ball screw vs rack and pinion?

[allen gohl]

Hi all, I'm looking at getting my first cnc machine and am debating over the shopsabre 23 or a camaster stinger. The biggest difference that concerns me is one has rack and pinion on the x/y axes and the other has ball screw on all 3 axes. I'm trying to figure out if it's worth it to pay more for ball screws on all axes and sacrifice table length for it.
My number 1 priority is precision, followed very closely by table length.
I've heard that ball screw are much more accurate than rack and pinion, while rack and pinion are more stable at longer lengths, which leads me to 2 questions:
1. Has anyone using a rack and pinion noticced any significant amount of play that would throw off the accuracy of say, precision joinery?
2. How long is too long for a ball screw before it starts getting vibrations/whipping?

[Michael Burnside]

Both are excellent machines. The two machines you listed both have a lot of great features in common, things that I believe are more important than XY movement technology. Both will be very accurate (sub 0.001 and in my experience realized on repeat milling and seeing it for myself - XY and Z). Wood, if that's your main use, will have more variation moving than the machine

I love my ShopSabre, but Camaster was a close second. I was not interested in aluminum extruded versions as my primary use is hardwoods and furniture design. Mill a chunk of 8/4 Padauk, Brazilian Walnut or Bloodwood and you'll see why high-end CNCs use steel. I just finished a 6/4 white oak coffee table with custom joinery, milled wooden feet of my own design and two custom drawer faces that mirror each other. There was zero play in the sides/top, the legs identical, milled from separate 6/4 6"x18" pieces and drawer fronts perfectly aligned. I built to dimension and hand sanded so the drawers had 1/16" gap because they were that precise. Either of these machines are capable of this. Anyone else that says differently is selling something.

Whip on the 23 is irrelevant, it won't exist. Even on a 10' machine it will not affect accuracy, unless the steel is cheap along with a cheaply made machine. Both machines spec 0.001 repeatability and I would not hesitate to purchase either. I felt both companies were great to talk with, both believe their machines are the best (expected) and both can fully support your machine for a long time and finally are made with quality components. My only complaint with either is that Camaster took longer (a lot longer) to respond to questions and ShopSabre's first sales guy was trying way too hard

[Ronald Blue]

Don't compare the rack and pinion on CAMaster with the belt drive found on some low end machines. They run gear reduction and a timing belt like found on mane vehicle engines. The accuracy is not compromised in the slightest. As Michael said the two machines are very comparable. Well made and dependable. Used ones come up for sale and in the majority of the time they are upgrading to a larger and/or faster machine often with ATC and a vacuum setup. They were unsure if CNC was for them, but once they got going realized how much one could do with one. Then they wanted to maximize their capabilities.

[Bruce Page]

My Stinger has R&P on all three axis. I run accuracy tests on it periodically and have not seen any accuracy deviations in the 10 years since it was delivered. I would not make BS vs R&P a deciding factor.

[Jim Becker]

My Stinger II SR-44 is R&P on X and Y and ball screw on Z. Honestly...it doesn't matter and both the Camaster and Shopsabre machines are extremely well built and "Designed for Purpose". They are not designed and sold based on "attractive price" and are more industrial in that respect. And they are made here in the US, largely of US and North American sourced parts where that's practical and possible.

[Wes Grass]

There *ARE* precision ground R&P drives avaliable. Gernan made, AFAIK ... cost????

As for 'whip' of ballscrews at high feed rates, the solution is a *fixed* and tensioned screw with a rotating nut. If you're looking for a 'cost effective' solution... you don't wanna go there.

[allen gohl]

Thank you all for your input and advice. One other contendor in my search is a probotix. Not sure how I feel about a screwed together aluminum frame vs a welded steel frame, but does anyone have experience with a both a probotix and camaster/shopsabre to see how they compare? Rigidity, accuracy, max cut speed? I saw that a stinger 1 has cut speeds of up to 400 IPM whereas that's the Probotix's max rapids speed.
The one thing I liked about the probotix was the ability to add an auto tool change spindle, effectively bringing the cost up to that of a camaster or shop sabre. Any thoughts?

[Bruce Page]

I don’t have any knowledge of the probotix machine but it is unlikely that the bolt together platform is as ridged as a welded platform.
One thing to keep in mind is the Stinger, and likely the Shopsabre are plug and play, e.g., no additional hardware or software other than router bits are needed to start making parts.

[Bill George]

I had a R&P and did some really good work with it I wish I could find a picture to post. I'd get a CAMaster today IF I was in the market, again.

The attached was done on my 4x4 foot welded steel frame R&P machine .


RouterArt.JPG and IMG_4600.JPG

[Michael Burnside]

I considered Probotix. I found them when I was originally considering an AVID type setup. I loved the idea they assemble, test and validate the setup before sending it my way. IMHO puts them a cut way above AVID for the hobbyist. I also liked that they use LinuxCNC (if I recall correctly) which is a pretty flexible and comparable to WinCNC. The controller isn't as popular in the industrial space, but I've heard it is pretty good. All in, it was going to be a little cheaper (didn't consider ATC) but I went another route. Cost wasn't my main factor and I'm happy with my ShopSaber, no questions. I would certainly consider these over AVID. Nothing against AVID, they're fine machines, just can't compare to a welded steel frame. I don't remember if Probotix offered/used HSD spindles or not, but that's something I would investigate. The Mitsubishi VFD and HSD spindle I have on my ShopSaber is top notch. Also Vcarve Pro is included with ShopSabre and I believe CAMaster too to emphasize Bruce's point above.

