Resolution and repeatability measurements, what's important?

[DIY-Guy]

Hello,
Newbie here, nice to meet you all!
Assume a gantry style CNC router/spindle sheet cutting type of machine for this example.

What is better for sheet cutting (woodworking and plastics)?
A. "Positional accuracy" of +/- .002"
OR
B. "Repeatability" of +/- .1 mm (0.0039")

Do the two descriptive terms mean the same thing?

Is +/- .002" really going to make a difference in woodworking?

If the tradeoff were between a 7 times gain in speed using .0039" repeatability, compared to gaining a positional accuracy of .002" but dropping down to 47 inches travel per second, which would you choose and why?

Is there a type of woodworking or plastic (sheet) product that requires the greater accuracy and cannot be made properly or sold with
the less accurate tolerances?

Thanks everybody!

DIY-Guy

[ger21]

Originally Posted by DIY-Guy What is better for sheet cutting (woodworking and plastics)? A. "Positional accuracy" of +/- .002" OR B. "Repeatability" of +/- .1 mm (0.0039")

Is the positional accuracy per foot, or over the entire machine?

Originally Posted by DIY-Guy Is +/- .002" really going to make a difference in woodworking?

Probably not, but you're the one that needs to answer that. I see guys all the time say that 1/32" is close enough.

Originally Posted by DIY-Guy If the tradeoff were between a 7 times gain in speed using .0039" repeatability, compared to gaining a positional accuracy of .002" but dropping down to 47 inches travel per second, which would you choose and why?

Do you mean 47 inches per MINUTE? If so, then I'd go for the 7x faster, as 47 inches per minute is really too slow for cutting wood.

[toastydeath]

Originally Posted by DIY-Guy Hello, Newbie here, nice to meet you all! Assume a gantry style CNC router/spindle sheet cutting type of machine for this example. What is better for sheet cutting (woodworking and plastics)? A. "Positional accuracy" of +/- .002" OR B. "Repeatability" of +/- .1 mm (0.0039") Do the two descriptive terms mean the same thing? Is +/- .002" really going to make a difference in woodworking? If the tradeoff were between a 7 times gain in speed using .0039" repeatability, compared to gaining a positional accuracy of .002" but dropping down to 47 inches travel per second, which would you choose and why? Is there a type of woodworking or plastic (sheet) product that requires the greater accuracy and cannot be made properly or sold with the less accurate tolerances? Thanks everybody! DIY-Guy

Positional accuracy refers to absolute accuracy in some range, either by some increment or over the entire machine envelope. You tell it to go to 1,1. Is it at true 1,1? If not, how far off is it? That's positional accuracy. Note, this is a static measurement and doesn't look at a moving machine/profiling accuracy.

Repeatability is how consistently the machine goes to the same spot. If you tell it to go to 1,1 twenty times, and move it away each time, how much variation in position at 1,1 do you have?

I'd take a better repeatability over absolute accuracy every time. I can easily compensate a worn out machine that's repeatable, but I can't do anything with a machine that goes to a new spot every time I try to get it back to some point.

[Mariss Freimanis]

I'm baffled how repeatability can be worse than accuracy. That makes no sense when you think about it.

Mariss

[ger21]

Originally Posted by Mariss Freimanis I'm baffled how repeatability can be worse than accuracy. That makes no sense when you think about it. Mariss

I thought the same thing, unless he's talking about two different machines.

[DIY-Guy]

Originally Posted by ger21 I thought the same thing, unless he's talking about two different machines.

Yes, two different machines. Thank you Gerry.
Just to clarify, the first specs are information from a custom supplier. I am trying to understand the two different terms, and their relationships to performance. Assume rails and actuators for the X and Y portions of a gantry style wood/plastic sheet cutter for the first example "A."

A. One set of X,Y rails and actuators are specified at +/- .0039 inch "repeatability" with a free traverse maximum speed of 47 inches per second. And I guess that with inertia and brakeing a lot of that number will drop down. "Repeatability" sounds like "positional accuracy" to me, but speaking as a newbie, all the terms seem new to me.

B. The other specs I'm using for comparison is the http://shopbottools.com/prSalpha.htm
which shows these specs on their website-
● Cutting speeds of up to 600 inches per minute (depending
on cutting bit and material) and rapid transit speeds of
1,500 inches per minute
● Step resolution of .0004"
● Positional accuracy of +/- .002"

The term "rapid transit speeds" used by ShopBot, sounds to me like the "maximum speed" for free traverse which is used by the custom supplier.

And if the "repeatability" spec of .0039 inches is the same concept as what ShopBot is talking about when they say "positional accuracy," I think I'd rather have the potential 47 inches per second that the custom supplier offered me in example "A."

