How to Clean An Enamel Butcher Tray Watercolor Palette

Butcher tray palette

If you paint with watercolors, you’ve likely tried out a butcher tray palette. Butcher tray palettes are usually made with steel or another metal, coated in a water resistant enamel or porcelain. They’re a great option, because they hold pigment well, don’t stain, and allow for easy mixing.

Many watercolorists leave pigments on their palette between sessions. As the colors dry out, they harden, and can be easily reconstituted the next time you sit down to paint.

If you leave your paints on a butcher tray for weeks or months, though, they might begin to harden. Once this happens, the pigments can be very hard to remove.

Here’s how to clean the enamel on a butcher tray palette.

Start by making sure that your pigments and colors aren’t toxic. Some watercolors are toxic, and you want to avoid cleaning these in a sink that you use for food prep, teeth brushing, etc. In general, you should use a utility or art sink for cleaning up pigments. Either way (but especially if your pigments are toxic), you’ll want to wear gloves for this cleaning step, so the pigments don’t stain you hands.

You also want to make sure your sink can handle the pigments without staining. Again, a plastic utility sink works great.

Turn the water on as hot as you can. Thoroughly wet the stuck on paints. You’ll see that a lot of pigment starts to wash away immediately, assuming your water is hot enough.

Washing away stuck-on pigment on a butcher tray

Make sure not to scald your hands. Work the water across each piece of stuck-on pigment, until they’re all soaked. The colorful runoff water looks pretty trippy, actually!

Next, leave a little water on the tray, and grab a small sponge with a scrubber. Scrub away at each piece of stuck-on pigment to begin to release it. Use bursts of water to wash it away.

Use a scrubber sponge to scrub away at pigment.

If chunks of pigment are still being very stubborn, you can also use a pallet knife or metal scraper to gently lift them off. Make sure not to dig into or scratch the enamel surface of the palette.

Lift up stuck on pigment with a palette knife.

Give the butcher tray a final wash in water, dry it off, and you’re good to keep on painting!

How Much money Can You Make on the Medium Partner Program?

businessman working on project sitting at office desk
Photo by Evelina Zhu on

Lots of people looking into publishing content on Medium for passive income wonder how much they can make through the Medium Partner Program.

The answer is potentially tens of thousands of dollars per month, if you’re one of the top writers on the platform. The best writer in September 2020 earned $49,705.40 on the platform, according to Medium.

That’s the equivalent of almost $600k per year, if they made the same amount each month. So the potential is definitely there for very high earnings.

That said, most people who write on the platform will not earn that much. Less than 10% of writers earn more than $100 per month—in September only 6.4% of writers earned $100 or more.

However most people who publish on Medium do earn something. 68.7% of writers on the platform earned something in September. Once you have at least a little earnings, there are plenty of things you can do to tweak your writing and earn more.

This article originally appeared as one of my answers on Quora.

A Solar Crypto Manifesto

Solar is a transformative technology. Solar panels take something free and readily available — sunlight — and transform it into a valuable resource, electrical power. They do this with no moving parts and no direct greenhouse gas emissions. And they do it for 20+ years, with almost no maintenance — most panels require a wipe-down every few months, and that’s it.

Even better, solar now performs this remarkable alchemy very cheaply. A combination of tech advances leading to higher efficiency, better manufacturing processes, and scale (largely from Chinese producers) has driven the cost of solar down by 70% in the last decade. You can buy solar panels for a few dollars per watt or less— and that’s before generous tax credits in many countries, including the United States.

With atmospheric carbon reaching new highs every year — and climate change quickly becoming a lived reality instead of a concept — solar may be our best shot at achieving a cooler future.

But solar has a problem. As I’ll explain below, solar currently work at large and small scales, but not in the middle. I believe that another emerging technology — cryptocurrency — has the potential to bridge the gap and make solar a truly game-changing technology. I’ve been testing that theory since early 2019, and I’m pleased to report that in my proof-of-concept, pairing crypto with solar led to a 500% increase in earnings versus solar alone. Read on to learn more.

A Problem of Scale

Solar is an amazing technology, but from a business perspective, it currently suffers from a problem of scale.

On a very small scale — at the level of individual homes — solar works great. For an investment of around $20,000, you can place panels on your home, and take advantage of a process called “net metering” from your utility. With net metering, when your panels are generating power (usually 5–7 hours per day depending on where you live), your electric meter literally spins backward. At night or when it’s cloudy, you draw power from the grid, and your meter spins forward.

Image for post
Solar panels on a private home. Credit: Gado Images

At the end of the month, you pay only for the “net” amount — the difference between what your panels generated and what you used. For the individual consumer — especially in a sunny state like California — this can work amazingly well. In many cases, you’ll pay nothing for electricity. And again, most panels today last 20–25 years. So for a reasonable initial investment (offset by tax credits in many cases), you can effectively eliminate your electric bill for decades.

Net metering is a great start, but there’s a problem. While you’re spinning your meter backward, your panels are effectively generating power at a very high price per kilowatt hour (KWH). Where I live in California, we pay up to $0.34 per kWh. So if you’re using solar instead of grid power in California, your panels are basically making you $0.34 per kWh. That’s enough to break even on the cost of the panels in just a few years and to save yourself a ton of money in the long run.

Spin your meter backward past zero, though, and everything changes. If you generate more power than you use, your utility will pay you for it. But the rates they pay are terrible. My utility, PGE, only pays $0.0315 per KWH for excess power generated during net metering. That’s a truly awful rate–you’d have to run your panels for decades just to break even, much less make a profit.

