I am really liking this. How smoothly does it roll?
Smoothly, just like butter. No carriage movement that isn’t supposed to be there at all. No hesitation or catching, no harmonic drag wobble, just pure quick back and forth motion where it is supposed to be.
My only potential concern so far is if the wheel tensioners will loose up over time. If they do, it’s really easy to re-tighten them. I could also use some blue loctite on the cam lobe section if necessary. So far with my very limited run time It hasn’t been an issue. I’ll know more after this weekend though since I plan to do a lot of printing with it. Component wear won’t be an issue, and if anything does wear out, it’s going to be just as easy or even easier to replace a part without dismounting the entire X carriage. Debris in the track is another unknown. It won’t be an issue for the X axis, but it might be one for the Y axis. I have plans for a cover of some sort for the Y axis when I get to that point but I need to get the holes cut in the bed plate before that can go on.
I would love to see a close up shot of your part sidewalls once you have some printed.
The jankiness of the IGUS bushings has long bothered me… and the flex of the rods. I have improved the tightness of my printer with hardened rails and bearings but, those seem to bring a certain texture to my prints, mostly in the Y axis… I am highly interested in something that rolls smoothly, like butter. 
Thanks for all the hard work.
Will do. I’ll use tonight to tune the system, make sure the integrated bed tensioner is working right, make sure I don’t have a leadscrew pinched, etc. Once this modification is dialed in and tested, if I still need it at that point i’ll bolt my Z axis mount mod up to it and see how that works. Y bed is probably going to be next week assuming I can either find a local shop to punch the 4 holes, or I get adventurous and order my own 7.2mm bits for the Bed plate. I’m going to be interested in seeing if this helped, and specifically if I have any offset in the Y direction left. My parts have always been pretty good with the Taz, and the rev 4 parts definitly helped, so any improvement over that is going to be a bonus!
Here’s a somewhat out of focus extreme closeup of the latest part under LED light at the angle that shows the absolute worst ridging. It actually looks a lot worse than it is in this shot. I still have some wobble, but its +/-0.01 in depth, which is about how much I can move my Z bearings back and forth on their rod. 'This project definitely improved and minimized it, somewhat, but where it really shines is the layering itself. I am getting the best layer faces I have ever seen. My Anti Z wobble project might clean up the rest of this, though I’m thinking I’m going to have to figure out how to mount a Z openrail section to fully mitigate it. That or find some bushings with closer tolerances on the rods.
Nice. Thanks for the pictures. I will have to source some metal X end plates. I am currently using the printed all in one pieces.
Nice mod! Curious to see more testing… speed on the v-slot wheels would be particular interesting.
The wheels are basically roller blade wheels with a modified profile. They have 2 bearing sets in the middle. So far I’ve been able to print at normal speed with no issues, no heat buildup and no need for readjustment. The belt tensioner also helps somewhat. I’m going to wait on pushing the speed envelope too far until I get the Y axis mounted as well so I don’t end up destroying my existing Y rod bushings.
FWIW, I asked our CTO his thoughts on using these:
The short answer is yes, built correctly, rails will be better than smooth rods.
The Kauri project actually included testing of rails for making the X axis stiffer. We have sections of this v-slot rails as well as other rail systems. Rails fall into two sub-categories, adjustable and non-adjustable. With the adjustable ones, you need to individually adjust all of the roller wheels and tighten down their axle again to get the maximum performance out of it. The non-adjustable as you’ve guessed the just have the factory level of slop to them.
We started with the non-adjustable ones and quickly found they have the same or more slop than we have already in the TAZ design.
The issue with the adjustable ones we quickly determined is that it takes a lot of time to adjust all the rollers. There will be 3 or 4 to adjust on every rail so somewhere like 16 or so to adjust per printer. Then there is the question of how often those rollers will need to be adjusted in the future. There will certainly be a cost jump when we go to these. We haven’t yet made a physical model with the V-slot rails yet. With the death of Kauri and the simplification to Lancewood the testing has been pushed to the next printer development.
-Jeff
Can’t wait.
