Home solar

  • Thread starter Thread starter bwilson4web
  • Start date Start date
  • Replies Replies 34
  • Views Views 5K
I've had DIY solar for two years. Best ROI is grid tie with microinverters. If you are going with batteries you want an AIO like a Growatt or EG4 with built in MPPT etc. The biggest change I made is trying to charge my car in the daytime when I'm not on battery. I have enough reserve to pretty much do everything else, the GF pretty much runs the TV 24x7. I don't have energy efficient appliances, and I don't care. I have sensors everywhere monitoring my loads, I'm in the middle of revamping the automated scaling for the various EVSE and Crypto devices. Should easily hit 100 SOC on the batteries before 3:30, Crappy day yesterday, I overdid it, batteries bottomed out around 15% last night. The big 4T will flatline at 5KW in July/August, my computer crap pushes my steady state threshold over 1KW.

upload_2025-6-2_13-28-52.webp

Electricity here is pretty inexpensive, off-peak rates around 0.08/kwh, but I went nuts anyway to prove I could do it, knowing what I know now, and with the plethora of new AIO's on the market, I could have done the whole endeavor quite a bit cheaper. Panel costs have plummeted since I bought my first set for $8K. The second, only slightly smaller (455x32 vs 410x32) set was $3K. I spent $6K each on my hybrid inverters because I needed the performance to start my 4T HVAC but now the same company has an off-grid unit with similar specs for $2500 ea. My initial 30KWH of batteries was $10K, the second $8K. At ~ $3K/yr for electricity, not a great ROI, but I've only run on the grid twice since January, the last time because of some maintenance to the gear, prior it was just 3 cloudy days, and a bit over-aggressive about charging cars, think it cost me an extra $2.00 overnight. My Solar is primary behind an ATS with the grid as my backup source. When I do tap the grid for a boost, I run a 5KW battery charger to the cells, basically I have a big solar/online UPS running my house.

upload_2025-6-2_13-2-20.webp

I have 6 strings in panels, totaling around 28KW. A-PV1 is west facing, to catch extra afternoon sun. I broke a personal record of 181KWH a week or so back. I have 60KWH of battery, and 2x12KW output inverters (~ 100A) that run my whole house pretty much all the time. I started running bitcoin miners as a dump load last month or three, my daily usage in July and August can hit well over 100KWH on it's own, so we'll see how that rolls thru the summer.

upload_2025-6-2_13-11-8.webp

The day I hit 180, I charged a car and ran 3 miners close to wide open all day, peak demand actually hit 99A, a bit more than I wanted, mined about $6 worth of BTC. I would not use an EV battery for storage, too much fiddling & twiddling with semi-custom gear. A 50KWH 48v battery rack can be had for around $12-13k, I'd much rather have standard LifePO4 packs, that don't suffer from thermal runaway.

upload_2025-6-2_13-44-48.webp

I've been slowly adding capacity to my system, the last thing I want is one of those 50KWH racks, I have two short racks of 6 now, I'd like to replace them with two tall racks of 10 for a total of ~ 100KWH. Unsure as there are some new battery technologies that might prove out better, we shall see. Now if I can just get my pretty smooth blue PV curves back. This overcast crap is killing me.
 
The thing people don't understand about Solar, and power distribution is the first thing you have to take care of is your peak demand. I goofed on the 23rd, I allowed the car charger, crypto-miners, HVAC, HWH all at the same time, the load hit 107A, which is 7A over the inverters rated capacity. Fortunately it was short lived. I also bumped 101 a couple of times in the following days, nothing fell over, it's good to know there is some headroom. I de-rated my sub-panel to 125A just above the inverter output, thankfully I still wasn't close to blowing the main, but my ATS would have cut in if the inverters fell over. I want one more inverter so I can just ignore it.

upload_2025-6-6_9-45-13.webp

My 'purdy curves are back for the solar, but you can see 'yestidy when I just ran out of places for the overproduction. Topped off the car (red) with excess battery and anticipation of the sunny day. Crypto (purple) generates a lot of heat, I'm still working on venting the output from the miners outside. Again, demand is going to go WAY up in about a month. The yellow (HVAC) line is going to go flat starting around 13:00 ramping up to 5KW until after 21:00. This means i must have a "full tank" in the batteries to meet overnight usage, and will not be able to "bleed" battery overnight and early AM to fill the EV.

upload_2025-6-6_9-54-19.webp

This problem is the same as the power company when it comes to renewables.You or I doing our own thing, we are going to try and do stuff when it's convenient for our production, the PoCo has no such luxury. Like I want to be, they need to have enough overproduction to meet whatever the demand is. Further if you 'grid-tie' since they cannot guarantee any particular output from you, they have to have hot-standby if your production ceases. This makes your backfeed pretty much useless to them since they can't rely on it.