Because rates for excess power are so low, sizing a solar system is a balancing act. You want to generate enough power to offset all your usage–effectively earning money at high retail rates (again, up to $0.34 per kWh). But at the same time, you don’t want to generate too much. If you do, you’re selling power back to the utility at next to nothing, and essentially losing money in the process.

Again, due to net metering, solar works very well at the scale of a single home, but beyond that, there’s no opportunity to generate excess power and turn a profit.

Big Solar

On the other hand, solar works great at a very large scale, too. If you’ve ever driven up Interstate 5 in California or spent time in any sunny state, you’ve probably seen an industrial-scale solar farm.

Image for post
Solar works great a large scale. Credit: DOE/Gado

These tend to consist of acres upon acres of solar panels, all oriented in the same direction and soaking up the sun. They’re kind of beautiful in a way, like a very high tech version of the farms or orchards of yore. Or something out of a science fiction movie — where the computers have learned to power themselves and have taken over the planet from their pesky, carbon-spewing human creators.

Solar farms generate huge amounts of power and are also hugely profitable. The power generated by solar at this large scale is now cheaper than power from many fossil fuels, and since panels last a long time, solar farms can generate lucrative power for decades.

The challenge with solar farms is that they’re incredibly hard to start. First, you need a ton of capital to buy land, and to cover it in thousands of panels. That’s tough to get, but not impossible. The harder part, by far, is getting permission to sell that power back to a utility.

Unlike with net metering, a solar farm can’t just plug in and start selling power back to the grid. Instead, they need a Power Purchasing Agreement with the utility in their area. This is a contract where the utility agrees to buy power from the farm for a certain amount of time, at a certain rate. Utilities are very stingy about handing these out — many have quotas or limits on the number of PPAs they’ll issue each year, and the process can require armies of lawyers, mountains of paperwork, and years of negotiation and permitting to get in place.

Image for post
One of the biggest challenges of monetizing solar at scale is working with a utility. Credit: Gado Images.

Once a PPA is in place, the owner of a solar farm is golden. The contracts last up to 30 years, and often pay rates of around $0.14 per KWH — more than 4 times the rate paid to net metering customers, and enough to turn a healthy profit over time. By getting a PPA, a farm owner has locked in decades of nearly guaranteed returns.

Again, though, getting these contracts is difficult and expensive. Securing them can take years, and cost thousands or millions of dollars. Sometimes solar projects never get off the ground at all, since it’s so challenging to get an agreement in place. And of course, you need the legal and industry expertise to even stand a chance of negotiating a PPA — this is not territory for the individual or even a small startup.

For that reason, solar farms really only make sense for companies with the funds and clout to successfully negotiate a PPA. And by the time you’ve won a PPA, you’ve likely already spent tens of thousands on the process. To recoup their costs, solar farms thus need to be very, very big. They really only work at a very large scale — megawatts or more. That’s way beyond what an individual, startup, or even mid-sized company can afford.

Why Don’t You Just Meet Me in the Middle?

Solar works great for individuals, and it works great for massive companies. But in the middle, there’s basically nothing. If you’re an entrepreneur or startup looking to profit from solar, the business model just isn’t there. You can’t make a profit selling power back to utilities at $0.0315 per KWH on a very small scale, nor can you afford the process of winning a PPA and locking in a $0.14 per KWH rate at the large scale.

Solar installations of 10–200 KW could be very affordable to build for small startups and even wealthy individuals. And they could be profitable, generating clean energy and recurring revenue for decades. But at the moment, if you build one, there’s no one to sell the power to.

For solar to truly take off as a transformative technology, there needs to be a middle ground. There needs to be a way for risk-loving startups, small companies, or individuals to install solar panels and earn a profit from the power they generate. This would open up solar to millions of entrepreneurs who would jump at a chance to create new passive revenue streams from their capital while helping to counteract climate change.

Ideally, there would be a way to generate power from solar, and immediately transform it into money. This method should work at a medium scale, with a sizable but reasonable initial investment — the kinda startup or well-to-do individual could raise. And this method should require limited to no permitting, and no need to negotiate purchasing deals with a utility or anyone else. You should turn the system on, the power should start flowing, and the money should start flowing, too.

Enter Crypto

With the birth of cryptocurrency mining, this possibility now exists.

Image for post
Bitcoin logo on ATM. Credit: Gado Images.

For those who aren’t familiar, cryptocurrencies like Bitcoin and Ethereum are modern monetary equivalents built on a technology called the blockchain. While there are hundreds of cryptocurrencies out there, many of them work in the same basic way.

People, called “miners”, use their computers to solve a very challenging math problem. As they solve these problems, they are occasionally awarded a cryptocurrency coin. The blockchain allows everyone to see and verify who has been awarded coins. And once the coins are awarded, they can be sold or traded with others in the system — often for very large sums of fiat (read: actual) money.

The catch is that solving these math problems (which are based on a concept called hashing) is extremely hard. It requires computers to toil, day in and day out, 24/7 for days, weeks, or longer, often in groups, just to find a single coin. Doing this uses a lot of electrical power — worldwide, cryptocurrency mining uses more power than the country of Switzerland.

What mining computers do, though, is perform alchemy almost as impressive as solar’s. They take in electrical power as an input and create money. There’s no need to secure a buyer or have a contract with a third party. A mining computer anywhere in the world can enter the system, and through the power of the blockchain, any coins it mines can be immediately verified and awarded to its owner.

