search

clough42 / electronic-leadscrew

Lathe electronic leadscrew controller
MIT License
322 stars 117 forks source link

Long threads Not accurate. #139

Open poulbo opened 3 years ago

poulbo commented 3 years ago

Long threads Not accurate.

Am I the only one who has discovered that 1:1 position A on Weiss WM280V-F, (Warco-280VF, PM1127VF-F), is not exactly 1:1, when making very long threads they do not become precise. It is not the leadscrew itself, but leverage in position A that is not exactly 1: 1.

you can correct on leadscrew parameter in Configuration.h. I have a Weiss 280VF with 3mm lead screw instead of 300 I have written 293 and it should be 292.5, but then you have to specify 1/10 of a mm = 3000. Is there a better way to put a correction factor into the program.

I use ELS ver. 1.4.0

 #define LEADSCREW_HMM 293 // (300) Weiss VM280V-F
 #define STEPPER_MICROSTEPS 3
 #define STEPPER_RESOLUTION 1000
 #define ENCODER_RESOLUTION 4096
 #define USE_FLOATING_POINT

Yours sincerely, Poul Bo Nielsen

cad-quark commented 3 years ago

Have you checked the RPM with a digital tach? My BEI encoder is 1000 even so its 4000 pulses per rev rather than 4096. Thats 2%. Just like 7/300.

poulbo commented 3 years ago

Hi cad-quark You just hit the problem what I thought was 1024 / rev taco counter was a 1000 / rev It was marked incorrectly. Thank you very much for your inquiry, it all now makes sense. Also could not understand that lathe was not 1: 1 in position A. Thanks again for your help. Yours sincerely Poul Bo Nielsen

cad-quark commented 3 years ago

The 2% error was too much of a coincidence. The clough42 ELS acts like an electric gearbox. Any lost or gained revolutions is either input or output. Since most users got good threads, it had to be your input setting that did not match.

One thing I have wondered is since you wire up the I (or Z) index, which gives you one pulse per revolution, TI processor knows when the revolution is complete, but the program must not use that index.

clough42 commented 3 years ago

Correct. It doesn't use the index pulse.

James

On Sat, Mar 20, 2021 at 4:23 PM cad-quark @.***> wrote:

The 2% error was too much of a coincidence. The clough42 ELS acts like an electric gearbox. Any lost or gained revolutions is either input or output. Since most users got good threads, it had to be your input setting that did not match.

One thing I have wondered is since you wire up the I (or Z) index, which gives you one pulse per revolution, TI processor knows when the revolution is complete, but the program must not use that index.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/clough42/electronic-leadscrew/issues/139#issuecomment-803472042, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAZLSHH6SA7TO327VSILFHLTEUN4TANCNFSM4YW2JBEQ .

clough42 commented 3 years ago

Actually, allow me to elaborate further. It doesn't use the index pulse because the index pulse doesn't necessarily indicate one full revolution of the spindle if the drive between the spindle and the encoder is not 1:1.

On Sat, Mar 20, 2021 at 5:06 PM James Clough @.***> wrote:

Correct. It doesn't use the index pulse.

James

On Sat, Mar 20, 2021 at 4:23 PM cad-quark @.***> wrote:

The 2% error was too much of a coincidence. The clough42 ELS acts like an electric gearbox. Any lost or gained revolutions is either input or output. Since most users got good threads, it had to be your input setting that did not match.

One thing I have wondered is since you wire up the I (or Z) index, which gives you one pulse per revolution, TI processor knows when the revolution is complete, but the program must not use that index.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/clough42/electronic-leadscrew/issues/139#issuecomment-803472042, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAZLSHH6SA7TO327VSILFHLTEUN4TANCNFSM4YW2JBEQ .

cad-quark commented 3 years ago

I had no criticism of the ELS design. I am thrilled we users are able to use the weeks of engineering and code writing that make up ELS you have given us.. I just suspect that somewhere buried in the TI documentation they explain how to use GPIO59, QEP1 to zero GPIO 35,37.

cad-quark commented 3 years ago

Great, If you cannot get 1:1 at the spindle, then the number of pulses for one revolution ( calculated with your actual ratio ) can be set in Configuration H and the program math still works. Wow, what a killer design.