Outbuildings: 2026-01-22 Update

Outbuildings: 2026-01-22

While almost all of the building work was completed before the end of 2025, there were a few items outstanding which have been attended to over the past couple of days:

  • The rainwater drainage inspection chamber in the North-East corner of the Courtyard (top-left of the overview photo) has always been holding 30-40mm of water – implying an inadequate ‘fall’ on the outlet pipework.
    • The fall was checked with a laser level on Tuesday, which confirmed the chamber was set at the right level and pointed to some sort of issue on the downstream pipe run.
    • Excavation of that pipe run showed there was a ‘hump’ in the pipe which was causing the issue – partially due to it passing over a separate underground pipe connected to an Aco doorway drain.
    • With the pipe run re-routed away from the other pipe and re-levelled at the correct fall, the rain is now draining out of the inspection chamber.
  • The ground in the Courtyard has always been prone to holding pools of standing water – which was evident before construction started. As a consequence, a network of ‘land drain’ pipework was installed, formed from four ‘legs’ running east-West (which will end up in the middle of future planting beds).
    • The expectation was that there would at least be no standing water in the immediate vicinity of those land drains – but that wasn’t the case, and in fact there were some pools of water directly above the drains.
    • The issue turned out to be the use of solid clay as the back-fill material over the land drain pipes – and in some cases inadequate pea gravel over those pipes.
    • The agreement with the build team was that they would remove the layer of clay overlaying the land drains and top-up the pea gravel cover over the pipes to the specified 100mm depth – and then I will fill the trenches with some more free-draining soil from elsewhere on the site, as part of the work required to prepare the soil between the drains for planting.
  • Some other finishing touches that happened to be scheduled for today were for the windows and doors.
    • One of the window handles was from a faulty batch, and the lock cylinder kept falling out – so that handle was replaced (from a new, good batch).
    • There was some install-time damage to one of the door handles, which was also replaced.
    • The original window sills turned out to be too ‘deep’ to fit properly so those were re-made and fitted today.
    • A minor manufacturing defect on the bi-fold door meant the bottom seal had been torn – and would be torn again if one of the hinge fixings wasn’t adjusted, so that hinge was re-fitted and the seal replaced.

Temperature and Humidity Monitoring in the Outbuildings – Thread Network Range Issues?

The GL-S20 Thread Border Router and the two original Timmerflotte temperature and humidity sensors have been working fine. Once they came back into stock at the local IKEA I bought two further Timmerflotte sensors. One of those didn’t want to Commission (needs further investigation) but the other one did Commission successfully, at the first attempt, and took the expected firmware update.

While that third sensor has mostly been OK, it’s periodically dropping off the Thread network. The first time that happened I intervened with a power-cycle, which brought it straight back online, but on a couple of other occasions it came back all by itself.

Periodic connectivity drop-outs for the Timmerflotte sensor in the Ground Floor Store room

The main difference from the other two sensors is the distance from the Thread Border Router – roughly 16m, with one 100mm concrete-block wall in the way. Might that be the issue?

Thread networks are expected to operate in a ‘mesh’ topology, with mains-powered Thread devices acting as relays for the radio signals, between the central Thread Border Router and the (battery powered) End devices, so maybe one of those would help?

As a first step I’m going to try moving the problematic sensor a bit closer to the TBR. If that doesn’t work I’ll move the TBR much closer to the sensor (there are limited options because of the availability of wired Ethernet network ports).

Update: Moving the sensor about 1.5m closer to the TBR (but still on the other side of a wall) prompted it to come back online, so now it’s recording readings again. I’ll see if that proves to be a long-term fix.

The implication is that the drop-outs were indeed caused by a Thread network range issue – which implies the plan to locate the fourth Timmerflotte sensor even further from the TBR is likely to be problematic without some sort of network extender device. While not yet available in the UK, the IKEA GRILLPLATS ‘smart socket’ looks like it might be suitable (and not too expensive).