Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Field Time
#1
Dear,
Firstly , congratulation for your work and the description of the protocol.

I work for a international company in Building Automation.
The system is very reliable and made for the industry.

1) Our SCADA is more or less the equivalent to a FHEM or OPENHAB. It's a server with SQL DB.
2) The Field CTRL are equivalent to Node (HVAC, SECURITY, ...)
3) A Gateway make the link between the SCADA and the CTRL (the Serial gateway or Ethernet gateway)

When the SCADA is shutdown for maintenance reasons the CTRL must stay functional (access control, HVAC, ...). It can also be down due to power outage or a network issue.

For my home system, I can also imagine that when I go on holiday, I can shutdown the server.
In case of an intrusion into my home, and if an alarm is generated off line, I need to know the real time when it occurred. Two weeks ago, I got a issue with a customer because a stealing had been made when the network was done. I was able to give the real time when it happened.

For my house, to monitor my gas meter, I build a CTRL equipped with a EEPROM AT24C32 and a RTC DS1307. It keeps +/- the last 60 days of data (one value at the end of the day) and the last week of measures (each 0.1 m3) (FIFO).
The goal is to send the measure when the server is back (I can have a pointer the the last data sent in the FIFO).

I can imagine (like our SCADA) to have two columns for date/time : a date/time when the data are recorded in the DB and a FIELD time that come from the CTRL.
The CTRL are synchronized each day at midnight (for those that have a RTC).

I didn't have seen a Timestamp in the protocol.
Is it something planned/possible?
Reply
#2
(12-17-2015, 05:14 PM)esa1966 Wrote: For my house, to monitor my gas meter, I build a CTRL equipped with a EEPROM AT24C32 and a RTC DS1307. It keeps +/- the last 60 days of data[...]

I didn't have seen a Timestamp in the protocol.
Is it something planned/possible?

A running smarthomatic system currently assumes that the base station is on all the time and receives and processes the data immediately. Therefore, there is no timestamp in the protocol / message format.

I don't see it necessary nor possible to let all devices record their data for later retieval (in case the base station is down) without HW modifications (RTC, EEPROM). And I think it wouldn't be adequate to have the possibility at every device when additional HW is needed, because on the other hand it's pretty easy to have e.g. a small RasPi running as FHEM server to record the data, with optional battery to bridge power outages.

Additionally, a user usually wants the system to react on events immediately (call the police when the window was broken and not after the vacation when the thieves left Wink ).
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)