Building Reserve and Using KWH

[DanS26]

Bruce, I take a somewhat different approach when trying to find an underperforming panel. Our solar plants are very similar. As you know, I have 96 ground mount panels made up of 8 strings feeding two Fronius 7.5kW inverters. Because its a matched set any time one inverter is performing poorly I can eliminate 4 strings and concentrate on the remaining strings. With my trusty DC Amp meter I can now identify the problem string of 12 panels.

Now here is where we differ on approach to identifying the problem panel(s).........

1. I shut down the AC feed to the lower performing inverter. That shuts down the DC load.
2. Next is to open the DC switch feeding the inverter.
3. Next I open all the combiner box 15 amp touchless fuses....all 4 of them.
4. Now that I'm sure the load is removed from the offending string which I double check to confirm no voltage, I proceed to detach all 12 panels from the string.
5. In full sunlight I now check the voltage and short circuit amperage in each panel with my trusty Fluke multimeter.

At this point I can usually find the problem, but if I still cannot find the panel then I have to check each panel under load.......

6. Using a 12v car battery and a cheap charge controller, I rig up a test system with my Fluke in the middle to measure amp production for each panel under load.

All along the way I'm also looking for poor, loose and corroded connections.

[bcroe]

The multiple string PV solar system here had a problem, here is my method to track
down an array issue.

This system is actually two complete 7.5KW inverter systems with identical panel
arrays. On a clear midday with good sun, the outputs usually match very closely. It
was noted, that the WEST system output was lagging the EAST system by several
hundred watts.

The ground mount array looked like this, all the strings operate at the same voltage
since they are wired in parallel. Variations in power will be caused mostly by variations
in string current. The first step was to use a clamp on DC ammeter to check the
current of each string. This gives best results under strong sun, and STRING 6 was
found to be lagging the others.

StringsTest.jpg

The normalized output of a solar cell looks like this, with about half a volt MPPT per
cell today. These strings have 12 panels of 60 cells wired in series. With 720 cells
in series the string MPPT total voltage is about 360V. With todays low sun and my
poor, rather low winter elevation, the MPPT current was about 5.5A. With a series
connection all cells in a string are forced to conduct the same current. Even with
matched panels the precise MPPT current will vary slightly, but each cell will vary its
voltage slightly to match the current of the others. This will normally stay pretty close
to the broad power curve MPPT top. The problem comes when a sufficient reduction
in voltage to match current, is far from the rated MPPT voltage. This fault condition
might run the panel (or a portion of a panel) all the way down to zero volts without
reaching a matching current. In that case a panel bypass diode will conduct to avoid
much reverse voltage.

PvMpp.png

The shift in voltage is used here to find an under producing element. A voltmeter can
detect the low voltage element, the rest of the string will raise its voltage to match the
other strings, but in sliding up the voltage curve, the current (and the POWER) drop.

My 12 panel strings looks like this, the big problem is actually connecting a voltmeter
to the weather proof wiring. To narrow it down, I would like to measure the voltage
across each quarter of the string, toward finding the under producer. The negative
end is at ground potential here, so the frame can be a test point. 4 more test points
are needed as indicated.

StringTest.jpg

I break apart the MC4s where indicated and insert a 3 way MC4 connector like this.
The 3rd port will be a voltmeter access. But WAIT A MINUTE! This array can deliver
enough power to cook 50 hot dogs at a time, and breaking any DC connection under
load can cause awful, equipment damaging arcs. Leaving any conductors exposed
to the weather can also cause serious damage.

3wayMC4.JPG

So here it must be PITCH BLACK to work on wiring, AC outlets at the array allow
use of work lights. The 3rd port is not left exposed. An MC4 with no conducting
metal is plugged in, its wire hole is plugged and the latches are ground off. The O
ring will retain it briefly. Then wait for good noontime sun.

Noon next day, the voltmeter reveals that three of the string sections accessed
are running good voltage, the section nearest the POS output is low. Just making
the measurements needs some special test leads. Long extension wires (no
metal exposed) are connected to the voltmeter, to reach any point on the string.
At the ends of each are both a male and a female MC4, because the 3 ways might
have either. Again, the latches are ground off, so the plug may be removed and the
voltmeter connected quickly without tools. The plug can be put on the unused
voltmeter MC4. In general, MC4s are kept mated to avoid weathering.

MC4TestPr.JPG


Once the offending section is identified, the test points are moved into that area
to resolve to an individual panel. WAIT FOR NIGHT AGAIN. The new locations
are shown here, near the POS end.

StrinSegTest.jpg

Here is a pic of the offending section, the top 3 panels, with the test points
already moved to bracket each one. The night work light is to the right, the
cement mixer not involved.

StringSegTst.JPG

At this point one panel might stand out with way too little voltage under good sun.
Weak sun is found to be inconclusive. HOWEVER, my situation was different. The
first attempted fix was to replace (IN THE NIGHT) the center panel with my unused
spare. Next day the center panel voltage was right where it should be, but those on
either side were not good. These panels had previously been washed, which had
no effect.

