Linear Bearing Mendal, any Thoughts?

I have been working on re-designing the bearing parts for mendal to accept linear bearings to try and improove accuracy and rigidty. (£20 for 12 off of these bearings!) and wanted to see if anyone has any thoughts or opinions on it? have a look at my work do far in this 3D PDF.

Thanks Guys/Girls.

If you watch the video explanation of the Mendel bearings they are designed to work with imperfect rod and alignment. Unless you get the bars perfectly aligned won't this design bind?

What makes you think the Mendel bearings are lacking in accuracy and rigidity?

What sort of linear bearings do you get for £20/12? If they are just sleeve bearings then don't you need some slack in them in order for them to move freely? The Mendel bearings are under tension so there is no slack.

---

[www.hydraraptor.blogspot.com]

They are a Ball type linear bearing, from a chineese supplier (part LM8UU ), and i do intend on getting alingnment bang on, if nec i was thinking of running a hand reamer though the bearing location bores (once assembled) for exact sizes, and using shim down to 0.02mm thick to get exact alignments, hopfully no more than 0.05mm of out of square through out the mendal frame. And 0.02mm of parallelism between 8mm ground rods.

But its a suck it and see, i geuss. It may turn out to be a white elephant.lol.

[myt-bearing.com]

I am VERY interested in this, I have plans to build my second Mendel with this process, or something similar (I am 25% complete on my first mendel, using a standard Mendel design.) I have all the frame pieces for a second unit, thought I would get the first one done and then make a second extreme R&D version. Have the 5/16 centerless stainless rod,k I wanted to try nylon sleeve bearings or something like your idea. As I finish my first system I will be watching your attempt with interest

Watch out also that your rods are perfectly straight.

It's generally dangerous to over-constrain anything, unless your tolerances are extremely tight.

Yeh i do agree with over constaining, but i think it is perfectly possible to get linear bearing to run smoothly, just need to spend some time making sure every is exact as you build, i plan to also use centerless ground ø8.0mm rod, we buy alot of this where i work so i know it very straight.
I have only done the 'X' axis carrage and 'Z' axis ends so far as shown in the PDF but i am starting to work on designs for the 'Y' axis now.
I will put another PDF up of the 'Y' axis when it is done. and then hopfully i can crack on with a build!
Any thoughts and ideas are always apreciated!

Steve

Nophead.
with the current Mendal bearings, as you said they are designed and set up for inacurrate rod and alingments, for this very reason i would say they cannot be accurate, due to the design intent of being inacurrate. (if that sentance makes any sense at all..lol).
Thus my design intent is to not have any inacurate rod or missalingnments, and so giving me an opertunaty to use purpose made linear bearings. This will also reduce the overall parts count for mendal, and actually make the RP parts alot simpler from a design perspective.

Thats my theory anyway, it may all end up as a good looking door stop.lol

Personally I was very skeptical of the bearing designs on the Mendel. But having assembled mine, I am surprised with how good they are. If you have not built one with the standard bearing design, I would recommend giving it a shot and seeing how well they work. They are designed to avoid binding, however they have a lot of room for adjustment. If you spend a few minuets getting them adjusted well, you can get very smooth movement. I can not feel any wiggle in them at all. I have purchased linear bearings from 8020 (mind you these are not amazing bearings either), and I think the Mendel ones are better.

Basically, if you are dismissing the Mendal bearing design without testing it, I think you are making a mistake.

Josh

just a little update with a pick of the 1st Linearing bearing bits that have been assembled.

WOW Very nice...

How much were the linear bearings, how long did they take to get in, and did you buy any extras that you would sell? (and are you in the US, Europe?). I am very interested in a linear bearing version, was going to try nylon sleeve bearings next, would love to try your bearing design out as well.

The 12 off linear bearings were around £25 i think..from a cineese supplier,ebay link below (ask for the 8mm version and a price for however many you need, i am using 12 off). the bearings took about 2 seconds to get in..(bores reamed to ø15mm to suit bearing) . The Plastic parts have slighting larger bolt holes to allow for alignment later.

All of the bearing housings i have drawn up can be retrofitted into an existing medal, except for the X-Carrage that has to be totaly replaced. i will be selling sets of the printed bits as soon as i have proof of concept and its built and moving smoothly.(only selling sets to fund Build costs, electronics etc)

i am in the UK

have you any pics/models of your bearing design? as i would be interested to see it from some one elses eyes..if that makes sense..lol

[cgi.ebay.co.uk]

Cheers Guys
Steve

Thought i would add another update, heres the opposite 'Z' Bearing side of the carrage now assemblemed, with linear bearings.

Quote
sweaving
with the current Mendal bearings, as you said they are designed and set up for inacurrate rod and alingments, for this very reason i would say they cannot be accurate, due to the design intent of being inacurrate. (if that sentance makes any sense at all..lol).
Thus my design intent is to not have any inacurate rod or missalingnments, and so giving me an opertunaty to use purpose made linear bearings. This will also reduce the overall parts count for mendal, and actually make the RP parts alot simpler from a design perspective.

