Another plasma component...
- tommylight
-
- Away
- Moderator
-
Less
More
- Posts: 17756
- Thank you received: 5901
17 Feb 2019 00:51 #126640
by tommylight
Replied by tommylight on topic Another plasma component...
Well, got it working with a mesa 5i25/7i77 and noticed some things :
twopass is very picky about where what is, and sometimes things just do not work, and no error is reported. The Gui works but the links in hal fail. Noticed that way earlier but did not have time to investigate, still do not know what is going on, but i can bet that it has to do with the order of things to load.
Reverse run runs only forward ! This was working on previous versions.
All the THC related stuff enabled as soon as Linuxcnc starts, even while in E-stop, has to be changed and some more interlocks added, although they are not explicitly needed.
Attached the config is someone wants to have a go at it. Velocities are way to high, scale at 40 steps/mm for X and Y, 500 s/mm for Z.
I have to attach some hardware to it very soon.
twopass is very picky about where what is, and sometimes things just do not work, and no error is reported. The Gui works but the links in hal fail. Noticed that way earlier but did not have time to investigate, still do not know what is going on, but i can bet that it has to do with the order of things to load.
Reverse run runs only forward ! This was working on previous versions.
All the THC related stuff enabled as soon as Linuxcnc starts, even while in E-stop, has to be changed and some more interlocks added, although they are not explicitly needed.
Attached the config is someone wants to have a go at it. Velocities are way to high, scale at 40 steps/mm for X and Y, 500 s/mm for Z.
I have to attach some hardware to it very soon.
Please Log in or Create an account to join the conversation.
17 Feb 2019 01:49 #126642
by rodw
Replied by rodw on topic Another plasma component...
Tommy,
Your comments on TWOPASS ring a bell with me as I had similar issues taking a sim to a live machine so I commented out the twopass command and things still seemed to work. But that was with a version of Master quite a few months ago. I thought I saw a bug report about this.
Anyway great progress!!
Your comments on TWOPASS ring a bell with me as I had similar issues taking a sim to a live machine so I commented out the twopass command and things still seemed to work. But that was with a version of Master quite a few months ago. I thought I saw a bug report about this.
Anyway great progress!!
Please Log in or Create an account to join the conversation.
- tommylight
-
- Away
- Moderator
-
Less
More
- Posts: 17756
- Thank you received: 5901
17 Feb 2019 02:18 #126644
by tommylight
Replied by tommylight on topic Another plasma component...
Another thing that has to be changed in the ini file is the steplength and stepspace. As it is set a 1 and 0, Mesa stepgens do not like it and remain high all the time. After setting them to a more reasonable 8000 it works properly.
Encoder has to be set in the correct mode ( it already is in the config attached ) and it is set to read the encoder velocity on encoder number 5. That means the output from THCAD has to be wired to the A and /A inputs on the spindle encoder on the 7i77.
@PCW, if you are reading this, when i set the THCAD to 1/1 i do not get any reading form the 7i77 even at 0V input? That seems odd, or am i missing something ? I have the 300 variety of it. Still did not test it at higher voltages, i need a 300V/10mA power source! I have 2 tha so 60V at 2A so that will have to do for now. Hmmm i do have some that do much more driving some Helium-Neon lasers, but they most certainly do much much more, so it's a no go! Mains i can choose between 220V and 380V but i am not going there. Will have to dig out something usable in the workshop.
Later.
Encoder has to be set in the correct mode ( it already is in the config attached ) and it is set to read the encoder velocity on encoder number 5. That means the output from THCAD has to be wired to the A and /A inputs on the spindle encoder on the 7i77.
@PCW, if you are reading this, when i set the THCAD to 1/1 i do not get any reading form the 7i77 even at 0V input? That seems odd, or am i missing something ? I have the 300 variety of it. Still did not test it at higher voltages, i need a 300V/10mA power source! I have 2 tha so 60V at 2A so that will have to do for now. Hmmm i do have some that do much more driving some Helium-Neon lasers, but they most certainly do much much more, so it's a no go! Mains i can choose between 220V and 380V but i am not going there. Will have to dig out something usable in the workshop.
Later.
Please Log in or Create an account to join the conversation.
17 Feb 2019 02:47 - 17 Feb 2019 02:51 #126646
by rodw
Replied by rodw on topic Another plasma component...
Tommy, from my 7i76e config.
FOUT+ to ENCA+
FOUT- to ENCA-
Steal some 5 volt power from a stepgen or something for +5v and -5V
I have my THCAD st to 1/32 division as PCW says its better for LinuxCNC. I think it also gives more time between counts to get a better sample. Remember to change your scale numbers to agree with 1/32.
So with that done, jump into halshow and drill down to the 7i77 encoder input. Watch it and you should see random counting around the Zero volt frequency.
I found a couple of 48 volt DC power supplies wired in series (96 volts) gave me a good test signal close to cutting volts.
Good luck!
FOUT+ to ENCA+
FOUT- to ENCA-
Steal some 5 volt power from a stepgen or something for +5v and -5V
I have my THCAD st to 1/32 division as PCW says its better for LinuxCNC. I think it also gives more time between counts to get a better sample. Remember to change your scale numbers to agree with 1/32.
So with that done, jump into halshow and drill down to the 7i77 encoder input. Watch it and you should see random counting around the Zero volt frequency.
I found a couple of 48 volt DC power supplies wired in series (96 volts) gave me a good test signal close to cutting volts.
Good luck!
Last edit: 17 Feb 2019 02:51 by rodw.
Please Log in or Create an account to join the conversation.
