[tkubic]

My machine has ~1500mm travel in x and y, I currently am using McMaster rack gears. The tooth to tooth tolerances is unacceptable for doing any jobs for people. It would be easiest to replace the poor tolerance pieces in my system with higher tolerance gear racks. Though, I do have a design 95% done for 25mm 10mm pitch ballscrews.

My issue here is that I have only found one source that posts tooth to tooth tolerances (http://sdp-si.com/ss/pdf/81001194.pdf) , are there any other sources for rack gears that specify tolerances, or should I convert to the ballscrew design?


Any suggestions?

[frare bear]

My machine has ~1500mm travel in x and y, I currently am using McMaster rack gears. The tooth to tooth tolerances is unacceptable for doing any jobs for people. It would be easiest to replace the poor tolerance pieces in my system with higher tolerance gear racks. Though, I do have a design 95% done for 25mm 10mm pitch ballscrews.

My issue here is that I have only found one source that posts tooth to tooth tolerances (http://sdp-si.com/ss/pdf/81001194.pdf) , are there any other sources for rack gears that specify tolerances, or should I convert to the ballscrew design?


Any suggestions?

I just ordered a rack and pinion kit from cncrouterparts.com and I don't have any experience with ballscrews. That being said though I think these from the same site are designed to fit together very well with tight tolerances to maintain the precision of the machine overall. I don't have any numbers or references, but you can check them out in action on the videos there.

I'm just getting into the CNC game, so don't take anything I say too heavily. I just replied because it seems there aren't always posters to help out with questions.

[ger21]

Most r&p users seem to be using rack from Moore Gear, and I've never heard of the issues you claim to be having with the Moore racks.

[CarveOne]

My second CNC machine (4' x 6' built into a 4' x 8' wooden work table) originally had two low cost 72" racks from McMaster-Carr. Those racks had cutter burrs on top of each of the teeth. I used a fine cut mill file to surface plane the burrs down just enough to remove the burrs - but not lower the teeth. That significantly quieted the noise while traversing but they were still noisy. This was way back just after CNCRP released their first r&p drives. I didn't notice any accuracy problems at that time though.

After a few months I laid out a steel box tube frame of 5' x 12' size and ordered four racks from Dianne at Moore Gear. Moore Gear's racks had no burrs and appeared to be made of much higher quality steel and machining finish than the low budget ones from McMaster-Carr (that I still have in storage). Don't skimp on money and buy the low budget racks. McMaster-Carr sells much higher quality racks.

I recommend Moore Gear as a very good source for the racks. Look them up online and call them to ask any questions you have. They have some useful specifications and documentation on their site as well. These racks have caused no problems with noise or accuracy since installation. However, I don't do anything with this machine where being off by a couple of thousandths over 6" distance anywhere on the table is anything to worry about.

Ballscrews will give you better accuracy, but with lower speeds and shorter travel lengths. There's always trade-offs that depend on what you want the machine to be capable of.

[wendtmk]

I'll second Gerry's and CarvOne's Moore Gear recommendation. I got my racks from them too. Dianne has also posted here occasionally. I talked with her at length about the racks I wanted, and their customer service is outstanding! No problems with the racks at all, came well packaged and got to my address quickly.

Very nice company to do business with.

Mark

[tkubic]

Hmm, I purchased my rack gear from McMaster in January 2011, which was back when cncrouterparts had a link to McMaster for rack gear. I just checked McMaster again and didn't see any sort of precision rack gear listed, only two steel types, 14.5 degree and 20 degree pressure angle rack gears. When attaching a 12" set of calipers to the machine and jogging around I mapped out a very inaccurate rack that had both positive and negative error throughout travel even when using nominal steps per and using Mach's built in steps per calculation tool

I had looked at Moore gear's website prior to posting and couldn't find any specifications or specific product descriptions. The best I could find was from their product brochure that states they are cut on CNC machines and have "tight tolerances". That doesn't really help me though when spec'ing out the machine, I will send them an e-mail message and ask them what sort of tolerances they will make claims to. I am very hesitant to purchase anything from companies now days who don't list specifications and product brochures.

I would like to be within +-.010 in a 6" square anywhere on the table and within +-~.015" for a combined positioning accuracy and precision. Ballscrews will slow my current design down from ~1000 ipm to ~300-400 ipm, which honestly 1000 ipm is a lot faster than my machine ever gets up to during most rapids, I would doubt I'd see even a 20% increased machine time with the reduction.

Keeping my cncrouterparts rack and pinion setup would definitely be cheapest, I just really hope I can get the accuracy desired from it.

[wendtmk]

Give Dianne at Moore a call. She's quite knowledgeable, and should be able to answer all your questions.

Mark

[louieatienza]

Maybe your machine will run a bit slower in max rapids, but you'll gain about 2.5X mechanical resolution. Unless you run programs that have many position changes your cutting feedrates should be close to the same. And if you did have that many position changes you might have to reconsider toolpath strategies.

The chinese ballscrew manufacturers claim an accuracy of C7 or about .003"/ft. They also claim zero backlash. Still, even without backlash compensation that's 3x the accuracy you're getting now...

[tkubic]

I e-mailed Diane and got the following response for accuracy "Typical tooth to tooth error is .0010” with a total composite of .0019.” I have not figured out yet how to translate this into something understandable.

@louieatienza
The C7 accuracy spec on the chinese ballscrews is why I have been working the redesign, Though the $163 shipped quote for the rack gear from Moore is quite tempting over the estimated ~$650 it will likely cost to retrofit the X, and Y axis.

[CarveOne]

That specification makes sense because of the way errors tend to add up algebraically. The tooth to tooth number is "typical" for their racks, meaning that some will be less and some will be more. The actual measured values over the specified length tends to vary between 0.001" and just under 0.002". I have seen specs of +/- 0.009" per foot of length. The numbers Dianne gave you seem quite good to me.

That accuracy can be ruined by a pinion gear that bottoms out in the rack teeth for any reason. There should be a small gap at the bottom of the teeth when the rack and the pinion gear are fully meshed. If the fit of the pinion gear teeth allows bottoming out in the rack teeth you will have backlash problems and noisy operation.

[victorofga]

I think you can go both way..
I believe the ""accuracy"" same for both product.. for gears and for ballscrew..

I have 6 feet working envelope and it is exactly 6 feet theres no error .. with the C7 ballscrew..

just because of accuracy you don't have to choose either..

for ballscrew the advantage, you don't have to count on backlash.. the 0.015 mm backlash for wood or general parts will works.. this backlash will be present with gears too..

for length, since you can measure with tapemeasure only over a few couple of feet, then you never find if you machine makes shorter or longer an 8 feet part with 0.02 inch..

the technology for both product ballscrew and rack require very precise equipment ..
they will make the 8 feet to exact 8 feet..