Voltage loss

hillside

Just cannot find this voltage loss. I've been trying for several days now and have got nothing.
Overnight last week the SRNE controller showed 1.0 volt in the batteries. Normally it's at 24.5V. The system is used for TV at night and washer/gas dryer during the week.
I keep usage low but constant so we can use the system when power is out for extended periods. Like hurricanes. 
This setup has been working perfectly for 3 years now. 2 strings of panels each to it's own CC, each like this:

The SRNE controller on the left string was the first thought of trouble. Morning voltage at CC (early morning pre-sun) shows 27 Volts.
But the Renogy CC on the right shows 60Volts, which is normal.
Several readings since then are similar:
Mid-morning 48 Volts vs. 63 Volts
Peak sun mid-day 34 Volts vs. 65 Volts
Best reading on left CC has been 55.5 Volts vs. 65.8 Volts on the right.

And it looks like this has seriously compromised the 4 3 year old L16 Deka batts. Specific gravity shows all in the fair range, 1200-1225 and I see no percolation anymore.
I thought maybe one panel had gone south. Doesn't seem so.
I've checked open circuit voltage on all 8 panels and all are consistent, (although below max.) at 32.5V.
I found a pile of ants in the pvc tube going thru the concrete wall on the left string and thought if they've eaten thru the + wire could that cause the voltage loss.
Then I checked that wire before it enters the wall and what really surprised me was that it showed 64 Volts. So, I thought that was the problem. But it isn't. A few minutes later it was back to 50 volts. Same reading inside at the CC.
So, could I have a shaky connection at one of the MC4 connectors? Or at the Y splice connection? All cables are secured and don't move around much. For 3 years this has worked so well.
I've reached the limit of my knowledge here. Now I think I'll need 4 new batteries but this voltage loss needs to be fixed first. 
Thanks for reading.

BB.

This debugging sounds like it may be a pain...

Both charge controllers are MPPT type.

The voltages you are telling us is the Varray voltage (50/60/32/etc volts).

First one array+controller had issues... Now the second sometimes has issues?

And you see these issues primarily as Varray voltage issues?

If you have a DMM... Can you measure the voltage at the Vbatt terminals of each MPPT controller? And measure the Vpanel voltages too?

If it was one panel or cable that was having problems... Generally Vpanel input for the controller should be consistent (Vmp-hot array). Just variation of current (1/2 string "dead", 1/2 current)... Since it is the full array with voltage issues--Then the common wiring between the two parallel strings would be where I would look most closely. It is not unusual for crimp connections to fail, water corrosion, poor crimps, etc.

With MC4 connectors, you have to pull them apart to see the pin and if the cable pulls from the crimp. Look for signs of overheating (metled plastic, browning of insulation) and corrosion.

One easier method to check parallel strings--The AC+DC Current Clamp DMM (digital multimeter). It is so nice to just clamp one wire on a panel vs its parallel neighbor--They should have very closely matched current. If you see three strings with 5 amps and 1 string with near zero--then you can focus on the low current string connections and panels).

https://www.amazon.com/gp/product/B00O1Q2HOQ (less expensive priced AC+DC current meter/DMM)
https://www.amazon.com/gp/product/B019CY4FB4 (mid priced clamp AC+DC current meter/DMM)

With "matched panels" (parallel strings)--You have a nice A/B/C/D test for current/voltage/wattage (same angle to sun, etc.). The "different" panel/string is probably where the issues are.

Also, don't forget to check the Battery Connections (Vbatt) for the controllers. If you have loose/dirty/corroded connections from controller to battery bank--Those can really confuse the charge controllers too (they need solid/stable battery voltage connections--Why I also asked you to measure Vbatt controller connection voltages).

And, the first thing to try is usually the "reboot".

1. Disconnect array voltage (only one + wire disconnect is needed)
2. Disconnect Vbatt voltage (only one + wire disconnect is needed)
3. Wait a minutes or two.
4. Connect Vbatt power to controller and let boot
5. Connect Vpanel power to controller

Never have Vpanel powered without Vbatt already being connected and powered by the battery bank. If you "boot" with Vpanel first, it is possible to "confuse" the charge controller and even possibly damage it in some cases.

The MPPT charge controller varies array current to keep Vmp-array roughly stable (typically ~80% of Vmp-std rating). A confused controller (bad boot, failed controller, etc.) could account for the readings. Unstable Vbatt voltage possible too...

Current Clamp (DC capable) meters are wonderful to use in debugging DC power. No damage to wiring, no taking connections apart to insert current meter, etc.

Voltage readings at different points in the wiring are useful for debugging too... But MC4 connectors do not lend themselves to checking voltages while connected--And I would avoid piercing the cable insulation with a sharp probe if I could (hole in insulation can allow water into the copper wiring).

Is Vbatt wiring to solar charge controllers from battery bank. Does each controller have its own +/- cable set "home run" back to bus bar (and breaker)? That is the "ideal" method. I would avoid (for example) daisy chaining controller... Vbus to Controller A to Controller B.

Your batteries--Need to get them charged (>75% SoC--And possibly EQed). Using a genset/AC mains charger... Or even if you had too, bypass the charge controllers and just connect array to Vbatt-Vbus to get some charge in the batteries ASAP (not ideal, but need to get batteries charged).

That is where I would start.

-Bill

hillside

One string has been working fine since this started. It shows 63-68 Volts so the batteries have been getting some charging while the other weaker side shows about 20% less voltage.
For now I've covered the 8 panels with tarps and began charging one battery at a time. It's a small charger @ 6 amps for the 6 V side. Will test at the end of the day and see what it looks like. thanks!

hillside

Update;
So, because life gets in the way, I haven't had time to look into the cause of the voltage loss.
This little system continues to work and do what we need and the batteries have mostly regained their capacity. 
Just measured voltage array and one was 66V and the other 65V so...first on the list is to remove a small tree that at this time of year blocks array number 1 early morning. Then on to the rest.
On the specific gravity test, hmmm. just 2 cells of the 4 batts. show a reading above "Fair" but just below "Good" while all other cells show "Good".
The tester is this--and it has an open space between Fair and Good.; sorry about pic size.


So, a couple of the batts are a bit compromised, correct? But still have enough power for what we're using it for.
Now to the question of the day.
Thinking the batts. were toast I bought 4 new Deka batts. 370 AH and I'm not sure what to do with them. One option is build another system like the current two, or wait a while and see how these existing batts play out over time.
So, how long can I store 4 new 370 AH batts. without using them? In a hot climate..that never gets below 67* overnight.

hillside

Edit, to add, voltage array at 65-66V is normal range since inception.
Batts. are L16

mike95490

Put a charge onto the new batteries.   They will need topping off at least monthly, but their age clock starts ticking when filled at the factory.