ELK M1 GOLD KEYPAD

[lemahoney08]

I bought an ELK M1 GOLD with the standard keypad. I also added an additional touchpad keypad. The touchpad for the house, the keypad for the garage entrance. My question is, do I run the cat5 from the elk box to each keypad, or from elk box to the first keypad and then on to the next keypad?

 

[torgrimj]

I believe you can do it either way.  I did all home run wiring for my keypads.  You just have to ensure proper termination, or you will have issues.

 

[tadr]

Homerun (from the box to each keypad) is best practice and reduces the likelihood of problems.

 

[Del91574]

Homerun (from the box to each keypad) is best practice and reduces the likelihood of problems.

Not always and can't be said as a blanket statement.
 
You need to consider if there's other bus devices to be installed in the field or on the same cable. Actually, a physical daisy would be the preferred method, with only a single exception: power must be separate. It actually makes it easier to troubleshoot a bus problem if you keep the addresses straight as well (address X has issues but Y does not). It also cuts down the overall and total resistance and voltage drop on the data A/B pair compared to homerunning.
 
The reason is the M1's bus is going to be limited to 4000' total length, so when you homerun devices, especially on a larger M1 install, you're going to cut into the total length significantly.
 
That said, in most installs, most people are going to (or hopefully think about) pulling enough conductors to the keypad locations.

 

[lemahoney08]

This will be a smaller install in a smaller apartment, just getting my feet wet on a more simple setup. So I will homerun each keypad. By proper termination, you mean setting the jumpers on the M1 board correct? nothing special needed at each of the keypads?

 

[lemahoney08]

Also one other question before I drywall, is the one speaker that comes with the M1 enough to transmit the sound over 700 sq ft? 
 
I was looking at the IntraSonic I600-BASICPAC Music Distribution Package. Is there a way to get the elk sound output to be an input into this system if the stock speaker is not enough? I assumed that you would have to leave the system on some auxiliary input at all times in order for it to output whenever the elk needs to say something. 

 

[lemahoney08]

So to confirm that I have all the wiring ran that is needed. Here is a description of my system:
 
ISY994i + ELK M1 running some insteon switches for each room. Then 3 door sensors, two motion detectors, and 2 keypads. 
 
The switches are wired as normal, so I am fairly sure everything there is handled. 
I ran 18/2 from each of all three doors back to the panel for the door contact sensors.
I ran cat5e to each of the keypads back to the panel.
I ran cat5e to each of the motion detectors back to the panel.
 
Does anything seem like it is missing or needed? Do I need a special wire for the speaker or does it connect to they keypad cat5 wire? 

 

[Del91574]

Speaker should be handled separately from the Cat X. While it *can* work, it's usually undersized and not in your best interest to do so.

 

[Neurorad]

I've never heard of Intrasonic. Is that unit good?

I try to avoid hardware that hasn't been fully 'vetted', i.e. multiple trusted reviews online.

 

[crossbar]

One thing to keep in mind with Cat cable and the M1 bus (Keypads, input / output expanders, etc ) is voltage drop for the power.  While the RS-485 bus has a maximum length of 4000' the voltage drop for the power will be much more limiting.  Take the KP keypad for example at 0.16 A. 13.8V at the panel will drop to  9.89V at 600' (with Cat 6 23 AWG cable).  You can use the extra pair of wires for power as well to double the length.  So Home runs are also better for power because each home run gets power from the box and there is only one load on the wire.  Daisy chains will have all the loads on the same wires and will be longer wires for power (in most cases).  While this will indeed increase the length of the bus substantially the extra length of the bus should not be too much of an issue as longs as you are well below the 4000'.   Also with home runs you can always upgrade to a  ELK-M1DBHR if you need to.
 
Update : This example is actually best case and you need to design around worst case which is the M1 running on Battery.  The M1 will run on battery down to 10.2V so this should be the value used in any calculations.  Check with the specific devices for their lowest operating voltage.  Again I do not endorse parallel power conductors if you choose to do that make sure you understand what you are doing.

 

[Del91574]

One thing to keep in mind with Cat cable and the M1 bus (Keypads, input / output expanders, etc ) is voltage drop for the power.  While the RS-485 bus has a maximum length of 4000' the voltage drop for the power will be much more limiting.  Take the KP keypad for example at 0.16 A. 13.8V at the panel will drop to  9.89V at 600' (with Cat 6 23 AWG cable).  You can use the extra pair of wires for power as well to double the length.  So Home runs are also better for power because each home run gets power from the box and there is only one load on the wire.  Daisy chains will have all the loads on the same wires and will be longer wires for power (in most cases).  While this will indeed increase the length of the bus substantially the extra length of the bus should not be too much of an issue as longs as you are well below the 4000'.   Also with home runs you can always upgrade to a  ELK-M1DBHR if you need to.

