Does a drainback loop need a particular roof-pipe layout?

If power stops the pump, a drainback system should let water in the collectors and exposed piping return to its reservoir, so freeze protection does not depend on circulation continuing. That only works if the installation drains completely: a sag, trapped section, or poor pitch can leave water behind and create a freeze risk. With a glycol loop, freeze protection remains during an outage as long as the fluid’s condition and concentration are adequate, though circulation and heat collection stop.

For a glycol loop, I’d expect periodic checks of the fluid’s condition and freeze protection, with service timing depending on the system and its manufacturer’s guidance; it isn’t necessarily a fixed annual fluid change. That upkeep is the trade-off for not relying on every roof section draining completely. If the roof layout makes a dependable drainback route difficult, the glycol checks may be the more manageable constraint.

I’d give local freeze severity weight, but not let it decide the system by itself. In a harsher climate, I’d favor whichever option can be made dependable at that particular roof: drainback if the piping can truly empty, or glycol if its protection can be maintained. A design that is marginally easier to service or install matters less if its freeze protection is uncertain.

Applying your roof-specific rule, if one section of a hypothetical roof makes complete drainback uncertain in a harsh-freeze area, I’d treat that uncertainty as a strong reason to favor glycol - provided its fluid checks and servicing are manageable. The extra upkeep seems like a more predictable burden than relying on a drainback route that may leave water trapped.

If the roof can be laid out so the collectors and exposed piping reliably empty, and freeze exposure is moderate, I’d lean toward drainback: I’d rather avoid adding glycol-fluid checks to the upkeep. In a harsher climate, though, I’d give freeze reliability more weight than that maintenance preference.