dmm/dyn2/acorn/taig pulse settings

[nicolasjolliet]

Hi folks, I'm new here. I've been through the forums a lot and feel that I already know you guys
As a beginner as well as not being the sharpest cutter in the shop, I was trying to research as much as possible before posting obnoxious questions all over the place. But now I think it's time to bother the wizards of acorn a little with my broken English.

To the point:

I'm trying to set up jfreshly arrived a Taig mill and a taig lathe (ballscrews).
I went for acorns, dyn2 drivers and dmm 57n-dht motors.

I've searched all over and watched the precious Martyscncgarage videos and got to the conclusion that I should use the 2000/8000 pulse settings for Gear_num and pulse count on the Acorn.

Then I came across this bit of text somewhere on a forum (forgot where, sorry):

Electronic gear reduction parameter. GEAR_NUM x 4 = number of pulses needed to rotate motor 1 revolution. For example, if GEAR_NUM is set to 1,000, then 4,000 pulses are needed to rotate motor 1 revolution.

Generally, GEAR_NUM should be set as high as possible for best resolution and smoothness. But GEAR_NUM also affects max motor speed depending on CNC controller pulse frequency so this should be balanced with the speed requirement:

Example 1: Controller max pulse frequency = 100kHz.

Machine target speed = 500IPM.

Motor direct drive on 5mm pitch ball screw.

500inch/min / 5mm/rev = 2540rev/min from motor.

2540rpm = 42.33rev/second.

100kHz = 100,000pulse/sec / 42.33rev/sec = 2362pulse/rev.

So set GEAR_NUM to 590.

Note 1. The GEAR_NUM parameter is only used to calculate the command from the CNC controller. The servo drive and servo motor always maintains position at 16-bit (65,536points/rev) resolution

Note 2. GEAR_NUM setting should allow at least twice the resolution of machine accuracy target. See example 2 below.


Example 2. If motor is driving a rack and pinion axis with a 2:1 reduction and 25mm pinion diameter. Target accuracy = 0.05mm. 25mm diameter pinion = 157mm circumference (linear travel per pinion revolution). 157/2 = 78.5mm linear travel per motor revolution. Twice target accuracy = 0.025mm. 78.5/0.025=3,140points/rev. So set GEAR_NUM to at least 785.

The LINE_NUM is the same as the GEAR_NUM but it controls the resolution of the encoder OUTPUT. The GEAR_NUM is the pulse ratio of the pulse/dir command input from the controller. The LINE_NUM is the encoder output resolution from JP5 connector. The MASSO doesn't read the encoder from the drive so this parameter is not used.

Servo direction, if a servo rotation needs to be reversed (as in a slave situation), the Dir+, and Dir- can be reversed on one of the drives without damage. The Masso also has an option to do this in the Slave axis window.

So from that I thought I would calculate the best settings for my ballscrew taig, dmm 57n and acorn. And I came up with this:

My taig has 2.5mm pitch diameter ball screw, I want to target a rapid speed of 100 ipm on the Taigs.

2.5mm = 0.0984252 inch pitch = 10.16 rpi (rotation per inch)

rpi x steps = spi (steps per inch on driver)/4= steps per inch on acorn

What ipm max speed can dmm 57n motor take (rated at 3000rpm)?



3000 rpm / 10.16 rpi = 295 ipm (motors rated at 3000 rpm)

I need 100 ipm

100 ipm x 10.16 rpi = 1016 rpm (lots of head room to 295 ipm)



So the motors can take 100 ipm (our target) easy = 1016 rpm (torque is consistent up to 3000rpm)

What maximum steps per seconds can Acorn controller do?

MAX frequency is 400 kHz (you have to set this in the wizard, default is 200khz) so that means 400000 steps per seconds

(the pmw can’t be shorter than 0.8us on dyn2 driver, and from what I understand acorn will give us just a bit over 1us pmw at 400khz)


What steps per seconds do I need at 100ipm (1016rpm)?

1016 rpm = 16.93 rps (rotation per seconds)

16.93 rps x steps = sps (steps per second)

controller 400’000 sps (capable) / 16.93 rps = 23626 steps per rotation are available

motor max steps per rotation is 65’536 (16 bits encoder) but we are allowed only 23’626 from acorn at a speed of 100ipm.

the motor has 2.73 times more res than acorn is capable of.
That's why I can't use the full 65'536 the motor encoder could take.



To be safe, I'll set my encoder settings (pulse per rotation) in acorn at 20480 steps per rotation

on the dmm dyn2 driver (GEAR_NUM x 4) it will be /4 = 5120

and our rpi is 10.16

For more headroom but less precision I could go 4096 and 16384



Does this look correct? or do you see obvious mistakes?
I'm gonna try this and see how it goes, but I'd be happy to see what you guys think about this to help me troubleshoot.
I've read somewhere that when people go over 2000/8000 they get problems even though it should work.

Cheers.

[martyscncgarage]

Try it....I've been happy with 2000 gear/line numbers.
Let us know how you make out.

