|« Insulation||Drywall »|
Our house is heated with in-floor hydronic heating a.k.a. radiant heat. It’s a fantastic way to heat a house, particularly in our relatively mild climate; however, I’ve never been happy with how our control system worked (or didn’t).
Radiant heat has a number of upsides:
Of course these can also turn into downsides:
All of the downsides can be eliminated with an intelligent control system, and ours was not that.
The water topology of the system is as follows:
The control topology is as follows (in reverse order to the above):
For an additional level of complexity, the call for heat from the thermostat causes the local zone control module to energize the telestat(s). When the telestat(s)‘s end switches confirm that the loop is open (several minutes later2), the local zone control module forwards the call for heat. Similarly, the central control board doesn’t forward a call for heat until the corresponding valve’s end switch confirms that the valve is open. I believe the reasoning here is that this prevents the circulator from trying to push into a fully closed system. However, when I wired my last house’s radiant system (my contractor at the time let me do the electrical work in return for a lower price) we didn’t bother with any of that and it probably isn’t really necessary.
So, easy, right?
When our house was built, our contractors placed the local zone control modules next to the corresponding local manifolds. They left most of the valves on the central manifold permanently open, relying on the local manifolds’ telestats to do the work. And finally, they failed to connect the thermostat in the garage to much of anything so our garage was un-heatable. (Having a selectively heated shop space in the winter is amazing, by the by, now that I’ve gotten it to work.)
This made it so that there was no central place where I could go into dad mode and say, “why is this system running when it feels like it shouldn’t be?“. The system mostly worked but there was no easy way to debug it, and we’d already had to do a bunch of debugging due to initially mislabeled and incorrectly connected valves.
And with the simple (albeit perfectly normal) thermostats we had, there was no way to view or modify their setpoints centrally or from a phone.
I hatched this dream of IoT-ifying the entire system; I’ve had my eyes on the Particle infrastructure for some time and this seemed like a perfect application. I know enough about tech to realize that this is not a particularly complex project. However, I also know enough about tech to realize that all tech will go awry eventually, especially IoT stuff, and I didn’t want to have my family murder me in a freezing cold house should my gadgetry ever fail.
As such, I designed a reflorgling of the system that would allow me to retain or move back to the simple thermostats whenever I wanted to, and that didn’t rely on any network connectivity to do its job.
I then spent the better part of a week redoing all of the wiring of the radiant heat control system:
Key to making this work was realizing that
I carefully constructed a wiring spreadsheet:
Then I took a few warm-ish days in early November to tear the system apart and rebuild it.
Here’s what one of the local manifolds looks like:
I also dealt with a number of sins of the past, including finally realizing that the reason the garage thermostat had never gotten hooked up was that our contractors burnt out the corresponding channel on the central control board so it didn’t work. Rather than debugging and replacing that they just disconnected it and said “fuck it”.
My aspiration with the Particle-based IoT-ified thermostats is not only to allow for more oversight and reign in any small fry fuckery, but also to integrate the system with our security system which should know when any room’s windows are open. That crossing the streams flow of information is really my nirvana of environmental conscientiousness, resource conservation, and just all-around extreme “turn off your lights!” dad mode. Someday.
For the record, here’s the equipment the system is using, in case it’d ever be helpful to anyone:
V8043E5061valves? Of course not.
V8043E 5000 Seriesvalves, because of course those are different from
40003916-526rather than SKU
40003916-026because fuck you.
|« Insulation||Drywall »|