Solar and Crypto — a Perfect Pair

Cryptocurrency mining — I believe — unlocks a totally new business model for solar.

Remember, solar’s big issue is the lack of a middle — a method for players other than giant companies to generate electrical power from solar, and immediately convert that electrical power into money at a favorable rate. Crypto provides that method.

Here’s how a solar/crypto project would work. A startup with some decent funds (say $500k) to spend would buy land in a sunny area, and install solar panels. They would connect those panels to cryptocurrency mining computers. The computers would use the power right onsite to perform a mining operation, creating something of value — cryptocurrency coins — which the startup could then sell at a profit.

With a solar system in the 10s to 100s of kilowatts, a startup or even a wealthy individual could make good money. And because of how crypto works, there would be no need to negotiate complex deals with utilities in order to unlock profit from their system. They could switch the whole thing on, and be earning passive revenue the next day.

My Solar/Crypto Proof of Concept

When I started thinking about this business model, I didn’t want it to just be a theory. I wanted to see how it would perform in the real world. So I set about building myself a solar/crypto proof of concept (shoutout to Chris C and Brett R, who helped me think through these ideas and acted as a sounding board during testing).

On the solar side, I worked with Renogy, a high-end solar panel company, to source some panels. I ended up buying 250 watts worth of panels — a tiny system, but enough for a test — as well as the charge controllers, cabling, etc. to run them.

Image for post
My Renogy panels. Credit: Gado Images.

I also bought a battery, to test the idea of storing power generated during the day to run the system at night (more on that later).

Image for post
My storage battery. Credit: Gado Images.

I hooked the whole thing up and placed it in the California sunshine. In my area, we receive about 6–7 hours of sun per day, which is enough for a solid test. In practice, my 250 watts of panels generated about 200 watts of power, mostly due to some long cable runs, a cheapo charge controller, and some shading.

On the crypto side, I started by building a computer. For cryptocurrency mining, you don’t need much — a motherboard, processor, network card and power supply are enough. I used a Platinum rated power supply, for energy efficiency purposes.

Image for post
A Platinum rated power supply improves the efficiency of my system. Credit: Gado Images.

What you’re really trying to do is provide a home for a graphics card, or GPU. These are the devices that do the heavy lifting of cryptocurrency mining. For my setup, I used an NVIDIA 1070.

Image for post
Installing my 1070. Credit: Gado Images.

To connect the two parts together, I got a 500 watt power inverter from Renogy. This takes the 12 volt output of the solar system, and transforms it into the 120 watts needed to run the computer.

Image for post
A 500 watt inverter was plenty for my proof of concept. Credit: Gado Images.

For software, I installed Nicehash. Nicehash is great because it doesn’t require you to mine a specific cryptocurrency coin. Rather, it’s a marketplace for hashing power — the computing power required for mining. Buyers come to Nicehash to buy large amounts of hashing power for their own mining operations, and sellers like me come there to sell it. Nicehash handles payment, configuring the algorithms, etc.

I liked it for this test because it’s easy — you open the program, and it starts earning you money. I also like that it has the potential to diversify in the future — if cryptocurrency suddenly goes away, Nicehash could pivot into selling hashing power for other purposes (like AI training), and sellers like me could still earn money.

My Setup By the Numbers

After tweaking power settings using a program called MSI Afterburner, my computer draws about 170 watts while running full out mining cryptocurrency with Nicehash. Another way to put this is that in 1 hour of run time, it consumes .17 KWH. At the same time, during my test in mid-2019, it generated $0.0258 in cryptocurrency earnings per hour.

This means that the system is making about $0.16 per KWH from the power it generates.

Stop for a moment and think about that. By using solar and crypto together, I can make 500%+ more than what I would make selling power back to the grid via net metering. And I can even make about 15% more per KWH than if I had a solar farm and a complex PPA in place with my utility.

This setup requires no contract, no prior authorization, and not even a grid connection. I just set the whole thing up, placed it in the sunshine, and started earning.

I ran my solar/cryptosystem for about 3 months, just to see how it would do over a slightly longer time period. The numbers stayed consistent, and it ran with minimal maintenance. I wiped the panels down every week or so, and that was about it.

One challenge I ran into was running the system after dark. My solar panels, in full sun in California, generate about 65 amp-hours of power per day. 170 watts of power consumption from the computer is about 14 amps at 12 volts (170/12). That means the panels can run the computer for about 4.5 to 5 hours per day.

Ideally, you would want to run the computer all day, so it’s always mining and earning you money. This would require more panels, but it would also require a way to store the energy. To experiment with this, I connected up a lead-acid battery and used it to run the system for about an hour each day after the sun went down.

The challenge with running on battery power is that you need a lot of batteries to run 24/7. Even if you live in a place with 7 sun hours per day, you still need to run on batteries for the other 17 hours. At 14 amps of power draw, running the computer would require 14 * 17 = 238 amp-hours of battery capacity.

But there’s a catch — for longevity, you don’t want to drain lead-acid batteries below 50% charge. So you would really need at least 400 amp-hours of capacity. That’s about 3–4 large marine batteries, which gets bulky and adds expense.

That said, the contract-free $0.16 per KWH from solar and crypto is very compelling. I can see two different ways to scale this up. I’ll look at each one, considering a hypothetical 10KW solar system, and then a bigger system. This would be a reasonable size for an individual looking to diversify by generating passive income from solar for under $100k, or a startup with more to put in.

