Overwintering in zone 6b (Massachussetts) - my solution for building a greenhouse

[harshadg]

I moved to Massachusetts this August, and this is my first winter here. In my previous location I overwintered everything in an attached garage — stable temps, minimal swings, essentially a cold room with no surprises.

Here, the only usable spot is a section of pavers in the backyard. No garage. No enclosed shed. So I bought a 10' x 10' greenhouse/polytunnel thinking it would be good enough for a typical New England winter.

Short answer: it wasn’t.
Long answer: here’s everything I ended up doing, and what actually worked — with real temperature data.

1. First Attempt: Polytunnel + Sensors = Basically a Windbreak​

I put the trees inside the polytunnel and added Govee sensors to monitor temps. Very quickly I learned that a single-layer polytunnel barely holds any heat. It cut the wind, sure (and this was critical when we had 38 mph winds recently), but inside temperatures were only 1–2°F above ambient. This wasn't going to work for truly cold nights that were anticipated.

On top of that, here in MA we've seen some awfully cold temps with dips into the single digits. With nowhere else to store them I had to adapt quickly.

2. Insulation Overhaul (Double Wall, Bubblewrap, Foamular)​

I upgraded the structure with:

• Double polyethylene skin
• Interior bubblewrap layer
• Foamular boards along the cold walls and base
• Corrugated cardboard to fill in the gaps

Currently the order is: Outside>6 mil PE sheet>6mil PE sheet> bubblewrap>inside but the better strategy for next year is: Outside>6 mil PE sheet>bubblewrap>6 mil PE sheet>inside).

This improved nighttime performance but created two opposing issues:

1. Daytime overheating — even on 40°F days, interior temps shot into the upper 60s or low 70s.
2. Nighttime heat loss still too steep — not enough buffer for nights in the teens.

So I needed a way to flatten both ends of the temperature curve.

3. Controlling Solar Gain: Shade Cloth + Reflective Bubblewrap​

To tackle the daytime spikes, I added:

• 50% shade cloth over the top of the polytunnel, and
• Reflective bubblewrap across the south-facing front section

This combination cut enough direct sunlight to stop the runaway midday heat while still allowing the structure to warm gently. After adding these, daytime temperatures stopped hitting 70°F and held in a much safer range < 50°F (closer to 40°F).

With predicted temps dropping to 6°F earlier this week, I needed to rapidly work on a solution for the nighttime lows.

4. Heater + Thermostat: ITC-608T + 900W Ceramic Heater​

I installed a 900W heater (max 1500 W, with tip-over protection) controlled by an Inkbird ITC-608T thermostat. This finally gave me:

• A stable nighttime minimum, and
• Automatic cutoff to prevent overheating

Simultaneously, working on insulating the gaps (between the doors and the rest of the paneling for example) really helped with keeping warm air in the greenhouse.

5. My Phone Wasn’t Logging Sensor Data Unless I Was Near Them​

Next, I needed to tackle the issue of realtime monitoring. My Govee sensors are bluetooth beacons and to gather the data my phone needs to be in proximity (10-15 ft). That meant I needed to go out in the biting cold to sync my phone to the devices. It get's a bit annoying and I don't particularly enjoy the cold, so I built my own solution that allows me to collect the data while staying indoors.

6. I Wrote an App to Pull, Store, and Visualize All Sensor Data​

To fix the logging problem, I wrote a small app that I can deploy on my computer that continuously syncs and plots data from multiple Govee H5075 sensors. Right now I’m running five sensors, each in a different microclimate:

1. Greenhouse interior
2. Three-season porch (attached to house, partial buffer zone)
3. West wall (sun-exposed side)
4. North wall (near furnace exhaust; surprisingly warm niche)
5. True ambient control (open yard)

Further, I can gather data from the closest weather station too and overlay it alongside readings from my garden.

7. This weeks Freeze (predicted 7°F Outside overnight): Actual Temperature Performance​

Here’s what the sensors recorded:

• Outdoor ambient (control): 10°F
• Greenhouse interior: ~25°F
• Three-season porch: ~25°F
• West wall (air-exposed): ~15°F
• North wall (air-exposed): ~15°F

So far so good. I’ll keep collecting data this winter and refining the setup, but so far this combination has made it workable even with single digit lows. If anyone wants details on the sensor layout, the app, or the insulation approach, I’m happy to share more.

Stay warm!

 

[hemmy]

To fix the logging problem, I wrote a small app that I can deploy on my computer that continuously syncs and plots data from multiple Govee H5075 sensors.

That is awesome! I have the 5179 sensors and have the 1min data archive to my email monthly. But I know the pain of having to stand outside in the cold to sync the data from the coldframes. Most of the time I can get them to sync from various rooms in the house.

What is the reason for having them on benches and not the floor? I’d think being on the floor and then mulched in could moderate the root temps.

 

[hemmy]

• Outdoor ambient (control): 10°F
• Greenhouse interior: ~25°F
• Three-season porch: ~25°F
• West wall (air-exposed): ~15°F
• North wall (air-exposed): ~15°F

I’m now on my 3rd winter in the Midwest 6B, but each year we’ve had some polar vortices take us down round -10F for a short period. When those temps are forecasted, the cold frames (already closed below 25F) get wrapped in moving blankets and thick black plastic.

The book Bonsai Heresies by Hagedorn has an unpublished table by OSU listing the root zone damage temps for numerous landscape trees.

Cheers!

 

[harshadg]

thanks for your comment. Happy to point you to my github repo if you are interested in the code at all (just DM me). With some tweaks it could be repurposed to read the 5179 sensor.

I pretty much built the greenhouse around the summer location for the benches. The first iteration was with the benches sitting directly on the pavers (which get cold pretty quickly), and I didn't have mulch on hand. So I reasoned that the benches should be warmer than the pavers (hot air rises = uninsulated floor should be coldest). In this iteration I now have the foamular boards on the pavers, and could in principle set them on the ground.

However, with the empty space under the benches, I can place buckets of water underneath when it gets really cold. The latent heat of fusion (ie, energy released when liquid water is converted to ice) should further buffer against temp drops.