Not to dampen your efforts because there is value to making mendel cheaper and faster to print like you say, but I would like to correct what you said about the standard bearing setup and accuracy.

Although the ball bearing design is made for less accurate parts like you say, and although bent rods will affect accuracy (vs. binding for linear bearings) I don't think bearing alignment effects accuracy at all. This is because the precision of the plastic holes for each bearing is a static concern which does not change during movement on each axis and so can be ignored from what I've seen. My Mendel uses this setup and I haven't noticed any problems (mind you I made sure not to buy any noticeably deformed bars).

tl;dr linear bearings should be cheaper using your source but a slightly increased accuracy is probably not going to increase build quality by any noticeable amount IMHO

PS. Sweaving if you could post the info on your linear bearing mendel + stl files on the wiki when you get a chance I'm sure it'd be greatly appreciated by everyone who wants to build one, best of luck

Thanks for the comment theOtherRob, i do agree that there will only be slightly increased accuracy, but i think the benifit lies in reducing the parts count, fewer RP parts and bought parts, And far less complicated assembly, also as part of what i am doing, i am also reviewing every mendal part to improove print times. I will definatly be posting the STL files on Wiki, but before i do i need to finish an actual build for proof of concept. Doing a mendel virtual assembly is alot easier than an actual build..lol..

Latest update for those who are following, i have now finished the design of the Y-Axis, and built up to test for alignment, so far things are running very sweetly on the ground rods, although i am yet to put it all together..lol

all mechanicals are built now! and running very sweetly, no binding or twisting, i didnt even have to adjust the frame at all! just bolted it all together, a few accurate measurments, and done! very pleasant surprise! just waiting for my steppers to arrive now!...woooo...pics to follow soon.

Good that it's working for you! At Maker Faire last weekend I was sitting next to some students with another Mendel design that was over constrained on the X and Y axis, and they were having a lot of trouble getting it to run smoothly. I think they'll get it running in the end, but it's certainly not the easiest way to build one of these machines.

Hi Wade, im just printing your extruder design to go with it! (nice work there). What was there method for running the bearings then?

ive actualy purposly put alot of twist into the frame now to test for binding. an the darn thing still runs smoothly..lmao. the only way i can get it to bind is if i purposly make the guide rails not parallel, even this has to be out alot(over 2mm), i think the bearings i have bought must be able to align themselfs somehow.

You can address alignment by making it adjustable. You may approach this like the cross slide on a lathe, or could even be accomplished with set screws to jack the fixture around within a track.

For the rigidity needs of a reprap: a printed to accuracy reprap using an interference fit could probably be made on an adjustable fabricated repstrap.

Adjustable means less scrap, but more hassle for the user. Non-adjustable means you're stuck with as good as you get. I think non-adjustable is generally a little bit more rigid.

Bump! lol, thought i would re-invigerate this post as i have finaly got motivated to get my linear bearing mendle sorted of and running! have just been going through a testing phase with the old Gen 3 electronics that i bought 2 years ago for it, and all seams sweet, except for the stepper extruder running from the H brigde that noisey as hell so i have opted to go down the Frankencable route to a 4 stepper board that i an putting together this weekend.

Hopfully this weekend i can make my 1st print...only taken 2 years...not long..lol...i will post up some pics at the weekend for those interested..PS all my linear bearing designed parts can be Retrofitted to an existing Mendel chasis. So anyone who wants to convert to the LM8UU's can do so with minimal parts to print.

Steve

Well, i got over excited and tested my frankencable with a 4th stepper board last night, and its running as sweet as can be, left it feeding filiment for about 15 mins to check on motor and board temps and all is good, just need to re-assembly route a few cables and get printing now!

Test prints completed over the weekend, with only one problem, the z axis is missing steps! when i am watching a print 90% of the z moves work perfect ok, and then the odd one will simply make the motor buzz with no movement.i have tried allsorts, adjust the pot to min and then slowly edged it up untill working, still has the problem, turned it to max still has the problem, swapped stepper boards still has the problem, slowed down the rapid move in skienforge still has the same problem..

Any thoughts or ideas of what to try would be apreciated..

Just thought i would post an update, i have now dumped all the old Gen3 Electronics, was causing more hassle than it was worth and looked like a train had just driven across it!..lol, now have in place Ramps 1.4 with marlin and running Slic3r with pronterface. i have to say this is a great bit of software! much better than repg, only pain i find is the manual control panel.

am now working on a new nozzle, as im still running old ni chrom wrapped M6 rod (alothough very reliable). is a bit outdated now and i fancy tring one of those barrel heaters off ebay! any tips on there set up?

Afternoon All, just thought i would give this a little bump as i have noticed people talking about durability of the linear bearings,

The Mendal that i have written about here has now been running since early 2010 and has been very reliable. Although it has gone though teething troubles, and a few upgrades, all the main mechancical parts have been sound! Its whirring away as i right this, lol

its even still running an old style nichrome nozzle!!!! (4 years and counting on that now.lol)

Any comments are always welcome.