Solution 1: Full Off-Grid

One way to scale my system up would be to keep the basic concepts, but make it a lot larger. A big advantage of my proof-of-concept system is that it’s totally off-grid. This means no utility oversight at all, and depending on your jurisdiction, limited permitting, as long as you’re not connecting to a structure.

To scale this up, a solar entrepreneur could buy some land in a high-insolation (read: very sunny) area. The great news is that these areas tend to be dessert, so land is cheap. You can buy a few acres for less than $10k. Here’s a parcel for $6,000 a press timein Kern County, California, the highest insolation place in America.

Image for post
Land in California’s high desert is sunny, plentiful and cheap. Credit: Gado Images.

You would then get off-grid panels, and set them up on the land. Let’s use a 10kw system to make things simple. The Solar Store sells panels by the pallet — three half-pallets of their Q Cell line would run you about $9,000, and provide 10.440 KW of power.

Kern County gets an average of 7.66 sun hours per day, which means in a year, you’d get 2,795 hours of sun. That means your panels would generate 10.44 * 2,795 = 29,179 kWh per year.

You’d then need enough mining hardware to take advantage of all that power. I throttled my computer down pretty far to stay within the 200 watts from my tiny solar setup, but running at more like 220 watts was ideal, and boosted my earnings up to about $0.185 per kWh. My computer, mining 24/7/365, consumes about 2,000 kWh per year. Most of that is to power the NVIDIA 1070.

To use all 29,000+ KWH generated by your off-grid panels, you’d need 15 1070s. At press time, they were running about $370 on Amazon. So that would be about $5,500. Add another $500 or so for a mining MOBO (the ASUS B250 can easily accommodate 15 GPUS) and some power supplies, and you’re at $6,000 total for computer hardware.

Remember, though, that you’d still have to run at night, so you’d need batteries as well. With 7.66 sun hours per day, you’d still need to get 16.34 hours of dark run time in each day. With 15 GPUs, your system would draw 220 * 15 = 3,300 watts or 275 amps at 12 volts. So to run all night, you would need a battery capacity of 275 * 16.34 = 4,493 amp-hours.

There will be conversion losses, power lost to charging and discharging, etc. And for maximum life, you don’t want to run your batteries all the way to empty. So let’s nearly double our AH requirement, adding 80% more capacity to leave plenty of margins. That means you’d need about 1.8 * 4,493 = 8,097 amp-hours. Big deep-cycle lead-acid batteries cost about $1 per AH of capacity, so that would add about another $8,100 to the system’s cost. Batteries are expensive!

The connect everything together, let’s throw in another $2,000 for charge controllers, cabling, inverters, mounting hardware, etc.

Image for post
Solar cabling. Credit: Gado Images.

At this point, you would have a system that was totally off-grid, and could mine 24/7 with 15 different GPUS. The total system cost is:

  • $9,000 for panels
  • $6,000 for hardware
  • $8,100 for batteries
  • $2,000 for other equipment
  • Total: $25,100

How much would your system generate? Remember that in my testing, the system generated $0.185 per KWH with more generous power settings. That means you’d make 29,179 * .185 = $5,398 per year from the system.

In a little over 4.5 years, your entire system cost would be paid off. Run for another 1.5 years or so, and you’d have bought yourself a free parcel of land. Beyond that, all the earnings would be pure, passive profit. Typical break-even points for a commercial solar farm are 7–8 years, so with a 4.5-year timeframe, you’re already ahead of the game — even with the extra expense of the crypto hardware.

The great thing about this setup is that it scales really well. Depending on how much you had to spend, you could simply buy more panels, more GPUs, and more batteries to meet your needs.

Say you were a small startup instead of an individual, and you had $500k to work with. You could build a 200 KW system (~$25k for our 10kw system X 20. The payoff period would be the same, but after the system was paid off, it would bring in $107,960 per year in passive revenue. In theory, you could scale the system size up or down to match whatever size you wanted.

Solution 2: Bolt-on to Existing Solar

Another solution is to add crypto to an existing residential solar installation.

As I mentioned before, many homeowners install solar panels to offset their own power usage. But because they basically get peanuts for any energy generated beyond their real usage, most solar installers recommend only installing enough panels to meet the customer’s actual needs. This leaves unused capacity on millions of roofs around the world.

Image for post
Rooftop solar. Credit: Gado Images.

What if using crypto, this extra rooftop capacity could be leveraged? Let’s look at my home in California, for example. According to Google’s Sunroof calculator, I have about 899 sqft of usable roof space. Paying for my whole electric bill would require a 9.5 KW system, which would take up 669 square feet. A solar installer would tell me to install a 9.5 KW system, to meet but not exceed my actual needs.

But this still leaves 230 sqft of unused roof space. That’s enough room for another 3.3 KW of panels. What if I bought the extra panels, and used that spare electrical generation capacity to mine crypto?

Since this would be a grid-tied system, I’d have much less flexibility on sourcing, and would likely need to use a professional installer to meet the standards of my utility, and zoning requirements. The average cost of installed solar in the US is $2.99 per watt, so the extra 3.3KW would cost about $9,867.

But wait — because this is a residential installation, I can likely get a 30% tax credit from the Federal government in the US (phasing out over time, starting next year). So the actual cost for the panels would only be $6,906.

To stay consistent with Solution 1, let’s pretend my house is in Kern County. That means I would get 7.66 sun hours per day, yielding 9,226 KWH per year from my 3.3KW of extra solar panels. I would need a computer with about 5 NVIDIA 1070s to use all that power, which would add about $2,350 to the system’s cost (5 cards at $370 each, plus $500 for other components).