The problem with using "Solar" to only charge a vehicle is you waste enough production to generally run a home when you are not charging. This is actually what started my journey. You don't have to charge at 7KW, but . . . The sun doesn't shine evenly thru the day most places, so you have to buffer your output somehow (generally with batteries). Also you really want to take maximal advantage of the window when the sun is shining, so you need to ramp up your charge rates to follow the sun. I think i discussed this in another thread somewhere forgive me if it was here.

A typical 6KW inverter put's out 25A, realistically you can charge reliably at 20A/240v = 4800W, and your going to need about 16-20 400W panels (depending on your sunshine amounts) and a 5KWH battery as a buffer. Now I have 20A or so I don't need when I'm not charging, and hey! I have a backup I can use if the power goes out, 5KWH would keep the fridges, lights, tv's running for a really long time, even run the microwave, and . . . So if i go for the 12KW inverter, a few more panels, a bit more battery, ...

You end up here.
 
The experiment continues:
solar_car_charging.webp
  • The Tesla L2 charger runs about 85% efficiency
    • Battery charge rate does not include overhead
    • Battery cabin cooling with A/C runs about 1.1 kW
      • Fan only is about 600 W
  • Maximum solar output is 20 A with any available charged battery 22 A
    • Tesla max battery rate 32 A means 38 A circuit current
  • Tesla at 16 A results in 18.9 A circuit load leaving 1.1 kW for house load or solar battery charging or Tesla cabin cooling
My plan is to monitor the solar panel output with a current sensor and adjust the EV charger rate to 85% of that value to the Tesla. This will adjust the charging circuit load to follow the solar panel power yet preserve the solar battery charge. I'll probably use an averaging technique to handle passing clouds.

There will be an optional, "MAX" rate that will signal 32 A. This will take power in the following order:
  1. Solar panel output
  2. Solar battery output
  3. Grid power (buying kWh for speed)
There remains the question of interfaces:
  • hard wired - will cause "angst" for electricians
  • Wire-less or combinations
    • LED lights and buttons
    • Bluetooth
    • WiFi
    • Zigbee
Bob Wilson
 
Last edited:
Passing clouds are going to kill you worse than you think. I think you said you were going with micro-inverters, but they are not very good unless you actually grid tie, so the grid can handle dips, valley's and demand. Just for the car as the original spec it will be interesting to see. With solar batteries you should be able to buffer output to your average solar input or a little higher depending on how well you float. Sell back is pointless, you definitely want to use everything you produce locally if you can.

But the main thing is ... Are you having some fun playing around with it. Cause if your not, it's a job. :)
 
Actually I bought the whole system from a contractor to "buy their expertise." They do a lot of solar for the oil patch and have been around for years.

I am grid tied and the micro-inverters are working well. Initially I was disappointed that they only provided 300 W from the potentially 400 W panels. Then I looked at the geometry and remembers a beloved, former GE manager, Dick Hicks who once told a software supervisor, "Hank, 90% isn't good enough, it is perfect! Ship it!"

Getting the software on to the integration and test team meant we could document problems, assign priorities, and get the software to a usable state. Many developers have little operational 'common sense' and will happily 'polish the musket balls' instead of delivering product.

Yes, I am having a ball as whole energy lifestyle is changing. I'm becoming better user of when it is available. Better identification of a parasitic loads, and better able to fully exploit my EVs and live cheaply.

Bob Wilson
 
Yea, on the panel side, I was dissapointed, then pleasantly surprised, disappointed to max out at around 75%, then pleasantly surprised to get output that extended beyond expectation, at least here. Good rule of thumb is 6 hours at 75-80% when things are optimal.
 
