Solar Diversion for Water Heating

Not really CNC related but there has been quite a lot of interest in electronics/PCB design so I thought people might find it interesting.

Mid last year I finally got around to having solar installed at our place. It needed a switchboard upgrade first which I had been procrastinating for a couple of years. Our system is 16x 440W panels in 2 strings on a Fronius 6kW hybrid inverter.

We have electric hot water via a 250L cylinder mounted under the house so the plan was always to figure out some method to preferentially use solar to power the hot water cylinder.

To that end I put together a basic PCB that has an ESP32, a zero-crossing detection circuit, control for an SSR and some other basic features. The ESP32 is running micropython and connects to an MQTT server to receive commands telling it what duty cycle to run. Based on the commands received, the SSR will turn on or off for a set number of half-cycles which avoids switching losses and minimizes electrical noise issues that can come from hard switched loads.

The PCB is designed to mount directly on top of the SSR via brass standoffs which provides both physical mounting and electrical connections. The SSR is mounted to its matching heat-sink for simplicity, although the entire thing is massive overkill and damn near any heatsink would work. There are also standoffs that serve to stabilize the screw terminals and also directly earth the heat-sink.

The decision logic is that my Home Assistant instance takes measurements of both the consumer mains and the power into the hot water cylinder via my whole-house energy monitoring system . It uses those to calculate what the potential export would be. It then sends a duty cycle to the PV diverter board based on that potential export. It’s pretty crude and could do with a proper closed-loop control system being implemented but the goal is ā€˜simplest idea that works well’, which I think this does.

That’s it installed under the house with some flexi conduit into the connection point for the hot water cylinder. It’s in a surprisingly solid little PETG enclosure that I printed. It’s a nice tidy looking little unit once everything is buttoned up.

Total cost was around $150 NZD all up, mostly because I went with a massive overkill SSR and the matching heatsink to make life easy.

That’s this morning’s power graph showing our main input overnight, the solar starting to come on at 6:30am and then ramping in. At around 8am there’s enough solar that we start to have export so the water heater turns on and starts soaking up that excess, keeping our mains input around zero. Once the cylinder has finished its full heat cycle at around 10am the thermostat turns off the load and we start exporting . It often kicks back on again in the afternoon for an hour or so.

I’ve got an automation set up with Home Assistant to notify me if the cylinder has gone for a full 24 hour period without going above a couple hundred W and a critical notification to let me know if we’ve gone 24h without significant export, which would indicate that the cylinder hasn’t 100% heated. On the to-do list is to add temperature monitoring to the cylinder, ultimately with the goal being to get a bit more flexibility over winter and potentially trigger a heating cycle overnight on off-peak power if a low temperature threshold is reached.

9 Likes

That is a great idea. Also really makes me want to play with home assistant more, sounds like a lot of fun.

Whenever my water heater goes out I will surely get a heat pump style next to take better advantage of my solar… probably not to this degree but I love the efficiency.

2 Likes

I bought a heat pump water heater almost 2 years ago now. It has been working great. Got lucky and found one at a ā€œscratch and dentā€ place. Home Depot wanted $2800 for the water heater, got it for $1200 with the same 10 year factory warranty because of a small dent in the side of it that I couldn’t care less about lol.

2 Likes

I’m a bit confused on where this is attached. I get that you prefer to run your wayer heater off of solar, rather than export the power to the transmission lines and then later import power for your hot water.

But does this skip the breaker box in your house and go straight from the inverter to your water heater?

The other thing I’d like to know is what the failure mode is if you don’t have network or home assistant fails. I assume the water heater defaults to a state where it work on the regular power if your project fails?

1 Like

wow, sounds great, but man, all those acronyms, for me to get alot out of it, I would be googling alot! You go get em!

Ok, I did it, just in case anyone else, (SSR) Solid state Relay (MQTT) I think is this: (Message Queuing Telemetry Transport) which is for his home automation.

now it makes alot more sense :slight_smile: Very Cool!

whoa edit 3, 16 x 440 w panels, that is AWESOME!

