For a printer, it seems like a Servo Belt type system would be far easier to implement.
According to Nexen, spring loading would require too much force.I wonder If spring loading the rack and roller pinion together would work?
How precise are your pinion locations? I'm guessing that any error will result in wear in the rack.
When I bought my Nexen pinions, I saw another set of pinions (not Nexen's) on Ebay that came with an aluminum rack, which didn't show any signs of wear.
What about hard anodizing the rack?
I was going to make my rack from phenolic (Garolite).
In the video, it doesn't sound that loud at all. The sound of the stepper is louder.
Is there any noticable backlash in your system?
Gerry
UCCNC 2017 Screenset
[URL]http://www.thecncwoodworker.com/2017.html[/URL]
Mach3 2010 Screenset
[URL]http://www.thecncwoodworker.com/2010.html[/URL]
JointCAM - CNC Dovetails & Box Joints
[URL]http://www.g-forcecnc.com/jointcam.html[/URL]
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
For a printer, it seems like a Servo Belt type system would be far easier to implement.
Gerry
UCCNC 2017 Screenset
[URL]http://www.thecncwoodworker.com/2017.html[/URL]
Mach3 2010 Screenset
[URL]http://www.thecncwoodworker.com/2010.html[/URL]
JointCAM - CNC Dovetails & Box Joints
[URL]http://www.g-forcecnc.com/jointcam.html[/URL]
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
Hi Gerry, the pinion locations are quite precise, I drilled the holes on my CNC mill. The Cam followers press fit in quite nicely.
There is some noticeable backlash in this system....
When I machined recess for the bushing that contains the grub screw to fasten on to the motor shaft I stupidly machined the diameter approx 0.01mm oversize.
This has resulted in the whole roller pinion head being slightly off center, I think this is the main cause of the backlash. I need to machine a closer fitting bushing.
The very first time I assembled everything and tried to move it by hand it kept sticking.
I had to take it apart and use some foil as a shim to make it more centered.
It works, but I don't think its really centered properly.
It runs fairly smooth but I can hear some of the rollers impacting harder than they should on each revolution. you can't really hear it on the video.
You can see in the following photo (oxide smearing marks) that only the top part of the rack is making contact with rollers.
I think this is due to the geared stepper motor shaft not extending far enough to the base of the roller pinion plate.
And I reckon the plate must not be properly square to the rack.
I think a Nema 23 geared stepper with a larger diameter shaft and additional roller bearing would work a lot better for this size RPS (it would not flex as much).
These issues are due to my amateurish design/machining skills.
Hopefully I'll sort this all out in my next version.
I'm sure your right about the Servo Belt type system being better suited for a 3d printer, only I've never used servos before.
And all the Arduino based 3d printer controllers I've seen are setup to work with steppers.
One of the reasons there are not many larger format 3d printers out there is because steppers and belts don't really work that well together once you reach a certain axis size.
I'm hoping to create a DIY RPS style large format 3d printer that works with the existing Arduino based stepper motor controllers that are out there.
Nikki
This system demands absolute precision. If the rollers aren't all making smooth contact with the rack, it's not going to work well.This has resulted in the whole roller pinion head being slightly off center, I think this is the main cause of the backlash. I need to machine a closer fitting bushing.
If you haven't already, you might want to read through Nexen's technical data. They (lightly) preload the pinion into the rack to eliminate backlash. But the pinion to rack interface must be perfect.
The Servobelt system is not your typical belt drive. It can be used with steppers, and has none of the issues typically associated with long belts.
http://www.cnczone.com/forums/linear...-ever-say.html
Gerry
UCCNC 2017 Screenset
[URL]http://www.thecncwoodworker.com/2017.html[/URL]
Mach3 2010 Screenset
[URL]http://www.thecncwoodworker.com/2010.html[/URL]
JointCAM - CNC Dovetails & Box Joints
[URL]http://www.g-forcecnc.com/jointcam.html[/URL]
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
Hi, I'll second that wholeheartedly....you must have precision if you want to work with cyclic movement.......the ultimate aim, apart from moving from A to B, is to move a specific amount each time and every time, otherwise you end up with parts that are dimensionally odd, which means you could have got better results with Althread and a bronze nut.
Now that you know the design as you have it works, I'd move on to the material choice as you have the program to cut the rack so material is just another trial and error approach and plastic is much cheaper than aluminium.
There's a good reason why the involute curve is used for gearing, whether it be a rack or pinion, and accuracy of movement is the prime reason.
A manual lathe relies on the accuracy of rack and pinion drive to generate the auto feed that gives you the finish on a turned cylinder.
BTW, going to a bigger diam pinion with more rollers will mean you lose the resolution due to one rev of the stepper travels a greater distance.......and gearing down will introduce more backlash.