Exactly which was what was said in simpler terms in post #4.
 
However, you do NOT want to get into the practice of doubling up pairs of undersized conductors to make up for an improper size AWG for the voltage drop or ampacity. This is a poor practice to get into, let alone suggest. Basically, it's putting the load in parallel, which is a no-no by code and as a secondary issue, it can also contribute to it's own host of problems.
 
Keep in mind, the 4000' is end to end, but once you do a feed and return topology, it is cut in half, so 2K' end to end is easily met in most common installations, so either a topology or wiring method change is needed (be it via a retrofit hub or daisy) but the retrofit hub also has it's own items to consider before saying it's a fix-all solution.

 

[crossbar]

Exactly which was what was said in simpler terms in post #4.
 
However, you do NOT want to get into the practice of doubling up pairs of undersized conductors to make up for an improper size AWG for the voltage drop or ampacity. This is a poor practice to get into, let alone suggest. Basically, it's putting the load in parallel, which is a no-no by code and as a secondary issue, it can also contribute to it's own host of problems.
 
Keep in mind, the 4000' is end to end, but once you do a feed and return topology, it is cut in half, so 2K' end to end is easily met in most common installations, so either a topology or wiring method change is needed (be it via a retrofit hub or daisy) but the retrofit hub also has it's own items to consider before saying it's a fix-all solution.

Post #4 refers to voltage drop of the Data bus not the power bus.  I was pointing out that in most cases you will run into issues with voltage drop on the power bus before the data bus.
 
Actually code is a non issue here. But if you are talking NEC power code that is exactly how they run 200A+ service panels (with parallel conductors).  Not that I was recommending it just pointing out that is something that can be done.   I would be curious to know what the other host of problems that might contribute to?
I too would recommend against it and run an additional properly sized cable.
 
Again the M1DBHR is just an option if bus length is an issue.

 

[Del91574]

I didn't address the specific item that power must be separated and not daisy chained, as that is clearly stated in the M1 documentation regarding the bus devices themselves.
 
Parallel conductors on an AC installation is far different than a DC LV installation, which has it's own set of rules that must be met.
 
Actually, the NEC forbids the parallel installation method on LV. I'd have to dig up the specific articles and notes, but it's in there.
 
In the specific case of doubling up a pair of conductors on a Cat X or DC power circuit, if you lose one, you lose your ampacity and current carrying ability. So if you have a broken or nicked conductor (we're talking solid Cat X here) you're going to have either an undervolt or under current situation which will get compounded as the conductor ages or if there is any corrosion that occurs. Once you have voltage issues, you're talking a lot of issues that become difficult to truly diagnose, especially with an M1 or other hardware that reboots upon low voltage. The other part, while someone is going to argue circular mils and the cross sectional dimensions, a pair of undersized conductors do not equal a single larger conductor.
 
The DBHR does not increase the total length available for the M1's data. 4000' end to end. Only changes the physical topology of the bus so that daisy chaining is not necessary.

 

[crossbar]

My example above is actually best case and you need to design around worst case which is the M1 running on Battery.  The M1 will run on battery down to 10.2V so this should be the value used in any calculations.  Check with the specific devices for their lowest operating voltage.
 
According to the M1DBHR manual each of the 4 branches can be up to 4000' in length.   It basically has 4 RS-485 buses that it bridges together with the main bus so you end up with 4 buses instead of just 1.  And you can add two M1DBHR for a total of 8 buses.

 

[Del91574]

My example above is actually best case and you need to design around worst case which is the M1 running on Battery.  The M1 will run on battery down to 10.2V so this should be the value used in any calculations.  Check with the specific devices for their lowest operating voltage.
 
According to the M1DBHR manual each of the 4 branches can be up to 4000' in length.   It basically has 4 RS-485 buses that it bridges together with the main bus so you end up with 4 buses instead of just 1.  And you can add two M1DBHR for a total of 8 buses.

Your overall length statement is NOT true. It splits the 485 bus into 4 managed branches. It does not regenerate or repeat the data. Look at the note on the second page of the installation document. Right in the diagram of the topology and connections, 4K total length, which I've confirmed with Elk before (I had a project that required an exceptionally long pull between buildings).
 
Your information about the M1, voltages it operates on and functionality and design criteria is inherently flawed. The main panel does a LBC at 10.2VDC, however the bus devices will start to reboot and lose connectivity far before that value. Once they start to dip below 12V, they start misbehaving, and right around 11.5VDC (memory) is when the reboots actually start occuring with bus devices.
 
The design criteria for the M1 bus devices should be to maintain 12VDC (or as close to) at the end of the run for VD calculations.