richard naninck wrote:All my scripts are open, but they are based on use with HouseBot so probably hard to use in your case.
Do you use other FS20 related stuff or just the FHT80b icw FHZ1300PC?
I have some FS20 stuff (a sound signaller and some switches and a 220v socket switch). They always have trouble not responding well. I have a light in my bedroom whicih is switched through FS20. When I switch it off via software 1in 5-6 times it does not. And the device is only 6 meters aways from the FHZ1300PC (through some walls though). I got sort of used to this, but when the heating refuses to operate this is quite annoying.
richard naninck wrote:Each FHT80b produces 3 messages per hour containing temperature settings, battery stats etc. Once every 2 minutes you should see a message containing Valve data. Not sure who produces this message. (The FHT80b or the Valve itself).
Okay, you mean that. Yes I know that too. I also keep track of the valve data message, because this is the trigger for me to send a new message to the FHT80B (about 2-3 seconds before the new valve data is about to be received). In this way the chance that the message is processed by the FHT80B seems to be the largest. Sending a new temperature also fails regularly (1 out of 4 times or so). My software therefore keeps a queue of transmitted temperatures and resends them when necessary (again when the receive window is about to open so every two minutes). This used to work reasonably well, but the last week the system has stopped to respond altogether.
richard naninck wrote:
Sometimes I don't get messages from the FHT80b anymore. I trigger myself by setting a timestamp on each received message. If the time from the last message is older than one hour, my system throws an error message and I up the FHT80b with 0.5 degrees. By doing this, almost Always the FHT comes back online. Actually it never goes offline because the valve continues to react to the FHT bu the FHZ1300PC simply doesn't get the messages anymore. Sometimes I also don't get the Valve messages anymore but as soon as a valve is controlled open again, the messages start flowing again. So often in summertime when the valve remain shut for months in a row I might loose the message stream. Mayby this is by design to save battery power. Not sure about that.
Indeed sometimes, (ahem), I also don't get messages from the FHT80B anymore. But when I then change the temperature that doesn't help because the system is totaly unresponsive. I have to stop controlling it, sometimes for more than a day, for it to come 'online' again. And ideed it never really goes offline, because you keep getting the valve messages. I don't think I overflow the system myself, because I always wait the 2 minute cycle to (re)send a message and always check the FHZ1300PC buffer to see if there are buffers available to send something in the first place.
I often see that the FHZ1300PC has few or no buffers available to send messages. But where this comes from, no idea.
richard naninck wrote:Flooding actually appears when too many messages are sent or received by the FHZ1300PC. This should normally not happen but while testing I have seen this a couple of times. I had to either wait a long time or unplug the FHZ1300PC to get it fixed again. That has something to do with the protocol itself because messages are buffered and communication with the FHT80b's is bidirectional.
Flooding seems to occur regularly in my case. The FHT80B then locks itself up in its cocoon and stops listenig for a long time. This is quite anoyying. I can't find any system in it. Nor can I find a reason why the connection between FH80B an the valve is suddenly lost. Perhaps this stuff only lasts a few years and the breaks
I have put the FHT80B's in automatic mode now with a weekly schedule, so the house is not entirely cold when I get home. But this is a real step back in relation to my google calendar controlled software
I'm strongly considering (I think the chance that I will migrate is nearly 100%

, but the financial aspect is also to be taken into account

) getting the MAX! components because at least they are fully bidirectional so I don't have to keep track myself if the thermostat has responded to a temperature change.
Also the MAX! seems to respond much faster when requesting a temperature change. In the case of the FHT80B, the 2 minute time window, and the fact it often 'misses' a change request, can cause a delay of 5-15 minutes before a temperature change is actually performed. I find this waiting time too long.
That point is also a nice opportunity to rewrite my software and use MQTT to transmit messages over the network so the system is less dependend of platform and implementation of the different components
