Full Revolution issues - Rotary Axis - Barrel Organ Machine

More
03 Jun 2020 15:47 #169954 by pijki
Hello,

I'm working on a special machine to cut cardboard for barrel organ (my parents are making these french musical instruments). It's a 3 axis machine and im using the MX3660 driver with Linux cnc 2.7. Here the beta version of the machine on my website : antoninodin.com/CARDBOARD-MACHINE ) .

I have an issue with the rotary axis A, I dont manage to do a full revolution in both direction : when I do a G91 G0 A360 it's OK but when I do a G91 G0 A-360 there is an offset ( I have marked my physicial axis to see it). I have tried to change the settings all the day but I dont underrstand why I dont have a full revolution in both directions...

Here the settings of the A axis in the Ini files :



Thanks a lot for your help.
Attachments:

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

More
03 Jun 2020 16:02 #169957 by Aciera
Interesting!

It's usually better to attach your HAL and INI files to your post.

Is the offset always the same and is there also the same effect when you use G90?

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

More
03 Jun 2020 16:42 #169961 by pijki
thanks !

here the HAL and INI files :

File Attachment:

File Name: test2020.ini
File Size:2 KB

File Attachment:

File Name: test2020.hal
File Size:3 KB



The offset seems to be always the same. here :
a video that shows the offset of this axis :
I do 2 times G91 G0 A360 and then 2 times G91 G0 A-360.

For the G90 i have an error : " absolute position invalid for the wrapped rotary axis A"
Attachments:

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

More
03 Jun 2020 17:44 - 03 Jun 2020 18:06 #169970 by Aciera
1. Can you rule out mechanical backlash?
2. I am unfamiliar with the "Wrapped rotary" function but the manual states:

WRAPPED_ROTARY = 1 - When this is set to 1 for an ANGULAR axis the axis will move 0-359.999 degrees. Positive Numbers will move the axis in a positive direction and negative numbers will move the axis in the negative direction.

So it would seem that 360 is then outside the allowed range. Is there a particular reason to use "WRAPPED_ROTARY=1"? I also use a rotary axis but have never used it. I didn't even know this feature existed. Have you tried without it?

[edit]
Ok, I just read up on the wrapped_rotary and it prevents the wind up of the angular value so that might make sense. Nevertheless it might be worth to try things out without it.
Last edit: 03 Jun 2020 18:06 by Aciera. Reason: learned about wrapped_rotary

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

More
03 Jun 2020 23:13 #170045 by pijki
I dont think that the problem is coming from a mechanical backlash.
I have tried without the Wrapped_Rotary and it's not changing anything.

What is weird is that I can send G91 G0 A360 many times and I will never have any offset, but when It's a G91 G0 A-360, it's seems that the motor is doing less steps that needed... By testing, i figured it out that a G91 G0 A-364.4 is actually doing a full revolution, but i dont understand why ?

I was expecting that a A360 and a A-360 will do a full revolution and having both at the exact same position...

thanks for your help!

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

More
04 Jun 2020 05:20 - 04 Jun 2020 05:28 #170071 by Aciera

I was expecting that a A360 and a A-360 will do a full revolution and having both at the exact same position...

Well that is certainly what we would expect.
So maybe try the following:
1. Jog the axis in the negative direction (-360°) to the mark
2. command G91 G0 A360
3. report if the turn was a full revolution or if there is an offset

[edit]
Since you made two revolutions in both directions in the video I'm wondering what happens if you do only one revolution. Is the offset only half and If you do four revolutions is the offset then doubled?
Last edit: 04 Jun 2020 05:28 by Aciera.

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

More
04 Jun 2020 08:06 #170079 by pijki
So from the mark (the origin) I have jogged the axis to -360°, and it came back to the mark correclty (so 1 revolution). But then the G91 G0 A360 do an offset. We can also notice that after these two moves, I can come back to the mark at 0.00 by doing again -360°. But, If I jog the axis to A360°, it still getting an offset. I have to reset the origin of the axis, to do be able to do A360° and do a full revolution.

So to sum up, once the origin is defined, If the first jog is negative (-360°), the positive jog (+360) will be offset. In the contrary, if the first jog is positive, it's the negative jog that will be offset.

The offset is always the same whatever is 1 or 4 revolutions.

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

More
04 Jun 2020 08:28 #170082 by bbsr_5a
it seams to be a SCALE and Homing issue
WHAT is the MICROSTEPPING of the MX3660 set
and what is the Drive system Timingbelt 5:1 ?

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

More
04 Jun 2020 08:57 #170086 by pijki
The microstepping is set to 1600 on the MX3660. So 8 in the StepConf wizzard Microstepping.

what is the Drive system Timingbelt 5:1 ?

I'm not sure about what you mean about "Drive system Timigbelt 5:1", as the picture posted below, this axis is a direct drive, and there is no timigbelt.

I have tried to change the microstepping and the scale, but i had still some offset...

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

More
04 Jun 2020 09:53 #170089 by Aciera
So if I understand you correctly the offset happens after a reversal of the direction which is what makes it look like a mechanical problem. I know you ruled that out but I strongly suggest to remove the motor and check again on the bare motor shaft.
If it also occurs with just the naked motor then it seems that you are losing steps while reversing the motor for which I have no possible explanation at the moment.

Have you tried to assign the rotary axis to one of the other motors in HAL and tested it that way?
The following user(s) said Thank You: pijki

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

Time to create page: 0.399 seconds
Powered by Kunena Forum