Understanding & Selecting gear pinions - setting the steps per mm[inch]

[kaartman]

Thank you Gerald
I have played with the settings this afternoon, used the 20.886 and it worked perfectly, Kobus i will have a look for that page and measure over long distance, learn-learn that is the only way, thank you for the input, much appreciated

[felix]

Originally Posted by Gerald D

Sean Reish is using 7.2 ratio and 20 teeth. Doc Tanner & Greg J are using 3.6 ratio and 30 teeth. Lots of people use 7.2 and 30 teeth. Nobody has cut quality problems. Greg J would like a bit more torque at high speed.

Hi all,

one might consider the power supply as well in all this drive selection process. Torque problems can come from inadequate PS as well as gearing.

Yves

[sailfl]

I am looking at every thing since I realized I had a problem with my cutting on my machine. I am looking at threads that I might have read before or skimmed over in the past.

If I understand the "Gear Speed Step Freq Calculator" correctly, for gear box of 7.2 and with 20 pitch, 30 tooth pinion gears, my motors steps should be set at 3055.775 which they are but my kernal should be set to 37 kHz which it is not. I am currently running at 25 kHz.

What I don't understand is what effect does this have on the way the machine runs? Further clarification, please.

Thanks

[bradm]

Nils, you may have put your finger on a problem here.

25 Khz, or 37 Khz is the number of update cycles your system can make per second.

If you are familiar with the concept of a 'carrier wave' in radio transmission, it's similar.

If you attempt a movement that requires more updates than your kernel can sustain per cycle, it will "jump" to catch up. This can cause stairstep like patterns.

Unfortunately the max kernel Khz is a function of the computer you are running on, and isn't always directly related to the clock speed; it's related to the I/O and interrupt handling capacity of the machine, which is much harder to predict.

If you push it too high you'll know, because the machine will hang, or you'll get errors reported.

[sailfl]

Brad,

Thanks for that explination and I will at least set kernal at the speed recommended by the Geard, Speed....Calculator. That is a good place to start.

[domino11]

Brad, Nils,
Correct me if I am wrong here, but isn't the nubmer for kernal speed in the calculator for MIN kernal speed for the gearing/pinions/max move speed? All I mean is that number is the kernal speed you would need to do the max speed move (rapids). When cutting at slower speeds , as long as the kernal is capable of covering the speed your at you should be fine. There is also a risk of running your computer at too high a kernal speed for the hardware you have also.

[bradm]

Heath, you are correct. However, I don't know what the max speed of Nils' machine is, and the fact that his setting is low in combination with the other symptoms makes me suspicious. It may turn out the the problem is that it's too high for the computer to sustain, and that's what causes the issue.

What I'm looking for at this point is some evidence that changing his computer configuration affects the periodicity of the cut problem.

[domino11]

Brad,
Thats a good point, it might shed some more light on the computer in use. I did not mean to discredit your test.

[sailfl]

Don't we have to look at what is being processed and what is being sent out.

Lets look at an extreme example. When a computer is running simulator software, it is taking in lots of data, sending out data or commands and processing lots of data so the kernal has to be set at a level that will handle all this processing or traffic.

If I understand how Mach works, G Code is processed one line at a time. A line of G Code is sent one after the other. I don't know if there is a message sent back from the PMDX that says I got the line send the next or if mach is just sending lines and assumes the PMDX is processing but I suspect there has to be some communication because if you push hold it takes a little time to process that command and send a message back that says "ok, we will hold".

This is simple communications and so the kernal is not going to be handling a lot of traffic. So the Kernel needs to be set at a very low level but it still needs to be high enough so it can handle the traffic. To low and we have a problem and to high, it might use processing that could be used some place else but we don't have any thing else going on so high doesn't matter as long as it is not too high.

Is that the way it works or is my thinking off?

I don't think that having the kernal too high will be a factor but too low could be.

[bradm]

We're definitely drifting off topic here in gear pinions (my fault).

The PMDX never sees the GCode. All it sees are the step and direction signals that are calculated in software. Generally software that handles GCode will pause after it completes a line. So if you start a long slow cut from one end of the table to the other, and press pause, the pause takes effect after that cut completes, not in the middle.

The kernel is handling a pretty substantial amount of traffic, in that to move an inch, it has to output thousands of step pulses, and each one has to come at precisely the right moment. You are correct that having it's setting too high is not a problem unless it exceeds the capacity of the computer in question.

[felix]

Nils,

I'm not sure about how Mach3 does it but here is what I understand about g-code. First, this is a command language which means it as to be interpreted and transformed into signals then you have what is called a trajectory planner which basically looks ahead for the next move(s) in order to keep the path as smooth as possible (set by config) and all this while taking into account accel. and decel. and a few other things as well.

All this planning and calculations take quite a bit of time compared to just maintaining the signal (pulse) stream at a desired level. There are also real time events such as an E-stop that needs to taken into account, the response time for this also affects the overall system performance.

Basically, you need to have a signal (pulse) cycle active time long enough so that the CPU as time to do its other (required) tasks with time to spare just in case it needs to stop on a dime and this active time needs to be short enough that the cutting speed is acceptable.

One other thing to consider is the rather steep decline in torque of the stepper motors as its RPM increases. This physical limitation also limits the maximum signal frequency required for the controller. In other words, even if you could generate pulse in the hundreds of Khz there will come a frequency beyond which the motor will stall and begin to miss steps.