I think as a DIY approach, a plastic compound for the rack and a small amount of spring loading to the pinion "could" possibly iron out the inaccuracy of the pinion/rack profile engagement.
As you have a rack already, you could make a pinion too but using an involute profile which would be heaps more accurate than the round pin and a rack form.
In that case it would work better if the rack was left as aluminium and the pinion cut from Nylon.
The Nylon would be self lubricating if you can get the graphite impregnated type.
Any form of spring loading would probably result in a backlash type of hesitation to the drive if you hit a tight spot and the drive reacts to it by forcing the pinion out of the rack slightly, whereas the resilience of the Nylon by itself would probably suffice, but concentricity and accurate alignment is of prime importance.
Ian.
the manufacturing very difficult for this system
I believe it can be manufactured with only machines, where glassscales ensuring the backlashfree system..
in the link at least 2 bearing will be implemented, one the holding the pinion, and other is the actually in contact camfollover..
theres no ""playfree"" bearing..
by bearings you have about 0.015- 0.02 mm play , and possible this is the point where nexen applied ""preload""
the whole system as moving along the rack also has deviances..
I mean it might not possible to keep absolutely parallel the path of pinion, and the rack..
overall I believe a simple rolled ballscrew will works better..
and the original nexen system where it could yielding on large applications, over 16 feet, like large gantry mills, there the system simply not sufficient rigid, compared with a decent ballscrew..
ballscrews are not ending at the 1605... there are screws with 50 or larger mm lead.. and their backlash also only 0.01 mm ..
this is my opinion..
it is a very interesting approach, but this system will be more expenses than a ground ballscrew, what today accepted for most application..
there are machines, using for feed back glass scales, so there the backlash on screws or rack n pinion wont be important..
so , overall beyond an interesting approach, its expenses, and way less effective than already in use solutions..
I see positional accuracy reports (no better than a precision ballscrew for sure) but I'd like to see a set up and run ballbar tests to see if it could also keep a constant feed through that rack. I betting its going to show cyclic error in large amounts.
I did not read the entire thread, but some of the posts on this thread are not O.K. by any means.
This is MY opinion from a long experience with mechanics:
1. roller pinion from nexen is a nice solution but to expensive for what it offers ( from the price found on the net, not sure where or how, but i remember thinking that is a high quality ball screw price territory )
2. servo belt is also a nice solution with the same problem mentioned above.
3. rack and pinion is proven reliable and easily set up with very little effort since there are cheap ready made motor mounts with reduction and springs and all. They do have the backlash and vibration problems ....
4.Anyone who thinks that ball screws are not precise ( i read here complaints on the subject ) never had experience with NSK or THK industrial ball screws. 0.001mm is the norm there and i will not even get into the other related subjects, just look at the best and most expensive and most precise machines ==> ball screws all over, still today in the age of linear motors.
5. The humble and simple toothed belt. reliable, quiet, durable, etc etc. To many good things to mention here.
FYI all of the plasma cutters built by me in the past 4 or 5 years use belts, very thin (8 or 10mm wide) belts, BUT i build the moving parts very light and sturdy and always 2 motors on the Y axis.
Later
Tom
Hi, as soon as anyone mentions resolution it immediately gravitates to .01mm accuracy, no matter how big or small the machine, whatever.
If you had a small 3020 type router made with Hiwin linear rails and top market ground balls screws, that would be "normal" or up to expectation, but if you are in the realms of a 5 meter long table router and got .01mm accuracy that would be stupendous.
The size ratio between small and big apparently is not a factor in the equation, but at 16 1/2 times the size, the scaled comparable accuracy for .01mm would be ,165mm, and you could get that from an aluminium built machine by expansion and contraction on a hot day or when frozen
Big ball screws are the way to go if your pocket is deep enough and .01mm is in the air, otherwise the chain is only as strong as the weakest link.........the other way is to use encoders and know where you're at even if you can't accurately step the way from end to end on long travels.
Backlash can't be cured by fitting an encoder, even if it adds movement to compensate for lost movement etc, and backlash is a bogey when climb milling anywhere.
I think there has to be a sliding graph for the expected accuracy of travel and applied to the size of machine when doing a DIY model, from small to very big.......large commercial models are expected to attain the .01mm out of the box......the mega bucks price tag ensures that.
Ian..
you re right..
that was only example.. but in reality hard to imagine any serious manufacturer use this roller system versus a ballscrew..
-----------------------------------------------------------------
probably fitting 150 pices of 100 mm gauge wouldn't work to checking the 15 meter length :-)
for precise machinery the only way is the linearscales..
rotary encoders will copy all error from screws..
But if its a known error, its easy to use pitch error compensation. All depends on what its being used on. Don't think it would be stiff enough on a high speed machining center, but may be fine on routers, lasers and waterjets that really don't need the stiffness.
