Journeyman 325 retrofit - Page 2


Page 2 of 3 FirstFirst 123 LastLast
Results 13 to 24 of 32

Thread: Journeyman 325 retrofit

  1. #13
    Registered
    Join Date
    Feb 2011
    Location
    USA
    Posts
    24
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Journeyman 325 retrofit-x-axis-setup-jpgJourneyman 325 retrofit-x-axis-tuning1-jpgJourneyman 325 retrofit-x-axis-vel-output-vs-time-jpgJourneyman 325 retrofit-x-axis-pos-error-jpg

    Second try with attachments...



  2. #14
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2173
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Hi Rick,

    500 counts of error would be 500/20000 = 0.025 inches. I would expect you can do better. What are your requirements? Under what conditions?

    You might add a 2nd order Low pass Filter as the last IIR Filter of 500Hz Q1.4. Then see if you can increase D gain and P gains for less error.

    You might add some I Gain. You will want to increase the Max I to 2000 as well.

    HTH
    Regards

    TK
    http://dynomotion.com


  3. #15
    Registered
    Join Date
    Feb 2011
    Location
    USA
    Posts
    24
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Tom,

    Thanks for the info.

    It's been a couple of decades since my study of S and Z domain math, FIR and IIR filters, so I'm a little rusty.
    I will never forget the brilliance and elegance of Maxwell though.

    Anyway, I implemented the filter as you suggested.
    I have tested numerous values of P, I, and D terms to try and reduce the position error.
    Increasing the P term beyond 4 to 5 results in increased ringing of the output with no benefit to the error. 4 seems to be the highest value for smooth response.
    The D coefficient seems to do very little. Increasing this term above 2 results in an offset between the output 0 and the position 0. I'm not sure if this is a problem or not.
    Adding Integral value seems to only cause more oscillation. I began with 0.0001 and increased. Higher values just oscillate more.

    Adjusting the feed forward values seem to have the biggest impact on reducing the position error.
    I have moved these each up and down. I think I've got the optimum values here.
    The position error is now around 100, instead of 500. Please give me some suggestions on reducing this further.
    100/20,000 = 0.005". If I understand this correctly, this will be 5 mils off during a move at maximum speed?

    As for requirements and conditions, The original controller claimed to be accurate to 0.0001".
    As this is my first and only CNC milling machine, I'm don't know what's normal, but I'd like to get it into a normal range for a good mill.

    Journeyman 325 retrofit-x-axis-pos-error-jpgJourneyman 325 retrofit-x-axis-pos-error-2-jpg



  4. #16
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2173
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Hi Rick,
    Increasing the P term beyond 4 to 5 results in increased ringing of the output with no benefit to the error. 4 seems to be the highest value for smooth response.
    The D coefficient seems to do very little. Increasing this term above 2 results in an offset between the output 0 and the position 0. I'm not sure if this is a problem or not.
    D Gain should help you to increase P Gain with less ringing. D Gain usually needs to be 10-50X higher than the P Gain to have an effect. D Gain will add a dampening effect an will slow down the response (unlike P and I Gain where increasing gains will tend to make the response faster). At some point the D Gain will cause the system to go unstable. Find that point then back off. I don't understand what you mean about an offset. It isn't clear if you added the 2nd Order low Pass filter that I suggested in the last email. It is required to reduce spikes in the output caused by High D Gains.

    Adding Integral value seems to only cause more oscillation. I began with 0.0001 and increased. Higher values just oscillate more.
    Try lower values.

    Adjusting the feed forward values seem to have the biggest impact on reducing the position error.
    I have moved these each up and down. I think I've got the optimum values here.
    It usually best to turn FF off until the feedback gains are optimized. The problem with counting on FF to reduce error is that it is open loop. If anything changes (load, mass, friction, cutting forces) the errors will increase.

    The position error is now around 100, instead of 500. Please give me some suggestions on reducing this further.
    100/20,000 = 0.005". If I understand this correctly, this will be 5 mils off during a move at maximum speed?
    I think your calculations are correct. 60000 counts/sec / 20000 counts/inch = 3ips = 180ipm. At Rapid speeds 5 mills of error may not be an issue. You might test at your max cutting speeds.

    As for requirements and conditions, The original controller claimed to be accurate to 0.0001".
    That's very unlikely.