1 Like

It’s well worth having a muck around with it, especially now as it’s SO much more dialed in. It has always been good but the entire experience is surprisingly frictionless. I got started with spare raspberry pi, then moved it to a dedicated newer one but both of those were always just kinda irritating in terms of SD cards going bad etc.

That’s definitely a great way to get started and dip your toes in but if you’re going to be relying on it for stuff then I strongly recommend the Home Assistant Yellow:

It’s purpose built hardware that uses a Raspberry Pi compute module and lets you use an m.2 SSD which solves 90% of the issues I’ve ever had with HA. It also has built-in Zigbee support which saves you having a dongle plugged into the pi etc. It’s intended to be passive which the newer Raspberry Pis kinda aren’t, at least not at temperatures I’d want to run at long term. The Power-over-Ethernet support is way better and more reliable, too.

I think for me the biggest thing is that once I had it in place and could trust it, it became my go-to for if I want to log anything long term or monitor anything. It’s not the best way to do that, but it’s good enough for my uses. If I had an LR4 running and wanted to do something like log the temperature of the steppers or router housing, I’d get an ESP32, put ESPHome on it, hook up a DS18B20 one-wire temperature sensor and be able to have it displaying in Home Assistant with high fidelity data logging with about 15 minutes work. For me that recipe has turned entirely into the software/electronics equivalent of making jigs out of plywood scraps, hot glue and brad nails.

1 Like

Yeah, a heat pump is definitely the way to go long term. I kinda made a mess of planning with my setup because my water heater corroded out (if you have an enamel cylinder, CHANGE YOUR ANODE!) and lightly flooded the downstairs around 4-5 years ago. At the time heat pump water heaters were SO much more expensive than an equivalent immersion heater ($6-7k at the time) that there was no real avenue for payback on them for me. They’re much larger cylinders for the same ā€˜capacity’ of hot water due to lower idle temperatures but then they periodically heat to higher temperatures using an immersion element. Our usage is so little hot water (2 people, roof catchment water so low flow showers and we’re careful about water usage) that it’s difficult to tell how much of that would be simply off the immersion element anyway.

So my logic was to get an immersion heater and put the money saved towards a solar setup to provide power for it. At the time, a modest 1-2kW solar system and the immersion heater were cheaper than the heat pump water heater.

Then I let feature creep get involved and got a large stainless steel cylinder so that I could store more energy in it as a dump load but also get one with a 2nd element so I could easily have a backup in case the solar wasn’t enough. All of that meant I spent more than I should have on the cylinder and it has much higher standby losses than I wanted. I went from around 1kWh a day for a 150L cylinder in a cupboard inside to 3-4kWh a day for a 250L cylinder outside under the house. Some of that was how the plumber wanted to do it but I really wish I’d spent more time figuring it out and gone with a smaller stainless cylinder in the same location. That would have allowed higher temps so more capacity, stainless so no corrosion issues and would have been way more efficient. Oh well.

Anyway, from running the numbers for a colleague, it’s still ā€˜close’ when choosing between a conventional cylinder plus solar and a heat pump cylinder. If you have a good cylinder already, the solar wins by a mile. If you need to replace the cylinder and it’s already one that’s outside, the heat pump wins. If you don’t use much hot water or spend a lot of time away from home, solar wins. If you use a ton of hot water then the heat pump wins.

There are a couple of heat pumps that seem like they have the ability to be remote controlled or run on a timer. Running one on a timer so it only runs off-peak or during daylight hours already gets you 50% of the way to a diversion setup. There was one that had an RS485 interface for a remote control panel. I’d really love to have a hack around with that and see if I could do a basic form of diversion with that.

2 Likes

No, everything is still connected exactly as normal. The feed for the house comes into the main switchboard, goes through the main switch and then gets connected to all the circuits via individual RCBOs (combination circuit breaker and residual current device, basically what the rest of the world calls the things Americans call GFCI units or whatever). The main feed and each individual subcircuit have current transformers on them which connect to an energy monitor unit that provides pretty much realtime (every ~8s) measurement of the power in each circuit with good accuracy.

