surprising really that steel is attracted to them. I'm not saying the tools fly from across the room. If I'm working on the machine, near the hotend for example, or the carriages, the tools will get slammed into the magnets on occasion if I'm not careful.

[626Pilot]

I did a test print with my new 300mm Trick Laser arms. It was a lot better than the old ones, but still off enough that it would make the screws impossible to drive in. I think one of the problems is that the ball joints in Traxxas ends are a little bit larger than 3mm, but are secured with caphead M3 screws. I can move the joints a little with the screws loose.

I guess I could get some countersunk M3 screws. If I'm careful, that should center the joints properly. If that doesn't work, I'll probably get the new carriages/arms/effector from SeeMeCNC. If they're using acetal, it's probably good.

[bot]

Try some SAE #5-40 screws instead. You'd have to drill and tap, or possible get new carriages, but they should fit those balls much closer.

[626Pilot]

I might do that. I don't have a drill press, so I'd have to do it by hand. I've tapped things before but I don't know how easy it would be to get it at a perfect 90-degree angle. I figured the conical profile on a countersunk screw would force the ball joint to center. I had also thought about fabricating some custom screws with M3 threading, with an unthreaded section near the cap having the proper diameter to fit the ball joints. Seems like a lot of trouble.

This guy has an interesting idea: http://www.thingiverse.com/thing:344408

Instead of magnets, he just uses steel balls on either end of the arm. Tension is maintained with Spectra line and stiff springs going from each side of the effector to its carriage. You can see it if you scroll far enough through the gallery. (I wonder if it makes weird creaky spring noises all the time?) He also has his effector designed with a built-in "hot end Z probe." The hot end swivels on a long screw, and springs hold it down on the other side. If it runs into the bed, it trips a microswitch.

If printable Delrin is a thing now, that would be a perfect material to use for the cups. Has anyone printed with this stuff? How is it?

[626Pilot]

Got a question for anyone who's tried making their own mag arms. Where do you get 1/2" or similar threaded steel balls? I didn't see any on eBay or Amazon.

[bot]

There's a guy on the delta bot google groups, Haydn, who sells them very inexpensively.

[626Pilot]

I'm gonna get some of those. I tried countersunk M3 screws on my Trick Laser arms and they don't help. Actually, I'm getting layer shifting now.

I also thought about giving the SeeMe arms a shot. I know the ball part is acetal, but what about the cup? If it's ABS, won't it wear down, just like the ABS in Traxxas joints wears down?

I'm starting to think the guy who uses Spectra line and a spring to hold the tension on the arms, without bothering with magnets, had the right idea. I'm thinking I might just design an effector with the SeeMeCNC hole pattern and CNC cut it out of a block of acetal. That stuff isn't too expensive. I would print one using acetal filament if I had an accurate printer.

[KAS]

These are the ones I have, not sure if they'll work for your application.

http://deltaprintr.com/shop/parts-and-a ... eel-balls/

[626Pilot]

Interesting. I have a DeltaPrintr but it's sitting unused because the auto calibration doesn't work. I had no idea they were getting into magnetic arms, but to be honest those acrylic arms they were using were super flimsy!

I just ordered a bunch of chrome steel balls and magnets from Haydn Huntley in the deltabot group. He answers quick! I also got some 12-inch "strong wall" rods from TriDPrinting.

EDIT: He sells a complete 288mm set of rods, ASSEMBLED, for $64. How about that! I ordered some. He also gave me some good advice on doing the jig for the ones I'll make myself.

[bot]

Yeah... had I seen his post before I ordered the berrybot parts, I would not have spend nearly as much. The berry bot parts are way overpriced, IMO. Look at his site. He re-sells a $30 stepper for $95. WTF? The build quality on some of the berry bot parts are not what I was expecting for $600.

The magnets are strong, tho.

Haydn's stuff is much more appropriate for most applications.

[626Pilot]

I don't have the parts yet, but based on the specs (M3 stud, 10mm hex collar) and a few measurements, I think there's a good chance these magnetic balls can be fit to a Trick Laser platform and carriages. They're getting here on Monday. I can't wait!

Berry's stuff costs a lot. I guess he's charging "machine shop" rates? Idunno. I just can't see spending six hundred bucks on the drivetrain. I also don't get why he uses such a honking huge stepper. A short-canned NEMA 17 with a planetary gearbox has loads of torque. That thing looks like it could punch holes in steel if you ran wire through it!!!

[bot]

The stepper is a nema 11, so it's much smaller than a 17.