Build Thread Tree325 Retrofit Started - Page 13


Page 13 of 23 FirstFirst ... 310111213141516 ... LastLast
Results 145 to 156 of 273

Thread: Tree325 Retrofit Started

  1. #145
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2805
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Hi PeterTheWolf,

    Today I loaded the X-axis Table up with about 140lbs of steel and tested the Error base on the current final working X-axis servo "Step Response" Values (P,I,D, V, A, J,) over a the 10,000 count move.
    The Error plot and Command/Position Plots look the same as the plots without the load. However, I did see less dithering at the pulleys (physically oscillating) but the error plots still is hunting at zero.
    I'm surprised the extra mass didn't affect the performance of how the feedforward reduced the error. It isn't clear what you mean by "looks the same". Its hard to see the amount of error on the Position/Command plots. Are you still only getting ~6 counts of error? If so I suppose the change in mass reflected back through the lead screw + belt reduction compared to the motor mass isn't as significant as I expected.

    The few counts of servo dither can easily be affected by mass and friction changes so that isn't surprising.

    I also noticed with a load on the table when I move with a count of 10,000 I no longer get an exact move of .500" on my dial indicator. It moves .493" and then back to zero.
    Can I assume that this will be addressed later?
    Well I'm afraid no. Again is the Error Plot showing small < 6 count errors in movement? If the encoder indicates the move went perfectly (or within 6 encoder counts or 0.3mils) then that is the best the servo can do. It is doing everything it can to move correctly. If the table/part doesn't move correctly then that would indicate some type of mechanical error. Backlash, leadscrew, belt compliance, or the guide-ways are are allowing some sort of roll, pitch, yaw of the table that is load dependent. This is where linear scale encoders right on the table would help but even then there can be roll, pitch, yaw, non-straightness, etc...

    I was able to barrow a Gates Sonic Tension Meter (model 508C) to check the tension on all three servo belts.
    Based on the information I found the X-Axis and Y-axis checked within a couple of Hz; however Z-axis needed adjusting.
    I don't fully understand this but it seems the tensions are reasonable.

    Set#1 - accurate set

    Once again, I am having a hard time discerning what I am looking for in the plots on this test.

    Here is a video on this test ... I am sure my confusion is pretty pronounced. (Zip File 43mb)
    The main thing to look for is the error plot and the maximum error. Your video only shows one plot at V=40000, A=100000. J=infinity where the error has increased now to ~18 counts. That's still less than 1 mill so it might be ok to move on.

    Are we talking about the same thing here?
    Somehow I think I am missing something. I do what to try and hold 0.001" accuracy with a 60-70 ipm feedrate (while cutting) and I understand how 20 counts = 0.001" but this 1 mil precision certainly isn't the same thing ... is it?
    Since we think we have tuned the servo optimally already the only thing we can do now to reduce errors is to slow down the system and/or use less acceleration. Your biggest source of error appears to be with the shock caused by sudden acceleration (when you stomp on the gas). This shouldn't change with top end speed (if you spill your coffee when we stomp on the gas it won't matter if we intend later to back off on the gas when we reach 60MPH or 100MPH). What matters is the acceleration used (how far we stomp the gas pedal down).

    The V=40000 is 120ips so you might want to change that to 20000 if you want to see the results for 60ipm.

    HTH
    Regards

    TK
    http://dynomotion.com


  2. #146
    Registered PeterTheWolf's Avatar
    Join Date
    Aug 2008
    Location
    USA
    Posts
    136
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by m_c View Post
    ......
    In imperial a mil means 1 thou.
    In metric, using it verbally would be short for millimetre.
    Thanks m_c for the clarity on this.
    In all my years working in manufacturing
    /machine shops..... I have never heard anybody refer to 0.001" (one-thousandth) as a mil.
    If units of metric were used .... then milimeters and/or microns were verbally/written accordingly.

    .....
    .....



  3. #147
    Registered PeterTheWolf's Avatar
    Join Date
    Aug 2008
    Location
    USA
    Posts
    136
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Quote Originally Posted by TomKerekes View Post
    Since we think we have tuned the servo optimally already the only thing we can do now to reduce errors is to slow down the system and/or use less acceleration. Your biggest source of error appears to be with the shock caused by sudden acceleration (when you stomp on the gas). This shouldn't change with top end speed (if you spill your coffee when we stomp on the gas it won't matter if we intend later to back off on the gas when we reach 60MPH or 100MPH). What matters is the acceleration used (how far we stomp the gas pedal down).

