As anyone looked into the RPS : Roller Pinion System ? In surfing the 'net for a better mouse trap, I found the RPS "rack and pinion" as offered by Nexen (and others). A roller pinion system is a linear drive approach that combines the best features of ball screws, rack and pinion, belt drives and linear motors while eliminating many of their short comings. It sounds like a dream drive system for our OXs and beyond ... Read the write up at http://www.designworldonline.com/roller-pinions-another-way-roll-line for some details (from Nexen). Pricing, although they keep saying it ends up costing less, might be an issue ... who factors in long term repairs, maintenance, etc. anyways ?
There was a pair of Nexen RPS pinions on sale a while back on eBay (I can't seem to get back to it). Seller was asking something like 150$ for the pair, claiming they usually go for as much as 800$ each !! I can't find any other pricing info (quick Google'ing). Did flash clicked to a UK site which seem to have pricing, but I can't find the URL - not even in my cache. Was I in a dream state ?
That's the one . Doing an eBay search for "Nexen roller", not "Nexen RPS" that I was doing (!), gives a few other active listings. It seems to be for much larger models (25+mm bores) with prices near 200$ and up per pinion (!) while industrial suppliers list same Nexen pinions for over 3 times the price (last price updated of over a year ago as well). Not sure where the eBay guys get theirs to drop the price by so much ... No eBay results for "Nexen rack" ... I guess I was dreaming.
The seller notes he can provide a .dxf for the track on a forum post. And as the pinions didn't sell, they may still be available. As far as the low prices from some ebay sellers, you'll note that all of these are used. They were probably discards from a system upgrade where these people worked. A lot of industry to industry businesses have no methods in place to resell small parts like this and the cost of doing it would exceed the value returned.
I don't see this as a viable OX upgrade. It sure is cool though. Maybe for another machine. Good find Serge
I was looking into this the other day. Any opinions on DIY-ability? It seems like a perfect drive system that isn't overly complex and doesn't rely on industrial measurement abilities (as opposed to ballscrews). Creating the "gear" outer flange forms and the wave-pattern rack seem to be fairly straightforward, even from theory alone. The question is, how difficult are those tiny needle bearings to either make or buy? McMaster-Carr has 8mm ID-15mm OD bearings, which would make this huge. With a supplier for the small ones, say 3-5mm OD, this could be a viable scratch-build. Then again, cutting the larger forms might result in a higher-tolerance build, assuming fixed machine precision? It might require a couple tracks out of phase with each other to provide good traction and repeatability on each axis though, at least for higher-precision, higher-power builds. You could even use the resulting machine to upgrade its own drive system, though creating steel track on an alu machine and then hardening it (and then maybe lightly touching it up for heat deformation if you have some nice carbide cutters) would be a slow process. But even all that could be worth it if the alternative is price-parity with middle-range ballscrews, certainly for larger machines. Even more so if the parts are rarely even on the market any more.
That's a lot to unpack Rob. It's def. doable. I think it isn't worth the trouble. A constant pressure rack and pinnion system like the one used on the strong OX and cncrouterparts machines should be able to achieve the same results with less components. But hey, if you're looking for a challenge then that's all you buddy. keep us updated.
My concerns with constant pressure/preloading systems are instantaneous/shock loading and excessive wear. I suppose it depends on the setup and the materials being milled, and really these are considerations for higher-end machines later down the road, but if it's possible to do at least one aspect of the machine "perfect" off the bat, it seems to make sense to me. The other question with rack and pinion drives in general, though, is how to downgear them to get higher precision without introducing backlash into the system. The manufacturer's site said these units are capable of 20 micron repeatability, but that's not gonna happen straight off a 200-position stepper! If you have to pay $500 for a set of these or $500 for a set of decent ballscrews, it seems logical to go with the ballscrews and not have to worry about precision. At the lower cost end, instead of rack and pinion, it seems like a multi-stage chain drive would be a better option. I'm really just absorbing and analysing for the time being. There's a good chance I'll end up making some kind of more basic machine in the shorter term, but I have grander aspirations than that
Well, got some basic quantity one pricing from State side and it is somewhat high (for my budget) : RPS 16 Versa rack 1 meter section # 966860 ($172.00 US each) and value pinion # 966826 ($351.00 US each). This is there basic version - not the 20 micron high precision version. That translates to about 2100$us, at their door for a typical setup. The pinion are the expensive part of this system. They are also bulky and heavy ... as Rob Taylor pointed out earlier. However, they seem to claim zero backlash (less than the 3.5 micron apparently accepted as 'zero' by the industry), low noise, low to no maintenance, etc. As Rob Taylor points out, the steppers will probably need to be upgraded as well to move these things with such precision anyway, not to mention the added weight. I was looking at the specs and each pinion could reach over 1 kg on their own! Still, someone might be interested in making their own as apparently precision of the parts is not as critical with such a design ... What do I know, anyways ?
