SWH project part 16 - up and running

It's going! I'll be leaving behind a short list of outstanding finishing touches, and a couple of suggestions for future improvements, but in the meantime it's working: preheating the water for the laundry so the geyser doesn't have to work so hard, and thus saves fuel.

Solar Water Heater opening ceremony.  The first overcast day in ages, but was
actually just as warm and twice as humid.  Sweaty.

Some of my main helpers.

Re-fitting the heat exchangers (see here for why)

Lid back on after much intricate wielding of the hacksaw.

The brazing chap from Lahore was back doing some other work in the hospital,
so I got him to attend to a minor leak.

One last piece of colourful frazzled copper.

Snug as a bug in a rug about to chug a mug of... hot chocolate?  Glug!

SolarBot doing his thing.

Modest temperatures on a cloudy morning.  In the middle of a clear day the
panels hit 80°C, and the tank gets to 65-70°C.

Again, on a clear day the output beats 50°C.  This could be improved by
lengthening the heat exchanger coils, or by moving to a pressurised tank
with no heat exchanger.  The latter is almost certainly worth doing.  When
someone gets round to it...

Finished system.  For a given value of "finished".

Tank insulated and roofed.  Among the final touches I'm
leaving for the workshop guys to do at some point is to stop
the rainwater collecting on the platform and soaking into the
bottom couple of inches of insulation.  Preferably before the
monsoon starts in earnest.

Thermal expansion coupled with cycles of thunderstorms
followed by hot clear days has done for these strips we used
to hold down the glazing.  They'll replace them with some-
thing better and redo the seals...

(This one's an extreme example - the other nine seem to be fine!)

...which should hopefully stop this happening.  They'll also put some metal
sheeting along the top edge of the panels.

SWH project part 15 - teething problems

Temperature sensors

The day after I mounted the temperature sensors on the panels, we had a big storm.  Had this happened after we had put the glazing on, this would not have been a problem.  In reality, despite the sensors being encased in a large blob of silicone, the water managed to do some impressive capillary action where the wires enter the silicone, shorting out 5 of the sensors.  One of the sensors inside the tank also failed, though this was more expected. 12/18 still working!

Water tanks

The original setup made use of an old blue plastic water tank, which acted as a heat store (unpressurised) through which passed a couple of copper coils carrying the water to be heated.

System diagram - with heat exchanger.

Old blue tank with heat exchanger coils.

While the heat exchanger method was giving decent enough results, after a few days of sitting at 70-80°C, it became clear that the surface of the plastic was not responding well to the hot water, so we bought a new superior quality (allegedly!) tank to replace it.

This tank had the advantage of having a screw top, hopefully allowing us to seal the tank. This would allow us to run the tank under pressure with direct connections to the water supply, dispensing with the heat exchanger coils.   The larger tank will also run cooler, meaning lower heat losses and thus better efficiency.

System diagram - without heat exchanger.

In a separate issue, it was found that the pressure loss caused by all the additional pipework meant that there was no longer enough pressure for the washing machine's liking (it's a bit fussy!).  So we raised the height of the water tower supplying the pressure by another 5 feet.

Newly raised water tower.

Anyway, yesterday we sealed the top of the tank, screwing the top on tight with silicone applied to the screw thread.  After giving it a few hours to cure, I opened the valve from the water tower.  The tank ballooned out slightly as expected, then the top just popped off!  Inspection revealed no cracks or tears anywhere; the lid had apparently just stretched enough for the screw threads to pass over one another.

Recalcitrant lid (with air release valve and sensor leads), having just popped off.

So, take 2: this morning we strung a bunch of jubilee clips together around the edge of the lid, with the hope of holding the screw threads in a meshing position.  This seems to hold okay, but the way the tank bulges makes me nervous about its future longevity.

Bulging lid.  The rest of the tank looks similarly bloated.

I think we'll revert to the heat exchanger plan (assuming we can get the kinks out of the copper pipe).

Tank stand

A stand which was adequate for the 600l tank was not sufficiently strong for the new 1200l tank, with a couple of dodgy welds breaking. Fortunately we spotted one of the legs beginning to bend in time to avert a complete collapse, and the stand was then beefed up considerably with an extra central support and some proper diagonals made of some nice new chunky angle iron to stop it flexing.

Beefed up tank stand with new 1200l tank.

These problems partly stem from the way people are taught to learn parrot-fashion and the general low standard of education here - the guys have an idea what a structure should look like, but often without really understanding why you need to put a particular strut in a particular place. Hence you will end up with something which looks similar to a strong structure, without necessarily having the function of one.

SWH project part 14 - glazing

Or possibly plasticking.