Hope this helps,
Yves

[sailfl]

Brad and Yves,

Thanks for the clarification.

[Claudiu]

Hello Guys,

I did some calculations on my Gears etc. As I use Module 1, OM7.2, gears 35 teeth I get a magic Number of 130.962. Everything looks ok up to here..but when I calculate with the max rot of the OM of 250rpm I get out a speed of only 3800 mm/min. One whole minute for the spindle to go one full x direction? Can this be really correct? How long will it take to surface the table?
Please tell me I did a mistake and the working speed will be higher...

[Gerald D]

I don't know where you see: "the max rot of the OM of 250rpm"?

Maybe the output shaft max speed is 250rpm while the actual motor turns 7.2 times faster?

If the output shaft can turn at 250rpm, the pinion has about 100mm circumference, the x-speed is about 25000 mm per minute which is fast enough!

[Richards]

Gerald,
Oriental Motor lists the permissible speed on page C-231 of one of their older catalogs. I haven't downloaded the current catalog, so that information might be missing.

Claus,
That is the output speed of the gearbox, which, with a 7.2:1 gearbox, means that the motor would have to be spinning at 1,800 RPM. At that speed, the motor only produces only 30 oz*in of torque, so it is not practical to run that fast.

Because I'm on the inch system, I haven't paid any attention to the metric spur gears, but a spur gear with 1.25" pitch diameter (25 tooth) moves about 3.925 inches per revolution, so 250 X 3.925 = 981 inches per minute, or 16 inches per second. Getting 50% of that speed would meet most needs.

[David Bryant]

Hi
I have a Mod 1 rack, 31 tooth pinions 5:1 gearbox microstepping to give 1600 pulses per rev of the stepper.
I set Mach 3 to 82.144 steps per mm.

I am getting about a 6% error in the movement distance.
Where have I gone wrong in my Calc?
Thanks
David

[Gerald D]

What drives are you using for the motors, and what are they set to for micro-stepping? (The geckodrive is fixed at 10 micro-steps).

The 5:1 gearbox 1600 pulses has me scratching my head....

[Gerald D]

Also count the number of rack teeth over 100mm to verify that it is Mod1. (a quick mental calc says 32 teeth?)

[David Bryant]

Hi Gerald
The drives are motion King pulse per rev set for 1600 pulses per rev (8 micro steps 1.8 degree per step) followed by a gearbox mechanical gear reduction by G5
32 teeth on the rack very close to 100mm.
Cheers
David

[Gerald D]

Then 82.144 is correct. Are you sure that you didn't type in 87.144?

[David Bryant]

Thanks for checking I will keep looking...? Not a typo.
Cheers
David

[Kobus_Joubert]

Use Mach3 calibration to get it spot on.

[David Bryant]

Hi All
I did further testing today. I isolated the Z axis from the machine and got a counter onto the pulses out of Mach3.
I found mach3 pulses out to be correct (No surprise.)
I then set pulses per unit to 1600 to match the stepper drivers setting. I then counted revolutions of the MotionKing stepper motor. Always about 6% revs indication more than actual moved.

On Testing I found that 1700 pulses per set into mach 3 give the correct number of turns (allowing for the 5:1 ratio on the gear box).
So my testing has shown that the Motion king driver needs 1700 X 5 pulses to drive the stepper motor with gearbox one revolution. According to the instructions with the driver and stepper motor this should need only 1600 X 5 pulses per rev.

Is anyone else using MotionKing drivers and motors?
Cheers
David

[David Bryant]

Hi
I have four 34HS9801-G5 MotionKing stepper motors
I have been trying to work out why my mech mate does not move the correct distance(accordind to theory). So I put a signal generator onto the step input and measured the time for a number of revolutions using the MotionKing driver. (tried DC supply and AC supply to MotionKing Driver, no difference)
I always needed 17/16 more pulses than theory said I needed.
I changed to a gecko drive and repeated the test with a bench power supply and wiring direct to the motor. Still needed 17/16 times more pulses than the expected 200 steps at the motor X 5 for the gear ratio.

My conclusion is that the gearbox is not exactly a 5:1 ratio.
I have not seen any other problems with the motor or drive.

[David Bryant]

Hi
Just redid calcs and set the new numbers into Mach3 put it all back together, still out by 1-2mm in 1000. ???

Did many of you guys get the exact number or does everyone set up by adjusting afterwards?

Cheers
David

[danilom]

Mine calculation was

200 steps * 10 microstep = 2000
2000 * 3 times belt reduction = 6000 steps per rev
travel per rev = 3.14159265 (tooth distance = Pi) * 30 tooth = 94.2477796

steps per mm = 6000 / 94.2477796 = 63.6619662

but using dial indicator and mach calibration from SETUP page for one axis it was lower than 62 steps per mm, and other greater than 64. It must be because of in precision in motor movement as I think belt gears are pretty much the same.

[KenC]

10 micro steps per pulse is not a practical setting, almost all the article I read says 1/8 is the highest practical setting. maybe this is the problem.

[David Bryant]

Hi Ken
10 micro steps is standard for the Geckodrives

Cheers
David

[David Bryant]

Thanks Danilo

That is very close to the variation I have found. It will be good to hear some more if they got it exact.
thanks again
David

[KenC]

Ooops, I don't use Gecko