Elk driving me nuts

[MarkJames]

I've had my Elk system in place for at least 15 years and it's been mostly trouble free.  I made some changes with I had to update my M1XRF2G's to M1XRFEGs.
 
I now have a recurring issue that I can't seem to resolve.  I hope someone can help me with this because I've been tearing my hair out.  I've taken apart every keypad (8 of them) and checked the wiring and termination, every M1XRFEG and M1XIN and M1DBH and checked and double checked dbaddr's and termination. It's all exactly right.
 
For some reason (and it just happened again right now) I all of a sudden get a voice message 'Kepad 8 is missing' 'Input expander 5 is missing'.
 
I'm assuming that the input expander it's talking about is an M1XRFEG as that is on dbaddr 5.  It's at the end of the daisy chain just after keypad 8.
 
So something is wrong there but I can't for the life of me figure out what that something might be.
 
Termination is all correct - I've gone over it a number of times.  
 
Any idea where to look next?
 
mark

 

[RAL]

The data bus devices can run into trouble if the power supply voltage drops too low.  One thing to check would be to measure the voltage at keypad 8 and the M1XRFEG.   Compare the voltage readings to what you measure back at the M1. 
 
How long is the wire from the M1 to these devices?  If the wire length to these devices is long, they might be seeing a significant enough voltage drop to cause trouble.
 
Are you running these devices from the M1's power supply?  Or are they powered by an aux supply?
 
How's the back up battery?  Has it been replaced in the last 3 years or so?

 

[MarkJames]

The data run from the M1 is probably betweejn 300 to 400ft running over cat5. I didn't intend to go that far originally.
 
There is a P112 about halfway that supplies 2 M1XRFEG's, and 4 M1KPAS's.
 
When I realized how long this was going to end up I doubled up the wires in the Cat5 to decrease the voltage drop.  Later I added the P112 so I don't think I have power issues.  I'll put my meter on at various spots today to confirm though The length of the data bus lines does concern me.  I for some reason haven't doubled the data lines - just the power - perhaps that might help? I have two remaining unused wires in the cat5 bundle
 
mark

 

[RAL]

The data lines don't need to be doubled. They carry very little current and don't need a heavier gauge wire.
 
The total length of the data bus (assuming you are not using a M1DBHR) can be 4000 feet, but using CAT5 home runs will cut that in half to 2000 feet.  Is the 300-400 ft just to keypad8, or is that your total data bus length to all devices?
 
If you didn't have the NEG terminal of the P112 connected to the NEG/- wire of the data bus, that could be part of the problem.  In theory, the RS485 data bus doesn't need a common ground connection between the devices.  But over longer distances, when you have different power sources, they can cause signalling problems when you don't have a common ground reference.

 

[MarkJames]

I just went out and finished doubling up the data wires just for a 'suspenders and a belt' solution.
 
Ground is tied together all the way through from the M1 all the way down to the last device.  The P112 ground ties in to that along the way.  +12 from the M1 feeds 3 channels on the M1DBHR.  The fourth channel on the DBHR gets power from the P112 down in the yard nearest the first keypad there with no connection back to the M1 on the +12 line - only ground.
 
I checked voltages along the long run (the troublesome one) and I have 13.4V at the M1KPAS nearest the P112 and 12.5V at the furthest point which is an M1XRFEG. On a side note one of the M1XRFEG's has no zones defined for it - hopefully that's not a problem for the M1.
 
I'm just about to reboot and see if the problem recurs.  Last night at about 2am I I got woken to the 'input expander missing' and 'keypad missing' messages.  I had to turn the Elk off.
 
Very frustrating problem
 
Possibly related - possibly not - for some reason I can't program certain addresses into some of my keypads.  The keypad that used to be 6 won't allow me to use 6 anymore.  All devices show in ElkRP with expected addresses and there's no 6 there.  All are accounted for.  It used to be 6 but it won't take it anymore.  Very weird.
 
Thanks for the help!
 
mark

 

[RAL]

Do you have a M1DBH, or a M1DBHR ? Just want to make sure I understand your setup.