    The V=40000 is 120ips so you might want to change that to 20000 if you want to see the results for 60ipm.
    ....
    .....

    OK ... I re-read this thread all over again (from about #108) from where I started tuning the X-axis servo. Making it this far brought more understanding of all the previously information Tom shared, which, at the time was not too clear to me.
    So I re-started the tuning of the X-axis servo with all the default setting in the “KanalogInitialPID.mot” file. Step by step do everything over. The new settings I have now I believe produce a stable, accurate enough for me, tuning for the X-axis. When I apply +20% more for the parameters I have “Cutting Values” and "Rapid Values” on the Motion Profile (V, A, J) the system remains stable, which I did not have with the first achieved setting I first came up with.

    Can I please get verification from somebody on my math here?

    For the final motion profile parameters (V, A, J): Accurate Cutting
    V=30,000, A=60,000, J=600,000 which produce an Error of 10 counts.
    So my accuracy should be 10 counts/20,000 counts = 0.0005”
    And
    With V= (30000 counts/sec) * (1in/20,000 counts) = 30K in / 20K sec = (1.5 in/sec) * (60 sec/1min.) = 90 in/min.
    I should be able to cut at 90 in/min with an accuracy of 0.0005” (0.5 mil)

    Based in the next two plots … Am I calculating this correctly?

    Tree325 Retrofit Started-x-axis_cutting_error_graph-png .... Readable Image
    ...
    Tree325 Retrofit Started-x-axis_cuttingcommandposition-jpg .... Readable Image



    For the final motion profile parameters (V, A, J): Rapid Moves
    V=60,000, A=300,000, J=600,000 which produce an Error of 14 counts.
    So my accuracy should be 14 counts/20,000 counts = 0.0007”
    And
    With V= (60,000 counts/sec) * (1in/20,000 counts) = 60K in / 20K sec = (3 in/sec) * (60 sec/1min.) = 180 in/min.
    I should be able to move in Rapid at 180 in/min with an accuracy of 0.0007” (0.7 mil)

    Based in the next two plots … Does this look correct?

    Tree325 Retrofit Started-x-axis_rapid_error_graph-jpg .... Readable Image
    ....
    Tree325 Retrofit Started-x-axis_rapid_commandposition-jpg .... Readable Image

    Here is the Video of Kmotion testing of the above.

    I also have a question concerning the "Max Following Error" setting on the "Configuration" screen.

    Since it looks like I have about 14 count error being the worst error on the X-axis with these parameters, would I make this "Max Following Error" something like 25-30?

    .....

    J325 Quick Links:

    Machine
    Existing Machine Schematics
    Electronic Cabinet-Right Side
    Electronic Cabinet-Back SIde
    Existing Drive Board SD1525-10
    J325 Servo Drive-SD1525 Manual
    3-Phase Rotary Convert Used
    RickB's J325 Retrofit Wiring
    KFLOP 5VDC/15Watt/3A Power Supply
    KANALOG Mounted & Connected
    Kmotion - Axis Encoder Manual Test of Position via Manual Movement
    Kanalog-Encoder Voltage High/Low Checks & 1KOhm Resister
    Kmotion Configuration Screens "RUN-AWAY"
    Tree Journeyman 325 Designed Specs."
    Tree Journeyman 325 Axis-Tension Frequency Settings"

    .....
    .....

    Last edited by PeterTheWolf; 01-07-2018 at 10:20 PM.


  4. #148
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2805
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Hi PeterTheWolf,

    Can I please get verification from somebody on my math here?

    For the final motion profile parameters (V, A, J): Accurate Cutting
    V=30,000, A=60,000, J=600,000 which produce an Error of 10 counts.
    So my accuracy should be 10 counts/20,000 counts = 0.0005”
    And
    With V= (30000 counts/sec) * (1in/20,000 counts) = 30K in / 20K sec = (1.5 in/sec) * (60 sec/1min.) = 90 in/min.
    I should be able to cut at 90 in/min with an accuracy of 0.0005” (0.5 mil)
    I think your math is correct. Of course keep in mind Servo Errors are only one of many errors in a system. The worst case sum of all the errors will be the final accuracy.

