Deployment notes: TRVs on 5G heat networks c/o Marko Cosic of COHEAT 2015/09/17 Standard TRVs are designed to operate with a differential pressure of the order 0.1-0.2 bar across them. They're usually a flat rubber disc against a flat metal seat. They allegedly work up to 0.4 bar or even 1 bar differential pressure. They allegedly close fully. With an earfing great wax/liquid filled and spring-loaded TRV head on top of them this might even be true. Enter the Chinese eRV. This device can muster up 10N or so: just about enough to close a TRV. The Chinese software calibrates once and once only. The Chinese mechanicals incorporate no spring loading or flexible elements to tolerate differences in zero position. The Chinese software counts rotations of a gear wheel rather than reading current when deciding where the zero position is. End result: these eRVs don't close. Let the plastic case move from 20C to 40C temperature, the gearwheels bed in, the casing stretch, and the rubber washer in the TRV body squish/bed in, and you'l end up with a flow at the zero position. Takeaways for OpenTRV: Make sure you re-home every now and again. Consider measuring current during this rather than counting wheel spots, if you're not already doing this. Zero points move and zero points can be important. Consider making valve % open a non-linear thing. The first 10% really matters. Up to 30% does something. 30-100% by linear motion does virtually nothing. (DHD note: OpenTRV experience with TRVs in normal domestic raidator systems has been that none are open at all below 6% on the FHT8V's scale, most are not open below 10%, values as high as 25% to open have been seen, but for most most of them 33%+ is fully open.)