Spent most of today troubleshooting QNAN errors on various I/O modules. We mostly saw them with SNAP-AITM-4i modules, but also had the problem with SNAP-OMR6T-C and a couple others. Then other modules (in Point Config in PAC Manager) would show up as AITM-4i modules! Then other EMPTY slots would show up as AITM-4i modules! It was mayhem. (FYI, using SNAP-PAC-RCK-16 with EB1 and EB2 brains – I saw same problem on two racks.)
After doing lots of unplugging and plugging modules back in to see if we had a bad module, I finally screwed all the modules down. Did at least one full download of the strategy and rebooted (turning power off and then on, not just doing a repower I/O unit command from PAC Manager). It now seems to be pretty stable on both racks that were having the problem. Cross your fingers for me – I will update this if it turns out to be something else.
I think part of the problem was stiff thermocouple wire on the AITM-4i’s, and very little clearance in the rack.
It just goes to show you that you don’t need to screw down modules… /until you do/. I hope this saves some of you a few hours!
Screwing the modules into the rack is recommended. But it can be a pain, especially in the SNAP-IT racks where you have to disassemble things to gain access.
I also treat the contact blocks of all modules and racks with DeoxIt D5 as well, and that is extremely helpful.
What would be very handy would be captive screws for the SNAP modules so you didn’t have to fumble trying to get the mounting screws started, potentially dropping them.
I wonder if an add-on captive screw accessory could be designed. It would need to mount entirely on the top side of the module so it doesn’t prevent intimate contact between the bottom of the module and the rack PCB.
Another option is to use nylon 4-40 screws for mounting so that if you do lose a few, at least you don’t have to worry about them shorting things out. You can’t torque them down much, though!
I have a screw starter that I got back in the 1970s that works pretty well for reaching in and starting these screws, but it’s still kind of nerve wracking.
CAPTIVE SCREWS! Genius. How about it, OPTO? Especially on Thermocouple modules, and really any module that is sensitive to being loose on the rack. (Is that all of them? Some seem more sensitive than others.)
I’ll look into the Deoxit D5 – I haven’t heard of that. We might have a screw starter somewhere… Luckily when I drop screws all in my panel, there aren’t any exposed metal contactors below them to short out.
Lesson learned! Our groov EPIC and groov RIO terminal blocks all come with captive screws. I know this won’t help with PAC gear, but for future designs we hope you will consider the groov Family of products. Maybe for your SNAP issue, a magnetic screwdriver could help you from losing screws that don’t hit their mark?
Thanks, @Dan_Opto, but I think the screws that come with all the SNAP products and modules stainless steel? Which is great, except when trying to use a magnetic screwdriver…
Yep. This problem is sometimes seen with the use of 4 channel analog modules, but it’s even more common on high-density modules, especially when the customer makes his own wiring harnesses to “save money”.
The weight of the cabling if not supported can pull down so hard on the module, that the pins can have problems in making full contact with the socket connector on the SNAP-PAC-RCK backplane. As you can see from the enclosed diagram, the address pins are actually closest to the bottom of the connector socket, so if you pull enough on the module, you can see how easy it is with those four little pins to create a “fake” or duplicate module address (0-15) on the ARCnet bus. You can feel the pain of the poor SNAP I/O Processor software trying to cope with this.
The moral of the story is always to use the screws for 4 channel SNAP analog modules and high-density SNAP modules to avoid ARCnet problems on the comms bus. Even if it’s annoying, due to the difficulty of getting those pesky screws inserted, it’s a lot less annoying than tracing QNAN or ARCnet reset errors in a running plant!
Fully agree with @ktgaston, nothing worse than the sound of yet another SNAP module fixing screw dropping down into the pits of your cabinet wiring, never to be seen again. We use something similar to the following, which although a bit crude, does the job as long as you can still get the screwdriver in place.
Use the din rail clips on the rack and then just remove the whole rack and remove or install the screws. I just did a migration where there wasn’t any room to mount everything, so the only way to work on it was to use din mount. Btw, the plastic din mounts are terrible, don’t use them, use the metal ones, they are perfect.
I know most people don’t use the screws, but I have most of the time, at least when project is finished. The screws make it harder for a customer to accidentally apply too much force to the module and pcb connector.
my bad. Didn’t dig deep enough on that one…
