Need Help! with my milling machine
08 Jan 2015 21:09 #54771
by nooob
Replied by nooob on topic Need Help! with my milling machine
Hi,
thanks for the fast response, i reversed the polarity on the encoder card to ged rid of the negative values but it didn't work, my problem is still how can i get rid of these oscillating?
how big should the P value be in my case?
thanks for the fast response, i reversed the polarity on the encoder card to ged rid of the negative values but it didn't work, my problem is still how can i get rid of these oscillating?
how big should the P value be in my case?
Please Log in or Create an account to join the conversation.
08 Jan 2015 21:18 #54774
by andypugh
Big enough, but not too big![:-) :-)](/media/kunena/emoticons/smile.png)
The Gains required are different for every machine. I tend to start very small (.0001) and keep moving the decimal point until something happens.
Replied by andypugh on topic Need Help! with my milling machine
how big should the P value be in my case?
Big enough, but not too big
![:-) :-)](/media/kunena/emoticons/smile.png)
The Gains required are different for every machine. I tend to start very small (.0001) and keep moving the decimal point until something happens.
Please Log in or Create an account to join the conversation.
08 Jan 2015 21:33 #54776
by emcPT
Replied by emcPT on topic Need Help! with my milling machine
On your picture, what is the oscillation from top to bottom of the oscillation?
If it is about 0.005mm is hard to go better than that due to your encoder limitation. If you make your calculation after quadrature one encoder count (your maximal physical measurement) will be about 0.001, because you will have 5000 pulses per 5mm. And it is normal that the motor will be oscillating at least about a few encoder positions.
What you can improve however is the medium value of the error (so that it will be around 0 when you move the machine to one side and to another). This is acheived by changing the FF values (FF1 if I am not mistaken) with small values.
I hope that I gave some help.
If it is about 0.005mm is hard to go better than that due to your encoder limitation. If you make your calculation after quadrature one encoder count (your maximal physical measurement) will be about 0.001, because you will have 5000 pulses per 5mm. And it is normal that the motor will be oscillating at least about a few encoder positions.
What you can improve however is the medium value of the error (so that it will be around 0 when you move the machine to one side and to another). This is acheived by changing the FF values (FF1 if I am not mistaken) with small values.
I hope that I gave some help.
Please Log in or Create an account to join the conversation.
09 Jan 2015 18:28 #54815
by nooob
Replied by nooob on topic Need Help! with my milling machine
Thx for the answers super forum members
, i will tune all axis now with belts on and post the results.
![:) :)](/media/kunena/emoticons/smile.png)
Please Log in or Create an account to join the conversation.
03 Feb 2015 17:37 - 03 Feb 2015 17:53 #55622
by nooob
Replied by nooob on topic Need Help! with my milling machine
after a few test runs and sample parts i still have a new big problem, the surface finish isn't good enough i think it probably comes from the PID settings
here a few pics from the milled surface.
I've tried to tune the Axis more accurately but after hours and hours the result was still the same all changes made it worse then it was.
so now i need your help guys again... sry for that but might one of you can help me.
Milled surface:
And here the Pics from Halscope and my settings (click to enlarge)
Rapid Move 60mm:
With Feed 4000mm/min:
Without movement:
Program:
(the part of the program that finished the outer diameter)
here a few pics from the milled surface.
I've tried to tune the Axis more accurately but after hours and hours the result was still the same all changes made it worse then it was.
so now i need your help guys again... sry for that but might one of you can help me.
Milled surface:
![File Attachment:](http://img3.fotos-hochladen.net/thumbnail/img53556cribxmhoq_thumb.jpg)
![File Attachment:](http://img3.fotos-hochladen.net/thumbnail/img5356iuvmerx50s_thumb.jpg)
And here the Pics from Halscope and my settings (click to enlarge)
Rapid Move 60mm:
![File Attachment:](http://img3.fotos-hochladen.net/thumbnail/rapid60mm7awu9e6ogt_thumb.jpg)
With Feed 4000mm/min:
![File Attachment:](http://img3.fotos-hochladen.net/thumbnail/whendrivingfeyiqtx2d6wb_thumb.jpg)
Without movement:
![File Attachment:](http://img3.fotos-hochladen.net/thumbnail/withoutmovemenbfgc4w7vjl_thumb.jpg)
Program:
(the part of the program that finished the outer diameter)
(2D-KONTUR2)
N26815 M8
N26820 G0 X21. Y-4.1
N26825 Z15.
N26830 Z4.
N26835 G1 Z0. F1000.
N26840 Z-10.5 F800.
N26845 G19 G3 Y-3.6 Z-11. J0.5 K0.
N26850 G1 Y-3. F1000.
N26855 G17 G3 X20. Y-2. I-1. J0.
N26860 G2 Y42. I0. J22. F1200.
N26865 Y-2. I0. J-22.
N26870 G3 X19. Y-3. I0. J-1. F1000.
N26875 G1 Y-3.6
N26880 G19 G2 Y-4.1 Z-10.5 J0. K0.5
N26885 G0 Z15.
N26890 G17
N26895 M9
N26900 G53 G0 Z0.
Last edit: 03 Feb 2015 17:53 by nooob.
Please Log in or Create an account to join the conversation.
03 Feb 2015 18:16 #55628
by andypugh
It _might_ be from the PID settings, or it could be other things.