So, from that system I can see instantaneously how much power is being consumed by each subcircuit, how much power is being produced by the inverter and how much power is being exported/imported from the grid.

The device I’ve made is down at the hot water cylinder just because it was convenient, it could be anywhere between that and the main switchboard easily enough. In fact in hindsight I probably should have located it with the rest of my infrastructure equipment but this way it’s obvious if I ever need to have a different electrician look at my stuff. I’m trying to be at least a little bit aware of a succession plan in the case I get hit by a bus and my fiance ends up having to find someone to remove all the BS I’ve done.

So all the device really does is act a bit like a form of dimmer switch, just that instead of turning the load on/off for parts of half cycles, it does entire half cycles less often. Testing it on a lamp was horrifying, it’s basically a 5Hz strobe, which is truly awful. At the moment I have a control resolution of about 10% power, with each 10% being another cycle active out of 10 cycles. 10% power is one cycle on, 9 cycles off. 50% power is 5 cycles on, 5 cycles off, etc.

So to follow the logic of the system, say the house starts out with no solar and background loads of around 1kW. My energy monitor sees 1kW coming in and a total of 1kW going out to all the different loads. Home assistant sees that 1kW import and says ā€˜ok, no export so no need to turn anything on’.

Then the sun comes up and the solar system starts producing 1kW of power. The energy monitor will see 0kW of power on the main feed, -1kW of load from the inverter and then 1kW total going out to all the different loads. Home assistant will see 0kW import/export and go ā€˜ok, no export, no need to turn anything on’.

Then the sun gets higher and the solar system starts producing 2kW of power. The energy monitor will see -1kW of power going on the main feed (1kW of export), -2kW from the inverter and 1kW total going to all the different loads. Home assistant will see -1 kW and go ā€˜ok, there’s 1kW of export, I should turn something on’. I have the power of the element hard coded so HA just says 1kW export divided by 3kW element power = 33%. Round that down to the nearest 10% and set the target how water cylinder control to 30% duty cycle. That then gets sent out via MQTT and seen by the PV diverter that will go ā€˜ok, on for 3 cycles, off for 7 cycles’.

When the next set of measurements comes in, assuming the sun is still in the same spot, the solar system is still producing 2kW of power. The energy monitor will see 0kW of power going on the main feed, -2kW from the inverter and 2kW going to all the different loads, one of which is the hot water cylinder with its 1kW of load. HA sees that 0kW on the main feed and if it were only reacting to that, it would say ā€˜oh no, no export, quick, turn off the hot water cylinder’ and it’d sit there and oscillate on and off. So instead I make a synthetic measurement which is the power on the main feed minus the load from the hot water cylinder. So in this case, even though there’s no power going out on the mains, the 0kW on the main feed minus the 1kW of load in the hot water cylinder gives us 1kW of ā€˜potential’ export, which is what I use as the control variable. Based on that, HA goes ok, 1kW of potential export means 30% duty cycle still and keeps the control variable the same.

Then say another load pops up and draws an additional 500W, we’ve then got 0.5kW coming in the main feed, -2kW on the inverter, 1kW on the hot water cylinder and 1.5kW on all the other loads. HA would see that ā€˜potential’ export as 0.5kW - 1kW = 0.5kW and turn the duty cycle down to 15%, which gets rounded to 10%.

Once the water is fully heated then the thermostat shuts off and no matter what the PV diverter does, the load will be zero and the excess will be exported.

I don’t aim for ā€˜perfect’ balance because so far I’ve always got excess and I don’t actually care if I’m at 0W or -500W while the hot water cylinder is heating. There are other systems out there that do this as full closed loop control and will perfectly zero it out in real-time, but they’re more like $1.5-2k to get installed and if I can get 90% there with $150 spent then I think that’s a better bet. Considering that we use around 1500kWh a year heating water, that’s $375 if we only heat it on-peak, $225 if we only heat it off-peak and $127 if we heat only using solar (the opportunity cost of not exporting that same energy). So if I do nothing, it’s about halfway between the on-peak and off-peak costs. If I spend $50 and add a simple timer I can get to that $225 so that’s my starting point. If I can spend $150 and save $100 a year, that’s glorious. If I have to spend $2k to save $100 a year, that can piss right off. That would all be a lot different if I were using 10x as much hot water (big family, teenagers etc.) or had other heating loads like a spa pool, but we’re relative low-use on that axis.

