X10 Community Forum

🔌General Home Automation => Original ActiveHome => Topic started by: dash on August 26, 2007, 05:46:53 AM

Title: Dawn/Dusk - calculated activation times - Tools/Statistics menu
Post by: dash on August 26, 2007, 05:46:53 AM
My very first schedule using the CM12 (UK) PC interface module ran OK last night. Not from the PC though; I dowloaded it all to the EEPROM in the CM12. I have 2 macros and 4 individual timed events running on selected modules.

I used the dawn/dusk option in one of my macros and chose a different offset (from dusk) for each of 3 modules.

However, the activation time did not coincide with the time listed in the Tools/Statistics menu (it switched about 10 mins later). Are different formalae being used in different parts of the system, or is there some other error?

Would be interested to hear anyone elses experience/suggestions....

  Dave  ;)
Title: Re: Dawn/Dusk - calculated activation times - Tools/Statistics menu
Post by: Charles Sullivan on August 26, 2007, 11:15:04 AM
My very first schedule using the CM12 (UK) PC interface module ran OK last night. Not from the PC though; I dowloaded it all to the EEPROM in the CM12. I have 2 macros and 4 individual timed events running on selected modules.

I used the dawn/dusk option in one of my macros and chose a different offset (from dusk) for each of 3 modules.

However, the activation time did not coincide with the time listed in the Tools/Statistics menu (it switched about 10 mins later). Are different formalae being used in different parts of the system, or is there some other error?

Would be interested to hear anyone elses experience/suggestions....

  Dave  ;)

The CM12U has only a 1K EEPROM and your schedule of events has to be made to fit into that space by using a stair-step approximation to the daily times of dawn and dusk.  My recollection from studying the ActiveHome 1.42 software several years ago is:
  1. It doesn't do a very good job of computing the stair-step approximation in the first place - it really falls apart in the last few weeks of the year.
  2. It's not actually using the stair-step it shows you.

A 10 minute error from the _actual time_ of dusk for your locality is not too bad considering you have 3 separate sequences of dusk times, however the daily dusk-time curve is still relatively flat at this time of year.

Presuming your events all occur on the same days of the week and over the same date interval, you can improve the precision by using only one sequence of dusk times and using a time delay for the other two offsets.  I.e., your macros are all executed at dusk (or dusk minus some offset), but each event is appropriately delayed from that execution time.

Title: Re: Dawn/Dusk - calculated activation times - Tools/Statistics menu
Post by: dash on August 26, 2007, 11:21:54 AM
"Presuming your events all occur on the same days of the week and over the same date interval, you can improve the precision by using only one sequence of dusk times and using a time delay for the other two offsets.  I.e., your macros are all executed at dusk (or dusk minus some offset), but each event is appropriately delayed from that execution time."

Hi Charles

Thanks for the comments. What you described in your last paragraph (pasted above) is exactly what I have done...  I've managed to squeeze the full 50 resolution for dawn/dusk onto the EEPROM

Cheers

 Dave