A careful examination of the removed panel revealed a dark, rough, somewhat
streaked surface. I remembered, for a time a pine branch had grown out above the
3 panels, though not usually shadowing. The panels were coated with PINE SAP.
Washing with soap did nothing, alcohol did not do it either. I took a straight blade
from a utility knife and SCRAPED off a black powder. Fortunately the glass
underneath was very smooth, so the blade slid smoothly across once cleaned. IN
THE NIGHT I swapped it with the panel on the right as seen above.

UtilitBld.JPG


Next day, it was obvious, 2 of the 3 panels were at voltage. From the front, the
remaining panel (now on the right) was dark, clearly affected. Sap can also be
seen running down the edge of 2 original panels.

3PanPineSap.JPG

This panel was swapped with another scrapped clean panel, and all was well
again. OK but have I removed any panel protective coating? Time will tell.
NIGHT AGAIN, remove all the MC4 test access points.

So why not wire test points in the day? First opening a working circuit could
damage connectors or worse. MC4s are not so easy to separate, and slips
could be fatal.

Then why not open the fuse and test with no current flow? The panels need
to be checked under load, near their MPPT point. Even a bad panel often looks
good unloaded. Bruce Roe

[DanS26]

Here I use the DROKs to monitor my genset, note the frequency is the box on top. Very useful to monitor loads during an outage.

There's a lot going on with this panel that a few readers may be interested in......

It's a 10 position critical load panel that I modified to accept 12 positions
An N1 switch was installed to exercise the generator under load without interrupting power to the house. See the red label.
The item with the blue LED lights is a MidNite Solar SPD that protects the panel from high voltage surges that might emanate from a submersible pump.
Below that is an SPD from Levitron which protects the main distribution panel.
Next below is the DROK frequency meter.
Next below is the DROK phase A and B load meters.

Inside the panel are a low voltage monitor device that protects the switch mechanism from low voltage burnout and additional SPDs to protect the generator AC and DC control circuits.

Finally the switch was wired using the Canadian code method for critical load panels. That method requires that all wiring for critical loads be removed from the main panel and routed through a junction box before landing in the critical loads panel. I went a step further and used a pass through block in the junction box rather than a mess of wire nuts.


Monitoring.JPG

IMG-173.jpg

[bcroe]

Those DROK meters certainly are compact and economical, having integration over
time (KWH) and frequency a bonus. They do draw power on the order of a watt, so
the V leads need to be attached on the supply side of the current transformer to
avoid becoming part of the readout. Before them I had wired this into 240VAC ckts.

ENEmeter.png

I expect to be using the frequency readout if I run the gen set. PF could be
useful on big motors. Bruce Roe

[DanS26]

Here I use my DROKs to monitor my EV circuit..........

EV Circuit Monitor-001.jpg

[Ampster]

I used the same Drok device and it worked well. I my case it was to monitor current and voltage while parallel top balancing a LFP pack.

[bcroe]

Did you build that 240V power meter yourself?

The readouts are DROK 80-300V 100A Voltage/Current/Power Factor/Frequency/Electric Energy Monitor
I picked up for $21 from Amaz. I use a pair because sometimes the load may be unsymetric, so then one for
each 120V line. The rest is just a box for a 4 wire 50A plug feeding a same socket, the hot 120V wires passing
thru the meters current transformers. Each has a pair of voltage wires, one of each meter is connected to that
wire. The neutral volt wires are connected together, and switched to the big neutral for 2 meters. The switch
can connect that node to a 120V line instead, then one meter monitors a single 240v load and the other is off.
Bruce Roe

4W50ADVM1.JPG

[nerdralph]

Still clinging to the old technology is my very reliable, easily repaired, electric stove.
At the beginning of my PV Net Metering, it actually did a lot more than heat food, and
there were a few days when every incandescent light in the house was left turned on
24/7. No need for that anymore. Most seasons my 240V KWH meters show the stove
only uses about 0.4 KWh a day, not even 1% of my energy budget. It, with some
standby drain, is an acceptable tradeoff.

240KWM2.JPG
Bruce Roe

Did you build that 240V power meter yourself?

[bcroe]

Tis the season to watch Wonderful Life, where a fantasy hero is played by a real
life hero, Major General Stewart. Even the dry days are pretty overcast and
cloudy, PV solar is way down. However PV still usually collects at least 25% of
optimum condition harvest, instead of dropping to more like 10%, thanks to the
extreme DC:AC panel ratio. Net metering accumulated reserve should carry the
property carbon free thru March, when average solar output once again usually
exceeds consumption.

One of the subjects of energy conservation here is drying clothes. No problem
in the summer, the sun does the job directly out on the line. Not so in winter, that
clothes dryer blowing all the inefficiently heated air outside was a matter of extreme
irritation until replaced by a heat pump, UNVENTED dryer in 2019. Moisture is
directed down a drain.

