Your Snowmelt Sat Idle All Summer. Check It Before the First Storm.
A snowmelt system sits idle for six months, then gets asked to do its hardest work the first night it runs. That's when I get the calls: a loop that stays white, a boiler that locks out, a control that never noticed it was snowing.
Almost every one of those would have shown up in a fall check. None of them show up in July.
Here's what I look at before the first storm, and which parts you can do yourself.
Start with the fluid
Snowmelt loops run on 50/50 propylene glycol. Tube buried in a slab can't be drained and can't be kept warm, so the fluid is the only thing between that slab and a frozen, split loop.
Glycol isn't fill-and-forget. The corrosion inhibitor in it wears out, and once it's gone the fluid starts working on the metal it was supposed to protect: the boiler heat exchanger, the pumps, the manifold. So every fall:
Freeze point. A refractometer reading shows whether the mix is still where it was designed.
Inhibitor. Test strips made for the glycol in your system. Low inhibitor means treat it or change it this fall, not next year.
Look at it. A sample from a purge valve should look like it did going in. Dark, cloudy or sharp-smelling glycol has broken down.
Plan on a full fluid change every few years. Your glycol's manufacturer says how often.
Don't top it off with the hose
If the snowmelt side is low on pressure, the easy move is to open a fill valve. On a glycol loop that feeds straight water in and thins the mix. A little every time, and nobody writes it down. A few seasons of that and the freeze protection you were counting on is a number nobody has tested.
Glycol systems get made up with glycol, from a feeder or a pump. And if pressure keeps dropping, the real question is where the fluid went.
Compare the pressure, don't guess it
Your system was filled and started at a set pressure. Compare today's gauge to that number, on the startup tag or in your paperwork. If it's lower than it was in the spring, something lost fluid over the summer, and you want to find it before the boiler does.
Walk the manifold
On a system I build, every loop gets a flow meter at the manifold, and I balance loops to ±5% and write the readings down. With the system running, each loop should read close to where it was set at startup. A loop reading low or zero is air, a closed valve, or a problem in the slab. All three are easier to chase on a dry fall afternoon than during a storm.
While you're there, look for weeping at fittings and crust on valve stems. Dried glycol leaves a residue that's easy to spot once you know to look.
Make sure the control can see the snow
An automatic system starts when its sensor says it's snowing. If that sensor spent the summer under mulch, paint or debris, the system won't start until somebody notices the driveway is white. Clear it off, and if the control has a test mode, run it. That tells you the sensor, the control and the boiler are all talking.
Look at the settings while you're there. Most snowmelt controls can keep the slab partly warm between storms so it responds faster when snow starts. That's a real trade: quicker melting for more gas all winter. It should be a choice somebody made, not a default nobody looked at.
Fire it before you need it
Run the system for real on a cool day. The boiler should light and stay lit, supply temperature should climb steadily, and nothing should bang or whistle. Snowmelt supply has a ceiling. I stop at 150°F, because hotter water risks thermal shock in the concrete. Check that the control's maximum supply setting is still where it was designed.
What's yours to do, and what isn't
Reading the gauge, watching the flow meters, clearing the sensor and spotting a crusty fitting are all homeowner work. Testing and treating glycol, purging air out of a buried loop and changing control settings are not. Get those wrong and the fix is under concrete.
If your system hasn't had a fall check, or you inherited one with no paperwork, I'll come out and go through it: fluid, pressure, every loop, the sensor and a real firing. You'll get the readings in writing so next fall has something to compare against.
More on how I design and build these: Snowmelt Systems.
— Ryan Van Inwegen, ByDesign Hydronics & Air, Littleton, CO
Comments