Posted by Jim Wilkins on March 30, 2019, 1:25 pm
<ads> wrote in message
I've been experimenting with a DC-powered refrigerator/freezer as an
extended outage substitute for the inverter and compact refrigerator,
since the always-on sine inverter consumes nearly twice the battery
energy of my fridge. Being a UPS it's perfect for nights when the
power may go out, but it won't last 24 hours.
The result so far is that the DC compressor-type cooler runs about 3
times longer on a battery charge but is inconvenient to live with
since it's small and doesn't have a separate freezer compartment,
28~30F isn't a good long-term setting as both frozen food and liquids
slowly become slush. However it works pretty well in the car on trips,
running from a jump starter pack when the engine is off.
The datasheet shows that my East Penn batteries should last 50 deep
discharges so I don't think the occasional run time test and immediate
recharge hurts them excessively.
Posted by ads on March 31, 2019, 5:16 am
On Sat, 30 Mar 2019 09:25:04 -0400, "Jim Wilkins"
The lowest power "fridge" is a chest freezer with a fridge conversion
thermostat. They are better insulated than a fridge and the cool
stays inside when you open the door. If you search for a model with a
fan cooled compressor coil (instead of them using the metal shell of
the freezer as a radiator) it's even more efficient as the outside of
the inulation isn't being broughjt up to 100F when the unit is running
(like a lot of upright freezers).
I did a couple of 50% discharge cycles in testing my AGM battery bank
when I installed it but they were monitored by an AH meter so I would
know when it reached 270AH and not just a voltage that indicates
"they're about 50% discharged". I only plan to take them down 50% in
actual backup use so that's how I tested them. I wouldn't take any
lead-acid battery (flooded or sealed) below 80% DOD. I have an AH
meter in one side of the wiring to the battery bank so I can track the
AH that's been used when we're on the solar generator. The AH meter
will also track the AH put back in the battery but it there's no way
to tell it that the battery isn't charged until 115% of the used AH
has been replaced - there's still a need for a human to handle battery
Posted by Jim Wilkins on March 31, 2019, 12:53 pm
<ads> wrote in message
I bought several import battery charge/discharge power meters from
Amazon to test and so far all except the SIN9020A have drifted far out
of spec or failed, while only one of the wattmeters that measure
unidirectional current in its negative power lead went bad. The Drok
200117 that shows +/-50A bidirectional current but not Watts or Ah
still works. The 2405S isolated DC-DC converter operates down to 5V
input at the low current a bidirectional ammeter requires.
I use the constant-current charge voltage and then the decreasing
current that the battery draws after it reaches float voltage to
roughly indicate SOC, based on previous discharge tests. I've read
that unless you know otherwise by measurement the battery can be
estimated to be at 70~80% SOC when it reaches float voltage and full
when the float current falls to 1% of the Ah rating, 1A for my 105Ah
batteries. The only decision SOC drives is whether or not to use the
The bad unidirectional wattmeter, a PZEM-031, was connected to rooftop
solar panels and may have suffered static voltage during a
thunderstorm, so now there are less vulnerable analog meters
permanently monitoring the panel voltages, partly to show if the
diodes in the controllers have failed and are leaking battery voltage
into the panel wiring at night.
My 45W Harbor Freight controller did that. They used two small diodes
in parallel instead of a larger one rated for the full current, and
one shorted. I replaced them with one much larger Schottky that hangs
over the edge of the board to dissipate heat better. It wouldn't
withstand shipping vibration as well as the originals, but it no
longer has to.
The solar panel volt meter also shows at a glance if my MPPT
controller is starved for current, working normally, or in float mode,
and whether the panels were left disconnected for a thunderstorm or
jumpered for 12V or 24V.
This is my meter calibration standard
which agrees to the last digit with my Fluke 8800A 5-1/2 digit bench
meter. I found a 1.0006 milliOhm lab-grade current shunt in an
electronic surplus store.
Posted by ads on April 1, 2019, 5:55 am
On Sun, 31 Mar 2019 08:53:48 -0400, "Jim Wilkins"
I have that Drok (or its twin) and the 10 volt output is listed as
9.99668 volts. I don't have anything that displays that many digits,
but adjusting the calibration on an inexpensive (but stable and
durable) DVM to 10.00 with that standard gives me readings of 0.005V
or better. The DVM was last checked in 2016 and when checked this
week it was within 0.03V after almost 3 years. That's very stable for
my "Toss it in the toolbox" meter from Harbor Freight, but I've found
that their very inexpensive DVM (in a red case) has an internal
calibration pot and it holds calibration well. I'm only interested in
high accuracy in the range around 12 volts so calibrating the meter on
the 20 volt range gives me acceptable accuracy over the range of
I have better meters, but it's nice to have some that can be left in
convenient places and you know they'll probably survive a 5 foot drop
to concrete and if the drop changes the calibration, you can correct
that in a minute or two. I also have one of these in each vehicle for
troubleshooting vehicle problems or when I'm at someone else's house
(I'm the family handyman and I always have some tools in my truck ;-)
Posted by Jim Wilkins on April 1, 2019, 12:34 pm
<ads> wrote in message
The HF cheapies have an input impedance of 1 MegOhm and don't work
with high voltage range multipliers or to show the voltage of the
megger I use to hipot appliance repairs and measure diode and
lightning arrester breakdown. Otherwise they are great as battery
testing beaters that might be zapped by high current. I found the cal
pot touchy and sensitive to knocks.