The cause of data gaps in CSV files

 

There have been occasions when data gaps occurred in the CSV files collected daily.

The problems are unique to the power file. The azimuth and elevation are printed, but nothing else.

Further, that data is collected outside the schedule. Somhow, the NOAA2.py script is being run outside of schedule, and printed to the CSV files. The EpD3Dy.php script is run by a service/timer. The NOAA2.py script ius called from within the php script.

The problem may disappear if the python script is run in a different service/timer pair.

I'm examining an epevermodbus python script to read and record settings separate from the php script. This will allow me to examine the data for problematic results, such as negative values or other data anomolies.

I reset the timer to a time after the sun has risen. Unfortunately resetting the timer startup very season is not practical. I may be able to set the startup time from within the service or the script.

This problem is relatively recent. Perhaps this is a result of the differences between the Epever Tracer and Triron charge controllers.

Autumn change

 

Autumn and the shorter day is coming.

I need to raise the panel array up higher since the sun will not be rising as high in the sky until next Spring.

The best solution would be a concrete base for the bottom, 1/2"+3/8" rebar reinforcemant for the steel axis pipe, a more permanent support for the high end.

I need to get short 'samples' of the rebar to insure the fit before buying the ten foot long pieces.

 

working files and folder structure

 I am making a working set of files and folders available for free.

I've done this before but so much time has passed, and changes made, I need to do this again.

Instead of posting each file on the blog, I'm providing a Dropbox public link.

Files link

Battery bank wiring modification

 

The battery bank has been wired in parallel with system connections at the closest battery. This creates an unbalanced load on the bank. 

The modification: instead of direct wiring to the closest battery, one connection to the closest battery (+) and one connection to the most distant battery (-).

The ideal arrangement would be separate connections to the system for each battery, a radial configuration.

Inverter final test

 

I finished the installation of the replacement lever actuated terminal connector WAGO 2606-1103.

This afternoon, I had occasion to run the inverter/generator. I plugged the unit in while the Xantrex inverter was on. The system was maintained. Notably, the RPPi4 experienced no discontinuities in the data stream.

I count this as a success.

There was a brief  lapse in the lightiing, about 1/3 second, but this was partly due to a setting on the iniverter. The transfer is set up for APL, appliances, but could be set up for sensitive electronics. The second option warns against transfers with inductive loads such as electric motors (water pump, fridge) running.

Power limit discovered

 

I have discovered the reason why my charts never rise above 560W in delivered power in my 1kw system.

My system has 24v panel system and 12v battery system. The charge controller can handle both but the inverter only takes a 12v input.

If I want to get the full 1 kw power delivery, I need to replace the 12v inverter with a 24v inverter. Then the charge controller can run at maximum capabilty. I will be able to charge the 5kw battery bank in 5 hours on a sunny day. Since the bank is never drained completely, the time will be significantly less.

'AC In' connector replacement.

I have received the WAGO 221-412 and 221-2401 connectors in the mail. Those have replaced the wire nuts traditionally used to tie two household grade wires together, typically 12/2.

I have removed the connector sub-pcb and unsoldered and removed the damaged connector using the solder iron, solder, solder sucker and flux paste.The pcb is ready for the new WAGO 2606-1103 connector. I added two push-fin cable strain relief clamps for the AC connections.

I received the new connector a couple of days ago. Preparations included setting up the new soldering iron plus solder, sucker and flux paste. The replacement went smoothly. After the replacement, I re-assembled the unit and removed the 1200W unit. I attached the 2000W unit to the wall after attaching the AC and DC cables.

I started the unit this morning and all is well. 

The final test will be the attachment of a running inverter/generator like the Predator 2000 to test shore power detection and switching.

WAGO 2606-1103 pcb connector for 'AC In'

Before

After

Repaired unit

Wire nut replacements, 'AC In'

Wire nut replacements, 'AC Out'

Inverter modification strategy.

 

The closest visual equivalent to the Xantrex unit's AC In connector is the WAGO 2606-1103.

After removing the connector pcb, I discovered the brand and general model series : WAGO 2606.

The inverter needs a modification.

An AI agent (Google) suggests adding stranded 12AWG as a stress relief strategy.

Stranded 'pigtails' go to the pcb then to WAGO splice connectors that tie stranded and solid wires together.

The existing lever connectors would be removed and stranded wire soldered directly to the pcb. Alternatively, the broken WAGO connector could be replaced and stranded wire inserted there.

The AI response mentioned an inherent weakness in using solid 12AWG wire in those connectors and that reports of loose or intermittent contact had occurred. Hence the stranded wire substitution. This is intuitively correct and appeals to experiential common sense.

