ok...according to Spanky at ELK the 1 second delay is normal. This suck pretty badly IMHO since beelzerob has already proven that when having CQC direct the ALC controller he can send commands MUCH faster than that.
In any case, I'm still having issues with my ELK<>ALC interface using the special ALC part that sits direclty on the ELK databus.
The symptoms are the following:
- presets levels are not consistently rememberred. Sometimes the dimmer go on to full sometimes to their presets, doesnt seem to be a pattern to it.
- I will unable to replicate the double OFF tap extensively discussed in the posts above
What works absolutely fine without missing a beat is that whenever ELK rules issue a command to ALC the lights will indeed do what they are told.
Other issues:
- ELK's lighting control screens allow the setting of ON/OFF and setting to a certain %. This is great and works well. When you do not use a percentage and just say 'ON' the lights come up fast and when you say 'SET TO 100%' they come up nice and smooth. However, it is not possible to use 'SET TO 0%' to turn the light off, the lowest value accepted by ELK is 1%. This means that there is no 'soft off' when using ELK to control the lights unless you double up on all rules and first use 'SET TO 1%' followed by a delay time and then 'TURN OFF'. This works but is silly complex to achieve something very simple.
General setup:
ELK M1G <> DATABUS HUB <> ELK/ALC INTERACE <> Distribution Module. Using a Cat5 / RJ45 from the 'branch 1' RJ45 on the ELK/ALC interface board to the RJ45 on the Distribution module. I also have an ALC expansion module (adds 3 branches) but other that the flat expansion cable it's not connection to anything.
All aux wiring for the dimmer is homerunning and cross connected on the distribution module. 'In the field' however multiple dimmers (in the same gangbox) may be on the same polling loop run.
TROUBLESHOOTING PROGRESS:
- When i remove the Cat5 cable between the distribution module and essentially make the system 'dead' but keep the LV wiring on the distribution module intact all the aux switches still control the dimmer, scene switches obviously don't work anymore BUT the double off tap produces a nice smooth transition to OFF and the presets are remembered without fail. This tell me the switches themselves are fine.
- No there could be an issue with the polling loop wiring, but since I've been very carefull to do this right i seriously doubt it. On top of that all command sent from ELK are executed without fail, so I think the physical communications links are good to go.
- I've found that the power supply provided with the ELK/ALC interface is absolutely useless. The controller gets it's power from the ELK DATABUS, even with the power supply plugged in and the databus connection removed the controller does not get power. The power from the ALC power supply doesnt seem to go anywhere. There is a jumper (JP3) right above the power connection which i suspect may determine where power is obtained from, but nothing is described in the documnetation about JP3 or any of the other jumpers. I don't like that it draws power from the ELK databus since this could potentially overload it. Ofcourse i can just take the 12V + /- and connect them to another supply, but then i believe i need to tie the negative together or something. If the stupid thing comes with a power supply, it expect it to do something. For the last 4 months it's been sitting plugged in the outlet warming up and using watts while doing nothing.
- When i provide power to the controller but no communication (i did that by disconnecting the databus comm wires but leaving the 12V +/- connected the scene switches work, the presets all remembered and the doubtap soft off works. In this configuration the ALC controller essentially works in standalone mode (without communication to ELK) and EVERYTHING IS FINE (except of course that i cannot use the ELK to control the lights).
- Based on this I'm tentatively concluding that the communication between the ELK and the ELK/ALC interface causes the system to misbehave. I have no idea whats causing this. The expanded ID is unique and the ELK databus is otherwise fully stable. I don't know how much power i'm drawing from the ELK or if this could have an affect, but the following is connected to the ELK without using an auxillary power supply (2 Input expanders, 1 M1XSP (for tstat control), ELK/ALC interface (logically treated as an M1XSP), 2 KP2 keypads, 1 KPAS decora size keypad, 1 strobe, 1 RT1 outdoor speaker/siren, 2 SP12F speakers).
NEXT STEPS:
- Disconnect a bunch of stuff from the databus to ensure that this is not an issue related to databus loading. I don't expect this to yield the solution, but it's the proper step to take in the process of elimination.
- Same step as above, particularly removing the other M1XSP (which talks to the aprilaire thermostates) from the system to ensure there's no conflicts.
Anybody have any suggestions? I'd like to hear if anybody using the 'old school' method of connecting the ELK to ALC (using a real M1XSP <> ALC serial module <> ALC controller) are having any of these issues. The failure to remember the presets is my biggest gripe.
I'm considering dumping the ELK/ALC special module and getting the kit needed for the 'old school' connection method. Hoping that this will work better and allow me to use a serial router (
http://www.avocationsystems.com/SR-2.html) to allow ELK and CQC simulatenous control. I've already established a 'heartbeat' communication between ELK and CQC so i would use CQC for all normal control and have simplified backup rules in the ELK which all first check if CQC is online). Before i jump all into that solution i'd like to try to figure out if I can at least make this work prooperly.
Tony, when using the standalone controller with serial control module. Does this allow the use of the Scenetech software and what options does this give us? Can i use that software to programm the scene rather than then press/hold the scene switch button and run around the house to set the desired levels? That method is getting old pretty fast.
Has anybody succesfully 'recalled' an ALC scene using the ELK?