ELK M1 - Output Expander Trouble in High Temp

[Mookmook]

Hi everyone, 
 
I've been a lurker on this forum for 15 years and finally created an account as I have never been so stuck before and would really appreciate any help.
 
I have an ELK M1 with 3 keypads, an input expander, wireless expander, output expander and an ethernet module. They all get home run to an M1DBH in a central wiring closet. The terminator is enabled on the M1 and I have inserted the terminating resistor in to the last RJ45 jack on the DBH. All devices on bus have their terminating resistors disabled. 
 
The system runs perfectly 99% of the time, however when the temp rises dramatically (like high 80s here in the PNW), panel starts troubling out with "Output Expander Lost Comms". I have checked the connections looking for loose grips in the connector blocks, bad RJ45 punch downs, but I am at a loss... 
 
I measured resistance with system off at the DBH across data lines and see a pretty stable 140 Ohms. 
 
Willing to follow any advice for troubleshooting as I am tearing my hair out... 
 
-- Many Thanks
-- David

 

[wkearney99]

Some circuit freeze spray might help you track down what pieces are overheating.  That and a cheap Harbor Freight infrared thermometer to get spot readings.

https://www.harborfreight.com/121-infrared-laser-thermometer-63985.html

https://www.amazon.com/s?k=circuit+freeze+spray
 
Going to be tough to get a baseline on normal temps given the heatwave going through the area now.

A starting point might be where the wiring comes into/leaves the devices.  Check if cooling there changes the error condition.  Read the current temp using the spot thermometer, cool it a bit with the spray, see if the error changes, note the temp.  

I would not just blast a whole box or board. Better to check in spots to help narrow down the problem. I mean, you could chill the whole board on each side of the connection and see if one or the other eliminates the error. But you'd still be back to doing spot tests to find the component.

Could be there's a capacitor or other component that's failing due to thermal expansion from the heat.  Could also be possible solder joints losing contact, again due to thermal expansion.

Don't rule out the grounding as well.

 

[RAL]

Where is the output expander located?  Is the expander or the M1 getting hotter than normal (e.g. say more than 90 degrees ambient temperature)?
 
Bad connection points are the most likely cause.  With RJ45s, you can't just look at them and be sure whether they are good or not.  The best approach is to cut them off and install new ones. 
 
When you measured the 140 ohms resistance, was that with the DBH connected to the M1?   The terminating resistors are 120 ohms, and with two properly installed, you should be seeing something around  60 to 65 ohms.
 
How long is each of the home runs to the DBH?

 

[Mookmook]

Thanks both of you! Can only make 1 post a day so will try to make it count. Haven't had a single expander trouble since the first post probably because ambient temps in PNW are 'normal' now. 
 
I have a temp gun, but that circuit freeze is an awesome idea next time it happens. I redid the wiring at the Output Expander last week and still no change once the heat came in. The expander is located in the Garage in an ELK branded enclosure. I checked the board for any defective/unsealed/leaky capacitors and couldn't see anything apparent. 
 
When I measured 140 ohms across Data A Data B it was at the DBH while everything is connected to the M1 and terminating resistors set. Just M1 turned off when measuring. Each run to the DBH varies - there are 6 + connection of DBH to M1. The longest is 100' and the shortest is 16'. Connection from DBH to M1 is 6'. 
 
Thanks, David

 

[RAL]

140 ohms isn't right. It should be around 65 ohms, so something isn't terminated properly.
 
Your wiring runs are relatively short.  I just wanted to be sure you didn't have a total distance that was getting into the thousands of feet.
 
You said in your first post that you had the terminating plug in the last RJ45 slot of the DBH.  Do you mean that you have it plugged into the J9 position?  That would be incorrect. It should go in the first unused position. So if you have 6 home run cables to the DBH, the terminator should be in J7.  If you have it plugged into the wrong place, that would explain why you don't see 65 ohms.
 
If the DBH terminator is plugged into the correct jack, then you need to figure out why you're not getting the right reading. To determine which terminator isn't working properly, connect your ohm meter to Data A and B, and then remove one terminator at a time and see if the reading changes.  If the reading changes, then that terminator is having an effect.  If it doesn't change, then that's the one that is the problem.
 
An incorrectly terminated data bus won't always cause immediate failures.  Depending on wire lengths and such, it can be just on the edge of not working, and then something like high temperatures can push it over the edge.

 

[Mookmook]

RAL, 
 
You are absolutely correct, even though it's in J7, resistance should be about half of what I was seeing. I'm a bonehead..2 x 120Ohm terminators...duh..  RS485 is versatile, but I think you figured out my problem; termination is the culprit. . 
 
I took out the terminator on J7 and terminated instead at the output expander since it's last in the line. Resistance now reads ~80 ohms instead of ~140 ohms. I tested terminating resistor that was in J7 and it reads 120. 
 
Enrollment found everything and no issues so far. My hunch is that the terminating resistor is going across a pair that is somehow broken since we are using 2 pairs for each bus connection and hence if I find the break or simply terminate the other pair at the DBH, I'd be fine.
 
Since I have no plans to expand, I think I'm going to leave it as-is with the J7 resistor taped to the DBH as a warning to future me. 
 
I'll wait for another hot day to come here in the PNW to see if the horrible output expander lost comms reemerges. 
 
Thank you RAL so much for the pointer... had to read RS485 documentation to see how termination can truly produce intermittent issues like this.

 

[RAL]

I took out the terminator on J7 and terminated instead at the output expander since it's last in the line. Resistance now reads ~80 ohms instead of ~140 ohms. I tested terminating resistor that was in J7 and it reads 120. 
 
Enrollment found everything and no issues so far. My hunch is that the terminating resistor is going across a pair that is somehow broken since we are using 2 pairs for each bus connection and hence if I find the break or simply terminate the other pair at the DBH, I'd be fine.

You're right, it's possible that the return pair from whatever you have plugged into J6 (the output expander?) has a break, and that's why plugging the terminator in J7 doesn't change the resistance.  Either that, or there is an open in the circuit board wiring on the DBH, which seems unlikely.
 
If you are terminating instead at the output expander (and it is what is plugged into J6) I would disconnect the return pair of wires at the expander, so the bus does not have an open length of wire dangling after the terminating resistor, which can create problems, too.  The RS485 bus really works best when the terminators are at the true end, and not somewhere before the end.

 

[Mookmook]

Per your post I set on a mission to fix this and OMG I found the root cause. I attached two cat5 cables in the attic for the last keypad in J5, the return channel for Data A was much higher resistance and when attic gets hot, it acts like an open. It must be a bad crimp up there and I'm not going up for this. Hence why J6 output expander got troubled out. 
 
I stole a spare conductor and repurposed keypad side and DBH side for Data A return and all has been fine. Terminating resistor is back in J7 and all the device terminators are off.
 
Thanks RAL.

 

[RAL]

Glad you found where the problem is!  Hopefully, things will be trouble free now.