 

[mikefamig]

It does sound to me like the devices in question are losing power but I think that it would be worth your time to test the data bus wiring and termination with an ohm meter.
 
A quick way to tell if you have the data bus properly terminated is to turn off all power to the system and connect an ohm meter to the M1 panel with one lead on the DATA A connector and the other lead on the DATA B connector. You should measure 65 ohms +-10 ohms.
 
Mike.

 

[MarkJames]

Do you have a M1DBH, or a M1DBHR ? Just want to make sure I understand your setup.

 
I have an M1DBHR
 
the first bus has 2 runs attached - one leads to a keypad that's terminated.  The other goes to my garage passing through an unterminated M1XIN and finishing at an M1KPAS that is terminated.  The jumper is off for this bus.  Resistance across a/b is 65 ohms
 
bus 2 has 2 runs that go to individual M1KP2 keypads.  Both keypads are terminated.  The jumper is off for this bus.  It has a resistance across A/B of 73 ohms
 
bus 3 has 1 run that goes to an M1XRFEG which is terminated.  The jumper is on for this bus.  Resistance across A/B is 74 ohms
 
bus 4 is the problem one - It has 1 run that goes through an M1KPAS then an M1XRFEG then another M1KPAS then another M1KPAS then another M1KPAS and ends at an M1XRFEG.  The only terminated device is the last one.  Jumper is on for this bus.  Resistance across A/B is 85 ohms
 
Hope that helps and thanks for the input!
 
[edit] I just got the keypad expander missing error again after turning it off and on to test the resistance.
 
mark

 

[MarkJames]

It does sound to me like the devices in question are losing power but I think that it would be worth your time to test the data bus wiring and termination with an ohm meter.
 
A quick way to tell if you have the data bus properly terminated is to turn off all power to the system and connect an ohm meter to the M1 panel with one lead on the DATA A connector and the other lead on the DATA B connector. You should measure 65 ohms +-10 ohms.
 
Mike.

 
Thanks - that's a handy bit of information.  I put the resistances I measured in my reply to the previous post.  They're all in the +- 10 ohm range except the troublesome one which is 85 or so ohms.  That run is quite long - 350/400 ft.  I'm figuring that's why the resistance is a bit higher in that run but maybe it's still a termination issue.  I see from this post https://forums.homeseer.com/forum/legacy-software-plug-ins/legacy-plug-ins/legacy-3rd-party-plug-ins/elk-m1-plug-in/general-m1-discussion/19210-elk-m1-data-bus-trouble-shooting that if I'm over 75 ohms I'm missing a terminating resistor.  I'll go back and look again.
 
mark

 

[RAL]

Your resistance measurements on bus 4 seem high to me.  First thing I would do is check for a bad connection at each end.  Another check would be to disconnect the A&B wires at both ends.  Then short the two wires by twisting them together at one end, and measure the resistance across between the A&B wires at the other end.  For a Cat5 cable of 400 feet, it should be about 10 ohms.
 
Even accounting for the long wire length and its resistance on bus 4, the calculated value you should see with the keypad and M1XRFEG connected and terminated should be 64 ohms vs the nominal value of 60.  So maybe a bit over 70 at the outside. 
 
Another thing you could try is swapping the connections for bus 3 and bus 4 on the M1DBHR and see if the problem moves from bus 4 to bus 3 along with the original bus 4 devices, or whether the problem stays with bus 4.  

 

[MarkJames]

Thanks.
 
What do you mean by a 'bad connection'?  As in not firmly connected to one of the devices on the bus?  I just went down and checked to confirm there is, indeed, a terminating resistor at the end of the line on bus 4.  Perhaps what I should do is go down tomorrow and break the daisy chain along the line checking for 65 ohms to see where my resistance goes wrong.  
 
I've got the diagnostics running on a keypad and in 30 minutes I got 13 data bus errors.  Is that too many?
 
I'll try disconnecting and twisting the ends together to check the resistance like you suggest tomorrow - it's pitch black outside already - damned daylight savings time!
 
