[jharvey407]

I need some input on micro stepping vs. gear reduction.

I am in the process of building a CNC router. This project started as a tool for me to use in my marine electronics business, but the interest from other contractors and carpenters has been so great that I will be running a job shop.

I know that for commercial work buying a turnkey CNC system would be better than building one, but with the economy the way it is and my business doing so poorly, buying a CNC is simply not an option.

So my goal is to upgrade my existing prototype so that I can do light commercial work. Accuracy and repeatability are my main goals with speed being second. To be honest, if I can bill an average of 2 hours a day it will completely turn my business around.

Here is my CNC machine as it stands now.



The motion control is via Keling KL23H286-20-8B 425 oz. stepper motors and KL-4030 drivers. I am currently using the micro stepping to give the X and Y axis 1420 spi. I am using #25 roller chain for the drive system but I am looking to upgrade to rack and pinion for better accuracy when I build the permanent 80/20 table.

My question is what are are the pros and cons of micro stepping vs. gear reduction for increasing stepper motor resolution?

I have seen some gear reduction transmissions such as the Motiontek GR-231010H as well as some timing belt gear reduction setups, and I am wondering if they are worth the time and money, or if simply micro stepping is the way to go.

Please feel free to post opinions, experience and links to other threads.

Thanks
James

[dertsap]

if your motors have enough torque then micro stepping is the way to go as long as you can get the right resolution , gearing it down will gain torque and resolution but will also greatly reduce your speed

[jharvey407]

if your motors have enough torque then micro stepping is the way to go as long as you can get the right resolution , gearing it down will gain torque and resolution but will also greatly reduce your speed

Thanks, that's the type of information I was looking for.

So there is no problem with microstepping? It does not degrade the accuracy of the machine or lead to other problems?

It sounds like I could use a combination of the two. Maybe some gear reduction to increase torque but not enough to seriously reduce the speed, and micro stepping to acheive the resolution that I need.

[dertsap]

it won't degrade the accuracy by any means , micro stepping will run the motors much more smoothly than running full step
rack and pinion will run fast so gearing it down some should help with gaining torque to get the acceleration up to a better point (if the motors are powerfull enough on their own) , running at a low acceleration point will be a pita when running any 2.5 or 3d stuff because the motors cant reach max speed quick enough so the feeds won't be running at the top end . most likely a combination of gearing and micro stepping will be advantageous

[Torsten]

Microstepping for precision has deminishing returns.
For example without microstepping one step may result in 0.001 inch of machine movement.
When microstepping X10 one step would theoreticly be 1/10 of of the previous distance resulting in 0.0001 of movement.
In reality the devisions archived be breaking up the original step are far less accurate where a Half step X2 would be expected to represent a 50 percent position between full steps it may only be 30% or up to 70%.
For precision your actual devisions become increasingly compromised as you go past halfstepping.
When calculating the design resolution of your System you will need to know what accuracy your Machine is mechanicly capable of.
If your machine is capable to repeat to within 0.010 inch you would want your system to be able to get a precise resolution using no more then halfstepping of at least about 0.005 inches or as high as about 0.001 inch.
You want about twice the mechanical or up to 10 times the resolution.
Now you may run this at any microstepping you want for best performance of your machine there are benefits at different settings, just be clear your not doing it to get better precision.
Mechanical gear ratios affect the resolution more directly, each having a smaller negative affect such as backlash and friction, if you need one pick one that gives a lot of range such as timming Pulleys.
No Mechanical gear is best but one is better then two acting on top of each other.
Finding a good compromise between the Range of rpm's used on your machine and the best torque-curve for the motors at the right resolution is key here.
Good Luck

[ger21]

if your motors have enough torque then micro stepping is the way to go as long as you can get the right resolution , gearing it down will gain torque and resolution but will also greatly reduce your speed

I'm going to disagree a bit, and say you need both. Most rack and pinion users are using 3:1 reduction and still getting 600-1200ipm. Using the reduction should get you quicker acceleration, which is very important.

And even with 3:1 gearing, you're still getting about 1" of travel per motor revolution.

