On Tue, 17 Jun 2014 13:06:57 +0100 (BST), "Dave Liquorice"
wrote:

On Tue, 17 Jun 2014 02:40:19 +0100, Johny B Good wrote:

Mindful of the comments I'd seen recently about these solar panels
having high reverse leakage sufficient to undo the benefit, I checked
and measured a mere 2 micro-amps (using both a digital meter and an
analogue one on its 30 micro-amp range as a sanity check).

Rescued mine from the garage, not a lot inside. Connected across the
PV panel is a blue LED and 5k1 resistor. The ouput has a 1N400? diode
in series, that's the sum total of bits inside. Don't know what the ?
is as that's against the tiny bit of PCB.

The diode is most likely a 1N4001, the lowest PIV rating of that
series, one or two hundred volts afair. It's worth keeping in mind
that this represents a volt drop equivilent to slightly more than that
of a single cell but it's a necessary sacrifice not just to prevent
backflow in darkness but also to stop the the LED drawing current from
the load (usually a battery) and giving a false indication. Very
likely the panel will be made up of 30 photocells in series so this
will represent about a 3.3% sacrifice of power output.

The 5K1 resistor rings a bell. I can't remember whether I opened up
the original unit or not but I certainly did for the second one on
account the indicator LED wasn't working (was seeing the same 20ish
open cct volts and able to draw about 80mA charging the battery).

I replaced the self flashing blue LED with an ordinary blue LED but
replaced the 5K1 resistor with, most likely, a 15K one to compensate
for the 100% on duty cycle and also because those blue leds can be
fiendishly bright on anything more than a couple of milliamps, they
use far less current than the older red and green LEDs to achieve the
same brightness.


The 1N400? range is supposed to have between 5 uA and 50 uA at max
reverse voltage and 25 to 125 C junction temp. Testing gave an open
circuit output voltage of about 19 V. Apply a load and it could
produce about 24 mA @ 12 V ish. Not measured the drak leakage, might
do that later.


I'm not sure if this is typical of a decent quality of SLA but if it
is, it would seem best _NOT_ to keep them on a float charge long
term.

They ought to be OK provided the float is within the spec of the
battery, taking into account the battery temperature. The SLA's in
the alarm panels lasted far longer than those in the UPS. APC UPS's
don't keep within spec of the batteries...

Well, the float voltage is so tightly specced that there are three
different voltages quoted for the classic 6 cell battery between the
AGM, Gel and flooded cell types. The optimum voltage being a
compromise to avoid excess corrosion and sulphation for each of the
three cell types although I've never seen anything other than 13.8v
being specified for any of the 12v SLA types (55.2v for a string of 4,
hence my choosing to set the float voltage at 55.3 open circuit but I
think, in the light of the wiki article I should revise this down to
54.1v).

According to wikipedia, the three voltages are 2.23, 2.25 and 2.32
volts per cell. For the 24 cell battery pack I need to select between
a low of 53.52 and a high of 55.68 volts depending on battery type
chosen and, afair, I think VR4 trimpot covers the whole range. The
last time I checked the charging voltage I think it was set to 54.3v
open circuit which may well explain the relatively short lif of the
SLAs (although not for the even shorter life of the car batteries).

Luckily, as well as having a cheapish Aldi DMM to hand, I've also got
a Fluke DMM that _was_ calibrated just before it was gifted to me when
I quit BT just over two decades ago.

When I bought the Aldi DMM, the first thing I did was to compare its
calibration against the Fluke and was assured to see they were within
a couple of millivolts measuring a 12v SLA (as was an even older Tandy
DMM). With three DMMs to hand, I think I can be reasonably confident
of their accuracy being sufficient enough to set the float charge
voltage correctly.
--
J B Good