diy solar

diy solar

My final version

hubieg

New Member
Joined
Feb 5, 2021
Messages
25
Greetings all, It seems I had been here quite a while ago but recently refound this site. I figured Id take a moment to share my system that is finally basically working. Ill also share my finding what went right and what went wrong. Since I knew it would be an ongoing project with many mistakes to make, I built it all on the store shelf I had in storage and got for free. This allowed to design as I went with plywood and shelves. This was after the initial system with the AGM batteries on the ground and plywood on the wall.

When I say this is the "final version" I did have it up and running before. When I started I started with a EP-Ever MPPT BN and used AGM batteries. 1st lesson, dont waste money on used batteries from a Facebook or Craigslist seller. I say it in this way because they usually seem to pick up their batteries from someone swapping out due to age or cycle limitations. I got about a year out of 10 200 AH batteries before all but 2 lost a cell. At $100 each it was not worth it overall. Better to buy newer ones from maybe someone who is fed up with the path to solar.

Speaking of batteries a big shout out to "Will Prowse" from You Tube and his series on building lithium packs. After seeing his videos then doing the math, I decided to go all in on Lithiums. The seem a bit intimidating but after running the numbers on cost/performance they are indeed cheaper than even FLA batteries. This carries one caveat however. They are cheaper if you buy them direct from China. I got mine 200 A/Hr 3.2v cells direct on Aliexpress. This leads me into the issue with BMS's.

I ordered the batteries and a set of 100a 8s Daly BMS's. Once they were put together the first big lesson of BMS's rears its ugly head. It seems Daly (and many other) BMS's will not run in parallel. This was confirmed after initial hookup and rehookup, and re rehookup. I then contacted Daly and they told me that thiers will indeed NOT run in parallel with each other. At this point, it was back to the planning.

It was during this I was turned on to "Overkill BMS's". Only one slight problem. The Same "Will Prowse" apparently caused Overkills business to increase 4x running them out of stock when I needed it most. Because of this I did some searching and talking and found the supplier (or manufacturer) of their supplier in China. Even though I ordered them from China the credit MUST go to Overkill! Their tech docs and data was instrumental in making sure that all was correctly setup.

After receiving my initial set of 16 cells, I guess I got scared and had too much money in my pocket. I decided to order a second set. This gave me a total of 19kw of storage. Im glad I did now that its up and running in the winter in Ohio. One other thing that had to be changed for the lithium conversion was the MPPT. EP-Ever said their BN will not work with LiFePo4 batteries. Because of the expense involved I decided not to risk it and ordered their recommended Xtra Series MPPT. With all this done and ordered if anyone is interested I now have 2 BN MPPT's and 4 100a 8s Daly BMS's surplus.

Long story short, the system is now up and running. One thing I should note. I am a machine controls engineer by trade. Because of this I am working into a heavily automated system. To do this you will see the Raspberry Pi on the platform. You can also make out 5 different boards on top of this. This is a Raspberry Pi 4 with a Codesys runtime in it. This turns the Pi into a "soft PLC". The extra boards on top are all TTL to RS-485 converters. If using industrial equipment the choice always boils down to some form of Ethernet, CAN, or RS-485 for communication. As of these photos, the Pi is recording and displaying data (to the world over the net) from the MPPT. This is because the MPPT speaks the industrial "Modbus RTU" protocol. The other boards are for the BMS's all with RS-485. Problem here is I need to write the protocol handler for the PLC.

Currently I have the PLC displaying read only status data out on the net. If you wish to take a look, the address is http://phantomworx.com:10111 . This is essentially an industrial HMI package that comes with the Codesys runtime. Note it is still a work in progress and at the moment every major change I make loses the short term data. The "trend" however stays. The Trend chart only contains the battery voltage and the MPPT charge state. Dont worry about pressing buttons when you look around. Anything that ca affect system is password protected. Also keep in mind this is a work in progress so it may change in its "look and feel" from one day to the next.

The last big thing of note is the result of a challenge I saw on Quora. Someone asked if a welder could be used as a charger. Everyone replied no. I couldnt theoretically see why it wouldn't work through an MPPT. Because of this I bought a cheap inverter welder from Amazon to use as a charger. This was initially tested with the BN MPPT and AGM batteries and found to work beautifully. If you decide to try this route just make sure of the duty cycle of the welder vs amperage and use a DC welder (preferably of the inverter type!). This particular welder I believe is 100% duty cycle up to 60 amps. As of right now the welder when being used as a charger puts out about 20 amps to the MPPT. Note the latter is what is regulating the current and not the welder. In fact the amperage dial on the welder seems to have no effect but the MPPT limits output to 40a.

