THC cannot keep up

More
09 May 2022 23:18 - 09 May 2022 23:20 #242422 by snowgoer540
Replied by snowgoer540 on topic THC cannot keep up
Apologies, I meant max speed.  It's derived from your velocities in your .ini file, and it's the speed the THC operates at.  Your max speed is 600ipm, which is fast.  Your P value for the PID is 25.  Given those two facts, I would expect the torch to oscillate when the THC is active as it is likely over correcting and doing it very quickly.  

Places to start with taming to oscillation would be: lowering the max velocity, lowering the P value (default is 10), or both.

To lower the max velocity for the Z axis to 300 ipm, for example, change the following:

Under [AXIS_Z]
MAX_VELOCITY = 10


Under [JOINT_3]
MAX_VELOCITY = 5
STEPGEN_MAXVEL = 6.25
Last edit: 09 May 2022 23:20 by snowgoer540.
The following user(s) said Thank You: Jalapeno56

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

More
09 May 2022 23:18 #242423 by tommylight
Replied by tommylight on topic THC cannot keep up
You are learning, we were all in the same position at some point in time.
:)
You are welcomed, always, and thank you for reporting back with the solutions.
The following user(s) said Thank You: Jalapeno56

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

More
10 May 2022 00:39 #242442 by Jalapeno56
Replied by Jalapeno56 on topic THC cannot keep up
Snowgoer: I made changes over the weekend after alot of research as ya'll recommended. Even went back to basics on some things.  I attached my new .ini file if you would like to take a look. My max velocity for z axis should be around 200ipm now.  Latency numbers wont allow me much faster and stay reliable....  I am "cutting" with 200 speed and 11 -p value.  Trying to fine tune things, getting better every time!

Tommy: Thank you for the words of encouragement!  I am always happy to share my failures, I have plenty of them, that's how we learn!  I cant tell you how many times I have looked up an issue that I had, only to see the OP never left an end result.... very frustrating.  I hope one day I can help someone else as you all have helped me!!!

Thanks again!
Attachments:

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

More
10 May 2022 00:44 #242443 by snowgoer540
Replied by snowgoer540 on topic THC cannot keep up
I meant to mention this earlier, but forgot...

You can change this line:
KINEMATICS = trivkins coordinates=XYYZ kinstype=BOTH

to this:
KINEMATICS = trivkins coordinates=XYYZ

Nothing earth shattering.

Everything for the Z axis looks good.

I am curious about your comment regarding latency... how bad is it?
 

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

More
10 May 2022 00:53 - 10 May 2022 01:00 #242444 by Jalapeno56
Replied by Jalapeno56 on topic THC cannot keep up
Latency worst case after 15 minutes of running was 24789ns. It was worse than that before I changed out my ethernet cable. I am running an Intel I3. I have had alot of issues with video. I can only get one really old dell monitor to work even after trying to load drivers. I have been thinking about Rods advice on another post (think it was Rod), about installing a video card to get the dependencies off of the motherboard and improve times and maybe be able to get a better monitor! Latency times hang around 11000 doing normal stuff, but for some reason if I switch my KVM switch to my design computer, it really spikes.

And changed the KINEMATICS, Thanks!
Last edit: 10 May 2022 01:00 by Jalapeno56. Reason: clarify wording

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

More
17 May 2022 23:55 - 18 May 2022 02:10 #243225 by Jalapeno56
Replied by Jalapeno56 on topic THC cannot keep up
Time for another update....
I started off with trying to cut a straight line up an incline of approximately 10 degrees over 10 inches. Slowly increased the PID-P value until the THC could keep up for the full length. In my case this number was 49. Then, I tried relatively flat cuts with varying curves and arcs increasing the PID-I value until oscillation was at a minimum and the THC was maintaining a good cut height. My value for PID-I was 39. These 2 links helped me considerably to understand how PIDs work:
www.pid-tuner.com/pid-control/
grauonline.de/alexwww/ardumower/pid/pid.html
I would also recommend cutting various shapes and sizes to put the THC through a good workout in order to fine tune after finding PID values that work.
Another issue I had was not being able to get the torch close enough to the work surface for a good cut without hitting the float switch. This was an oversight on my part. I was using a fine cut tip with a normal deflector. This resulted in incorrect torch heights (all too high). I found fine cut tip deflectors and I cant wait to put them to use!! I am currently running without any deflector (bad, I know) and making fast and clean cuts!

Thank you again everyone for all your help! I also hope my failures can help someone along the way!

Edit:  Something I forgot to mention...  Run Halscope in the background during fine tuning to keep an eye on your pierce and THC delays, also encoder velocity will give you some hints for fine tuning.  They will change as your cut improves and needs to be adjusted accordingly.
Last edit: 18 May 2022 02:10 by Jalapeno56. Reason: Add info for delays
The following user(s) said Thank You: phillc54, tommylight, rodw

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

Moderators: snowgoer540
Time to create page: 0.165 seconds
Powered by Kunena Forum