We got this twin-wall polycarbonate shipped up from Karachi.  Bizarrely, it was
only available in sheets 19' x 6'11", so all of the panel sizings were designed
around this.  We had them slice them into 5 equal parts for easier shipping.  And
on the right panel at least, easier mounting too, as we've had one or two minor
alignment issues since we first put the frames up.

This is the satisfying bit, like peeling off a 7-foot plaster.  The lighter blocks are to
keep the glazing off the pipes, should it bend/sag once it's warmed up.  Yes, we
probably should have painted them black too, but that would have wasted another
half-day while we waited for it to dry.

Whether by learning from experience or from fluke, we did a better job of
keeping things perfectly straight when we constructed the left panel's frames,
so all the glazing just dropped neatly into place.  Aah.

All glazing in place, secured down by wooden slats and liberally silicone'd.
And now ready for its first wash by the looks of it.  This is a generally dusty
place anyway, but possibly being near the incinerator doesn't help.

Specially procured polyurethane insulation for the backs of the panels.
The Pakistanis' favourite cheap styrofoam is no good as it has a tendency to
melt at the temperatures we'll be asking of it.

3 down, 7 to go.  We've also insulated all the hot pipe runs with fibreglass wool.

Trying to screw stuff into place while holding it upwards onto a slope of 50° is not
easy.

SWH project part 13 - plumbing

Pipe-laying in the trench.  Since it was dug a few weeks ago
up to 3 feet of vegetation has sprung up along its edges.

Every joint requires you to cut 2 screw threads with this die, which is no mean
feat even when you're not trying to do it in the trench.

Cold line, sensor cable conduit, hot line (which needs some
attention with the tape...)

Coiling the heat exchangers.  Strangely, they coiled beautifully
 in this direction, but when we started trying to bend the ends
 upwards, it was threatening to kink at every opportunity.

So, how complicated can we make this?

Copper coils successfully installed, and due a little de-kinking with the pliers.

Exchangers connected up.  Incidentally, you should never
do what's been done here and connect the copper directly
to the galvanised pipe.  I predict corrosion before too long,
but there's only so many times you can tell your boss "We
should REALLY do it a different way".

Still, at least all the parts likely to suffer from galvanic
corrosion are cheap, on the outside, and easily replaceable.

This was a good bit. We're using that big black tank to provide
 the pressure to run the water to and from the SWH, but when
we came to disconnect it so that we could alter the pipework,
 we discovered that no one had thought to put a stop valve on
 the outlet.  Which means you can't work on the pipes down-
stream without 500 gallons of water coming out.  This is what
we call a DESIGN FLAW.  Solution:  just start unscrewing
the pipes (fast!), working backwards from the nearest union,
then quickly screw on a new valve, all the while being sprayed
by this forceful jet of water.  Refreshing on a hot day, mind
(we're up in the mid-to-high 30s these days).

We have water!

Data logger up and running.  Had a few short circuit problems to sort out when
I discovered a couple of the sensors weren't as waterproof as I might have liked.

I'm still in love with the colours of frazzled copper.

SWH project part 12 - welding*

*or in actual fact, brazing.  But everyone here seemed to call it welding...

This has been a busy week, which may or may not have had anything to do with Mr P's out-of-the-blue announcement the other week that "I think we should have it up and running by the 15th". Perhaps a little over-optimistic given that that's tomorrow, but nice to have some sudden ambition introduced to the proceedings.

So the panels have been fitted into the frames, and their tubes welded together and leak tested. Pipes to connect the system with the laundry have been purchased, along with glass wool sheets to insulate them with. Polyurethane foam sheets for the backs of the panels have been ordered, produced and delivered. My temperature monitoring system is pretty much complete, and semi-installed (more on that another day).

A good week's work.

Fitting the panels required a spot of saw-wielding.

First three in place - some dropped in beautifully; others had to be beaten into
submission with planes, saws, hammers and a little brute force.

Even before paint and glazing are applied the panels are around 50-60°C, and
visibly bending.  This week it has been high 30s outside, and in the sunniest spot
on the compound, in front of what is essentially a 7ft x 40ft mirror, it's HOT.

A couple of chaps came up from Lahore to do our welding.  After they'd been
connected together some of the panels required re-beating-into-submission.

Lugging those oxygen cylinders across the orchard is no joke...

The pretty colours of tortured metal.

Finally all welded together and in place.  Time for a leak test, which in typical
Pakistani style involves about 5 hoses shoved into each others ends, and a guy
holding it on to the tap at the other end.

A leak!  But actually in one of the previous welds from when the panels were
fabricated.  Still, 1 out of 164 ain't bad.

Take that pesky leak!