    However two things:

    #1 - I see your I Gain is extremely low (1e-7). That is why the 10 counts of error maintains for a full second or more without any correction attempt. The 10 counts of error is still small because of well tuned Feed Forward so it would probably work ok. Like I tried to explain before, I find it best to rely on feedback rather than feed forward. You might think about it like running through a maze blind folded vs feeling the walls to determine your path. If you learn the maze well you can do it blindly accurately and quickly, but if anything changes you'll do badly. Feeling the walls would be more forgiving to any changes and is a more robust approach.

    You might turn off Feed Forwards, turn I gain up to reduce the errors, then Turn Feed Forward back on.

    #2 - when testing cutting you don't have the luxury of setting Jerk to whatever value you wish. Set it to 1000X the Acceleration value to simulate no Jerk limit. In your case where A=60000 then set J=60000000 or 6e7. Then test again. The higher Jerk may cause larger errors.

    For Rapids you can set Jerk to 10X the Acceleration like you have or to whatever works best.

    I also have a question concerning the "Max Following Error" setting on the "Configuration" screen.

    Since it looks like I have about 14 count error being the worst error on the X-axis with these parameters, would I make this "Max Following Error" something like 25-30?
    Yes that might be a good value. It isn't very critical. If you ever get following errors under normal conditions you can then increase it some more.

    Regards

    TK
    http://dynomotion.com


  5. #149
    Registered PeterTheWolf's Avatar
    Join Date
    Aug 2008
    Location
    USA
    Posts
    136
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Quote Originally Posted by TomKerekes View Post
    #1 - I see your I Gain is extremely low (1e-7).
    You might turn off Feed Forwards, turn I gain up to reduce the errors, then Turn Feed Forward back on. ....
    Ok, I started tuning the Y-axis and I have to say the Y-axis is tuning so much more easier than the X-axis. It makes me think that maybe the point "mmurray70" brought up on page #131 of this thread may be valid. I certainly do not have the knowledge and/or experience to answer this correctly; however, I am sure someone on this forum would know. The SD-1525 driver board provides for multiply Potentiometer setting for each axis according to the manual (see Quick links below). I am wondering if it would pay to try and adjust this setting to help get better tuning out of Kanalog on my X-axis?

    As it stands I believe this is the best I can get out of the re-tuned X-axis after re-visiting and raising I-Gain I had. Tom, Can I please get you to once more look at these X-axis plots video for feedback?

    Video of tuned X-axis (with out Weight on table) (ZIP file 14mb)

    Video of tuned X-axis with 125lbs on the table (Zip file 23mb)



    Here are the videos of the first tuning of the P,D,& I-Gain on the Y-axis which, as I said above, tuned so much better than the X-axis. The second video is the with the D-Gain and added filter. I just did not see any improvement with the D-Gain. In fact, I left it at "2" because if I increased it any more I just got more error.

    Am I interpreting these plots correctly?

    ....
    ....
    Video -- P-Gain Tuning on the Y-Axis (Zip file 45mb)

    Video -- D-Gain Tuning on the Y-axis (ZIP file 23mb)

    Video -- Final Tuning of the Y-axis with P, D, I Gain & Feed-Forward (Zip file 9.6mb)


    Here is a video of the final tuned Cutting Parameter for the Y-axis (ZIP file 8.3mb)

    Here is a video of the final tuned Rapid Parameter for the Y-axis (Zip file 6.2mb)

    .....
    .....

    J325 Quick Links:

    Machine
    Existing Machine Schematics
    Electronic Cabinet-Right Side
    Electronic Cabinet-Back SIde
    Existing Drive Board SD1525-10
    J325 Servo Drive-SD1525 Manual
    3-Phase Rotary Convert Used
    RickB's J325 Retrofit Wiring
    KFLOP 5VDC/15Watt/3A Power Supply
    KANALOG Mounted & Connected
    Kmotion - Axis Encoder Manual Test of Position via Manual Movement
    Kanalog-Encoder Voltage High/Low Checks & 1KOhm Resister
    Kmotion Configuration Screens "RUN-AWAY"
    Tree Journeyman 325 Designed Specs."
    Tree Journeyman 325 Axis-Tension Frequency Settings"

    .....
    .....