[Richards]

Congratulations on your new mill and lathe. I recently purchased a CNC ready Taig mill and a manual Taig lathe that I plan on converting to CNC with a different configuration than Taig offers.

You've done a lot of homework and thought things out. As Marty said, just try it out. See if it works for you with those settings.

I'm running my Taig mill with the DM542 stepper drivers set to 800 steps per revolution, which is the same resolution as the ClearPath servos that I'll install as soon as I buy and bore out the .25" Tiag motor connectors to .375" to fit the servos.

At 800 steps per revolution, the Taig's axis will move just 0.000123 inches. The Taig is advertised as being accurate to 0.0002 inches, so even at 800 steps per revolution, each step is about 2X finer than a well-tuned mill needs. Unfortunately, the only DMM servo I have is the NEMA 34 size 750 watt model, which is too large for the Taig, so I can't experiment with a DMM product on my Taig. On the test bench, I had that motor set at 500 pulses per revolution, which, with the quadrature encoder, became 2,000 on the Acorn's Wizard settings. Currently, that motor is tucked away on the shelf until I buy/build something large enough to require that size motor.

Good luck on your project. Keep us informed of your progress.

[nicolasjolliet]

I'm running my Taig mill with the DM542 stepper drivers set to 800 steps per revolution, which is the same resolution as the ClearPath servos that I'll install as soon as I buy and bore out the .25" Tiag motor connectors to .375" to fit the servos.

At 800 steps per revolution, the Taig's axis will move just 0.000123 inches. The Taig is advertised as being accurate to 0.0002 inches, so even at 800 steps per revolution, each step is about 2X finer than a well-tuned mill needs. Unfortunately, the only DMM servo I have is the NEMA 34 size 750 watt model, which is too large for the Taig, so I can't experiment with a DMM product on my Taig. On the test bench, I had that motor set at 500 pulses per

I didn’t know that only 800 steps would already match the taig’s precision. That’s why the 2000/8000 formula works and is what people do. Now it makes sense.
I know not to try to reinvent the wheel, but sometimes, I can’t resist the temptation of making things as difficult as possible .

I did try my fancy settings last night and it seemed to work very well. Over a 1 inch move I get consistent precision of better than 0.001 from what I can see. I seem to have a backlash on the ball screw of about 0.0005. I don’t have the spindle, headstock or motor yet, (waiting for glockcnc.com to ship) so I can’t do real work test yet. I’m far from breaking chip.
The 0.0002 precision of the Taig must be calculated over the whole travel of the machine, I guess, and not accounting for any backlash.

Thanks.

[Richards]

In order for anything to move, there has to be a little clearance. Right now, I have no way of measuring the true repeatability of my Taig mill. I only have some 0.001" per division dial indicators. 0.0002" is just 1/5th of a division on those dial indicators. One of these days I'll buy a 0.0001" dial test indicator.

What I learned years ago with my Shopbot was that advertised repeatability and actual performance were two different things. Even though the Shopbot (when clean, greased, and well adjusted) could repeat a single axis move to 0.002" (yes, 10X worse than the advertised repeatability of the Taig), the combined errors of the axes, along with the flex of the cutter and the possible give and take of the vacuum hold-down all added up. I considered 0.0156" (1/64 inch) a perfect cut on that machine. Even that tightness was lost on wood. Wood moves a lot as the humidity changes. Here in Utah, where 20% humidity is high humidity, we have to leave at least 1/8" gap on all doors. Wood moves enough that when the swamp cooler is running on the 100-degree summer days, that the 1/8" gap disappears and some doors begin to stick.

A metal cutting mill is different. It's meant to be more repeatable. But, when you consider just the difference that temperature makes on metal, holding tolerance to 0.0002" seems to be something to talk about around the water cooler rather than expected performance. I've watched a lot of YouTube videos from NYC CNC and Pierson Workholding. Those guys run $150,000 machines that they keep in top condition. It seems that they joke about 0.0002" repeatability more than expect it.

Anyway, I try to remember that a typical piece of copy paper is about 0.003" thick. Except for bearing pockets, there are few things that I cut that need even 0.003" repeatability. We've had a mild winter here in Utah where the coldest temperature was just above zero. This summer the temperature might hit 105 degrees. A high precision cut in aluminum in my 60 to 75 degree work shop might not measure the same when measured at zero degrees or 105 degrees in the garage.

It's nice that Centroid makes an Acorn that can send 400,000 pulses per second, making extremely fine cuts possible, and its nice that DMM sells servos that can take advantage of that many steps per second and its nice that Taig makes a mill that is carefully assembled with tight tolerances; but, my actual requirements will never approach that level of ultimate accuracy. Back in the 1970s, I often had a local machinist make parts for Kodak photo printers. I asked him how accurate the parts could be. He smiled and said that he could hold 0.001" on the size of parts that I needed, but that I could never be able to afford them. Then, he and I looked at what really needed to be done. His "quick and dirty" rate matched my needs (+/- 0.0025"). The parts fit perfectly and my wallet still had a little money left after paying him.

[nicolasjolliet]

Thanks Richard,
That was going to be my next question. What tolerances to expect in the real world of small machines.