    HTH
    Regards

    TK
    http://dynomotion.com


  5. #17
    Registered
    Join Date
    Feb 2011
    Location
    USA
    Posts
    24
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Thanks,

    I'll remove the feed forward values and tinker with the PID values some more. I wasn't aware that the D values should be that high.

    Do you have any ballpark targets that I should be trying to get with the following error?



  6. #18
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2173
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Hi Rick,

    Its hard to put a number on error as it is dependent on your system's components and your optimization goals.

    As you can see the error peaks during acceleration. Your acceleration is somewhat high 30 in/sec2 ~ 0.08G. Reducing acceleration is likely to help accuracy while compromising productivity.

    I now notice you are saturating the amplifier (Green plot hits 2000 DAC Counts) during accelerations which is likely the major cause of error. Notice deceleration at the same rate causes little error. It may be that you can change settings in the Amplifier to get a larger range of power.

    Rapids can take advantage of limited Jerk vs the infinity like value you have specified. But that is only useful for rapids not feeds.

    HTH
    Regards

    TK
    http://dynomotion.com


  7. #19
    Registered
    Join Date
    Feb 2011
    Location
    USA
    Posts
    24
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    OK, I think I've got her now...

    As you suggested, I reduced the acceleration. This made a good reduction in the error.
    I also tried reducing the jerk until I could feel a difference in the motion, then raised it back some. This also reduced the error some more.
    Tweaking the feed forward velocity and acceleration down to the ten thousandth, made it even better.
    This error holds as I reduce the velocity from 60,000 to 10,000. Still has an error less than 2.
    This should equate to 0.0001 inch at 20,000 counts per inch.

    When the machine is actually milling, the speeds should be less than 10,000 per second.

    Journeyman 325 retrofit-x-axis-pos-error-3-jpg



  8. #20
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2173
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    WOW!

    TK
    http://dynomotion.com


  9. #21
    Registered
    Join Date
    Feb 2011
    Location
    USA
    Posts
    24
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    After a few hundred tries tweaking and re-tweaking; This is what I have for the axes...

    Journeyman 325 retrofit-x-axis-pos-error-3-jpg Journeyman 325 retrofit-y-axis-pos-error-1-jpg Journeyman 325 retrofit-z-axis-pos-error-2-jpg

    I wasn't able to get the Y or Z quite as accurate as the X.
    I assume this is because of the extra mass / inertia on the Y axis and the asymmetrical force of gravity on the Z axis.
    It's still a total error of half a mil, I think it will be fine.



  10. #22
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2173
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Hi Rick,

    Looks very good to me. Thanks for posting back.

    Of course where the encoder says the table is and where it actually is are two different things.

    Regards

    TK
    http://dynomotion.com


  11. #23
    Registered
    Join Date
    Feb 2011
    Location
    USA
    Posts
    24
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Another question regarding the spindle.

    I have a 5hp spindle motor with a Yakasawa drive.
    The drive uses +-10vdc to turn the spindle forward and reverse, It also ouputs a spindle_0_speed and spindle_at_speed bit to kanalog.
    The motor has feedback to the drive, but not to the Kflop / Kanalog, so I was going to run it open loop.
    I have a handheld tachometer, and have manually set DAC4 in the console to test the spindle operation and measure the actual RPM.
    This all works fine.

    I have configured channel 4 to no input and channel 4 DAC.
    I need to set the parameters for the spindle using the step response screen, but this will spin it one direction and immediately the other direction.
    Is there a way to have it test only one direction?
    I understand I need to use the FF setting to drive the motor, but how should this be calculated?

    Also, I would like to have a program intercept spindle commands, so that a calculation can be made to scale the speed setting to the DAC value.
    Assuming it's linear, DAC values of 0 to 2047 = speeds of 0 to 6250 RPM. a multiplier of 3.05.

    Thanks,
    Rick



  12. #24
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2173
    Downloads
    0
    Uploads
    0

    Default Re: Journeyman 325 retrofit

    Hi Rick,

    For simple open loop DAC control you should be able to use the basic configuration described here:
    KMotionCNC Spindle Control

    But actually that won't allow negative DAC Voltage so if that is needed you should follow the Advanced Configuration using C Programs controlling an Axis. See:
    KMotionCNC Spindle Control.

    HTH
    Regards

    TK
    http://dynomotion.com


Page 2 of 3 FirstFirst 123 LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


About CNCzone.com

    We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!

Follow us on


Our Brands

Journeyman 325 retrofit
Journeyman 325 retrofit