But you do want to use microstepping as well, as it gives you smoother running motors.

[jharvey407]

It sounds like I am on the right track. Here is what I have right now.

I am running 425 oz. steppers and 1/16 microstep. #25 chain drive with a 9 tooth sprocket for 1422.22 spi.

I am capable of rapids of 750 ipm, but have been running 500 because it is more than adequate for my table of 32" x 54". The weakness seems to be the torque of the motors. I ran a cut test with a 1/4" carbide 0-flute endmill, .125" depth of cut in 6061 plate. Running the spindle at 27500 rpm I started at 10 ipm and ran up to 100 ipm in 10 ipm increments. The cut improved markedly up to 50 ipm, but then the tool path started to wander and I got the impression that the motors were stalling or missing steps.

Perhaps if I sacrifice some speed for torque I could push the endmill a little faster. The specs for the endmill specify 120 ipm at 18,000 rpm and .25 depth of cut. The endmill is rated to 35,000 rpm.

I'm not after ultimate speed, my goal is more a balance of accuracy and repeatability and speed. I need to have a reliable machine for job work.

I guess I need to do some math and figure out a good compromise.

[jharvey407]

I'm going to disagree a bit, and say you need both. Most rack and pinion users are using 3:1 reduction and still getting 600-1200ipm. Using the reduction should get you quicker acceleration, which is very important.

And even with 3:1 gearing, you're still getting about 1" of travel per motor revolution.

But you do want to use microstepping as well, as it gives you smoother running motors.

Is there a particular reason that they are using 3:1?

I'll have to do the math, but 1" per revolution? What pitch rack is that?

Could you guys check my math on this and make sure I understand this concept.

Stepper - 200 steps per rev.
Micro steps - none (for clarity)

Rack pitch 20 tpi (0.05)
Gear teeth 10
Gear reduction 3:1
Inches per rev 1.5
SPI 133.33

Thanks,
James

[ger21]

191 steps per inch with no microstepping.
CNCRouterParts

[jharvey407]

191 steps per inch with no microstepping.
CNCRouterParts

I've taken a look at this and I like the idea. My drive system is set up just a bit different so I'll have to engineer a system of my own, but I feel like I am on the right track.

Based on all of the comments above it seems that I should be shooting for 2000 spi using 1/2 micro stepping, but it is going to take some testing to find the best amount of micro stepping.

[mike_Kilroy]

James,

looks like you are ready for some REAL what-if figurin instead of just guessing.

Most of the rules of thumb and suggestions given so far are great help for you to understand the pros and cons of chain vs rack/pinion vs direct drive vs microstep vs gearing, and I would say about everything said is ok data - no glaring errors so you should take all these good comments into account.

Now go put all this data into a good free motor sizing program and let it show you which is really best for your needs with the hardware you have available.

Our motioneering motor sizing program lets you put in all your mechanical and motion movement data and see how it fits together. Pick rack/pinion design, save it, pick chain design, save it, play what-if with changes adding gearing, removing it, etc.

http://kollmorgen.com/website/com/en...tioneering.php

good luck!

[ger21]

Using 1/2 stepping would only be about 400 steps inch. Where are you getting 2000?

200 steps per revolution, 1 inch per revolution.

[jharvey407]

Using 1/2 stepping would only be about 400 steps inch. Where are you getting 2000?

Not where... how.

The question is how do I get 2000 spi using a combination of gear reduction and micro stepping to acheive a high level of resolution.

[ger21]

Use drives with 10x microstepping.

[jharvey407]

Use drives with 10x microstepping.

I'm trying to design a system around the components I already own.

I'm on the right track now I just need to design and test it.

[Diane-MooreGear]

Is there a particular reason that they are using 3:1?

I'll have to do the math, but 1" per revolution? What pitch rack is that?

Could you guys check my math on this and make sure I understand this concept.

Stepper - 200 steps per rev.
Micro steps - none (for clarity)

Rack pitch 20 tpi (0.05)
Gear teeth 10
Gear reduction 3:1
Inches per rev 1.5
SPI 133.33

Thanks,
James

The rack used for most CNC R&P tables are not 20tpi, but 20DP which is pi/20 = .15708"CP or Circular pitch, which is distance from same spot on tooth to same spot on tooth.

