Major update

 

A lot has been done since the last post.

Previously, I had 580 watts of solar panel power. One of the panels, on examination, had lost a significant amount of power. That panel was replaced by a new 100 watt panel, making the system 600 watts.

More power was needed. I ordered two Renogy 100 watt panels from ebay and added them. A few weeks later I decided to buy a couple of HQST 100 watt panels and add them.  That's ten panels in all.

The current power total is 1000 watts, or 1 kw (kilowatt). There is a noticeable differenc in the charge rate of the LiFePo4 batteries. I run the mini-fridge and the internet connection more often now, and, dependiing on the weather, overnight.

The second half of the upgrade was the mounting hardware for the additional panels. steel struts used in electrical construction in factories and institutions is the framework I chose. The necessary additional pieces along with nuts, bolts, washers, lock nuts, couplers, and strut nuts were purchased.

I believe I have reached the maximum array size for the strut framework. I'm uisng half channel Superstrut. Full channel struts would approximately double the capacity, but the phyusical size of the rray is starting to become awkward. I need to be able to single-handedly move and manage the array. A single array that size could become potentialy dangerous.

A second array would be in order.

Another option: fiberglass struts. The load bearing characteristics would have to be examined and comppared to the array loads and to the steel strut characteristics.

I neglected to add an image of my revised webpage for the solar array monitor.

In case the link on the new charts page is not working, here is that image:

RPi4b storage memory: SD to SSD conversion

 Ordinarally, the RPi uses an SD card for memory storage. Those cards have gained a reputation for unreliability. My suspicion is oxidation of the contacts.

The RPi is now capable of USB drive storage. I setup an SD card and an SSD hard drive with the RPI operating system. I followed the instructions for use of a USB memory device and now use a 250Gb SSD for memory storage and operating system installation. The SSD is much faster than an SD card and has enough space to contain the operating system and any programs I wish to install.

New webpage and chart(s)

 

The new web page, for monitoring the solar power station activity, has two charts, one above the other. There are links to a copy of my old website, a more touchscreen friendly version of the page, and a sunpath chart web-app by Dr. A.J. Marsh.

The new chart software was more difficult than I imagined, still, I got the work done.

The charts are synchronized with a vertical crosshair. A vertical line appears wherever the mouse is over the charts. A similar line occurs, simultaneously, over the second chart and is synchronized with the first. The page is updated, approximately, every 30 seconds.

The top chart shows panel array and battery bank, volts and amps. The bottom chart shows panel array power output and battery bank  power input. The solar elevation (altitude) and azimuth (heading) are charted too.

There are two solar path chart images at the bottom. One in Cartesian format, the other in Polar format. These predict the sun's elevation and azimuth for any day throughout the year. The images are large and can be downloaded for easier and more accurate reference.

           The new charts - live 

February 2025 update

 Over a year ago, a lightning strike destroyed or damaged most of my solar power station.

Only the panels were spared, except for one which had a leaking diode, which I replaced. The lightning came into the house over the aircraft cable guy wire carrying signal and power cables from the panel array and into the station inside the house.

I replaced the Epever Tracer 3210 charge controller with an Epever Triron 4210.

I replaced the inverter with a Xantrex Freedom X 1200. The thing I like about that product is the auto switch off when the presence of an external power source is detected. If  I connect a generator to house electrical system, the inverter automatically disconnects.

My lead-acid batteries had a much shorter lifetime or were destroyed outright. The battery bank went through a major upgrade. LiFePO4 batteries had become affordable, and were, therefore replacements for the Duracell Ultra Marine lead-acid batteries. Each of four batteries is a 12VDC/108Ahr battery. Three are made by WattCycle, and one is a LiTime product.

I also chose to, at this time, upgrade and update my Raspberry Pi and associated software, always a formidable task.

I updated the RPi3b with a RPi4b-2Gb. The prevoius operating system, Debian Buster, was replaced by Bookworm (Debian 12). 

PHP and Python programs were both updated to current versions as well.

The Exar driver, always a pain and always taking days, was a pain and took days. However, the programmer of the original driver software, Kasbert on github, provided software to send and receive messages from the Exar driver, now an integral part of Raspberry Pi OS, labeled as 'xr_serial.' I needed software that would work to get the driver running but without any interaction. The work-around lasted for a week or two. I tried everything to use the code, but the minimal use resulted in two failed tries before USB0 would run consistently, for each use. 

I tracked down the linux source code and discovered some code in an article on serial rs485  communications. I extracted the code and made minor modifications to get a runnable executable. I compiled the code in the Raspberry Pi, installed a service to run at boot, and that solved the Exar driver problem.

Currently, everything is up and running. My next task is to change the online graph software. 

I have been using HTML, PHP and D3. The new software is Dygraph 2.2.1. That software seems so much easier than D3. And more versatile, but not without problems.

The queue also includes a new array support system made from 'superstrut.' The old patched aluminum ladder plus 2x6 and 2x12 is gone, as is the large concrete slab used as an anchor for the ladder and array.

February 28, 2025., spell check April 06, 2025.

RPi update & battery calculations

 The RPi is online and running 24/7 now, but not collecting data or moving the array. 

I decided to bite the bulket and update to Bookworm OS and python's virtual environments.

I still haven't gotten the network set up the way I want and the OS included Exar driver is not working, and while the Maxlinear driver compiles, it is not inserting ('insmod').

More study is needed.

I still need two more batteries.

Battery charge time calculations.

I have 585 watts of solar panels that I will round off to 600w.

I also have 4.2kW of battery storage.

4.2kW/0.6kW=7hr to full charge

But, the charge controller has a ma charge rate of 20A for the configuration.

20A x 12V = 240W=0.24kW

 This results in a charge time of 

4.2kW / 0.24kW = 17.5hrs to full charge from complete discharge.

A complete discharge is not likely since system shutoff is set at 10.5V.

Four batteries

 I now have four LiFePO4 batteries in place and balanced. The batteries store 4.8 kWhr of energy. 

The RPi is now set up and running a 24-hour schedule.

New Batteries: Three and counting

 There are now three LiFePO4 batteries wired in parallel, connected to my system.

   There is a fourth on order.

   Eventually, I want to have six for a total of 7.2kWh of stored electrical energy. 

   The catch is, there is a high probability tariffs wiil be levied against Chinese products by the incoming President. Prices will go up. There's no way to know how high prices will go and if availability remains steady.

   For that reason, I'm considering buying the remaing two before next year (2025). Another, more positive incentive, is a 30 day price guarantee for $147.16 per 12V/100Ah battery.

Currently, the retail is $169.99. I will have to think about this some. The guarantee ends on the 23rd, I believe. I'll have to tighten my belt, if I do.