Here’s the great part. Unlike with an off-grid system, I wouldn’t need to include any funds for batteries. Since this would be grid-tied, I could use the grid itself like a giant, free battery. During the day, my extra panels would churn out lots of power, both running my mining computer and sending electricity flowing back to the grid to spin my meter backward (remember, net metering).

At night, my 1070s would continue to churn along and spin my meter forwards. In essence, I would build up a surplus of power during the day and then draw from that surplus to run at night. In the end, the extra power generated and the power taken from the grid at night would “net” out to zero. It would be just like storing power in a battery to run at night, but with no battery expense.

Another great feature of this system is that it would be very easy to set up. The cost of cabling, inverters, etc. is already included in the $2.99 per watt figure for professionally installed solar, so I wouldn’t need to account for these. I could literally plug my computer into the wall and let it run.

Ideally, I would probably go a bit beyond this, setting up monitoring software on the mining computer to throttle its power consumption and earnings up and down depending on the surplus power my panels were generating.

Image for post
With monitoring software, I could scale my system’s consumption up or down on demand. Credit: Gado Images.

If it was a really sunny day, and my panels generated extra power, I could throttle power up to the 1070s, and earn a slightly higher rate. If it was cloudy for a while and I risked drawing too much power from the grid, the computer could throttle down the 1070s until the sun started shining again and I built up a surplus.

How would this system’s earnings look? The total system cost would be:

  • $6,906 for the extra panels
  • $2,350 for the mining computer setup
  • Total: $9,256

Again, the system would generate 9,226 KWH per year of extra power. At $0.185 per KWH, that works out to $1,706 per year. The system would be paid off in about 5.5 years and would generate profit from there out.

The challenge with this solution is that it’s harder to scale. It would work fine for one house, but to scale up, you’d either have to buy houses or work with other homeowners. The best way to scale it would probably be to work with a company that already does solar installations, like Sunrun or Sunpower. Homeowners who were already buying solar could opt to add a few extra panels, install a mining computer somewhere in their house, and reap the benefits of some extra passive income from their system.

Risks, and Ways This Could Fail

The biggest risk with all this is that it’s essentially a bet on crypto, and crypto is volatile. A drop in crypto prices means lower market prices for hashing power and thus lower earnings. Indeed, since my testing in mid-2019, crypto prices have dropped a bit, and my system today is only yielding about 9 cents per KWH. That’s still 280% more than the net metering rate, but it would mean a longer payoff time.

Image for post
This is a bet on crypto, and crypto is volatile. Credit: Gado Images.

Of course, by the same token, if crypto prices increased, your system could be paid off much faster and could earn more. If crypto went back up to $20,000 per Bitcoin, where it was a few years ago, your system could be paid off in 1–2 years and yield up to 100% more going forward. But again, this requires at least some faith in the stability of the crypto markets and making a speculative bet that the coins will remain valuable long enough to break even on your equipment and earn a profit.

As crypto bets go, though, this is likely less risky (albeit way more labor-intensive) than just buying coins. If crypto prices plummeted (or markets were regulated away), you could still presumably sell your solar panels, GPUs and land, and recoup much of your cost. That’s harder to do with the coins themselves if things go sour.

Another risk is the depreciation of your equipment. Solar panels have a known lifespan, and it’s long. They’ll dutifully provide power for 20–25 years at least. And batteries — if well cared for — should provide power for at least 10 years, and possibly much longer.

Crypto mining hardware, though, is much more of an unknown. Those 1070s are great now, but in five years or more— as technology advances and crypto gets harder to mine — they could be much less efficient, and you could end up needing to replace them with newer GPUs, which adds cost.

Image for post
Replacing mining hardware adds costs. Credit: Gado Images.

The good news here is that the breakneck pace of hardware innovation in crypto is slowing down a bit — GPUs are holding up for longer, and some coins are becoming more resistant to dedicated mining hardware tailored to a specific coin, which gives more market share to GPUs.

Some miners have also had success “trading up” on their GPUs every few years by selling the old cards to gamers and using the proceeds to purchase new ones. Even after a year or two of use and depreciation, most GPUs are still plenty powerful enough for gamers, who apparently pay nearly full price on this kind of used hardware — and sometimes more, if the cards are scarce. This creates work, of course, but could provide a way to upgrade hardware as crypto advances without too much additional cost.

The systems I’ve laid out are pretty basic, and also pretty conservative on estimates. With tweaking and optimization, you could almost definitely get them yielding more. For example, you could mine coins directly rather than relying on Nicehash, cutting out their fees. You could tweak power consumption, and get a better mix of efficiency and mining profit. You could use dedicated ASICS or another coin-specific mining technology like an Antminer. Or you could add another clean energy technology — like wind — to bump power and profits up further.

New Models, New Opportunities

Solar feels like a magical technology — it’s that irresistible alchemy of getting something (valuable electricity) for nothing (free sunlight). That the panels are nearly maintenance-free and last for decades is the icing on the cake.

Again, though, solar has a problem of scale. There are lots of ways to use it today, including to save yourself money in your own home. But there are relatively few ways to make a profit from solar unless you’re a giant company with the people and time to pursue a PPA and develop a megawatt-scale solar farm.

Pairing solar with crypto opens up profit-making opportunities for much smaller players. A risk-loving startup, or even a private individual looking for a new source of passive income, could pair solar with crypto in the ways I’ve described and scale to their heart (and wallet’s) content. These systems are modular, so one could always dip their toe in, see how it works, and add capacity as the earnings materialize, or exit gracefully if they don’t. Crypto opens up solar as a business model, not just a savings tool for homeowners.

