Valentines Day weekend the overnight temperatures dipped to -13F. I was mostly worried about how cold it would be skiing. While I was off skiing, Tyson discovered another problem — the pipes to the solar panels froze.
Tyson first noticed water running out of the high temperature dump. He found the solar panels at 224F despite the pump attempting to circulate. He turned off the hot water dump since it is ineffective in the winter heating configuration. All it does it dump electrically heated water out of the domestic hot water tank. Then he turned his attention back to the pump. Presumably the pipes carrying solar fluid through the unheated attic had frozen, or at least congealed enough the pump couldn’t push the fluid up. However, before Tyson could debug more, the pump started circulating water and all the temperatures equilateral. Either the bright sun or the warming air had defrosted the pipes.
The next morning was almost as cold. Tyson headed off to work. I had President’s day off, so I ran up and down the steps every three minutes checking on the solar panels. They started out frozen. I preemptively disabled the hot temperature overflow. While waiting, I inspected the old glycol tub for any clues. It had a sticker converting from refractometer reading to freeze point. Our first winter in the house, the pipes froze once, but the plumber fixed it by adding more anti-freeze. I had no record of the final concentration. Depending on the concentration, the freeze protection might be anywhere from -10F to -100F. Around 9:30AM the fluid finally defrosted enough for the pump to circulate water.
The temperatures seemed reasonable, but the pressure was low — only 12 psi. Tyson remembered the system new holding more like 30 psi. So, I drained the pipes and shut off the pump in case there was a leak.
That evening I called customer support for RadiantSolar. The technician seemed to think it was unlikely the pipes had frozen or burst. He also thought it was unlikely our solar fluid had degraded significantly in 5 years of operation. He suspected our problem was we had neglected yearly maintenance to re-pressurize the system.
We needed a known mix of propylene glycol and water to refill the system and we needed a pump. To get the right mix, we either needed a refractometer or premixed solution. I called around to every hardware store and automotive store looking for a refractometer. Nobody had one in stock. Tyson found a premix available at some of the local Home Depots. Friday evening he raided all the stores in Nashua and came up with 11 gallons and a sump pump rated to 30 feet. The premix rating is: freeze protection to -22F, pumpable to -27F, burst protection to -80F.
Saturday morning we poured 5 or so gallons into a bucket, hooked up the sump pump, and waited for the temperatures on the solar controller to equalize. They didn’t. The sump pump just wasn’t strong enough to push the fluid all the way to the top of our house. So we turned on the standard solar pump in series with the sump pump. The temperatures changed rapidly, and the water level in the bucket dropped. We poured more fluid into the bucket.
The new fluid was blue whereas the old fluid had been pink. We tested the old pink fluid with a PH test strip from the hot tub. It’s an imprecise measurement, but the fluid didn’t seem excessively acidic. We kept pouring jugs into the bucket. We ran out of jugs, so we moved the sump pump to a smaller bucket to keep it submerged. Finally the fluid level stopped dropping and the fluid changed color from blue to pink. Tyson had bought barely enough anti-freeze to fill the pipes from the solar panel to the utility room. I had not drained the radiant piping in the floor. The rest of the morning, we progressively flushed air from each sub-system connected to the panels.
Then it was time to re-pressurize the system. The sump pump just couldn’t make the pressure budge above 12 psi. Not surprising considering it couldn’t circulate fluid without help from the solar panel pump. We tried elevating the bucket for just a little bit more pressure.
Next steps: buy a stronger pump, a refractometer calibrated for glycol, and PH strips.
With the panels turned off, the sand and slab temperature dropped significantly despite several sunny days. Since turning them back on, the mass temperatures have bounced back.