Inverter modifications.

 

I tried plugging the generator into the external outlet but the inverter had an unusual display.

On re-reading the Xantrex manual, I discovered the generator has to be connected to the 'AC In' terminal.

Two problems exist.

1.) There has to be a circuit wired for the inverter, from the back porch outlet, to the inverter.

2.) One of the 'AC In' terminal blocks is bad.

I discovered the connector is a WAGO device.

That block (VAC Line) has to be replaced or the wire soldered in place after removing the bad terminal block.

google AI: Is there a way to balance the voltages in a prallel array of PV panels

 

google AI: Is there a way to balance the voltages in a prallel array of PV panels

Yes, you can balance and manage mismatched voltages in parallel solar panel arrays. Left unmanaged, parallel panels force the entire array to drop to the voltage of the lowest panel, which severely limits your total power output. [1, 2, 3, 4]  

Here are the most effective ways to balance voltages and maximize your yield: 

1. Use DC-to-DC Optimizers (Best Solution) 

Installing smart module-level optimizers (like those from  Tigo Energy 

 or  SolarEdge 

 on each panel conditions the voltage of individual panels before they combine. This allows every panel to operate at its maximum power point (MPP) independently, completely eliminating the voltage penalty. 

2. Independent Charge Controllers 

If your panels have vastly different voltages (e.g., mixing a 12V panel with a 24V panel), the best approach is to wire them into separate, dedicated charge controllers. Route each panel or string to its own  MPPT Charge Controller 

 before connecting them to the battery bank. 

3. String Inverters with Multi-MPPTs 

If you are using a grid-tied string inverter, look for a model that features multiple Maximum Power Point Trackers (MPPTs). You can wire your mismatched strings to separate tracker inputs, allowing the inverter to individually balance and optimize the voltage for each parallel group. [10, 12, 13, 14]  

4. Direct Diode Protection (Emergency Use Only) 

In a pinch, if you have two panels with similar but slightly mismatched voltages, you can use blocking diodes (or Schottky diodes) on the positive lines. This prevents the higher-voltage panel from "dumping" current into the lower-voltage panel. Keep in mind that this is not recommended for permanent setups, as the lower-voltage panel will still fail to produce its maximum potential power. [12, 15]  

Rules of Thumb for Parallel Arrays 

• Keep them close: If you must wire panels directly together in parallel, ensure their $V_{mp}$ (Voltage at Maximum Power) ratings are within $0.5\text{ to }1.0\text{ Volts}$ of each other. 

• Match series strings: If you are building parallel strings (multiple panels wired in series, and then paralleled together), ensure every single string is identical in both panel count and model type. [11, 13, 16, 17]  

AI responses may include mistakes.

[1] https://www.reddit.com/r/SolarDIY/comments/1dqtao7/can_i_connect_different_voltage_solar_panels_in/

[2] https://community.goodsam.com/discussions/technicalissues/series-parallel-configurations-for-solar-panels/3561082

[3] https://www.facebook.com/groups/447643471600275/posts/682772848087335/

[4] https://xantrex.com/about-xantrex/blog/marine/understanding-the-impact-of-series-parallel-and-mixed-solar-wiring-on-system-efficiency/

[5] https://diysolarforum.com/threads/unbalanced-pv-in-series.64527/

[6] https://diysolarforum.com/threads/can-shading-be-dealt-with-by-parallel-strings.54317/

[7] https://solarpowernation.com.au/power-optimizers-explained-enhancing-solar-system-performance/

[8] https://www.linkedin.com/pulse/how-fix-voltage-mismatch-solar-pv-systems-causes-solutions-barua-jtkuf

[9] https://www.anernstore.com/blogs/diy-solar-guides/rapid-shutdown-strings-microinverters-optimizers

[10] https://diysolarforum.com/threads/volts-series-amps-parallel.10654/

[11] https://diysolarforum.com/threads/solar-panels-with-different-voltage.75463/

[12] https://www.youtube.com/watch?v=OnV8UYBqzD4

[13] https://www.youtube.com/watch?v=b2H8vpj8rQg

[14] https://www.facebook.com/MultiTechInfoo/posts/when-connecting-multiple-strings-in-parallelvoltage-must-match-closelyif-one-str/1408124434688499/

[15] https://diysolarforum.com/threads/how-do-i-match-solar-panels-of-different-wattages-in-a-parallel-array.6581/

[16] https://solartechonline.com/blog/how-to-wire-solar-panels-in-parallel/

[17] https://www.facebook.com/groups/2573968699280898/posts/8156362984374747/

Inverter replaced plus more

 

The most recently purchased inverter that seemed to have random overnight shutdowns was returned.

