Rack 'n' Pinion vs. Precision Rod?

[CnoEvil]

Have there been any discussions on here reguarding these two motion transmission methods in comparison to one another. I'll be starting a machine this year.... I believe, (sheesh... finally!) and really would like to get the most out of my x-axis, span-wise, but I understand that rack and pinion isn't very precise compared to a ball screw mechanism. Are there ways to enhance the precision of rack and pinion drives...or will I have to bite the bullet and suffer the expense of a long precision rod to get the accuracy?

Thanks,
Mark

[jcc3inc]

Sir,

You will need to ask the rack supplier what his accuracy specs are. We used to make height gauges using precision rack; it came in 11" lengths and we were shooting for .0005" accuracy or better. When using short sections of rack, you can correct for error every 11" and have an overall accuracy that is very good.

The commercial rack that we used on our cutting machines came in 6 foot lengths. I checked overall accuracy of a piece just for general information, and it was within .020" in 6 feet. We did not specify the accuracy however as the rack was used on plasma and oxyfuel machines that did not require milling machine accuracy.

Regards
Jack C.

[LeeWay]

For the cost difference, the rack @ .0033" per foot accuracy isn't too bad. Better than some rolled ball screws. There are other accuracy considerations with rack. That is how well the pinion stays in contact with it. Easily done in one direction, you have to design for both. Most use spring loaded designs to achieve this. I have seen some better than others.
Having a large diameter ground ball screw will likely get you the best accuracy, but you will pay dearly for those extra mils.

Some have been using smaller screws and rotating the nut instead to prevent whip. That might be another consideration for you.

[harryn]

Hi jcc3inc - you point out a very critical aspect of rack - how precisely it is made. Mostly, I see specs on pressure angle, etc, but very rarely for tolerances. The metric rack suppliers tend to include this info more often, and the racks they offer are sometimes much more expensive than inch version (like std steel), but I wonder if this is just a reflection of the improved machining or just more expensive. Example - is boston gear's rack better, or just a bit more money ? No idea.

Did you find any rack suppliers that offered inch based rack in improved precision ?

Thanks

Harry

Originally Posted by jcc3inc Sir, You will need to ask the rack supplier what his accuracy specs are. We used to make height gauges using precision rack; it came in 11" lengths and we were shooting for .0005" accuracy or better. When using short sections of rack, you can correct for error every 11" and have an overall accuracy that is very good. The commercial rack that we used on our cutting machines came in 6 foot lengths. I checked overall accuracy of a piece just for general information, and it was within .020" in 6 feet. We did not specify the accuracy however as the rack was used on plasma and oxyfuel machines that did not require milling machine accuracy. Regards Jack C.

[jcc3inc]

HarryN:

Sources of precision rack are, among others, Allied Devices, W. M. Berg Inc.,PIC Design.
Overall accuracy of 11" sections can be .0006", precision 3 designation. ($$$!!)

If I were to recommend an approach to making a machine, it would be to use a 6 foot length of rack from someone like Standard Structural Steel, determine its accuracy by either measurement or asking the supplier, and ask myself if it is possible to establish (1)perpendicularity of the X to Y to Z axes, (2) to maintain deflection due to loading, (3)
to precisely position the machine while in motion, and then see what the effect of the rack error is in relation to these other sources of error.

Another option is to use the error correction available with some CNC controllers to correct the resultant errors, if you are able to measure them.

Jack C.

[jpgdesigns]

The best rack and pinion:

http://www.nexengroup.com/nexen/index.jsp