Troubleshooting routine (dehumidifer control)

Hi all,

I have set up a couple of routines for a basic dehumidifer control - that is, turn the dehumdifier on when the external humidity sensor reads 52%, turn it off when the sensor reads 50%.

It has worked pretty good for a few weeks now, managing the dehumdifier for 15-20 cycles per day.

Except, that twice it has failed to operate. In one instance, the humidity on the sensor got well about 52% and the dehumidifier didn’t turn on. In the other, the humidity on the sensor fell below 50% and got as low as 46% before I caught it (6 hours of continuous operation) and turned it off manually.

How do I go about determining why the routine failed? It’s not reliable enough for me to trust it at this point.


The usual issue with ‘greater than’ and ‘less than’ in Routines is that they actually trigger the Routine on ‘rises above’ and ‘falls below’. That can be an issue if you have other preconditions or conditions in place that aren’t true when the trigger occurs.

There might also be issues if the hub reboots at the ‘wrong’ moment. I haven’t thought about what they might be but there might be some. There are probably other things that can go wrong. The bottom line is you are dependent on a single event as the humidity rises, and a single event as the humidity falls and if that is missed you are hosed.

Personally I’d be assuming the Routines could fail and installing a few safety nets.

1 Like

Thanks. I’ve put in some high humidity and low humidity alarms to force me to check if they should fail to turn on/off again. Thought there might be a more advanced way to troubleshoot that I should consider.

There are several threads which refer to the way the ST automation engine triggers routines. One of these is here
and you might find it interesting to scroll through.
Several users, myself included, use a rule to run a check switch every so often, while designating the conditions, ie your humidity levels, as preconditions.

I am not really aware of one. I think it is just a case of checking the history to build a timeline of when things actually happened.

I know it’s not the solution you are looking for but maybe add two more rules for above 53% and below 49%? Just as a safety net?