Project Summary
My desire is to test the viability and benefit of using an off grid solar array to offset my utility costs of powering bitcoin miners at home. My goal is to produce 12 hours of mining power(130 Terahash) from solar and use my utility for the remaining time. At bare minimum, I simply want to avoid paying the increased demand rates my utility provider charges during on peak times.
Here is a rough schematic of my electrical design and soundproofing + ventilation of the miners. Being in the Arizona desert I know I have potential temperature challenges ahead of me. I am interested in submerged mining but I wanted to complete this portion of the project first. I also have ideas for utilizing the exhaust heat if I maintain air cooling but for now I am using it to keep baby chickens warm at night.
What is Bitcoin Mining?
Bitcoin mining is the process of converting electricity and computation power into network security of a global decentralized monetary network. Bitcoin miners are Advanced Speciality Integrated Circuits which perform SHA 256 Hashing to achieve network consensus and confirm pending transactions into the timechain, aka blockchain, or more simply ‘the distributed ledger’ for everyone to see.
Across the globe, millions of these speciality computers are all competing to solve a very specific solution to a very specific problem. When that solution is found, that miner is rewarded with newly minted bitcoin + transaction fees of all pending transfers, and the network immediately begins attempting to solve the next problem. In very simple terms, MY bitcoin miners are guessing 100+ trillion times per second trying to solve a problem along with 200 million trillion guesses from computers across the network.
Project Details
My electricity rates are $0.08 kWh during 3pm – 8pm weekdays and $0.05 kWh all other times. I do have a one time surcharge per month of my highest usage during 3pm to 8pm so that complicates my calculations a little. I would still mine at these rates because I believe bitcoin is highly undervalued at current average price of $40,000 and I am securing the monetary network I utilize. I am looking to show that solar power can alternatively be used to mine bitcoin at the time when energy is cheapest and help balance the grid demand vs ‘selling’ it to the utility.
My system components were purchased over the course of a year and cost me approximately .56 bitcoin, which was converted to dollars at an average between the price points of 38k and 54k for all equipment except the miners which were purchased for bitcoin. My desire is to make this investment back, measured in bitcoin, within 36 months. I believe the dollar value investment will be recouped sooner.
14 x 435 Watt Solar Panels (USED)
Growatt 6000 Offgrid Inverter
3 x LifePower 48V 100AH Batteries
Bitmain S19j PRO 104T running Vnish
2x Bitmain S9 (USED) running Braiins OS+
Axios 8inch Inline Duct @ 800 cfm.
Various fuses, wires and components.
Used lumber, steel and 8 gauge wire.
1.5″ foam & 1/2″ plexiglass for miner box.
My Reasons & Resources.
Following March of 2020 my wife and I woke up to the fact the system is stacked against us. After 12 wonderful years owning and operating CrossFit Fury and helping the community find better health, we realized our money is broken. There is no way our kids will repay the massive debts we have seen incurred in our lifetime, let alone theres. We need to separate money from the state and bitcoin represents the most viable solution. We nearly immediately sold our business and completely exited the fiat system and adopted a bitcoin standard.
I have always been a building in the physical world, working in general contracting before opening the gym and pursued metal fabrication as a hobby. I however had very little computer science and technical training besides physics and mathematics in college. I began immersing myself in linux systems, command line and python programming. My first step in the bitcoin ecosystem was to begin running my own full node to understand the importance of being your own bank. I then stepped into the lightning ecosystem and built the routing node AZHODL.
Along my way I have had various mentors that may or may not have been aware of their significance in my learning, including my not limited to:
STEMr42, Sevi & LNRouter, Openoms, Rene Pickhard, Alex Bosworth, US Representative Thomas Massie and most importantly Chaincode Labs.
I participated in the January Bitcoin Development Residency offered for free by Chaincode labs. This was a tremendous learning experience over 5 weeks which helped me realize that I have an incredible amount to learn if I wish to contribute to bitcoin on a programming level but there are various ways to build on bitcoin outside of computer programming. That is where I realized I can utilize my strengths of physical building to improve bitcoin.
As for the solar energy resources I dusted off my Physics E&M notebooks along with the highly valuable resources offered at DIY solar Forum and the various videos published by Will Prowse. I would also like to thank the individual help I received from various forum members including Raurre and Sunshine_eggo. Also I am incredibly fortunate that my father in-law is a retired industrial electrician and helped me run the circuits and overlooked my schematics.
