[davtuner]

Hey,

I built a Joes 4x4 hybrid last year and have had accuracy problems with it ever since. The thing cuts fine, except for getting things like interference fitting parts and through hole PCBs (cutting on both sides, when you have to flip the board over). It's got a repeatability (precision) of 0.001", but an accuracy of +-0.002"

I traced the inaccuracies to the racks, which were purchased from Moore Gears. There is a random local error of around 0.003" every 0.1". I was wondering if anyone is able to get accuracies of 0.005" per linear foot or less (like a leadscrew) with a rack and pinion set up, and if so, where did you buy your racks?

Thanks,
David

[ger21]

$100 acme screws are only accurate to ±.009"/ft. To get the accuracy you're looking for would be extremely expensive, if it's available.

[Al_The_Man]

You are going to be hard pressed to get the accuracy of a pre-loaded ball screw from a Rack and Pinion, even precision machined rack.
Al.

[Diane-MooreGear]

David,

I took your comment to our QA Manager since the "random" local error makes no sense, especially not in context with .003" within .1". If the error lay with the rack, it would not be random.

We take quality very seriously, so I'm always on the look out for ways to improve. If the problem truly lies with the rack, we'd like to know about it.


[email protected]
573-486-5415

Hey,

I built a Joes 4x4 hybrid last year and have had accuracy problems with it ever since. The thing cuts fine, except for getting things like interference fitting parts and through hole PCBs (cutting on both sides, when you have to flip the board over). It's got a repeatability (precision) of 0.001", but an accuracy of +-0.002"

I traced the inaccuracies to the racks, which were purchased from Moore Gears. There is a random local error of around 0.003" every 0.1". I was wondering if anyone is able to get accuracies of 0.005" per linear foot or less (like a leadscrew) with a rack and pinion set up, and if so, where did you buy your racks?

Thanks,
David

[harryn]

David,

I took your comment to our QA Manager since the "random" local error makes no sense, especially not in context with .003" within .1". If the error lay with the rack, it would not be random.

We take quality very seriously, so I'm always on the look out for ways to improve. If the problem truly lies with the rack, we'd like to know about it.


[email protected]
573-486-5415

Hi Diane,

Separate from this particular case, could you help us understand what is a reasonable expectation for the results from a rack and pinion arrangement ?

For example, as Ger21 points out, a lead screw is accurate to about 0.009 inches / foot. Ball screw results vary with price.

In Europe, I see Mod 1.5 rack precision quoted, but in the US, the typical rack appears to be sold without a spec. (example, McMaster and some others I have looked at)

It would be very interesting to know what is expected from a rack + pinion setup, example, 14.5 degree pressure angle, 20 pitch, 20 tooth pinion. I assume this depends on if the rack is machined or machined and ground, but perhaps you can help us understand what is reasonably feasible ?

Thanks

Harry

[thefu]

I am getting repeat ability and accuracy of .0004 with my rack and pinion system. I use Atlanta helical cut gears with servo and 10:1 reduction. my inaccuracies reside in the reduction gears i use.

[mactec54]

davtuner

A C5 grade Ballscrew is .0005 per foot, so it can be done quite easy with the right build, you can also get a ground or helical cut rack & pinion which will give you better than .001 per foot, but cost is more than a good ballscrew

[harryn]

I am getting repeat ability and accuracy of .0004 with my rack and pinion system. I use Atlanta helical cut gears with servo and 10:1 reduction. my inaccuracies reside in the reduction gears i use.

That sounds about right for Atlanta Helical cut gears - they are impressive, especially the hardened and ground ones. As they claim, they really can match ball screws in many applications.

Those German made rack and pinion setups are real craftsmanship.

I can't afford them for my DIY hobby router, but they sure are nice.

[thefu]

You can afford regular straight cut from atlanta...the helicals are pretty pricy though

I paid $240 per rack (60" long i believe) and $65 or so for each pinion (this if all helical stuff)

I have 4 pieces of rack total on my gantry table.