I was also going to try removing the final M1KPAS and M1XRFEG by disconnecting them from the M1KPAS before them in the chain and then terminating that device to see if the problem is in the wire that runs to the last building.  I can't get to that one as it runs underwater to a dock.  It's conceivable that it got damaged and I can't check.
 
Thanks for your ongoing help!
 
mark

 

[RAL]

Thanks.
 
What do you mean by a 'bad connection'?  As in not firmly connected to one of the devices on the bus?  I just went down and checked to confirm there is, indeed, a terminating resistor at the end of the line on bus 4.  Perhaps what I should do is go down tomorrow and break the daisy chain along the line checking for 65 ohms to see where my resistance goes wrong.  
 
I've got the diagnostics running on a keypad and in 30 minutes I got 13 data bus errors.  Is that too many?
 
I'll try disconnecting and twisting the ends together to check the resistance like you suggest tomorrow - it's pitch black outside already - damned daylight savings time!
 
I was also going to try removing the final M1KPAS and M1XRFEG by disconnecting them from the M1KPAS before them in the chain and then terminating that device to see if the problem is in the wire that runs to the last building.  I can't get to that one as it runs underwater to a dock.  It's conceivable that it got damaged and I can't check.
 
Thanks for your ongoing help!
 
mark

 
Here's a post from Spanky, the former chief engineer at Elk, that says less than 1 per minute is ok. Personally,  I would consider that high, but maybe it's good enough for the system to keep working.
 
Does your long cable run to the problem devices run outside?  Is it underground?   If it is underground, is it cable that is rated for direct burial and wet locations?   Cable that is not rated for use in wet locations can/will cause problems, even if enclosed in conduit.

 

[MarkJames]

Thanks - from that post it seems like my error count is reasonable.  My last post was about an hour ago and the error count was 64.  It's now 94 - so 1 every 2 minutes.  Seems like a lot to me but within what Spanky was happy with (hard to type Spanky in a sentence lol)
 
My cable is cat5 but it's direct burial rated.  It runs in a conduit all the same though there are a few spots where it is exposed as it enters and leaves buildings.  
 
I'll recheck some more things related to cable resistance tomorrow and see if I've missed something.  In my experience it's always my fault. :-(
 
Mark

 

[RAL]

Direct burial rated cable in conduit should be ok.
 
In my earlier post, I made a mistake when I said that when you twist the A&B wires together and measure the resistance from the other end, it should be about 10 ohms.  It should be about 20.  10 ohms would be the one-way resistance of each wire.  I did take the resistance of both wires into account when I said the terminated bus should be about 64 ohms.  At least I got that part right. 

 

[MarkJames]

So an update...
 
I disconnected the last two devices (an M1KPAS and an M1XRFEG) from the daisy chain that was causing trouble and terminated the chain at the preceding M1KPAS.  That change immediately brought the resistance across the data lines on that bus to 65ohms.  It's been 24hrs now and the 'expander missing' message has not come up.
 
So... looks like the problem is in the last cable run.
 
It goes from a utility building out through a conduit under water to a dock.  I'm guessing the rising and falling of the dock may have damaged the cable.  For now I'm going to leave it alone and deal with it in the spring.  The  alarm area is still active as it's all wireless on the dock.  I can't disarm it from there but I can do it from my phone or from another keypad.
 
Is there a wireless keypad solution available?  I have 110VAC on the dock - a wireless keypad would make for an easy fix.
 
I have one last issue that hopefully someone can offer a suggestion to. 
 
I have an M1KPAS in my garage that has always been device 4 on the bus.  For some reason it will not take 4 anymore.  I can make it 1 and reprogram the 1 device to 4, I can make it 8 or 11 or any other unused number - just not 4.  I've been wondering if I'm programming it incorrectly.  
 
Hold * till keypad flashes
Press F then 1 to display address
Press 4 then Exit to program it to 4
Press F then 1 to verify address
Press Exit when done
 
For some mysterious reason 4 is just 'not an option' lol
 
Thanks for all the help with this!