17 Feb 2019 02:50 #126647
by rodw
Replied by rodw on topic Another plasma component...
Forgot to mention that PCW said it had to be on Encoder A as when I blew up my 7i76e inputs, he built different firmware that put ENCA on ENCB connectors. Very nice if him!
Please Log in or Create an account to join the conversation.
17 Feb 2019 03:04 #126649
by rodw
Replied by rodw on topic Another plasma component...
Tommy
You have
I have
I think the mode is just for MPG though. Check what it says about the position of the W links on the board....
You have
sserial_port_0=00xxxI have
sserial_port_0=2xxxxI think the mode is just for MPG though. Check what it says about the position of the W links on the board....
The following user(s) said Thank You: tommylight
Please Log in or Create an account to join the conversation.
17 Feb 2019 04:32 #126657
by phillc54
Can you do man plasmac and see if V0.18 is the latest?
Reverse does work on the sim here.
What THC stuff is enabled that shouldn't be, or do you mean the config items in the left hand panel?
Cheers, Phill.
For some strange reason I am not getting any emails when this thread is written to.
I do have the subscribed box checked...
Replied by phillc54 on topic Another plasma component...
Well, got it working with a mesa 5i25/7i77 and noticed some things :
twopass is very picky about where what is, and sometimes things just do not work, and no error is reported. The Gui works but the links in hal fail. Noticed that way earlier but did not have time to investigate, still do not know what is going on, but i can bet that it has to do with the order of things to load.
Reverse run runs only forward ! This was working on previous versions.
All the THC related stuff enabled as soon as Linuxcnc starts, even while in E-stop, has to be changed and some more interlocks added, although they are not explicitly needed.
Attached the config is someone wants to have a go at it. Velocities are way to high, scale at 40 steps/mm for X and Y, 500 s/mm for Z.
I have to attach some hardware to it very soon.
Can you do man plasmac and see if V0.18 is the latest?
Reverse does work on the sim here.
What THC stuff is enabled that shouldn't be, or do you mean the config items in the left hand panel?
Cheers, Phill.
For some strange reason I am not getting any emails when this thread is written to.
I do have the subscribed box checked...
The following user(s) said Thank You: tommylight
Please Log in or Create an account to join the conversation.
17 Feb 2019 15:43 - 17 Feb 2019 15:43 #126674
by PCW
7Ii77 muxed encoders have a default sample frequency of 6 MHz, if you have the filter enabled it requires 15 counts of this 6 MHz clock to be recognized (2.5 usec). Since the THCAD outputs 1 usec pulses in 1-1 mode these will not get past the input filter.
There are a couple different ways to fix this, one is to disable the encoder filter, the other is to slow down the THCAD output frequency
Using a slower output frequency has the advantage of better noise immunity (you do need to scale your scaling numbers when you do this)
Note that you do not lose resolution when you use the divided frequency output because the encoder counter does velocity estimation
based on both counts per unit time and measuring then pulse period
Replied by PCW on topic Another plasma component...
@PCW, if you are reading this, when i set the THCAD to 1/1 i do not get any reading form the 7i77 even at 0V input? That seems odd, or am i missing something ? I have the 300 variety of it. Still did not test it at higher voltages, i need a 300V/10mA power source! I have 2 tha so 60V at 2A so that will have to do for now. Hmmm i do have some that do much more driving some Helium-Neon lasers, but they most certainly do much much more, so it's a no go! Mains i can choose between 220V and 380V but i am not going there. Will have to dig out something usable in the workshop.
Later.
7Ii77 muxed encoders have a default sample frequency of 6 MHz, if you have the filter enabled it requires 15 counts of this 6 MHz clock to be recognized (2.5 usec). Since the THCAD outputs 1 usec pulses in 1-1 mode these will not get past the input filter.
There are a couple different ways to fix this, one is to disable the encoder filter, the other is to slow down the THCAD output frequency
Using a slower output frequency has the advantage of better noise immunity (you do need to scale your scaling numbers when you do this)
Note that you do not lose resolution when you use the divided frequency output because the encoder counter does velocity estimation
based on both counts per unit time and measuring then pulse period
Last edit: 17 Feb 2019 15:43 by PCW. Reason: sp
The following user(s) said Thank You: tommylight
Please Log in or Create an account to join the conversation.
- tommylight
-
- Away
- Moderator
-
Less
More
- Posts: 17756
- Thank you received: 5901
17 Feb 2019 18:26 #126681
by tommylight
Replied by tommylight on topic Another plasma component...
@Rod,
i set mine in mode 3 always for experimenting so i can use analog inputs also.
@Phill
Yup i have the version 0.18.
@PCW,
thank you, that explains a lot.
Thank you all.
i set mine in mode 3 always for experimenting so i can use analog inputs also.
@Phill
Yup i have the version 0.18.
@PCW,
thank you, that explains a lot.
Thank you all.
The following user(s) said Thank You: phillc54
Please Log in or Create an account to join the conversation.
- tommylight
-
- Away
- Moderator
-
Less
More
- Posts: 17756
- Thank you received: 5901
17 Feb 2019 21:04 #126693
by tommylight
Replied by tommylight on topic Another plasma component...
Added some more configs to the original threat, a full servo system, and with THCAD
Just in case:
forum.linuxcnc.org/plasma-laser/34978-th...-configs-with-how-to
Just in case:
forum.linuxcnc.org/plasma-laser/34978-th...-configs-with-how-to
Please Log in or Create an account to join the conversation.
Moderators: snowgoer540
Time to create page: 0.162 seconds