Keith Smith said:
Yea, on the panel side, I was dissapointed, then pleasantly surprised, disappointed to max out at around 75%, then pleasantly surprised to get output that extended beyond expectation, at least here. Good rule of thumb is 6 hours at 75-80% when things are optimal.
Click to expand...
A retired engineer, I am fully aware about how important it is to manage "User Expectations." Sad to say, my contractor could have done better. I had to ask Mr. Google and read the product specs to find out what a good job they did. As always, the customer is the ultimate "integration and test" engineer.

My girlfriend was over today so I showed her the Apps that monitor my system: (1) Franklin system controller; (2) SPAN breaker panel, and; (3) solar panel monitor. Seeing the numbers has converted "theory" into "reality."

Bob Wilson
 
Last edited:
It's not just your contractor. You expect the NMOT numbers on the panel spec sheets you are dreaming. BUT, it really is pretty amazing when I realize my solar setup can finally charge a car, and run my house all day without me even paying attention. So I do what you are doing for stoopid stuff like running bitcoin miners. After doing all this, I've realized this will eventually be the norm for 99% of a homes power needs in the not too distant futre for a large number of homes in areas where it makes sense like here, but really anywhere with a reasonable abundance of sunlight.

I've never seen my 455W panels actually put out 455W. But with a 400W panel running close to $100, if you have the space, just put up a couple more panels.
 
Woa, I just caught 'SPAN Breaker panel". You dropped a few shekel's on that puppy. I cheaped out with CT's and an aquisition device. for 8, and I have some $10 stand-alone energy monitors, all RS-485. As a programmer I found the software challenge fun. Most engineers program only at gun-point, er, I mean as proof of design concept ;).
 
In my case, time was important: GOOD, FAST, CHEAP

I bought their expertise since the price was same order of magnitude as the TSLA stock sold. The 30% tax credit nicely balanced the capital gains taxes.

Truth be told, I was not impressed with the other contractors who didn't return phone calls and were imprecise in what they were offering. I got the impression of "cottage industry" versus the experienced professionals I hired.

Regardless, I got what I wanted and am in the process of learning how to use it best. For example, load shifting to maximize direct solar power to load instead of taking the battery detour:
  • 90% loss charging the battery (typical for AC to DC)
  • 90% loss discharging the battery (typical for DC to AC)
  • 90% * 90% = 81% round trip loss
By matching loads to the solar panel output, I gain ~19% more power from my home system.

Bob Wilson
 
I tried to explain this to others. 10% loss every time it moves. What that really means is you either maximize like you are playing with, or you over-provision to meet you needs. When you are the power company, you don't have the former luxury, so they have to work from the latter. At the low cost of your power, it will be interesting to see if you can hit your savings goals. OTOH, I can be confident in saying that your power is not going to get any cheaper over time. Batteries as a minimal buffer is the most ideal. I agree this is still a cottage industry. Knowing what I now know, I could do a much cleaner job from scratch. Too many fly-by-nights in the Solar business. Worse than roofers.
 
LESSON'S LEARNED (near perfect day)

A productive day:
perfect_day_03.webp
  • "A" - early morning running on battery reserve for home lights, computer, TV, e.t.c.
  • "B" - battery reserve exhausted, using grid power
  • "C" - sun rises and solar panels begin taking over the house load and charging the solar battery
  • "D" - started charging BMW i3-REx at 16 A from 10 miles remaining range. During the day, clouds passed and reduced solar power that was filled from the battery.
  • "E" - BMW charging completed, full battery (28.3 kWh max.) Charger connected to previously, fully charged Tesla Model 3 that had been kept in "cabin heat protection" with A/C. The earlier 152 mi range had dropped to 137 mi. This reflects the cooling energy needed to compensate for keeping the cabin at 88 F in the sun. The Tesla started at 12 A but midway, I increased it to 16 A, like the BMW, even though it drew grid power.
    • Tomorrow, I will test an external windshield screen and roof-top sun screen. The windows will not be covered.
  • "F" - all of the battery reserve was consumed as the sun set. The grid assumed the house load.
My planned operational scenario was to store as much excess solar in the EV cars, 2017 BMW i3-REx and 2019 Tesla Model 3. This day showed the BMW can be 'fully' charged with some power left over. Then I was able to quantify how much energy the Tesla used during the day to maintain an 88 F internal cabin temperature from the solar panels and solar battery kWh.

All in all, close to a perfect day.

Bob Wilson
 
Last edited:
You must log in or register to reply here.
Back
Top