Host Engineering Forum

General Category => CTRIO and CTRIO2 => Topic started by: DBeck on March 19, 2013, 12:39:48 AM

Title: CTRIO counter increment error at word boundary?
Post by: DBeck on March 19, 2013, 12:39:48 AM

Hello,

I apologize in advance if I failed to search the forum effectively for an answer to the unexpected behavior I'm seeing with a CTRIO (2.1.18 firmware).

I'm using the capture input of a counter and the values being captured indicate that the counter isn't correctly handling the capture of a count that crosses a signed 16 bit boundary of the low word of the counter.

Here's the event sequence of an example I've logged:

0) A stream of correct captured values
1) Another correct value, high word=0x0BA1, low word=0x779D
2) Next expected captured value should be, high word=0x0BA1, low word=0x8016 (i.e. the counter was expected to increment by 2169 (0x0879)
3) Actual captured value is, high word=0x0BA0, low word=0x8016 (oddly, the high word decremented by one)

I saw a second example using these values:
Pre-error event, hw=0x0CCB, lw=0x779E
Expected/correct capture value, hw=0x0CCB, lw=0x8017 (again the increment between captures is supposed to be 2169)
Error result, hw=0x0CCA, lw=0x8017

In between these two anomalous events, meaning every 108.5 ms as the increment of 2169 occurs, the capture of the counter produces the correct value.  To elaborate, the next captured value after the erroneous one, has the expected value as if the sequence of 2169 increments had been correct all along.

Thank you in advance for providing any insight into the cause of this observation.

David Beck

Title: Re: CTRIO counter increment error at word boundary?
Post by: Greg on March 26, 2013, 02:18:25 PM

David Beck, I have tested this extensively in my lab and I cannot duplicate what you are seeing. I am using a DL260 and an H2-CTRIO with firmware v2.1.18. The only thing I can tell you is to try and "jiggle all the handles", which just means reset everything that is possible (e.g. PLC clear & reload, CTRIO clear & reload). Let us know if this doesn't work.

Has this system been running fine for some time and just start doing this? Or is this a startup? If it is a startup, then, of course, make sure there are no overlaps in your CTRIO mapping and program elements. A simple one-bit change could cause the symptom you are seeing.

I have, at various times in days gone by, seen bizarre behavior similar to something like this but I have never been able to pin anything like that down. And it happens so rarely that I've never heard it reported from the field.