[harryn]

I will look into their offerings again. The last time I looked, I was perhaps over hot - rodding the specs to see what was possible. That put me into Mod 2.0 for sufficient tooth strength.

Rack, Module 2.0, Quenched & Tempered, L = 2010.62 mm with mounting holes: $507.00 each
Pinion, Module 2.0, Hardened & Ground, , d = 36.00 mm, plain bore: $262.00 each

Both, high precision versions.

36mm pinions put me into serious gear reduction, which was the real deal killer, about $ 1 K / each. Needed 5 sets, so a bit crazy for a DIY / non revenue hobby.

Now that I have reduced my build size and specs, it should be more reasonable.

[mactec54]

harryn

And these only have the spec accuracy if they are mounted on a machined surface

Mounting to a none machined surface as most do, will lose all accuracy on any rack install

[thefu]

this is correct. My rails are mounted on angle iron that has been ground flat on both outside and inside edges then tapped with a datum "ledge", .0001" flat over a 10 foot span. When i originally built this table i planned to use it for hidef plasma only but now that prices are more affordable im seriously considering a 2k fiber laser

[Diane-MooreGear]

Hi Diane,

Separate from this particular case, could you help us understand what is a reasonable expectation for the results from a rack and pinion arrangement ?

For example, as Ger21 points out, a lead screw is accurate to about 0.009 inches / foot. Ball screw results vary with price.

It would be very interesting to know what is expected from a rack + pinion setup, example, 14.5 degree pressure angle, 20 pitch, 20 tooth pinion. I assume this depends on if the rack is machined or machined and ground, but perhaps you can help us understand what is reasonably feasible ?

Thanks

Harry

Our rack is machined and results in AGMA Class 8 Specifications. While there is no current AGMA standard for racks, we refer to the pinion specs for acceptable tolerance. For 20DP-20PA, Tooth to tooth composite tolerance is within .001" with a total composite tolerance of .0019". Keep in mind, this is the gear rack ONLY and does not take into consideration any other factors such as pinion precision, mounting, etc.

If you need more precision, then your next option is ground rack.

Diane

[thefu]

moor gear makes a mighty fine rack. I would be interested in seeing your helical cut prices.

[Diane-MooreGear]

this is correct. My rails are mounted on angle iron that has been ground flat on both outside and inside edges then tapped with a datum "ledge", .0001" flat over a 10 foot span.

How do you even measure that?

[Diane-MooreGear]

moor gear makes a mighty fine rack. I would be interested in seeing your helical cut prices.

Thank you!

You'll have to get me the details of what you need since that would be a custom item.

[thefu]

lets just say i have a friend that owns a very large brown and sharpe gantry cmm

the racks i currently use are 2 module, 2000mm long per rail, 300 teeth per rack and are from atlanta P/N 29 25 200

total pitch error is < 0.030 mm
per meter

case hardened and teeth ground
16mncr5 material
ground on all sides after they hardened it


wasnt too bad to align the 2 pieces of rach either using their assembly kit.

[Taflin]

lets just say i have a friend that owns a very large brown and sharpe gantry cmm

the racks i currently use are 2 module, 2000mm long per rail, 300 teeth per rack and are from atlanta P/N 29 25 200

total pitch error is < 0.030 mm
per meter

case hardened and teeth ground
16mncr5 material
ground on all sides after they hardened it


wasnt too bad to align the 2 pieces of rach either using their assembly kit.

It is impossible to measure with certainty 0.0001 on a CMM over that size.

[thefu]

i beg to differ

PMM-G - Brown & Sharpe

[Taflin]

For example,

The rated uncertainty of the machine I use, 1.7+3L/1000 µm where L is in mm.

A machine of that size will likely have a much larger rated uncertainty.

Now calculating for the length you have provided 3048mm, and using the uncertainty from my machine yields a result of 0.0004.

4x under BEST POSSIBLE conditions what you claim to be measuring.

I call BS.