plate heat exchanger help please!

[ptoddf]

Any self educated opinion very appreciated here, don't need exact engineering analysis by a long shot.

I'm running 2 each Grumman 3' X 7' glazed solar HW panels with copper collector tubes thermosyphon into a 55 gallon HW tank in a box with about 8" of fiberglass around it. Panels are at ground level and 45 degrees, the tank is on about a 4' pedestal. Not quite high enough but it thermosyphons just fine. Worked 3 years, can get up to 160 degrees in tank in summer, perfect for direct feed to dishwasher with heating element disconnected in this off grid mountain cabin. (5000'+ altitude East of Mt. San Gorgonio, on way to Yucca Vy, California.)

I put in a home built auto drain out system to prevent freeze damage. Servo valves cut off lines to/from tank, a drain valve opens, vacuum breakers open above collector and water in collector, looks like 1 1/2 to 2 quarts, drains onto ground on freezing nights. Worked fine until it failed and burst copper lines in collectors as expected. Damage can be fixed again but what a PITA!

Now I want to convert to a glycol loop through collectors and use a heat exchanger at the tank top. Question is, WHICH heat exchanger?

I'm looking at the small 10, 12 and 20 plate units on ebay like item #250963082494, and like the vendor pextubing sells on ebay as for example item #390667900898. These exchangers all look the same and are no doubt Chinese made. The 3/4" fittings match the 3/4" copper I'm running to the collectors so that seems OK..

Will a 10 or 12 or 20 plate one of these small units work for me do you think? All considered opinions greatly appreciated!

Todd F.

[Naptown]

since the amount of heat transferred is a function of three things
1-temperature differential between the fluid (More is better)
2- heat exchanger material and thicknes (thin, within limits and copper are best)
3- surface area of heat exchanger.

Since you can' control #1 in a solar application as it is constantly changing and all things being equal material wise
which do you think would be best.

[LucMan]

Any self educated opinion very appreciated here, don't need exact engineering analysis by a long shot.

I'm running 2 each Grumman 3' X 7' glazed solar HW panels with copper collector tubes thermosyphon into a 55 gallon HW tank in a box with about 8" of fiberglass around it. Panels are at ground level and 45 degrees, the tank is on about a 4' pedestal. Not quite high enough but it thermosyphons just fine. Worked 3 years, can get up to 160 degrees in tank in summer, perfect for direct feed to dishwasher with heating element disconnected in this off grid mountain cabin. (5000'+ altitude East of Mt. San Gorgonio, on way to Yucca Vy, California.)

I put in a home built auto drain out system to prevent freeze damage. Servo valves cut off lines to/from tank, a drain valve opens, vacuum breakers open above collector and water in collector, looks like 1 1/2 to 2 quarts, drains onto ground on freezing nights. Worked fine until it failed and burst copper lines in collectors as expected. Damage can be fixed again but what a PITA!

Now I want to convert to a glycol loop through collectors and use a heat exchanger at the tank top. Question is, WHICH heat exchanger?


I'm looking at the small 10, 12 and 20 plate units on ebay like item #250963082494, and like the vendor pextubing sells on ebay as for example item #390667900898. These exchangers all look the same and are no doubt Chinese made. The 3/4" fittings match the 3/4" copper I'm running to the collectors so that seems OK..

Will a 10 or 12 or 20 plate one of these small units work for me do you think? All considered opinions greatly appreciated!

Todd F.

Btu output for those 2 panels should be in the range of 40-50K BTU per day.
HX's are rated in btu per hour or flow rate depending on the style. I would select a HX with a capacity of 8-10K BTU per hour or 3-4 gpm .
Or you can make a tube in a tube HX with some 3/4 & 1 1/4 copper pipe and some T's.

[Naptown]

Btu output for those 2 panels should be in the range of 40-50K BTU per day.
HX's are rated in btu per hour or flow rate depending on the style. I would select a HX with a capacity of 8-10K BTU per hour or 3-4 gpm .

I disagree with that assessment.
You also have to look at the temperature differential the heat exchanger is rated at. If it is 10KBTU @ a 150d temp differential it will actually be much less than that in this application where most temp differentials are in tie 10d -4d range.