Hopefully that answers the question? Basically there’s no ā€˜interception’, it’s more like having a water tank that can be filled by the inverter or the main grid and drained by the loads. If you have too much power from the inverter, the tank fills until it’s overflowing back onto the grid. If you have too little power from the inverter it drains until it needs to be topped up from the grid. By watching the water level in the tank I can keep it in the middle by varying the outflow to something that doesn’t really care about the flow rate (like watering a garden, say), only that it gets ā€˜enough’ during the day.

That analogy is actually kinda interesting, now that I think about it, because it actually explains how the utility power meter works. The power meter doesn’t meter instantaneous power because that’s not really kinda how it works in an AC system anyway. Instead it measures that power and constantly integrates it until it reaches plus or minus 1Wh/3600J. Once that threshold is reached, it adds a count to the energy consumed counter (assuming the counter overflowed in the import direction) or the energy produced counter (assuming it overflowed in the export direction). It will also flash the LED on the meter when it does this, which is where you get the 1000 flashes per kWh display that’s common on utility meters. That’s what allows us to do that cycle by cycle control without constantly flicking between import and export. 1Wh is 3.6kW for 1 second so if my only load was my hot water cylinder, the light on my meter would be flashing a little less than once a second. So in practice, my 1.5kW export without the hot water cylinder turns into 1.5kW export 50% of the time and 1.5kW import 50% of the time. As long as I never draw more than 1Wh or export more than 1Wh in a row, it never increments that counter. I’m doing it at around 5Hz so in reality it goes up/down by around 0.15Wh at a time. As long as I’m balancing it to be slightly exporting (say 100-200W or so, depending on rounding etc.) it’ll slowly count down and maybe trip the counter to say I’m exporting 1Wh every 30 seconds or something, but that’s ok by me.

Yeah, that’s definitely a careful consideration here. In my case, it’s only hot water and my partner is patient and supportive so if we end up with no hot water, oh well!

Part of this has been that I’ve been running HA for years now and have made huge improvements to its reliability. It’s on a HA Yellow box which has been 100% reliable over the past 2 years. It’s powered by PoE out of my Ubiquiti 24 port core switch which is running on a hacked-up APC 1000VA UPS that’s running from an external 300Ah deep cycle battery and gives me about 30 minutes of backup. So I’m as close to 100% confident in my HA instance being reliable as I can be.

We also actually use the HA instance for a lot of obvious stuff around the house now, which I think is key. It controls grow lights, the A/C, some smart blinds, the coffee machine etc. So if we get up and the coffee machine is off or the grow lights don’t turn off at 10pm etc., those are all pretty good canaries that something has gone wrong. My fiance also uses it a lot to check temperatures around the house etc.

Where possible, everything is ethernet connected and, ideally, PoE to make use of the UPS. In this case it isn’t because ESP32 + ethernet kinda sucks to do but we’re pretty rural so my 3x Ubiquiti dual-band pro APs provide rock-solid wireless to everywhere in the house and pretty reliably in the surroundings. Under the house it’s a bit more 50/50, hence the external antenna. It’s at about -62dBm which is enough for tens of mbit at 0% packet loss.

Another thing is that both the energy monitor and the HA instance have Pushover setup (an awesome service that I pay for that allows reliable push notifications to both my phone and my fiance’s phone via a simple and stable web API). I will get notifications from either HA, the energy monitor or both depending on some pre-set criteria. One is that the cylinder hasn’t drawn any power for more than 24 hours. If the PV diverter were to die, I’d get notifications every 24 hours that something was wrong.

One thing that I could do is make it so that if the PV diverter doesn’t see a valid duty cycle update message then it defaults to 100% output, which would basically make it transparent and default to a ā€˜conventional’ water heater install. I’ll do that if I run into issues but for the moment I’d rather have it err on the side of ā€˜don’t heat’ so I know something is going wrong.