Do the position errors seen on Halscope have a similar magnitude to the surface finish imperfections?
Looking back through old messages it looks like this is the Chiron machine? That certainly ought to be able to achieve a good surface finish, and it is unlikely that the moving parts are light enough for chatter to be a problem.
One experiment that might be interesting would be to reduce the P term by a factor of 2 and see if the surface finish changes. I wouldn't expect it to necessarily improve, the question is whether it changes, giving a different scale of roughness, for example.
Replied by andypugh on topic Need Help! with my milling machine
after a few test runs and sample parts i still have a new big problem, the surface finish isn't good enough i think it probably comes from the PID settings
It _might_ be from the PID settings, or it could be other things.
Do the position errors seen on Halscope have a similar magnitude to the surface finish imperfections?
Looking back through old messages it looks like this is the Chiron machine? That certainly ought to be able to achieve a good surface finish, and it is unlikely that the moving parts are light enough for chatter to be a problem.
One experiment that might be interesting would be to reduce the P term by a factor of 2 and see if the surface finish changes. I wouldn't expect it to necessarily improve, the question is whether it changes, giving a different scale of roughness, for example.
Please Log in or Create an account to join the conversation.
03 Feb 2015 19:56 #55635
by nooob
Replied by nooob on topic Need Help! with my milling machine
[quoteLooking back through old messages it looks like this is the Chiron machine? That certainly ought to be able to achieve a good surface finish, and it is unlikely that the moving parts are light enough for chatter to be a problem.][/quote]
Yea i think so too i will try your proposal about the half Pgain and make another test part.
u think my I value is to high? cause its such a huge number.
Yea i think so too i will try your proposal about the half Pgain and make another test part.
u think my I value is to high? cause its such a huge number.
Please Log in or Create an account to join the conversation.
03 Feb 2015 21:17 #55640
by andypugh
I have no opinion. The absolute magnitude of the numbers is hard to predict and is what it is for any given system.
The experiment was just to see if PID tuning does have an effect on the surface finish.
There are other things that might have to be tried too, such speed/feed conventinal/climb etc.
Replied by andypugh on topic Need Help! with my milling machine
u think my I value is to high? cause its such a huge number.
I have no opinion. The absolute magnitude of the numbers is hard to predict and is what it is for any given system.
The experiment was just to see if PID tuning does have an effect on the surface finish.
There are other things that might have to be tried too, such speed/feed conventinal/climb etc.
Please Log in or Create an account to join the conversation.
03 Feb 2015 21:23 #55642
by emcPT
Replied by emcPT on topic Need Help! with my milling machine
My small opinion:
Run the code with the spindle in a high position with the spindle off and place you hand in the table. Feel the table.
If you can feel it jerky you will know that is not from the machine spindle or cutting conditions.
Then proceed depending on this test.
Run the code with the spindle in a high position with the spindle off and place you hand in the table. Feel the table.
If you can feel it jerky you will know that is not from the machine spindle or cutting conditions.
Then proceed depending on this test.
Please Log in or Create an account to join the conversation.
01 Apr 2015 03:40 #57354
by nooob
Replied by nooob on topic Need Help! with my milling machine
Hey guys,
finally i have time to get further with my retrofit but still have a few questions![:) :)](/media/kunena/emoticons/smile.png)
the PID Axis settings are done the only thing left is the Spindle.
1. The problem is that the Mesa card didn't send 10 volts the the servo amplifier when i set the Rpm
in MDI to 7100rpm wich is the maximum spindle speed the motor can handle, i get only 8,76 Volts instead of 10V so the spindle doesn't run at full Rpm.
I mesured with my Laser Rpm tester and ill get only 6900 Rpm thats ok because the amplifier gets the 8,76V but why is the voltage not 10V??
What i tried so far:
external power for the opto card wich is a 7i37TA but still no 10Vlot
i think my hal file is the problem maby u guys find the problem
2. The second problem is when i shut down the Pc but have the machine Power still on all Outputs getting true signal spindle brake also, and thats really bad the Z-Axis runs down then.
How can i avoid that??
hope u can help me.
sry for my bad english![:) :)](/media/kunena/emoticons/smile.png)
greetings nooob
finally i have time to get further with my retrofit but still have a few questions
![:) :)](/media/kunena/emoticons/smile.png)
the PID Axis settings are done the only thing left is the Spindle.
1. The problem is that the Mesa card didn't send 10 volts the the servo amplifier when i set the Rpm
in MDI to 7100rpm wich is the maximum spindle speed the motor can handle, i get only 8,76 Volts instead of 10V so the spindle doesn't run at full Rpm.
I mesured with my Laser Rpm tester and ill get only 6900 Rpm thats ok because the amplifier gets the 8,76V but why is the voltage not 10V??
What i tried so far:
external power for the opto card wich is a 7i37TA but still no 10Vlot
i think my hal file is the problem maby u guys find the problem
2. The second problem is when i shut down the Pc but have the machine Power still on all Outputs getting true signal spindle brake also, and thats really bad the Z-Axis runs down then.
How can i avoid that??
hope u can help me.
sry for my bad english
![:) :)](/media/kunena/emoticons/smile.png)
greetings nooob
Please Log in or Create an account to join the conversation.
Moderators: piasdom
Time to create page: 0.106 seconds