The openrail design only needs 2 adjustable wheels per axis, they press down against 2 fixed wheels on the opposite side of the rail, so you would be looking at 8 (edit - 2x Z axis, so 8 not 6)adjustment points per printer. The Adjustment process is just grabbing the barrel of the cam adjuster, and turning it to increase or decrease tension with a pair of needle nose pliers. I find it takes me about 10-15 seconds per wheel to dial them in, and once dialed in, the carriage can be removed from the rail, and even transferred to a second identical rail without having to re-adjust. You should be able to theoretically make a bunch of pre assembled carriages, adjust them on a bench fixture mounted rail segment, then place them on the line until needed without having to readjust them again. The amount of adjustment required is going to also be a constant. You could print a wrench with an angle stop on one side as a precision adjustment tool: Line up the face with the adjustment mark, then turn wrench until you hit the rail for perfect adjustment every time.
I would agree that the adjustment frequency is a big unknown at this point. I’ll keep you posted on how often mine need adjusting. I suspect that application of some blue locktite to the cam segment itself would make it so they would never need adjusting again within the service life of that printer axis.
Cost is also going to be a big factor. The V rails are more expensive than the rod system they replace, mainly due to the wheels. That’s one of the reasons I wanted to make this a bolt on upgrade. You could theoretically offer a “basic” version of the printer with just rods and the rod carriage, and at the same time offer a “Pro” model with the rails, or as an upgrade kit after the fact. I’ve already had 10 people message me from various locations asking if I will sell them a kit of the X axis upgrade, so there might be a market for that. Raw rail part cost for the Rail system in single quantities are around $95 for the X axis, a bit more for the plastic, $89 for the Y axis (not counting the machining costs) and probably around $129 for the Z axis unless that changes. You could simplify the X and Z axis design with some major modifications to the frame and maybe get that overall cost down a bit, but it wouldn’t be a bolt on modification to the TAZ anymore. Probably replace the back frame rail with an openrail segment, move the X motor outboard to the edge of the X extrusion, etc. I would expect a fully rail equipped printer would be around $300 or so more (Edit - minus the cost of the rod hardware and bushings these replace) just on parts cost alone, but the people who need that level of precision might be willing to pay for it, especially if there was a less expensive option for those who don’t need the added level of precision.
If you had access to plate cutting fabrication facilities, you could make the X axis plates smaller and lighter than the out of the box plate I ended up using to save some weight, though it seems to move fine as is. The existing Y bed plate would just need the 4 extra holes bored in it, so that wouldn’t add much to the manufacturing cost I would imagine. if you kept the existing rod mounting holes you could offer a basic and pro model of that axis as well.
If you want any changes to any of the parts I’m putting together to accommodate space for any new features you have planned, etc, let me know and I can change things up accordingly.
Cool, thanks for the info. We have various rails around here and have ordered the one you used to take a look at it. Igus, Misumi, Thomson, et. al. give us sample stuff to check out.
Thanks,
-Jeff
Neat! Let me know if there are any other particular systems that look interesting and cost effective.
I went ahead and started a poll thread over in the openbuilds forum to see if anyone had any imput on how often the cam adjustors need adjusting, and I didn’t get a whole lot of responses. I printed about 40 midsized parts of things this weekend on the X, 10 of those with the X and Y in place, and so far they haven’t moved at all. Not really a great long term test yet, but it is promising.
http://openbuilds.com/threads/how-often-do-openbuilds-v-rail-wheels-need-adjusting.2969/
The few people I did hear from over there seem to indicate an occasional need to adjust the cams, and those people were operating all up CNC mills with them instead of a comparably lighter duty printer.
I’m working on the full assembly instruction guide for both the X and the Y, I should have those done sometime this week hopefully.
Thanks for taking a look at it!