[J.P.M.]

since the amount of heat transferred is a function of three things
1-temperature differential between the fluid (More is better)
2- heat exchanger material and thicknes (thin, within limits and copper are best)
3- surface area of heat exchanger.

Since you can' control #1 in a solar application as it is constantly changing and all things being equal material wise
which do you think would be best.

One other thing not only necessary but without which the 3 mentioned become mostly irrelevant is the overall heat transfer coefficient (often called "U zero", or "U overall", or just plain "U"). The amount of heat transferred is U*A*Delta T. Those 3 are all interconnected and dependant on one one another to a greater or lesser degree depending on the design. Contrary to common, incorrect thinking, that value "U" is NOT constant. It's determination is actually the goal of heat exchanger thermal design, which is a very iterative process when done correctly. I'd suggest that putting a heat exchanger into a thermosiphon loop (unless it's something called a thermosiphon reboiler - which this is not) is an excercise in futility. The velocities obtained will mean poor heat transfer - low "U", thus requiring more area, thus additional lowering of "U", etc, etc. Besides, it'll probably be a plumber's nightmare. Without pumps in the system, the heat exchanger will basically control the temp. difference across itself, with that differential tending toward zero, slowing the fluids more, providing more resistance to the driving force from the collectors. It's a lot more complicated than that, but still an ignorant idea. Not evil. just uninformed. I'd build a breadbox heater or a batch heater w/ a glycol (propylene) loop directly in the storage vessel.

J.P.M.

[LucMan]

I disagree with that assessment.
You also have to look at the temperature differential the heat exchanger is rated at. If it is 10KBTU @ a 150d temp differential it will actually be much less than that in this application where most temp differentials are in tie 10d -4d range.

Here is a rating chart based on 180 degree input at delta T. Of course different manufacturers will vary.
A delta T controlled pump would overcome of most of the variables on the primary and secondary sides.

[ptoddf]

One other thing not only necessary but without which the 3 mentioned become mostly irrelevant is the overall heat transfer coefficient (often called "U zero", or "U overall", or just plain "U"). The amount of heat transferred is U*A*Delta T. Those 3 are all interconnected and dependant on one one another to a greater or lesser degree depending on the design. Contrary to common, incorrect thinking, that value "U" is NOT constant. It's determination is actually the goal of heat exchanger thermal design, which is a very iterative process when done correctly. I'd suggest that putting a heat exchanger into a thermosiphon loop (unless it's something called a thermosiphon reboiler - which this is not) is an excercise in futility. The velocities obtained will mean poor heat transfer - low "U", thus requiring more area, thus additional lowering of "U", etc, etc. Besides, it'll probably be a plumber's nightmare. Without pumps in the system, the heat exchanger will basically control the temp. difference across itself, with that differential tending toward zero, slowing the fluids more, providing more resistance to the driving force from the collectors. It's a lot more complicated than that, but still an ignorant idea. Not evil. just uninformed. I'd build a breadbox heater or a batch heater w/ a glycol (propylene) loop directly in the storage vessel.

J.P.M.

Thanks for informative reply. What is a breadbox heater or a link to this info?

I had no clue that thermosyphon thru HX problematic. Certainly will be low velocity as it is in present system. Though for $50 I can get ebay HX and plumb it in to existing system and try it pretty easily. I also have a standard cast iron circ pump for solar HW panels never used I can stick into the glycol loop if no good. Then, I can hope that the other side of the HX will thermosyphon thru existing water tank hookups which do thermosyphon with direct hook up to solar water panels. Who knows? it MAY work. Seems like it should with sufficient velocity forced by circ pump through glycol loop. I don't want to use 2 pumps, one in each loop, but it sounds like worst case this would be sure to work. I am planning to have glycol loop unpressurized with a tin reservoir tank above to take expansion. HX membrane failure would result in overflow from pressurized side into vented reservoir. Just water in glycol loop for initial tests to see if it works of course. I strap little circular meat spike thermometers on input/output lines to see what is happening.

Thanks for useful pointers on physics of these systems, what I needed to know. Just the controlling parameters, conceptually. Interlocking systems never are never as obvious as they naively seem to be! This is not going to be an fully engineered system, as it was not when built. Though it works fine as is, I can't stand the freeze problems anymore.