Hi, DIY people are not really in the class of serious manufacturers in the hobby sense, as in they only want to make a one off and hope to us it as such, but as soon as you start going big, the exponential curve of lack of accuracy become the biggest problem too.......how do you measure anything over 300mm with any degree of confidence without having the facilities of a temperature controlled environment.
I think the reason why short pieces and lots of them are in this design, is because you can measure and make short pieces relatively accurate and also align them one to another over a length to be accurate more than you could cut a long rack and hope to have it anywhere near to the ideal for close tolerance production.....but provided you can measure the overall length to detect accumulated dimension creep etc.........you'd still need to have measuring equipment to detect the overall length of the racks.
Ian.
Hi....pitch error compensation only works when you know if the overall length of the segmented parts is in error.
You could measure one segment and find it had an error of .05mm overall.......joining segments together does not eventuate in an exact accumulation of the ,05mm times the number of segments........that's the ideal World.
I think the Nexen system is costly and for those that need it as opposed to rack and pinion or long ball screws etc.
Making either the rack or the pinion to the Nexen design would be a compromise in accuracy......it could be done but the end result is not guaranteed.
A lot would depend on the alignment of the pinion to the rack segments, and over a long length that means in line and perfectly pitched.........measuring the pitch over a long length is a problem of metrology, whereas the alignment is a total "fits where it touches" scene......the carriage must ensure that a linear contact with the pinion and rack is maintained 100% at all times or you'll lose pitch if the pinion moves away from the rack by even a few hundredths of a mm.
I think even a simple steel rack and pinion needs a resiliently mounted back up roller to keep the pinion meshed with the rack or it'll have loose and tight spots that defeat the object of the exercise.
Ian.
Pretty easy to measure long distance accurately. Just need a laser system. Ive done 20ft travel machines myself. Not a diy solution though.
Heh Heh......I would think that anything 20 ft long would be difficult to measure accurately by any means......you'd first have to have a standard that was set at 68 deg C in a laboratory to ensure what you were measuring with was able to actually give you an accurate reading and were talking in terms of .01mm accuracy as said before.
As I also said, large items probably don't require to have the tolerances that small items need......certainly not to + - .01 mm sizes, and working with a router those tolerances most probably will never occur for any work anticipated.
As we're discussing the Nexen drive as opposed to actual measuring means, if it can be made less expensive then possibly more people would use it by choice.....more people....more turn over....a less expensive product.
Ball screws would be in the gold class if only a few people used them instead of the Acme threaded screws for transmission, so in the first instance, anything used in place of the mass produced ball screw has a tough act to follow, and the Nexen drive system is in that category.
So, do you gain any more accuracy or travel speed over a long distance with a Nexen system as opposed to a plain and simple rack and pinion or even chain drive with the chain being the static member.
If the answer is "not all that much"......where is it going and for what reason?
I don't think being potentially backlash free has a big tick in the must have box if the cost is also on the high side.....making a rack and pinion fit tighter will also reduce any backlash.
When I work with a manual mill I always nip up the lock on the dovetail slides to ensure that the slides don't wander, and for fast traverse it just runs as free as the slide way dovetails are adjusted to......the same applies ,to rack and pinion....tight set for feed work and slack for fast traversing etc will ensure the ideal World without spending mega bucks..
If you want a perfect World you need to have a perfect system, so tight for feed work and slack for fast traverse, and how that is accomplished is the price you have to pay, but you have to accept compromise and draw the line somewhere.
Ian.
We shoot for 10 microns over that length. Lasers have weather stations, so they can comp for small deviations. Shop is at 68, this shop is +/- 2 degrees usually, not the precision shops that hold within 1.
I still think this design will show some camming or cyclic error trying to interpolate a circle.
Hi, I don't think the manufacturers would have marketed it if it showed design faults.
You would actually need to go out and buy into the system to find out if all the blurb was true and then do extensive testing to prove otherwise.
The object of having this design as opposed to a regular rack and pinion or ball screw is relying on the integrity of the manufacturer to take advantage of the attributes and justify the cost.......extensive testing should never come into the picture, that is..... fit it and forget it.
Ian.
I'm not sure why this is even being talked about, it's close to the same price as a Ground Ballscrew, has some backlash, should never be compared to a Ballscrew, it is however better than a rack & pinon but the hobby user will use the cheaper rack & pinion, I have seen them working on a large scale, and they work very well, it's like all means of motion they have there place, and that is where they are being used
This drive system could never be used where a Ballscrew is needed, for precision positioning so don't compare the two, there is no comparison
Mactec54
Hobby router: 100 um
Hobby mill: 10 um
Precision mill: 1 um
Optics machine: 100 nm
Semiconductor fab: 1 nm
You pays your money and gets your choice.
But that roller system - I think it sank in the market. Haven't heard it mentioned for years.
Cheers
Roger
overall beyond an interesting approach, its expenses, and way less effective than already in use solutions.