Mounting Structure
I have a freestanding back patio I built from steel a few years back that provides southern exposure with a roof angle of 3 degrees. Below are the pictures during construction.
Solar Power
For my panels, I created a very simple and adjustable racking system that increases the pitch to 6 degrees during winter and -2 degrees during summer. This is a very simple metal frame with threaded rods to welded nuts on the underside of the patio frame where I can adjust the north or south side elevation based on the time of year.
These specific panels are a 128-cell model producing 74 volts and 5.9amps with a series fuse of 20amps. My inverter/MPP is rated for 250v and I wanted to use 10 gauge solar cable for all my roof runs. With this in mind and having fourteen panels was essentially limited to running a 2 series 7 parallel configuration.
I opted to build my own combiner box so that I could fuse each series individually before combining in parallel. This allowed me to run a 32 amp and 144v line (under ideal situations) to my inverter. This simple process of building the combiner box not only saved me a few hundred dollars but also provided a firm understanding of the circuitry that I had created with the solar panels.
My Suppliers
The panels were bought second hand but here is the manufacturer.
My batteries and Inverter were bought from Signature Solar in Texas.
My ASIC miners were bought from Kaboom Racks.
Bitcoin Miners
We ran a 30 amp line to a 38″ wide x 30″ tall x 26″ deep corner in my office which had a prebuilt dog door (for a dog we never got). I took the sound and heat control of the miners pretty seriously as I use this office every day and my father sometimes stays here. I created two baffles in the box to separate the air intake from the main compartment and lined the main compartment with 1.5″ foam.
My air box in 8″ wide, 30″ tall, 26″ deep and has two 4″x10″ filtered vents on the front pulling air from my office. I drilled eight 2″ diameter holes along the top of the intake box into the main compartment above the final baffle. I have enough room for the S19 and 2 x S9s stacked on a small rack I welded together.
The doggy door exhaust vent is 11″ x 16″ which allowed me to add another cubic foot or two of space while directing the airflow towards a vertically mounted 8″ inline vent from Axious moving nearly 800 cfm through the box. I have a 24×16 viewing /access window in the main compartment which I covered with half inch plexiglass.
For baseline I run Braiins OS+ on the S9’s and have them under-clocked to 905 watts producing 11Th/s at 77 watts per Th/s. I installed Vnish on my BeagleBone S19 and under-clocked it to 3.1kWh producing 101Th/s, right at 33 watts per Th/s. I have been experimenting with further under-clocking from 2pm to 8pm and then overclocking through the night to balance energy demand, supply and utility rates. Will update further with more data.
I am currently sending 96% of my hash to Slush Pool and the remaining 4% is attempting to solo mine a block, just for fun.
I monitor the temperature closely and I am currently maintaining a 12 degree difference between my office temp and main compartment temp.
Mining Pools
Because bitcoin mining is a global competition a home miner has two options to consider. You can pool your hashing power with other miners all over the world and split the compensation if and when a block is found by your pool. This provides the opportunity to earn consistent reveneue equal to your contribution to the pool and the size of the pool determines your odds of finding a block.
Another option is to attempt to find a block yourself and recieve the entire reward which is currently 6.25 bitcoin. The odds of this are very slim due to the amount of hash power of the network. If you think of this as purely a guessing game, my miners are performing 130 trillion guesses per second and I am competing against 200 million trillion other guesses per second… those odds are not in my favor!
Initial Results
On the first few days of operation I just let the system run with solar as first priority. I was only able to power the miners from 10am to 3pm off the grid, then utility kicked in. Since then I have been fine tuning the energy usage and a few settings on the inverter to achieve my desired results. Yesterday I ran off grid from 8:15am to almost 6pm with miners tuned down by 15%. I will continue adjusting my consumption while also ‘hoping’ that the increase daylight hours of summer will outweigh the decreased efficiency of hot solar panels. Also my utility provider informed me the peak demand time will be decreasing to 4pm – 7pm but they did not give a date.
I have earned 426k sats (0.00426048 bitcoin) over the past 7 days while utilizing an additional 62 kWh per day (estimated mid billing cycle).
While I took the sound deadening of the miner box very serious and am happy with the results, I did not foresee the noise the inverter makes while it is converting DC to AC power. While I will attempt to resolve this in the future, it provides and easy way to audibly track when my loads switch from utility to solar/battery. My immediate priority is to utilize and extra RasPi as an energy monitor for all these components for more accurate data & decisions.