Lastly you see the knife switch on top. This selects (manually) the source to the MPPT. Source is either panels (900 watts) or welder. This is part of my next change. It will be replaced by electromagnetic contactors. In this way I can add a relay to the welder to turn on and off and to changeover the source remotely. Sequence will be something like disengage panel contactor, wait 10 sec, engage welder contactor, then turn on welder (just a rough sequence overview). Eventually this will automatically happen if the batteries are too low. Onle last thing to note is that it seems the MPPT will not throttle back as it is currently setup. I will have to take data from the MPPT to know when to stop charging by shutting down and disengaging the welder (or so I think).

One side note of an earlier experiment. I built a motor driven alternator rig a while back. I gutted the diodes and VR from the alternator (burnt those out on my first test running directly to the batteries). I then brought the wires out and connected a 480v 3 phase bridge rectifier to the output. For field I used a Modbus or manual controlled power supply. I then ran everything into the BN MPPT. Worked beautifully without any filtering of the output and ran for 6+ hours without issue. This is a secondary project though now that I know it works and doesnt damage the alternator or MPPT. Also because of the shift away in the alternator from amperage to voltage, it ran cool to the touch the whole time.

I welcome all comments, questions or observations. Please no arguments about the alternator or the welder. Here it works. If you want to talk about it thats fine but please dont get into things like :that cant or doesnt work". It can and it does. On the other hand it could screw things up if not properly done. Note I also tested running two MPPT's in parallel. This seemed to work as long as one was simply connected but not making power. They were also BN's and not the Xtra's. Also they will fight each other if they are both making power.

Check it out and I would love to hear your comments and questions!
 

Attachments

  • 20211212_163231.jpg
    20211212_163231.jpg
    200.7 KB · Views: 84
  • 20211212_163124.jpg
    20211212_163124.jpg
    198.2 KB · Views: 80
  • 20211212_163050.jpg
    20211212_163050.jpg
    125.8 KB · Views: 76
Cool adventure! Sounds like you fell into the rabbithole, got chewed up a bit, but made it out ok, and lived to tell the tale :geek:

Have you ever looked into playing around with 'Solar Assistant' on the Raspberry Pi? Perhaps you don't need it, sounds like you have the other solution working well enough with the PLC...

I haven't tried Solar Assistant myself, but have read a bunch about it, and seems it might help me on my setup anyways. I'm going to try it.

Anyways, happy trails, thanks for sharing your story...
 
Cool adventure! Sounds like you fell into the rabbithole, got chewed up a bit, but made it out ok, and lived to tell the tale :geek:

Have you ever looked into playing around with 'Solar Assistant' on the Raspberry Pi? Perhaps you don't need it, sounds like you have the other solution working well enough with the PLC...

I haven't tried Solar Assistant myself, but have read a bunch about it, and seems it might help me on my setup anyways. I'm going to try it.

Anyways, happy trails, thanks for sharing your story...
Thank you, its been a long road and I can say the experience has NOT given me a lot of newfound hope for the solar future of the country :-D . All I m trying to do is run a normal sized refrigerator 100% of the time on 900 watts of solar. So far though looking at what was accomplished I think I have found an emergency power systen much better than running a generator 24/7. Once the controls are fully in place and automated, Im hoping to power my house and have to run a generator only a few hours a day at its most efficient power vs fuel consumption level. In fact a 3000 or so watt electric start inverter generator is high on my priority list now. Ultimately the panels would be for maintenance of the system and to give a few sunny days of quiet :)

I also thank you for mentioning "solar assistant". The PLC though is essential to the overall plan. A future purpose will be to stop and start the generator as needed in addition to the other things I mentioned. From there it also looks like sun tracking will be a big improvement looking at my charge curves (as can be seen with the "peakiness" of my charging curves). I already have functions for the PLC to build and run the tracker in addition to anything else I can think of :) Throw Node red into the mix and I can control smart plugs and switches on an as needed basis via an industrial comm protocol.

Lastly using the PLC helps keep me sharp in my profession when I am stuck in a contract job where my skills are not being fully used.
 
Yeah, I get you on the practicing for work thing, we use a lot of Python and Java in our product code for work, so I also like to do labs and things for home that involve those two, so it can help me be better with it at work, totally get that, it does help though.
 
Back
Top