Yeah. If I'm gonna pay $500 a unit, I'm going with ballscrews and dealing with any whiplash as necessary. Obviously, I assume most of us on this forum don't have $500 an axis to throw around. The zero backlash may be due to the constant contact of the roller system- probably three rollers are in contact with the rack at all times, depending on pitch, each with a different- and self-cancelling- pressure angle. It definitely has value in that way. The drive system is my main concern, but it may be that this is simply a rack and pinion upgrade system, and not something that could viably be used as an alternative to ballscrews. It would probably be possible to 3D print a basic functional version for iterative testing, maybe someone with a 3D printer could try that. Shouldn't a 3D printer be able to lay down the wave form smoothly as an outer wall, without filling bumps? Seems like it might work well for light duty work just in plastic. Interesting that precision isn't critical... Got a source on that? I'm still generally interested in this drive system for certain purposes, maybe a large format laser or something.
Lots of documentation on Nexen's web site on the advantages and nearly no disadvantages ... according to them at least. The installation guide is probably a good one to look at. The whitepaper on the principle is here. Indeed, the way the rollers contact the rack makes for the near zero backlash. Its ...from memory, with a touch of frenglish... a cycloid idea : the track tooth follow the pinion's roller with little need for mechanical clearance, ... It results in less friction, less noise, less wear, ..., less backlash, etc. Their FAQ is here. My mistake, there is precision involved in the installation but comments are that installation "is easier than alternatives". However, rack segment alignement error tend to cancel each other over "infinity" : "To transfer the positional accuracy to the following rack segments, a special tool was developed that uses the rack tooth profile to set the rack spacing rather than simply butting racks end to end. The alignment tool uses two teeth on each rack averaging out some of the transferred error. This small transferred error is ± each time and statistically trends to zero, so an infinitely long run would, in theory, have zero cumulative error. This allows runs of virtually any length without loss of system accuracy. In many cases, a linear encoder is no longer needed and the servo encoder can be relied on reducing cost and complexity." Page 4 of WhitePaper (link is at beginning of this post) The track could probably be done 'easily' by someone in the know with the right tools. The pinion looks tricker to get done (at least for me). At price received to date, yes, the RPS are out of many's budget at our scale of machines. The quoted 'Versa' are plastic (of some sort). It's the low end / entry level with still many of the advantages (except the very high precision) : "Versa Rack is the ultimate in plastic rack and pinion. With backlash less than 3.2 um and positional accuracy of ± 500 or 580 um depending on pinion selection. The Value pinion is ideal for use with the Versa rack giving the lowest cost system or higher performance can be obtained with any of the premium pinions. Versa rack is great in dirty environments, clean rooms, or food/pharmaceutical production since it can operate without lubrication." Their high end is supposedly able to move tons with high precision. Their very high precision (5 micron repeatability) is for "lighter" loads ... Both, I suspect, are at (much) higher costs. The Versa was suggested to me by the "Canadian" rep after basic info from me by email. If cost is close to ballscrew alternatives, it may be worth looking into for overall advantages offered by RPS (?) Just don't shoot the Messenger.
Yeah, the non-friction non-backlash cycloid theory is as far as I'd gotten other than some basic pricing and the installation stuff, so thanks for the extra links! I wonder if Delrin (acetal?) would be a good candidate for manufacturing from scratch? It would be relatively straightforward to manually mill (perhaps with a jig) the wave form, and the rollers wouldn't have to be steel bearings. It could be very cheap. It may even be possible to use the cycloid action itself to mill its own guideway, thinking about it. If you replace the bearings with spinning mill bits in a roller of the appropriate size. I think personally I'd use it in conjunction with linear encoding still- not that I disagree with them and think rotary encoding or step counting would necessarily be inferior, but that's just the direction I want to aim at in the evolution of my machines. Not that any of this matters for the time being, but I do like a good thought experiment, even while I can't build.
Indeed, Indeed, just don't over think ... As I tend to do, otherwise not much get's done to the ideas.