20 teeth on a 20DP gear has a 1.000" pitch diameter. One rotation gets you 1 times pi in distance.

I hope this helps.

[email protected]

[jharvey407]

The rack used for most CNC R&P tables are not 20tpi, but 20DP which is pi/20 = .15708"CP or Circular pitch, which is distance from same spot on tooth to same spot on tooth.

20 teeth on a 20DP gear has a 1.000" pitch diameter. One rotation gets you 1 times pi in distance.

I hope this helps.

[email protected]

That does help, thanks Diane.

[Torchhead]

1. Microstepping should not be counted on for resolution. At higher RPM the effect gets canceled out and some drives even automatically change to 1/2 or even full step at higher RPM in order to increase the torque curve. It is absolutely essential for smoothness and reducing mechanical resonance.

2. The final drive topology should dictate the step down ratio. Your goal is to get the optimum torque at the cutting speeds you expect. If you use R&P (or chain for that matter) the transmission is such that you increase speed and decrease torque and resolution. That is always the case with a transmission. You trade one for the other. With a chain drive your speed multiplier is the PI times the diameter of the final drive pulley. Lets say it's 2" in diameter. It offers then 6.28 times speed increase and the same amount of torque and resolution decrease. Even if you don't use the speed you have already made the trade. Since microstepping is not a "given" you can't really say your "true" resolution on your current system is 16 times the raw resolution of the motor and final drive combo.

The 3:1 belt reduction moves things in the other way and trades speed (even speed you are not using) for torque and resolution. For the example above a 6:1 would work better simply because it cancels out the final drive pulley effect. A 3:1 is a good fit for something like a 1" pinion.

In routing applications torque and resolution (including smoothness of motion) is important.

Even with the 3:1 ratio and your steppers, the pack & pinion you will still be able to hit 500+ rapids but have a lot more torque so you can make deeper faster cuts.

TOM caudle
www.CandCNC.com

[jharvey407]

Thanks for the explanation TH. This really gave me a few things to think about. Right now the router is working well enough for my current application. This allows me to take the time necessary to properly engineer the R&P upgrade.

I am going to do some tests with the Y axis. I've ordered a rack so that I can start the design of the new system.

Your explanation is going to allow me to calculate the best gear setup based on the torque curve and cutting speeds.

1. Microstepping should not be counted on for resolution. At higher RPM the effect gets canceled out and some drives even automatically change to 1/2 or even full step at higher RPM in order to increase the torque curve. It is absolutely essential for smoothness and reducing mechanical resonance.

2. The final drive topology should dictate the step down ratio. Your goal is to get the optimum torque at the cutting speeds you expect. If you use R&P (or chain for that matter) the transmission is such that you increase speed and decrease torque and resolution. That is always the case with a transmission. You trade one for the other. With a chain drive your speed multiplier is the PI times the diameter of the final drive pulley. Lets say it's 2" in diameter. It offers then 6.28 times speed increase and the same amount of torque and resolution decrease. Even if you don't use the speed you have already made the trade. Since microstepping is not a "given" you can't really say your "true" resolution on your current system is 16 times the raw resolution of the motor and final drive combo.

The 3:1 belt reduction moves things in the other way and trades speed (even speed you are not using) for torque and resolution. For the example above a 6:1 would work better simply because it cancels out the final drive pulley effect. A 3:1 is a good fit for something like a 1" pinion.

In routing applications torque and resolution (including smoothness of motion) is important.

Even with the 3:1 ratio and your steppers, the pack & pinion you will still be able to hit 500+ rapids but have a lot more torque so you can make deeper faster cuts.

TOM caudle
www.CandCNC.com

[jharvey407]

The new drive system is installed and working well. I used the Moore Gear racks and the 3:1 gear drive system from CNC Router Parts.

Now I have a choice of 8x micro stepping and about 1500 spi, or 16x micro stepping and about 3000 spi.

I am leaning toward 8x.

Any comments or suggestions?

Thanks,
James