  6. #150
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2805
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Hi PeterTheWolf,

    The Y tuning looks great. I don't know why Y is different from X. It is likely to be mechanical differences - mass, balance, stiffness, compliance, etc... Or I suppose it could be amplifier tuning. Again I'm not familiar with those amplifiers and I worry if you start adjusting those 6 potentiometers if it messes things up it might be hard to get back to where you were. I suppose you could record where they all are using an ohm meter. Or make careful changes to one pot at a time in a way that you can set it back if necessary. I'd suggest leaving things as they are and coming back to it only if it turn out to be necessary.

    Its great to see you now grasp so many of the fundamental concepts.

    But I think you are missing the point on D Gain. In the video you start adding some small D Gain and the error increases slightly and you think it is a bad thing and you stop, D gain by itself is like dragging the brakes or submersing the machine in honey. It tends to kill oscillations and makes the system more stable. But this by itself will tend to slow down the system, decrease performance, and increase errors. But then it is likely that the P Gain can be increased to increase performance to a point higher than was possible previously and still have a stable system.

    The D Gain tends to be a value ~ 10X times or more of the P Gain to have a significant effect. At some point it will make the system unstable.

    With high D Gain a Low pass filter is usually also required to reduce spikes in the output. Your Videos didn't show your Filter settings.

    But with all that being said I doubt if it would make much difference in your system. Your results are already quite good and your aplifier is probably already doing a good job of adding dampening with the Tachometer feedback.

    Regards

    TK
    http://dynomotion.com


  7. #151
    Registered
    Join Date
    May 2012
    Location
    canada
    Posts
    270
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Looks good. If the machine is powered down can you turn the pulleys by hand easily? Do the X and Y feel the same? Seems to be less oscillation with the Y axis for sure. As Tom said its probably not be worth messing with. But the possibility of getting the X a little better by adjusting the amp would be very tempting for me.



  8. #152
    Registered PeterTheWolf's Avatar
    Join Date
    Aug 2008
    Location
    USA
    Posts
    136
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Quote Originally Posted by TomKerekes View Post
    The D Gain tends to be a value ~ 10X times or more of the P Gain to have a significant effect. At some point it will make the system unstable.
    With high D Gain a Low pass filter is usually also required to reduce spikes in the output. Your Videos didn't show your Filter settings.
    Yep, had Low Pass 2nd filter on "The second video is the with the D-Gain and added filter" (#149) I also increased the D-gain on Y-axis to D=20.

    I also tuned the Z-axis and I believe I now have all the three axis tuned.

    Video of the Z-axis Plot of tuned , cutting parameters, and rapid parameters. (Zip file 55mb)


    Tom, Can you please advise as to the the next step I need to take with KFLOP/Kanalog?

    .....
    .....

    J325 Quick Links:

    Machine
    Existing Machine Schematics
    Electronic Cabinet-Right Side
    Electronic Cabinet-Back SIde
    Existing Drive Board SD1525-10
    J325 Servo Drive-SD1525 Manual
    3-Phase Rotary Convert Used
    RickB's J325 Retrofit Wiring
    KFLOP 5VDC/15Watt/3A Power Supply
    KANALOG Mounted & Connected
    Kmotion - Axis Encoder Manual Test of Position via Manual Movement
    Kanalog-Encoder Voltage High/Low Checks & 1KOhm Resister
    Kmotion Configuration Screens "RUN-AWAY"
    Tree Journeyman 325 Designed Specs."
    Tree Journeyman 325 Axis-Tension Frequency Settings"

    .....
    .....



  9. #153
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2805
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Hi PeterTheWolf,

    That all looks good. Just one minor detail I wanted to point out mainly for others. A common mistake is to not test your settings at full speed. Here is an example where max Velocity is obtained:
    Tree325 Retrofit Started-maxvelocityachieved-png


    Here is an example where Max Velocity is not tested. Because the plot above needs about 800 DAC counts to go half the speed. We should see about 1600 DAC counts to go at full speed. Your test size is too short to reach full speed. A quick look at the Velocity Plot would also show this.
    Tree325 Retrofit Started-nomaxvelocity-png

    I have no doubt your settings will work fine on longer moves so there is no need to go back and do more tests. I only point it out because I find this to be a common mistake where the max velocity is set way too high but a big enough move is never tested to show the problem.