[russ]

Take a lppk at http://www.builditsolar.com/Projects...ting.htm#Batch

Lots of DIY type water heating projects

[ptoddf]

Yeh, I've seen that site. Problem is I'm modifying/upgrading an existing working system. If I were starting from scratch I'd be using a static unpressurized water vat for heat storage and just throw in coils of PEX tubing for input and output heat exchangers, using my existing Grumman panels with pump and glycol, but unpressurized. But too much work already invested in existing system. I just want to freeze proof collectors in simplest possible way. If I have to add a pump to get enough flow thru heat exchanger from panels I'll do that, much as I enjoyed thermo syphon function in existing system. (Except for gradual scaling, it's been flawless, silent, no moving parts, it just works. Of course freezing breaks it.)

The gradual accumulation of scale in collectors will be eliminated with glycol/distilled water mix in them, but presumably not in tank side of heat exchanger. To be determined. Long as I can remove threaded couplings and pull out HX to descale it if necessary shouldn't be a killer problem.

I'm going to spring for an ebay bargain on small plate heat exchanger and see how it goes. Will report results. Will most likely need pump, which I have available if no thermo syphon thru heat exchanger, which seems to be the expert opinion here. I can jumper in a heat exchanger in a couple of hours. Has to be tried. Parallel, redundant heat snap switches on upper collector tube will be initial pump on/off control if pump is needed as expected.

Thanks, Todd F.

[J.P.M.]

Yeh, I've seen that site. Problem is I'm modifying/upgrading an existing working system. If I were starting from scratch I'd be using a static unpressurized water vat for heat storage and just throw in coils of PEX tubing for input and output heat exchangers, using my existing Grumman panels with pump and glycol, but unpressurized. But too much work already invested in existing system. I just want to freeze proof collectors in simplest possible way. If I have to add a pump to get enough flow thru heat exchanger from panels I'll do that, much as I enjoyed thermo syphon function in existing system. (Except for gradual scaling, it's been flawless, silent, no moving parts, it just works. Of course freezing breaks it.)

The gradual accumulation of scale in collectors will be eliminated with glycol/distilled water mix in them, but presumably not in tank side of heat exchanger. To be determined. Long as I can remove threaded couplings and pull out HX to descale it if necessary shouldn't be a killer problem.

I'm going to spring for an ebay bargain on small plate heat exchanger and see how it goes. Will report results. Will most likely need pump, which I have available if no thermo syphon thru heat exchanger, which seems to be the expert opinion here. I can jumper in a heat exchanger in a couple of hours. Has to be tried. Parallel, redundant heat snap switches on upper collector tube will be initial pump on/off control if pump is needed as expected.

Thanks, Todd F.

I'd suggest finding a way to arrange any tubing in such a way as to maximize the natural convection inside the tank. Simply throwing it in the tank haphazard may make already marginal heat transfer on the tank side of the coil worse. Also, I'm not sure I'd be using PEX.

The last time I checked, you'll need a pump for each (all) side(s) of the HX, otherwise you'll be back to thermosiphon performance, i.e. not much.

Fouling (Scaling) is a fact of life everywhere including the tank and glycol systems if not maintained.

BTW, probably not a problem here, but make sure the pumps are properly grounded. Ungrounded/poorly grounded pumps are an often ignored cause of corrosion in industrial piping systems.

If the vat is an open tank I suppose you could always simply put the HX inside the vat and in effect have yourself a thermosiphon plate HX. Not sure the inspector would like it and I'd question the setup for health/sanitary reasons.

One thing not talked about is plumbing codes. Some codes specify double wall heat exchangers depending on a lot of things like system pressures, fluids, system size, etc. Exceptions abound, but I'd check that out. Naptown is likely much more knowledgeable than I in that area.

[Naptown]

Single wall heat exchangers are allowed here if the following conditions are met
The glycol must be food grade propylene glycol.
In some places they require back flow preventers on the domestic side. Most only require a 9 D back flow preventer.

Back in the old days I was putting in Reynolds systems. These were double wall external heat exchangers. There was one jurisdiction that required a watts 900 back flow preventer.
When installed on a well they leaked ( normal operation) every time the cold water was run in the house.
Towards the end we installed the same one on about 50 systems and removed after inspection