Hot Fault Sensor comms fault, there seems to be a pattern

[bruce_miranda]

This isn't something new in terms of the Hot water sensor comms fault. But I just saw the logs and noticed a pattern in the time it takes to restore the fault. Clearly can't be random. See pictures attached.

20170114_155656.jpg20170114_155635.jpg20170114_155623.jpg

[HenGus]

I have seen this as well - not sure of the significance.

[bruce_miranda]

To me the significance is that it doesn't appear to be random comms failure and then a reconnect after some other random time. The reconnect is definitely happening at a predefined interval even if the failure time is random.

Come on Honeywell, why is this happening only on the WiFi controllers?

[HenGus]

I have had random HW comms faults, hot water sensor:

This is typical:

FAULT 16-09-2016,21:13

RESTORE 16-09-2016, 22:03

[bruce_miranda]

Are your times correct? The restore seems to be in advance of your fault. If the typo is in the hour then you are seeing exactly what I noticed. The restores seem to happen in either exact 50 mins or multiples of 50 mins.

[HenGus]

Are your times correct? The restore seems to be in advance of your fault. If the typo is in the hour then you are seeing exactly what I noticed. The restores seem to happen in either exact 50 mins or multiples of 50 mins.

Good spot - I have amended it.

[bruce_miranda]

So maybe the fault times are indeed random or externally influenced. But the controller appears to only restore that fault 50 mins later.

[HenGus]

Not always. I have now found an earlier set of HW sensor faults

28 Aug 11.42F. 12:32R

4 Sep 12.43F. 13.32R

6 Sep 22.22F. 00.12R (on the 7th)

All my faults have occurred when there is no HW is selected to OFF

[DBMandrake]

To me the significance is that it doesn't appear to be random comms failure and then a reconnect after some other random time. The reconnect is definitely happening at a predefined interval even if the failure time is random.

Come on Honeywell, why is this happening only on the WiFi controllers?

The restore time being a fixed number of minutes after the fault condition is reported makes sense - when there is no change in hot water temperature (or only very small changes, outside of the differential range) the sensor only sends a temperature update about once an hour as a kind of "keep alive" message.

If that update isn't received then after a while the controller will report a fault, and then on the next transmission the fault will be restored. Because of the fixed timing of these "idle" transmissions you'll tend to see the same delay between fault and restore - in reality all that happened is one (or maybe two) of its hourly "idle" transmissions was not received, then eventually one is received and the fault restores. I've had this happen to me once.

The problem is also MUCH more likely to happen when hot water is turned off and is cold - when the sensor is outside the set differential range the updates are far less frequent and small changes of temperature go unreported for a long time. On the other hand while it is actively heating within the differential range the transmissions are much more frequent and the chances of the fault occurring is very remote.

I hate to say it but if you have already tried tightening the battery contacts on the back of the PCB then positioning of the CS92 is probably the issue. As you know I had a lot of trouble with my hot water overshooting due to poor comms from the CS92 - greatly improved by tightening the battery contacts on the rear of the PCB but it still happened sometimes.

After moving the CS92 to a different wall with a different orientation the problem went away even though the new location gives a slightly lower signal reading...

As to whether the Wi-Fi controller is more susceptible to this or not - not having used the older one I can't say, however I will throw one thing out there - it's possible that the 868Mhz antenna in the newer controller has a different physical orientation than in the older controller. In a low power (10mW) system like this this could easily lead to some devices receiving a weaker signal while others receive a better signal due to the way reflections combine in the house. Your CS92 could be the unlucky one whose signal path is not favoured by the new antenna orientation, with reflections combining to produce a localised null where it is located.

Assuming the controller is wall mounted and thus can't be re-orientated, you could try moving the CS92 - you don't have to move it a long way, something like 30-50cm is probably plenty. I spent ages debating why it shouldn't be necessary but in the end I gave up and moved the CS92 about 40cm from the rear wall to the side wall in the boiler closet. (changing orientation 90 degrees in the process) Presto, no more hot water comms issues and no more hot water overshoots. Shouldn't be necessary I agree, but I'd rather just have it working. Give it a try if feasible.