Linuxcnc with RLS rolin encoder.

More
13 Mar 2020 12:42 #160006 by Roiki
So i'm looking if it's possible to integrate RLS Rolin encoders and linuxcnc. They're incremental encoders and that have SSI/BISS outputs for their counts. Since mesa cards are too slow to count them themselves i was looking if you could bypass that with the serial interface. I was thinking of using 7i74 as the serial card(with 7i76e possibly as the main card).

From what i've read it's possible to use SSI/BISS for encoders but can you do it with incremental counts too? you would then need to convert that number to distance in HAL? And would you want to use a RS422 ic(like MAX22028) on the encoder side since the mesa has differential lines as well?

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

More
13 Mar 2020 13:00 #160009 by tommylight
Mesa can count up to 10MHz per sec. So not slow by any standard.
I have used Serial Absolute Encoders with Mesa 5i25/7i77/7i74 on a retrofit some 3 or 4 years ago, never had a single issue with it, it is an industrial machine weighing in at over 5 tons and in use every day for at least 10 hours a day.
I am sure the SSI will work, not sure for BISS.
Also, most Mesa cards have encoder inputs that can be configured as diferential or single ended.

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

More
13 Mar 2020 19:39 #160070 by Roiki
I didnt mean slow in overall, just in this instance since the encoder has a count frequency of 15Mhz. Also the documentation says the max rare for 7i74 is at 2-3 Mhz for multiple encoders so I got the impression that's not really useful in my case.. I was also planning to make connector boards for the encoders with compatible rj45 pinout for the 7i74 and differential transceivers for better noise immunity in longer cables. I was wondering if this is possible and how it could be done. And also how they can be used inside linuxcnc.

I know absolute encoders work but do incremental ssi encoders? The increment count would need to be converted (in HAL?) to mm for use in a control loop.

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

More
13 Mar 2020 20:16 #160073 by tommylight
PCW might shed more light on that.

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

More
13 Mar 2020 20:25 - 13 Mar 2020 20:25 #160074 by PCW
What count rate do you actually need? That is, what is the resolution
of the scale and the maximum velocity?

It may be that quadrature is OK

Incremental SSI is probably OK, you would treat the SSI position just like
a quadrature encoder position, _but_ I think you lose the ability to home on
index so quadrature may be better in that case
Last edit: 13 Mar 2020 20:25 by PCW.
The following user(s) said Thank You: tommylight

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

More
14 Mar 2020 18:46 #160199 by Roiki
The max the encoder can do is 8192 counts / 2mm, which at 1,82m/sec gives count frequency of 15Mhz. (Which gives 0.244 mikron resolution). That's what id like to use. By dropping the interpolation to 4096 you get 8Mhz max frequency at the same speed. Which should give decent rapids.

I got the impression from the docs that since the multiplexing limits the maximum count rate well below that 10 Mhz.

I'm fine using limit switches for homing so index isn't really necessary. Though the encoders can be configured with variety of index pulses.

Also ic-haus has qudrature counter ics that can count multiple encoders and puput the counts over BISS, could these be also used? www.ichaus.de/MD

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

More
14 Mar 2020 19:00 #160200 by PCW
What kind of machine are you making?
Do you have linear motors?

Otherwise that combination of resolution and velocity doesn't make much sense

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

More
14 Mar 2020 19:10 #160202 by Roiki
I wish I had. :P
Currently it's a pick and place system with belt drives so the speeds can exceed 2m/sec. Though it won't be used that fast.
Its also general in nature for various CNC projects. There's a big cap between 8 and 2 Mhz which gives rapids at 2000 and 500(approx) IPM respectively. While 2000 isn't necessary, 500 is too slow and I wouldn't like to drop accuracy for precision. Around 6 Mhz count rates might be enough but can Mesa boards do that?

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

More
14 Mar 2020 19:35 #160205 by PCW
Not sure how better than 1 or 2 u resolution going to be very useful on belt
driven systems with a scale accuracy of +-40u

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

More
14 Mar 2020 19:41 #160206 by Roiki
It's 40um/m with short range accuracy of < 10um/10mm. They're plenty accurate and in my experience, more so than the specs claim.

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

Moderators: PCWjmelson
Time to create page: 0.093 seconds
Powered by Kunena Forum