Help me out please, I'm just a computer guy who's just starting to look at specs for off the shelf vs DIY components. Thank you!

[toastydeath]

Repeatability and accuracy are not the same measurement. Never have been. I would be shocked if someone who was in the machine tool business confused the two. But if prompted, I wouldn't place a wager on it.

Cutting speed is the maximum speed the machine can keep track of the tool. Beyond that, it has to open up the allowable error.

Rapid/Maximum speed/transit speed are all terms for the very fastest the machine can go while only keeping loose reference to where the tool is in transit. This does not take into accel at all; many inexpensive machines claim large rapid speeds. In reality, they never get there because the accel/decel ramping is too slow. A 25 inch per second (1500 IPM, ipm is the more standard measurement) machine with a higher acceleration is far, far more useful for multiple reasons (tool wear, cycle time, chatter) than having a machine that can rapid at 47 inches per second but never actually gets there. But you'd need to know the accel value to make that determination.

Step resolution of .0004" is crappy for a machine tool built to cut metal. I don't cut wood or use routers, so I'm not sure what the standard is for that. It's probably fine, considering that it's wood and all.

[DIY-Guy]

Originally Posted by toastydeath Repeatability and accuracy are not the same measurement. Never have been. I would be shocked if someone who was in the machine tool business confused the two. But if prompted, I wouldn't place a wager on it. Cutting speed is the maximum speed the machine can keep track of the tool. Beyond that, it has to open up the allowable error. Rapid/Maximum speed/transit speed are all terms for the very fastest the machine can go while only keeping loose reference to where the tool is in transit. This does not take into accel at all; many inexpensive machines claim large rapid speeds. In reality, they never get there because the accel/decel ramping is too slow. A 25 inch per second (1500 IPM, ipm is the more standard measurement) machine with a higher acceleration is far, far more useful for multiple reasons (tool wear, cycle time, chatter) than having a machine that can rapid at 47 inches per second but never actually gets there. But you'd need to know the accel value to make that determination. Step resolution of .0004" is crappy for a machine tool built to cut metal. I don't cut wood or use routers, so I'm not sure what the standard is for that. It's probably fine, considering that it's wood and all.

Toasty, thank you for expounding. Sometimes suppliers get so used to dealing with experienced buyers they forget how to talk to a newbie. That's why I came here to cnczone.com and you, Toasty, came through for a stranger. Good job!

I found this definition for "repeatability" online.
"Repeatability:
The deviation in actual final position when repeatedly instructing a device to move to a target position, approaching from the same direction every time."
http://www.zaber.com/applications/?tab=Tech%20Notes
Now I'll add that page to my bookmarks file.

So is maximum speed a function of maximum acceleration? This linear axis unit with (47 ips) 1.2 meter per second maximum speed, has a maximum acceleration of 6 m/s2.

DIY-Guy

P.S. You've also taught me that I need to phrase specifications in terms of IPM (Inches Per Minute) to use standard terminology.

[toastydeath]

Originally Posted by DIY-Guy So is maximum speed a function of maximum acceleration? This linear axis unit with (47 ips) 1.2 meter per second maximum speed, has a maximum acceleration of 6 m/s2.

That's pretty decent. 2820 IPM, .6g accel. Takes about 3 1/2 seconds to go from still to moving full tilt.

Max speed is not a function of maximum accel, they're independent stats. A little like repeatability vs. accuracy. Related, but independent.

[GregSr]

Hello. This is my first post. I've been in the PCB drilling and routing business for 28 years. I started out in field service and wound up as a senior apps engineer.

Anyway, a good analogy for comparing repeatability and accuracy is to think of being at a shooting range. You fire 10 shots at the bullseye and all ten of them hit the upper left quadrant in a tight pattern but far from the center of the target. This would be described as having good repeatability and poor accuracy.

Until a CNC machine has good repeatability, its accuracy can not be dialed in.

I look forward to dropping by this forum regularly to see what interesting discussions are going on.

Take care,
Greg Sr.

[Mariss Freimanis]

Ah,.. 1G of acceleration is 23,040 IPM per second squared so 0.6G is 13,824 IPM / sec^2.
Time to speed is velocity divided by acceleration so 2830 IPM / 13,384 IPM / sec^2 works out to just over 1/5 of a second (204.7 milliseconds), not 3.5 seconds.

Mariss

[toastydeath]

23,040 in / 12 = 1920 ft/sec

That's a little faster than 32.2 ft/sec^2. Go look your units up, you did the math as though we were on Jupiter.

1g = 386.09 in/sec^2

Edit:

I'm retarded, disregard this.