Providing solar with a “middle” makes is much easier and more profitable to adopt. And that’s good for everyone — solar equipment makers (who could produce cheaper panels with greater scale), solar entrepreneurs, and the planet.

This last piece is key. Again, crypto mining uses as much electrical power as the country of Switzerland each year. Moving more crypto mining over to solar is not only a great way to realize profits from solar power but also a great way to reduce the environmental impacts of cryptocurrencies as their reach and scale continue to grow.

Solar needs a middle, and I believe crypto could provide it. What do you think? Who’s tempted to build this?

This article originally appeared on The Startup on Medium in November 2019.

Review of the Hello Baby HB178 Audio Baby Monitor

Lots of people looking for a baby monitor for their nursery look at devices with built-in video–or try to repurpose an IP camera, like the Nest Camera from Google. The issue with standalone video monitors, though, is that they’re relatively easy to hack, the video is usually low quality, and you often have to charge the devices’ portable screens every night. With a Nest camera, you usually have to pull up the video feed on your phone to check in on your nursery.

In many cases, though, you don’t actually need video from your nursery–you just need to be able to hear what’s happening there, so you tell if your baby is crying, your toddler has gotten up and needs help, etc. If you just need a live audio feed from your nursery (or another room in your house), the Hello Baby HB178 is a great option.

The HB178 comes with two units–a transmitter and receiver. You install the transmitter in your child’s room or nursey, and plug it into the wall. It then pairs with a receiver, which you place in your room (plugged into the wall) or carry around the house (with an optional battery). The receiver has a volume setting, as well as an indicator light to show how much sound is coming from your nursery (the device indicates five levels of sound with color-coded LEDs).

From a technical perspective, the HB178 works like a traditional audio baby monitor, but with some major upgrades over the ubiquitous walkie-talkie like devices from the 1980s and 1990s. For starters, the device uses digital audio, so the sound is crystal clear–there’s no crackly static in the background to deal with, or weird phenomena like hearing audio from passing aircraft (which sometimes happened with 1990s-era audio monitors). The audio feed is also encrypted, so people outside your home can’t spy on the sound in your nursery (or so Hello Baby claims).

The HB178 also does a great job of suppressing sound when nothing’s actually happening in your nursery. That means you can have the volume turned up to max, and you won’t hear a lot of background noise or static. I usually set my B178 to nearly full volume. That way, when there actually is sound coming in, it’s quite loud and will definitely wake you up if you’re sleeping. But if all is quiet, there won’t be any crackling or other background sound to disrupt your sleep.

The HB178 is also economical, at around $25 on Amazon as of press time. Another advantage of audio-only transmission is an increased range–the HB178 has an advertised range of 1,000 feet. In practice, I find it’s a bit less than this, but it still has plenty of range to monitor your nursery from across a big house, or even outside in the backyard if you use a battery in the receiver.

If you really want to see what’s happening in your nursery, a video monitor is probably a better bet. I find the best overall solution is to pair the HB178 with a Nest camera. I can then immediately hear if anything is happening in the nursery without having to load up my phone to view the Nest cam. But if I hear something I want to see further, I can always pull up the Nest video feed on my phone or Alexa device and take a closer look.

If you’re looking for a modern take on the traditional audio baby monitor–or if you’re concerned about security and convenience with video monitors–the Hello Baby HB178 could be a great fit.

ECG on the Fitbit Sense is Finally Live

I’ve been covering the new Fitbit Sense extensively on my own YouTube channel, and also on Medium’s new consumer tech publication, Debugger. When I first got the watch–which is Fitbit’s most advanced smartwatch to date–I wrote a detailed initial review on Debugger. I’m also working on a longer-term review (also for Debugger) which will cover what it’s like to actually use the watch in daily life, including using it to track stress, sleep, exercise, and more.

One much-touted feature of the Fitbit Sense which I wasn’t able to evaluate when the watch first shipped, though, is its electrocardiogram (or ECG) functions. These were only cleared by the FDA in the United States and CE in Europe about a week before the watch shipped, so it appears that Fitbit didn’t have time to actually roll the feature out when the Sense began to ship on September 25.

The company promised that the feature would be live by October, and they’ve delivered on that promise. As of yesterday, ECG functions are now live on the Fitbit Sense. If you have a Sense already, you can now begin to use the ECG functions, with a few setup steps.

Firstly, what is an ECG, and what is it used for? An ECG tracks the electrical activity of your heart. Fitbit’s PurePulse 2.0 heart rate tracking–which has been present on its watches for years now–only tracks the physical beats of your heart. It doesn’t look at your heart’s electrical activity. Issues with your heart’s electrical activity can be a sign of Atrial Fibrillation, a serious condition that increases the risk of stroke, heart failure, and other complications. Checking for the condition using your watch is a great health-oriented feature for Fitbit’s most health-centric smartwatch.

According to an email I received from Fitbit, “Before you can initiate an ECG reading, you must complete the brief introduction on the Heart Rhythm Assessment. This can be found by going to the Discover tab in the Fitbit app and selecting Assessments & Reports at the bottom. Once the introduction is complete, the ECG app will automatically be installed on your watch after you sync.”

I completed my own assessment, which required answering a few questions about my age (you have to be 22 or older to use the feature), etc. My app didn’t install automatically, so I followed Fitbit’s recommendations to “manually download the app from the Fitbit App Gallery (under the My Apps section) in the Fitbit app.” After the install, an ECG app appeared on my Fitbit Sense.