The final thing is that the PV diverter isn’t doing anything other than controlling the available input power from 0-100%. The hot water cylinder still has the same thermostat control on it as before so even if everything were to crash at 100% duty cycle, it would be a normal situation, not an ā€˜element stuck on’ situation. In future I plan to add some more temperature control related smarts to it, but I’ll still retain that thermostat as a max temperature cutout. The thermostat itself has 2 controls in it, one is the main on/off thermostat to control the temperature, the other is a single-shot overtemperature thermostat that needs to be manually reset by a button on the unit, so even in the event that the first thermostat fails, the over-temp one should shut it off. Even if that fails, there’s a pressure relief valve that’s sized for the 3kW element that will vent the extremely hot water out via a copper relief pipe that runs to the ground under the house. It’d be a bit of an experience if that went, but not hazardous to people, animals or the house in any real fashion.

2 Likes

Oh no, my bad! I tried to stay away from anything that I thought would be unfamiliar to people here.

You’ve got it in one. The tl;dr of the whole thing is that I made a bizarrely rube-goldberg like contraption that checks to see if my solar panels are sending power out to the grid and then lets me dial the heating element in my water heater up and down so that I’m using that power at home instead of exporting it. It’s focused on trying to do it ā€˜good enough’ with simple/cheap hardware so that it actually makes financial sense, as opposed to the very expensive off-the-shelf designs that can do it pretty much perfectly.

That’s them. It’s tough because there’s that tree on the north-east corner of the house (the tall straight one that you can see a bit of the white-ish trunk) which casts a long shadow, especially in winter. It’s a protected native tree (New Zealand Kauri, Agathis australis) that’s only found in the northern half of NZ’s North Island.

They grow slowly but very tall and very wide with dense, straight grain that’s pretty tough The timber is a beautiful deep yellow with a kinda golden shimmer to it in the light. Can reach 50m/165’ tall and the largest recorded one was supposedly 8.5m/28’ diameter. They were used by the Māori (the indigenous people of New Zealand) to make waka (traditional canoes used for long distance travel, fishing, battle, all sorts) and then logged damn near to extinction by early European settlers for ship masts then later general use timber. Not super dissimilar from the story of the giant sequoias, except that the I think the wood is a bit better and more useful so got turned into damn near anything. It’s not uncommon for houses to be framed in kauri, which to me would be like finding out that the rough sawn 2x4 holding your ceiling up was walnut… I had a colleague whose house was clad in kauri weatherboards. He joked that he should have stripped and clearcoated it instead of repainting it. I kinda wish he had!
Anyway, there aren’t a ton of them left and they have a root fungus that is attacking them and killing them slowly throughout a lot of the remaining areas that they can be found in. Our property has hundreds of these trees on it. The one nearest the house is a huge pain in my ass with dropping leaves into the gutters and shading my solar panels. Everyone I’ve talked to says ā€˜you should really fell that, or at least top it out’, to which I take great joy in pointing out that despite it being quite a young kauri at maybe ~100 years old, it was there before them and, as long as I have a say in the matter, it WILL be there after them.


That’s our ā€˜back yard’… Our property extends 500m or 1/3 of a mile in that direction, all forest. Most of the tall pointy trees out there are kauri. The flowering one in shadow on the right is a manuka. If you’ve ever heard of manuka honey, that’s a big NZ export… Plenty of bees in it during that photo!

3 Likes

Definitely. I was confused by the word ā€œDiverterā€. You did it exactly how I would have.

My projects most often fail because of my lack of maintenance or urgency. So I am always looking to decouple them. That seems to make the whole system more reliable. I run HA on a dedicated machine running proxmox. I bought a cheap hp workstation off of ebay for $75. It has been pretty reliable. But there is always 1-3 devices not working perfectly at any given time…

I was going to suggest that. I have a HA connected thermostat and temperature sensors throughout the house. I want the average home temperature to be the set point for the thermostat. But I didn’t want to have them that coupled. So I compute an average temperature for the house in HA and I compute an offset from the thermostat’s thermometer from that average. The thermostat gets the offset from HA and adjusts its set point based on the offset. The offset is low pass filtered (because it doesn’t change much in 15 mins) and if the offset is ever missing, the thermostata goes back to being as reliable as any other thermostat.