As far as IPM goes, what are you going to be cutting? I typically push my machine when I cut hardwood for my furniture builds and target chiploads of 0.007-0.011" for 1/4" bits and a bit higher for larger. I vary my RPM to try and cut at the slowest speed possible while maintaining a nice cut (drives me nuts so many think fastest is best - not realizing heat is the enemy of bits). Recently I was an idiot and used the wrong setup and burned up 2 bits at high RPM...oops. Anyway, that usually puts me anywhere from 160-300 IPM. The Probotix would probably need to slow down on larger bits.

As far as ATC is concerned, I'm not a production shop so the cost/complexity/maintenance just wasn't something I needed/wanted. What I would get is a tool-length sensor. Those are way more useful than the XYZ zero plates, which I don't even use 99% of the time. The tool-length sensor is awesome...poor mans ATC. You put in a bit, hit the measure button, zero to machine bed or work piece and start cutting. When you change a cutter just hit the tool length button, remeasure and start cutting again. You NEVER need to use the "XYZ" zero plate ever again. For me, that is fast enough and it's ridiculously accurate (at least the one on my ShopSabre is).

[allen gohl]

I am hobbyist for now, but would like to step up to making small batch production runs, hence CNC. I will mainly be carving in maple, but occasionally in other various hardwoods (walnut, bloodwood, purple heart, wenge, etc). I don't plan on taking a finished piece right off the machine, there will be some details that will have to be done by hand as I will have sharp corners and also don't want to see any semblance of mill lines so I plan on finish sanding regardless. I mainly want the CNC to do about 95% of the work for me and to do so quickly and reliably. Ultimately though I don't know how realistic it is to cut hardwoods at 2-300 IPM without getting massive tearout.

Bearing that in mind, IF the material can handle those kinds of speeds, would a machine like the Probotix be able to handle it? Money is a factor for me, and I only want to buy a machine once so I want to get the best I can and don't mind saving up more for the right one, but I also don't want to get more machine than I will ever be able to make use of.

After reading some yall's and other comments on ATC, I don't think I'm going to go with it and instead just get the tool length sensor.

[Michael Burnside]

Keep in mind that IPM really has nothing to do with tearing. It’s all about chiploads, maybe a little with RPM, but mainly chipload. That is, how thick the material is you’re cutting for each flute. Bigger chips draw more heat away from the bit, but if you go to the max size chip recommended for a bit diameter you may see a little tearing. To reduce the chip size, spin a little faster. For hardwood and a 1/4” bit I try to go anywhere from 0.007 (smaller chip) for figured wood to 0.012 for general hardwoods. For softer woods/ply, it’s a little higher and for MDF even higher.

Basically you should cut at the slowest speed you can at the appropriate chipload. Usually for me that’s 12000 RPM and between 180 and 275 IPM. For softer woods and ply I spin faster and have higher IPM, but I’m watching my chipload.

For bigger bits beyond 1/4” you need more machine and this is where steel wins. It’s just a fact. The Probotix CNCs look like great machines and it’s probably the one I’d get if I went aluminum. IMHO it is a compromise to the other machines but that’s why it’s less expensive. Whatever you do, get a spindle. You’ll find it much easier to hit proper targets because you can push the IPM and lower the RPM without compromising torque too much.

[Gary Campbell]

Your questions:
1) neither option would have a perceivable difference to a new user doing woodworking type projects.
2) ballscrews whip when driven too fast. The cheaper the screw, the lower the rpm that this will happen.

Some other random facts:
1) the highest priced, most accurate CNC machines made will use both ballscrews and rack and pinion
2) rolled ballscrews are the cheapest way to achieve mechanical reduction on small platforms, i.e., I can buy decent ballscrews for less money than I can build a clone of the self adjusting belt reduction on the stinger or a buy a planetary gearbox
3) ballscrews offer more mechanical reduction. Usually ~2.5 to 5 motor turns to one inch of travel. Rack and pinion systems usually range from 1.25 to 2 turns per inch. More reduction is usually more resolution and less speed.
4) both systems are capable of delivering more torque to the cutting head than sub $20k frames can withstand
5) most of the people that have bad things to say about rack and pinion are selling ballscrews.
6) there are a number of differences between the 2 machines you mention that are more important than the screw/rack one you bring up
7) should you actually purchase either machine, none of the CNC related things that you labor over prepurchase will be important to you in a year

[Richard Coers]

Probotix used to be located here in Peoria, IL. My friend traded a table for one of their machines. In those days he bought anti-backlash nuts to move the axis in mulitples because they wore out so fast. He also had a lot of trouble with electronic noise from the VFD for the spindle and the controller. The longer the machine ran, the further out of round circles would get. He finally mounted the VFD 20' away and it was better. Their machines were just really cheap hobby machines in those days, no idea what they are making now