New ECG app on the Sense

I opened the app, which walked me through the process of taking an ECG reading. To do this, you have to sit still and not speak or move your arm for 30 seconds. The app instructs you to put your thumb and index finger on opposite corner’s of the watch’s bezel. The bezel has integrated electrical sensors which are used for the ECG reading, as well as other readings like Fitbit’s EDA scans.

After 30 seconds, the watch finishes gathering data, and spends a few seconds crunching it. It then returns its analysis, which Fitbit says is

There are three possible outcomes, according to Fitbit’s email:

  • Atrial fibrillation: Your heart rhythm shows signs of atrial fibrillation, an irregular heart rhythm. You should talk to your doctor about this result.
  • Normal sinus rhythm: Your heart rhythm appears normal.
  • Inconclusive: If your heart rate is below 50 bpm or above 120 bpm, the ECG app will be unable to access your heart rhythm. This can happen for many reasons, such as moving too much, not resting your hand on the table when recording, wearing your band too loose, or due to certain medications or having excellent aerobic fitness. 

My own reading came back showing a normal sinus rhythm! At least as of when I took my reading, my heart was working well.

One of the challenges with AFib is that is can be “difficult to detect”, according to Fitbit, so it’s wise to take multiple readings over a period of time–especially if you think you’re experiencing symptoms like heart palpitations. Of course, if you think there’s any chance you’re in imminent cardiac danger, stop futzing with your watch and call 911.

Another helpful feature of the ECG app is the ability to download an ECG report to share with your doctor. This includes the raw output of the ECG reading, which you doctor should be able to interpret better than an algorithm, based on their medical training.

Fitbit’s report also shares some details about the accuracy of their system (in case your doctor isn’t convinced), and the fact that it’s comparable to a 1 lead ECG. Again, because it has been FDA cleared, the device should be accurate enough to provide medically useful data, even if it can’t direct diagnose you with a disease.

With the new ECG feature, Fitbit has taken yet another step towards making the Sense the premier health tracking watch on the market today.

For more coverage of the Fitbit Sense–including my upcoming detailed review–follow me on

Solar for Renters (Part 2)

When I left off, I was describing the first part of my solar power system designed for renters. You can read about it here.

As I described, I placed 250 watts of Renogy panels on my roof. Given how much power they output, the panels are pretty reasonable in size. All the panels together are about 6 feet by 4 feet. You could definitely fit this on the balcony of a city apartment, and could probably even go bigger, using Renogy’s 170 or even 300 watt panels. I connected the panels to a railing using metal safety chains used for theater lighting, which work great.

A quick note here; shading makes a huge difference in panel output. With even a tiny bit of shade–like from the railing of a balcony–my panel’s output was HALVED. That’s a huge loss, and I ended up moving my panels to a different part of the house with full sun to avoid the issue. If you’re designing system for a small balcony, make sure to avoid panel shading as much as you can.

From the branch connectors, I used a 40ft 10awg cable from Renogy to get power down from the roof to my garage. That length and gauge is not ideal for a 12v system; really it should be 20ft max, and I’m probably losing about 10% power with the long cable. But it was a long run down, so again, oh well. If you’re able to put your battery right by your panels, you could avoid this and get a big efficiency boost.

In the garage, the cables go into a 30amp Renogy Adventurer charge controller. I bought an extra Bluetooth module, which allows you to monitor the system in real time from Renogy’s app. The Adventurer plugs into a 12v lead acid battery, with about 120 amp hours. Sealed lead acid or even lithium would be better, but this works fine for initial testing.

From there, I use a 500w Renogy inverter (the lowest wattage available today is a 700 watt unit) to turn the panels’ output into 120v house current. Renogy’s inverters are really good, and have safety settings like low voltage protection and overcurrent protection. They’re also able to adjust their fan speed and other settings up and down based on load, so you can leave them on all the time and they don’t draw that much current at idle. You even get a little car key style remote to turn the inverter on and off remotely.

Here’s how the whole thing looks in action:

So how does it actually perform? Stay tuned for Part 3 to learn more!

Solar for Renters (Part 1)

A few years ago, I wrote an article about my dream of building an apartment-scale solar power system for renters. I wanted the system to be something you could install on a balcony or in a backyard, that could generate power off-grid without having to install traditional solar panels on your roof. My vision was for something renters could use, in anything from a small city apartment all the way up to a rented suburban home.

Well, several things have changed since I wrote that article. For one, solar panels have gotten a whole lot cheaper and better. You can now get off-grid panels for well under $1 per watt, and batteries and charge controllers have both come down in price. There’s exciting new battery options, too, like lithium batteries which might destroy your bank account but eliminate the need for ongoing maintenance and have a much lower risk of catching fire/exploding than old fashioned lead acid.

So with all these things coming together, I finally went ahead and built my system! This article covers the basics, and I’ll go into a lot more detail on specifics, too.

First a word of caution; if you have any doubt about working with electricity, hire a professional. And although this system is designed for renters, make sure your landlord allows this kind of thing before installing! 

Okay, with that out of the way, here’s what my system looks like. Ideally, I wanted to have at least 200 watts of generation capacity, as I wanted to run a computer off solar power. So the first part of my system is 250 watts of panels from Renogy; three 50w panels and a 100w panel. Ideally all your panels would be the same wattage, but I had the 100 watt panel already, so oh well. These are wired up in parallel using a four port MC4 branch connector, and placed on a balcony on my home’s roof.