You could even go a step further and make wire the SSR normally closed. So if your device lost power for some reason, or the microcontroller died, you could still have running hot water. I know my wife wouldn’t be very happy with cold showers. Especially in the winter.

I don’t want to lecture you though. Your solution seems awesome and I am jealous of your energy monitoring and solar set up. We are in the market for a new water heater and a heat pump version is on the shopping list. Our current one burns natural gas. It is a good trade off to look at doing solar instead of optimizing the efficiency of the WH. One of my tasks this week is to figure out how our taxes are going to change if we get said heat pump heater.

1 Like

I’m mostly copying industry standard stuff, just with trying to offload as much cost/complexity to stuff I already have in place. The sad thing is that it makes it much more custom and less ā€˜valuable’ in a generic sense, but I still think I could distill it down to an HA instance, a power meter and my diverter for less money than a commercial system. And you get an HA instance to play with.

I considered going this way but general purpose computing hardware is so inefficient that I’ve tried desperately to avoid it where possible. The HA Yellow draws about 3-4W off PoE so 4-5W total while even my lowest power celeron NUC is in the 20-30W range. Prior to getting solar I was using the rule-of-thumb that 1W of power is around 2 dollars per year cost (average of around 20-21c/kWh across the day, 8,800 hours in a year, 1W = 8.8kWh, etc.). So if I can save 50W then that’s worth 100 dollars per year to me. My colleague runs his instance in a docker container on his reasonably grunty Intel-based Synology NAS. We think that thing draws around 100-120W vs my Realtek based NAS at 14W measured. He’s trying to consolidate everything onto that where I’m trying to go lower power and distributed, etc. So far I think I’m around a couple hundred $ a year lower than where he’s at, although that’s before considering the solar.

Edit: I just checked and it cost me ~$250 for my HA Yellow with a 4GB RAM CM4 and a 1TB SSD. So assuming that workstation is ready to run, add a $25 for a USB Zigbee/BT stick and that $150 extra is probably a year’s worth of electricity, for me, 2 max.

Obviously depends a lot on personal preference, what else you’re doing with that hardware, what energy costs are, etc.

Hmmmm, that’s odd. What’s the reason your stuff isn’t working? For me the only time I lose sensors etc. is if the batteries in the temp/humidity sensors run down… It used to be a bit weirder where I’d sometimes lose stuff on startup but it has been rock solid for maybe a year now…

I’ve not seen a cheap general-purpose NC SSR but that’d definitely be an option if it were something critical. I guess you could always add the NC contact of an electro-mechanical relay across it if you absolutely wanted it to fail-safe in the case of a power failure within the unit itself. At that point you’re really only protecting against the SMPS which should be pretty reliable. The unit is powered by the same supply as the cylinder itself so a breaker tripping will take the whole thing out, anyway.

There’s also the consideration of fail-safe and fail-obvious. I’m actually not really worried about fail-safe on everything, I’d rather have fail-obvious. If I know it has failed, I’ll fix it or bypass it. If I’m not available to do that, it’s an easy job for an electrician.

In our case, we have enough hot water that we can go about a week of ā€˜normal-ish’ use before showers get noticeably cold. We had to do that a couple of years ago due to a storm damage related long term power outage. So even in the worst case of the thing outright dying, it’d also need the pushover service to die unnoticed, which it probably wouldn’t because I have it alerting me any time the water pump has been running for 10 minutes or more as a ā€˜just in case’ thing after we lost most of our water tank to a ruptured garden hose. That means I usually get a notification for anyone taking a shower etc. and I’ve left it at that as a heart-beat type notification.