UVdryer.JPG

HOWEVER in practice it has been found, anything (not only clothes) can be quite
quickly dried by hanging near the output of one of my mini-split Heat Pumps. This
I believe not only saves a bit of energy, but helps boost the quite low winter indoor
humidity. Although I have gotten by with 3 HPs in the house, for more even temps
and plenty of reserve in case of -20F weather or an HP failure, the number has been
boosted to 5. They are located near the 4 corners, and one at the exposed end of the
basement. 2 are doing regular drying service, the thought is to put hardware in place
to extradite the practice for most of them.

MiSpIn8.JPG

I suspect when the weather does hit -20F, the Coefficient Of Performance of my HP
setup may drop to around 1, the same as the old resistive heating scheme used here
half a century ago. But that temp is only temporary, and as it soon rises, so does my
COP. It appear I am at about the practical cold weather limit of the current air to air
heat pump technology, probably could not carry thru the entire year at a lot higher
than my 42 deg latitude. The original 70s HP was COMPLETELY and TOTALLY
outclassed by my current units.

Still clinging to the old technology is my very reliable, easily repaired, electric stove.
At the beginning of my PV Net Metering, it actually did a lot more than heat food, and
there were a few days when every incandescent light in the house was left turned on
24/7. No need for that anymore. Most seasons my 240V KWH meters show the stove
only uses about 0.4 KWh a day, not even 1% of my energy budget. It, with some
standby drain, is an acceptable tradeoff.

240KWM2.JPG
Bruce Roe

[bcroe]

Once again in Nov the net metering winter energy reserve has peaked near14,000KWH, and
is trending back down with short days and 20 F temps.Energy collected for a year is at the
same 30,000KWH as in the past with this solar 15KW AC array. Energy use is up a bit with
conditioning the shop building, which is only done for very hot or cold months, and the winter
use ventless clothes dryer. But PV system efficiency is up with the lower loss AC wiring.

The first 3 mini split Heat Pumps here are starting their 4th winter, the next 2 their 3rd winter,
and the last one its second. This technology teamed up with PV is quite a game changer.
Besides being far superior in every operational respect to the old fixed speed compressor
HPs, they are just much more pleasant to be around. It is cold enough I sometimes spot a
defrost cycle. I just cleaned their air filters the first time using the powerful central vacuum.
Of course they are supposed to be cleaned every year, the filters were dirty.

I just got word from NYLE that the residential hot water model e8 air-source heat pump will
become available in April. Of course they said that a year earlier. I can connect it to my
existing water heater, if it works out I would change that heater from propane to electric. Not
sure I want to use it in mid winter, since it absorbs inside heat that my other HPs bring in.
Maybe best would be the ability to change water heat sources a couple times a year.

From another source I read, some of those electronic KWH meters like the PoCo uses for
net metering, use a battery. No battery needed in my ancient but reliable spinning disc
meter. Bruce Roe

[bcroe]

Another clean up, need to get the backup gen set ready. 25 years old, less than
10 hours running, never at this location. My problem is the PoCo breaker is in the
house, the solar inverters in the shed, which must both be shut off. Then go to
the car repair garage and push the gen set out the motor home sized door. I have
a 6 gauge extension cord.

My old suicide cord gets replaced with this unit, which has a big switch to be closed
after every thing is shut off, cords plugged in, and the gen set started.

GenSCord.JPG

To monitor both lines and even the frequency, I can insert my dual meter unit near the gen set.
I used to have the old 3 wire stove cord stuff, but have now upgraded most everything to
the 50A 4 wire setup, separate neutral and ground. Bruce Roe

4W50ADVM2.JPG

[bcroe]

Thanks for the input, the clamps found so far are in the $5 -$10 range. That is enough
to have me round up some stainless and make my own. There is quite a range of styles,
need to match mine.

Noting strain, the actual antenna IS mounted with thru holes, but there are cables (coax,
ant rotator, etc) that need to not just be left lying on the roof. They run off the roof and
down to my ground rods for lightning mitigation, before entering the house. With enough
clamps, these should not be a problem. thanks, Bruce Roe

[Ampster]

And in your best engineering judgement what areas may that be and what types of additional loads might you have in mind ? ......

I answered Bruce's question about sourcing clamps and offered some insight which others have discovered about attaching panels to seamless roofs. I think Bruce probably has enough engineering skill and knowledge about the size and location of his antennas that he can take take that conceptual information and translate it into a plan.

[J.P.M.]

Look at racking manufacturers sites or Google "standing seam roof clamps".

The initial fastening of the standing seam roof may have to be modified to accommodate any future uplift loads in areas that might have additional loads.

And in your best engineering judgement what areas may that be and what types of additional loads might you have in mind ? Your advice is so general that it's useless.
At least it's probably not dangerous.

[Ampster]

.......
Is anyone able to steer me to sources of those clips? ......

Look at racking manufacturers sites or Google "standing seam roof clamps".

The initial fastening of the standing seam roof may have to be modified to accommodate any future uplift loads in areas that might have additional loads. You would know what those areas are and what the loads would be based on your antenna needs.