This is part one of a series of posts about the system—please follow my blog for the next post in the series!

All About Cryptocurrency (Sort Of)

Okay, so I realize that crypocurrency mining is not exactly a home automation topic. But I think if you do one, you may well have an interest in the other. And many of the skills you learn from HA (optimizing power consumption, designing systems to run on their own, building computer that run 24/7 for years, etc.) might make you very valuable as a miner.

So what is cryptocurrency mining anyway? Unless you’ve been living under a rock/away from any kind of overblown financial/tech news, you’ve probably heard of Bitcoin and something mysterious called the Blockchain. You’ve probably also witnessed people pitching Blockchain or Anti-blockchain views with fervor bordering on the religious. Blockchain will solve hunger! It will make the financial industry obsolete! Or, Blockchain will destroy the world’s economy! It’s the next gold rush/tulip panic/insert bubble metaphor here!

So what is the Blockchain anyway? Basically it’s a technology for transactions between different people/computers. To make it work, people have to use their computers to solve very complex math problems. When they solve the problems and keep the whole thing afloat, this is called “mining”, and they get rewarded with cryptocurrency. So basically, you volunteer your computer’s time to solve math problems very, very quickly, and you get cryptocurrency in return.

How do you solve these math problems? And whose math problems should you solve? It gets very complex, but essentially, you use very powerful computer hardware–either a GPU normally built for gaming, or a specially-built, very expensive, and ultimately quite silly device called an ASIC. Let’s stick with GPUs for the moment.

Say you have a home automation computer that you also use for gaming, and it has a nice GPU already. Or say you want to see what this cryptocurrency thing is all about, so you buy a nice GPU and drop it into your HA computer. What next?

To do cryptocurrency mining for real, you’ll need a mess of other things–a “wallet” to store your coins, a “pool” of other miners to work with, etc. If you want to skip that and just see what it’s all about, I recommend using one of a few new services which abstract away all the nitty-gritty of mining and reduce the whole operation to a few clicks. 

The best of these are Honeyminer and Nicehash. They scan your hardware, use it to mine any number of different cryptocurrency coins (and there are lots), and then pay you in Bitcoin which you can readily convert into real (sorry, “fiat”) money. If you want to peek under the hood, you can also use these tools (especially Honeyminer) to learn about what’s actually taking place, and then swap out the tools for the actual underlying software they’re built on. Or you can just keep using them, and enjoy the slick interfaces, etc.

So how much can you make? It depends on what the (crazy) markets for cryptocurrencies are doing at any given moment, but with a decent GPU, you can usually make between $0.60 and $3 per day with your computer running in the background.

What’s the catch? You’re paying for the electric power it takes to run a crypocurrency mining device. So factor that in as you plan your profits. The less your power costs per kWH, the more you’ll make mining. And that’s where it gets interesting. Just like optimizing lighting costs with LEDs, or using a Nest thermostat to switch off the AC when you leave, or designing lights that are only on when you’re home, cryptocurrency mining is all about POWER CONSUMPTION. And as an HA person, you’re already good at thinking about that!

Why not try underclocking your GPU, with a program like MSI Afterburner? Switching on your computer via your HA system and mining only when you’re on your utility’s off hours rate? Dropping in a Platinum+ PSU to eek out a few extra watts? Grab a Kill-a-Watt and start optimizing!

Let me know what you think of crypo in the comments. If you want to try one of the programs I mentioned, here are affiliate links, or you can visit their websites directly.



Google Wifi Speed Test

For my Home Automation system, I have a ton of hubs. There’s my Smart things hub, Hue hub, Rainforest Eagle power monitor…the list goes on. Most of them require a wired Ethernet connection, but who wants to junk up your office floor or tv area with a bunch of hubs and cables.

I wanted to put my hubs in the garage, but running Ethernet out there would be a pain. Enter Google Wifi. As I showed in this video, I use Google Wifi to have a wired Ethernet connection in a remote location without having to run cable.

When I published the video, tons of readers wrote in asking how fast my Google Wifi wired connection really is. Is it fast enough to stream media and run a real PC, not just basic hubs? 

The answer is yes, it’s plenty fast. Check out this speed test. Spoiler alert–I get 80+ mbps down and almost 10 up. 

Nest Cam as Baby Monitor

So how well does the Nest cam work as a baby monitor? The quick answer is “just okay.”

Lots of new parents want a way to monitor their kid’s crib, Pack n Play, etc. There are cheapo standalone video solutions out there on the market, but they’re from companies that don’t usually do video, and it shows in the quality and overall features.

Nest does video really well, so a lot of parents end up using Nest cams as baby monitors.

There are a lot of upsides. Nest cams can see in the dark, have Wi-Fi (so you can see your little one remotely or from any computer or phone), store video history for up to 30 days, and are reasonably priced.

The biggest downside, though, is that in order to view the stream from the camera, you have to pull up your phone or a computer. Dedicated baby monitors usually come with a little video tablet you can carry around. The quality of the tablets aren’t great, but they’re fairly reliable and always ready. If you want to monitor your baby at night with a nest camera, you’d have to leave your phone open and streaming all night.

How well a Nest Cam works for you, then, probably depends on what you want to use it for.   If you want to be able to watch your baby from the computer while you’re in a different room of the house, or you want to have really good nanny cam (ask permission!),  or you just want to check in periodically, Nest is a great baby monitor.  If you want always on video, though, you might want to look at a dedicated baby monitor instead.