    Lets move on!

    I'll assume you have set reasonable Max Following Errors. If not do so.

    #1 create an Initialization C Program with all your Rapid settings. See:
    KFLOP C Programs - Dynomotion
    Note you might open the Init3Analog.c example and simply merge all your screen settings into it. There is a button on the Config/Flash screen for this.

    #2 Assign the Initialization C Program to a User Button in KMotionCNC
    Tool Setup Screen User Buttons

    #3 Configure KMotionCNC for your XYZ Resolutions
    Tool Setup Trajectory Planner

    #4 Convert your cutting Velocities and Accelerations to inch units (divide by resolution) and set the axis parameters for cutting as well.

    #5 Set reasonable Jog Speeds and step increments for your system
    Tool Setup Trajectory Planner

    #6 test Jogging and GCode G0 G1 Moves

    Regards

    TK
    http://dynomotion.com


  10. #154
    Registered
    Join Date
    Feb 2010
    Location
    USA
    Posts
    10
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    3 phase power is not rare in the US, but it is only available to customers whose property is zoned industrial. A person cannot just decide to have 3 phase power brought into their residence, it isn't allowed. Most machine tools run more efficiently on 3 phase power and almost all new machine tools of any size are purchased for an industrial shop, so they are designed for 3 phase.
    As a result, most people who want a machine tool need to convert it from 220 3 phase to 220 single phase (expensive), or use a phase converter.



  11. #155
    Registered PeterTheWolf's Avatar
    Join Date
    Aug 2008
    Location
    USA
    Posts
    136
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Quote Originally Posted by TomKerekes View Post
    Hi PeterTheWolf,

    #3 Configure KMotionCNC for your XYZ Resolutions
    Tool Setup Trajectory Planner

    Regards
    Question:

    In the documentation on Trajectory Planner it reads:

    "The third parameter is the maximum allowed acceleration for the axis in inches/sec2. The G Code Language has no provisions for specifying acceleration rates. Therefore the acceleration (and deceleration) along a vector used will always be the largest acceleration such that each axis's acceleration is at or below the specified limit."

    Is this calculated on the Cutting Parameters or the Rapid Parameters established in the testing of the servos parameters?

    .....
    .....

    J325 Quick Links:

    Machine
    Existing Machine Schematics
    Electronic Cabinet-Right Side
    Electronic Cabinet-Back SIde
    Existing Drive Board SD1525-10
    J325 Servo Drive-SD1525 Manual
    3-Phase Rotary Convert Used
    RickB's J325 Retrofit Wiring
    KFLOP 5VDC/15Watt/3A Power Supply
    KANALOG Mounted & Connected
    Kmotion - Axis Encoder Manual Test of Position via Manual Movement
    Kanalog-Encoder Voltage High/Low Checks & 1KOhm Resister
    Kmotion Configuration Screens "RUN-AWAY"
    Tree Journeyman 325 Designed Specs."
    Tree Journeyman 325 Axis-Tension Frequency Settings"

    .....
    .....



  12. #156
    Gold Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    2805
    Downloads
    0
    Uploads
    0

    Default Re: Tree325 Retrofit Started

    Hi PeterTheWolf,

    Question:

    In the documentation on Trajectory Planner it reads:

    "The third parameter is the maximum allowed acceleration for the axis in inches/sec2. The G Code Language has no provisions for specifying acceleration rates. Therefore the acceleration (and deceleration) along a vector used will always be the largest acceleration such that each axis's acceleration is at or below the specified limit."

    Is this calculated on the Cutting Parameters or the Rapid Parameters established in the testing of the servos parameters?
    On this page - the Tool Setup | Trajectory Planner the "cutting" parameters should be entered. These parameters will limit the Velocity and Acceleration when feeding with G1 G2 G3 codes. They will not be used for G0 Rapid commands. For Rapids the Velocity, Acceleration, and Jerk settings in KFLOP will be used instead.

    HTH
    Regards

    TK
    http://dynomotion.com


Page 13 of 23 FirstFirst ... 310111213141516 ... 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

Tree325 Retrofit Started

Tree325 Retrofit Started