Instead, I paid  little more and got a newer version that has been installed and running without event for over a week.

I ordered and received fifteen Staubli MC4 connector pairs. Staubli is the original designer and manufacturer of MC4 connectors. This will, I hope, resolve any intermittent connection errors due to cheap, marginally standardized, connectors.

Another improvement is the addition of fifty feet, each, of red and black #8 AWG cable that will reduce any line losses due to existing #10 AWG cables running to the array from the house. 

A hand crimper, for smaller gauges, and a hammer crimper, for #8 AWG and larger sizes, have been added to my tool set.

A series disconnect switch has been placed in the PV array-to-charge controller circuit. A series 40A fuse has yet to be installed in the charge controller-to-batteries circuit.

A second Epever Triron charge controller was purchased as a replacement/backup for the existing unit. For a while, random fluctuation in csv recorded data occurred in the first 30-45 minutes of operation, in the mornings. I reset the start time to 6:30AM and the fluctuations have stopped. The repair mentioned previously, seems to be holding up.

After some calculations and some thought, and a consultation with Epever, I came to the conclusion the reason my system was not charging at a full 1KW was the batteries are in parallel. If I wire them in series-parallel with a resulting voltage oof 24VDC, instead of 12VDC, the controller will increase the rate of charge. for the inverter input, at a maximum current of 170A.

However, I would have to use an inverter that takes a 24VDC input and mine are rated for 12VDC. The 24VDC verison of my inverter, Xantrex Freedom X 1200W/2000W, is hard to find, especially on ebay, and is very expensive.

I could add a 'buck', or step-down, converter, but they are expensive. The converter would convert 24VDC to 12VDC

For the time being, I'll settle for a 600W maximum charge rate.

Used Xantrex inverter potential problems

 There have been a couple of instances where the inverter may be failing or creating anomalies.

I'm gettinga new RPi4 and using an SD card to eliminate the existing RPi4+SSD as a source of problems.

A new inverter may be needed. The csv files are not recording as usual.

A long standing problem solved

 

For a long time, The graph display was not reflecting the change to Daylight Savings Tme.

The solution lay in editing the php.ini file to set the date.timezone to "America/New_York".

The setting in the EpD3Dy.php file was commented out.

The display is now correct when the spring/autumn changes occurs.

However, the graphing now ends at 7PM instead of 6PM. A minor edit will correct that error.

Spring forward-Fall back

Triron repair finished

 

The Triron repair has been completed. 

I replaced the two 40A blade fuses using some from a local auto parts chain store.

The Triron, like all similar Epever products (Tracer, Xtra, etc.) have screw down terminals that are too small and are difficult to use. I spent more time getting the cables into the input termiinals than repairing the unit.

The unit seems to be running normally.

RPi4 standalone wifi - Redux

 I have come to the conclusion that the simultaneous use of a RPi4 as WiFi Access Point and Hotspot is too unstable to use. The attempt at the Hotspot has been abandoned.

Spring overview and changes

  The system has been running well.

I decided to get another Xantrex Freedom X  for the system. This one is the 2000W version. The previous one is 1200W. The 1200W model is now a backup component.

The Epever Triron stopped working. On opening the unit up, I discovered one of the battery terminals was burnt and one of two 40A fuses had blown. The unit was not detecting the PV input. The fuse probably goes to the PV bus.

The unit is in the lab and will have both of two 40A fuses replaced. I didn't use the Load terminals, so I removed them and replaced the burnt Battery terminal with one of the good Load terminals. In the interim, I'm using the Schneider C60 PWM charge controller. I may add an Epever 4210BN as primary and keep the Triron as a MPPT backup.

If none of the semiconductors is damaged, the unit may work once the fuses are soldered in and the unit re-assembled.

I decided to replace most of the system's #10 AWG wiring and connectors, with #8 AWG wiring and Staubli MC4 connectors. I purchased a crimper for this purpose and ten pairs of Staubli MC4 connectors.

The biggest reason for the last change is that my system is capable of 1KW output but has peaked at 600W. Cheap connectors and wire may have resulted in loose connections and resistive loads.

The finished update will have new higher amperage wire and connectors with improved output. 

At some point in the future, I hope to relocate the array into the old garden area. I recently slid the array to one side to uncover what was becoming a bare spot.

One last possible change that needs to be thought through is replacing the one inch steel pipe, in the array frame, with a steel strut. Perhaps a two inch steel pipe would be better and easier to install.