[Greenbuilding] Solar Water Heater architecture

[Lawrence Lile]

Getting ready to install a solar water heater in new construction, and have been agonizing about the arrangement. 
 
I have 3 used collectors, and El-Sid PV pump http://tinyurl.com/echz6, and an idea.  
 
The El-Sid PV pump is ingenious - it is a DC pump connected directly to a solar panel.  When the sun shines, it pumps.  
 
Friend #1 has installed drainback systems, and swears they are pretty straightforward to install.  We have hard freezes around here, so the consequences of the drainback system failing are severe.  I have read that drainback valves installed outdoors can ice up, then fail to work, then the collectors freeze and burst.  I have also read that PV pumps may not work well in drainback systems.  Picture this:  It is 20F below zero, your collectors are nice and dry because they drained back last night when the sun went down.  The sun comes up, nice and bright, but your collectors haven't warmed up and are still 20 below.  The PV pump turns on, and becomes a 3GPM icemaking machine.  Maybe the collectors would warm up later and thaw out, or maybe the whole thing would just ice up and bust. Also drainback systems with open tanks can be a constant humidity source in your house, which would be bad in the summertime.  I'm trying to avoid more complex controls, considering compex controls to be the achilles heel of the solar colelctor. 
 
Friend #2 swears by glycol systems.  They can't freeze, period. But Ethylene glycol is poisonous, and propylene glycol, while not poisonous, is expensive.  Glycol in the collectors means that the foolproof controls on the El-Sid pump will be able to function.   
 
Originally I had planned to have about 50 gallons of propylene glycol in a loop between a solar storage tank and the collectors, with a single wall heat exchanger (a length of soft copper tubing) immersed in the solar tank that preheats cold water to the domestiv water heater.  Simple!  But that's $800 worth of Propylene glycol, and it has to be replaced every 5-10 years.  Yow!  
 
So I am imagining a different arrangement - still relatively simple.  A smaller glycol loop from the collector to a tank heat exchanger (another loop of copper pipe in the tank)  Tank contains plain water.  Domestic water also passes through the tank, through another length of copper pipe.  This achieves a double wall heat exchanger, which allows the more dangerous but cheap ethylene glycol.  
 
Now you can buy a tank set up to do this, for another $800 or more.  I was going to fabricate the heat exchanger tank from scratch, since it is a relatively simple system.  
 
This problem has been solved a lot of times before.  What arrangements do you folks recommend? 
 
--Lawrence Lile