C-Axis (Angular) and sprockets configuration..

More
03 Feb 2016 13:54 #69511 by andypugh
If you are using that INI then you are using a scale of 123.

Are you sure that the HAL file is using the correct INI entry?

You don't get 180 degrees. What do you get?

Please Log in or Create an account to join the conversation.

More
04 Feb 2016 02:22 #69557 by pkasy

Are you sure that the HAL file is using the correct INI entry?

yes, if
17  float RW       123.8289  stepgen.0.position-scale

You don't get 180 degrees. What do you get?

if stepgen.0.position-scale=113.5802
g1 b10 makes 9.0467 degree

Please Log in or Create an account to join the conversation.

More
04 Feb 2016 04:32 #69563 by pkasy
So, the correct answer is 124.583585
exactly at this value error for one revolution is 0.0243 degrees.

Please Log in or Create an account to join the conversation.

More
04 Feb 2016 09:12 #69569 by andypugh

So, the correct answer is 124.583585
exactly at this value error for one revolution is 0.0243 degrees.


How did you determine that number?

These numbers are not random. If you can't explain the SCALE number then something is wrong. (missed steps, slipping coupling).

I could easily believe 127.777777 because that would be right for the very common 90:1 gearing on rotary tables.

Please Log in or Create an account to join the conversation.

More
04 Feb 2016 10:11 - 04 Feb 2016 12:10 #69571 by pkasy
>How did you determine that number?
the exact value I measured by a probe.
I measured the height of the steel bar, clamped to the rotational axis in two places, X1 and X2. rotated 360 degrees axis, again measured with the help of algebra is easy to find the angle between two segments.
then changed in the right direction
halcmd setp stepgen.0.position-scale XXX
and measured again. it was a long, but very accurate.
so now I know the answer with accuracy to 0.0243 per revolution axis.
>These numbers are not random.
I'm in shock
> If you can't explain the SCALE number then something is wrong.
No, I can not.
> (missed steps, slipping coupling).
No. high repeatability, 0.002 mm in height to a radius of 45 mm.
>I could easily believe 127.777777 because that would be right for the very common 90:1 gearing on rotary tables.
label says Harmonic Drive 17-80

this is why I asked - how to calculate the SCALE!
Last edit: 04 Feb 2016 12:10 by pkasy.

Please Log in or Create an account to join the conversation.

More
04 Feb 2016 11:19 #69572 by andypugh
Can you carefully count motor revolutions and check that it is exactly 80:1?

I think that, depending on what is held stationary and what is rotating there is more than one potential gear ratio with a harmonic drive.

For example, I think that the rings either have 79 and 81 teeth or 80 and 82 (I have never counted) so depending on which is the static ring you can have 80:1, 78:1 or 82:1 gearing.

Working backwards, and assuming that the pulley tooth counts are right, your number suggests a harmonic drive ratio of 87.75:1
The following user(s) said Thank You: pkasy

Please Log in or Create an account to join the conversation.

More
04 Feb 2016 12:05 #69573 by pkasy
>87.75:1
yes, about. but as far as I know, the harmonic gearbox with such a gear ratio does not exist. so the scale must be 124.583333333333

what can affect stepgen.0.position-scale? I saw no source, so I do not know, that can affect the SCALE
maybe something operates within the LinuxCNC?
it is used directly or is there more factors on which it is multiplied or divided?

Please Log in or Create an account to join the conversation.

More
04 Feb 2016 12:27 #69574 by andypugh

what can affect stepgen.0.position-scale?


Anything you want it to but it would be very unusual indeed to alter it in a running system. It is a pin that you are in control of.

The step generator is sent a new position command every servo cycle (through the stepgen.0.position-cmd pin)
It compares that position command with the current position it has reached (often but not always exactly the same as the previous position-cmd value) then multiplies the difference by the value of the position-scale pin to calculate how many step pulses to send in the next servo period.

Please Log in or Create an account to join the conversation.

More
16 Feb 2016 14:00 - 17 Feb 2016 00:38 #70231 by pkasy
Today I decided to measure more accurately. and got the result -124.5425
This corresponds to the gear ratio - 87.7212391304348 :blink:
I changed a little bit of height measurement unit from a reputable verser .
I performed a series of tests on the bar clamped in a cartridge.
distance from the center to the measuring point - 82 mm.

measurements were made three times, and then a slow rotation, then three more times.
1th z = 206.881650 mm
2th z = 206.882983 mm,
diff = −0,001333 mm

I have a self-made probe, like renishow. but not so accurate, so why stop there.

Now check the text of test, may be someone will need it.
#<z> = #<_ z> (save current Z position)
G90G1B0F300
G91F50G38.2 Z-50
G0 Z0.5
G4 P0.5
F5G38.2 Z -50
#<Averz> = #5063
G0 Z 0.5
G4 P0.5
G38.2 Z -50
#<Averz> = [#<averz> +#5063]
G0 Z 0.5
G4 P0.5
G38.2 Z -50
#<Averz> = [#<averz> +#5063]
#<Averz> = [#<averz> / 3]
(Print, average z = #<averz>)
G90 G0 Z #<z> (back to start point)
G90G1B360F300
G91F50G38.2 Z-50
G0 Z 0.5
G4 P0.5
F5G38.2 Z -50
#<Averz2> = #5063
G0 Z 0.5
G4 P0.5
G38.2 Z -50
#<Averz2> = [#<averz2> + #5063]
G0 Z 0.5
G4 P0.5
G38.2 Z -50
#<Averz2> = [#<averz2> + #5063]
#<Averz2> = [#<averz2> / 3]
#<Diff> = [#<averz2> - #<averz>]
(Print, average z = #<averz2>, diff = #<diff>)
G90 G0 Z #<z> (back to start point)
M2
Last edit: 17 Feb 2016 00:38 by pkasy. Reason: typo

Please Log in or Create an account to join the conversation.

More
25 Jan 2018 13:30 #105001 by pkasy
Hello.
today the topic received an unexpected continuation.
I removed the stepper motor from this axis and installed a servomotor with an encoder of 2500x4=10000 ppr .
after a series of measurements I came to the conclusion that the gear ratio of this reducer is 112.5.
something like that.
strange, right?
nameplate is 80. not 12.5
I used rawcounts and f-error.

Please Log in or Create an account to join the conversation.

Time to create page: 0.282 seconds
Powered by Kunena Forum