Tim
Cam adjustment frequency
I’ve designed/built/operated systems with cam-adjusted guide bearings ranging in size from smaller than the Openrail units, to significantly larger units. The need for adjustment depends primarily on three factors:
- Good design – making sure that the bearing assemblies are properly aligned with the primary force axes is critical in larger machines. Designing so that cams can be locked in place once adjusted is, obviously, the key to success. Usually this accomplished by making certain that cam fit in the mounting plate is just shy of snug - free turning but with a hint of drag. Locking is most often done by adequately torquing the cam mounting bolt (this is the weak point that results in most cases of the bearing going out of adjustment). More critical designs can incorporate a recessed groove around the perimeter of the cam that will catch a setscrew mounted through a frame element or collar. Finally, the materials used for the cam and mounting plate should be adequate to handle static and dynamic deformation loads (minimal on a 3DP if everything is adjusted properly). As a rule of thumb you will use different materials for the cam and mounting plate (i.e. mill aluminum cam mounted in a mill aluminum mount = friction welding - that’s one reason why brass cams are common.)
- Proper assembly, adjustment, and tightening – Cam mounted bearings should be installed so that matching pairs are counter-rotated equally into final position. On something like the X-carriage, the two top bearings would be fixed position bearings and the two bottom bearings would be adjustable. The cams would start in the same relative positions (usually at maximum gap or top-dead-center). Then the cams would be counter-rotated; the left bearing would be rotated clockwise into adjustment and the right bearing would be rotated counterclockwise into adjustment (or vice versa). If everything else is perfect (unlikely) both should be rotated an equal number of degrees. In the real world - strive for equal rotation and take up the remaining error by fine adjustment. Tightening the cam bolts once in position must be done carefully and incrementally in pairs (like torquing an engine head). This will keep the assembly from skewing out of alignment. Ideally the cam will have flats on it that can be held with a thin wrench while the bolt is tightened. NOTE that cam bolts must always be re-tightened any time that the cam is adjusted!
Assuming all of that is done as diligently as possible, the final factor affecting required adjustment frequency is:
3. Operation within design limitations – the most common cause of bearings going out of adjustment is sudden dynamic loading - shock loading. Very unlikely on a 3DP unless you drop it.
After several weeks and a very heavy amount of printing, the allignment is still perfect and rock solid on the X and Y axis. I’ve got the parts on the way to add this same mod to my AO-10x now as well. it already has a Y and Z rail system though.
Great work on these mods! Did you finalize the number and length of the screws as well as the rail length? I’m printing the parts now and want to be able to put my order in for the hardware.
I haven’t finished writing the instructions up yet, but I have the basic numbers. You will need the following:
X motor mount block:
6 M5 heat set inserts (I used the longer ones)
10 M5 10mm long bolts and
4 M5 T nuts
The existing M3 motor mounting hardware (4x bolts and washers)
X idler block:
4 M5 T-nuts
2 M5 heat set inserts (the shorter ones)
2 M5 12 or 14mm bolts (for the tensioner)
the existing bearings, washers, nut and large bolt from the Taz Idler
Mounting ends (2x):
3 each M5 heat set inserts (the short ones)
6 each M5 10mm bolts (you can use the existing mounting hardware for 3 of them)
3 each T nuts
Rail, Carriage and Carriage plate
9 M3 heat set inserts (the short ones)
4 M3 6mm or 8mm bolts and 4 m3 washers
4 M3 10mm bolts
1 (Size: 20mm x 60mm, Length: 500mm) Openbuilds linear extrusion, You will cut a section off of this, I’ll post the exact length later once I measure the piece
http://openbuildspartstore.com/v-slot-linear-rail/
4 sets of bearing wheels (nuts and washers come with them http://openbuildspartstore.com/xtreme-solid-v-wheel-kit/ or other sets) ,
2 spacers (6mm probably) http://openbuildspartstore.com/aluminum-spacers/
2 eccentric spacers http://openbuildspartstore.com/eccentric-spacer/
one of their Gantry mounting plates. http://openbuildspartstore.com/v-slot-gantry-plates/
Bolts (M5 low profile, I can’t remember if I used the 30 or 35mm ones)
2 extra 30 or 35mm low profile bolts for mounting the backplate
plastic spacers (or cut aluminum tube) to space the back plate out. You may need to use a couple of washers as well.
It would be cool if you can get instructions together and release a kit like what was done with the HOTBOX enclosure.
The instructions are almost done. I was playing with the Z axis instead.
As far as kits go, all I can tell you at the moment is stay tuned, there may be some movement along those lines at some point in the near future.