So even if the unit failed at the same time as the pushover service or energy monitor, it’d still be at least 3-4 days before we went ā€˜that’s weird, the hot tap isn’t super hot anymore’, which would then be 2-3 days before showers were actually uncomfortably cold.

The solution to that is a bit agricultural, to be fair. I’d have to go down to the hot water cylinder, isolate it and wire the phase connection to the cylinder across to the terminals of the incoming feed. If I were worried I could add a bypass switch to the diverter itself. I wouldn’t be able to easily add a bypass at the existing connection point without adding an extra switch because it’s required to be an isolator, not just a functional switch.

Actually, if I were worried about it, I’d use the top element position in the cylinder to add another element via another thermostat with a lower temperature setting such that in the event of any failure with the main element then I’d still have a backup/secondary. Honestly, I’m more worried about the thermostat failing than my PV diverter, I’ve had 2 fail on me over the years. One in a rental flat and one on this cylinder, funnily enough.

Anyway, that’s all SO far within the acceptable bounds for success with my household that I’m actually pretty stoked with the system. Obviously opinions will vary, though!

Not at all, it’s interesting going over all these considerations. As long as it’s a ā€˜this is what I’d be worried about so this is what I’d do’ then it’s a great discussion to have. Also as long as you’re able to see that I actually have seriously considered all of these possibilities and accept that my coming to a different conclusion is due to different circumstances and attitudes, not due to a lack of understanding or faulty information! :slight_smile:

Yeah, there are a lot of people here trying to bail out from their on-demand gas water heating due to rising costs. A colleague is in that same position and trying to determine between solar + immersion, heat pump or solar + heat pump.

It was a toss up when I considered it a couple of years ago so I suspect it’ll get pushed more and more towards heat pump as time goes on, especially if there’s something like a run inhibit or timer function that can get you to ā€˜maybe mostly solar’, assuming the re-heat time isn’t horrendous.

1 Like

This is cool I’m currently trying to figure out how to make more use of the excess solar we produce, at the moment that’s focusing on running the AC to stop the place getting too hot so we’re not hammering it in the evening when the cost of electricity goes up.

Here in Western Australia we’re probably living in the future a bit for north Americans, we have so much renewables in line now the cost of electricity between 10am and 3pm is A8c a Kwh add to that a 6.6kw solar system that’ll cost about A$6,700 currently you want to shift as much of you demand as you can to that time period.
Obviously this creates an incentive now to get a battery, a 20Kwh battery here is currently about A$17,000 they’re getting cheaper all the time. Though it may be better to roll that cost into an electric car and use vehicle to home, do that and providing you can charge at the cheapest rate you’ve effectively got all your electricity at the cheapest rate, you still have to pay for the battery or car though :wink:

When in California last year looking at the prices the solar systems looked to be about double the price we pay, installed? Thats a bit sad if correct.

1 Like

Yeah, that’s an interesting consideration, for sure.

I have personal misgivings about house-scale batteries at the moment (edit: specifically house-scale, I think grid-scale with proper controls have a much different set of considerations to work with and can be significantly more environmentally positive given how they allow higher percentages of grid-connected renewables). Here in NZ we have a remarkably low carbon grid so it took a long time before I was comfortable that getting solar was actually a net improvement, environment wise. Now it’s kinda the same thing with batteries. 50-200kgCO2eq per kWh (depending on which study you read) is pretty rough when considering that it doesn’t really do anything to generate more, it just time shifts the energy to allow you to use it yourself. Taking the middle-of-the-road numbers it’s 1tCO2eq for a 10kWh battery. Assuming 10 years of life at full depth of discharge each day that’s adding ~30gCO2/kWh to the carbon intensity of my energy usage just for the time shifting, which really doesn’t seem great to me. Of course that’s less of an issue if it lasts more than 10 years, which it should, or if there are good recycling options at its EOL, which there should be, but it’s just not so black and white for me, at least.

Here the best I can do is 25c/kWh daytime to 8.5c/kWh solar, so ~16.5c/kWh, assuming I use ALL of the energy from the battery in the hours between the solar no longer producing and going to off-peak. In reality, the time-weighted average price for my ā€˜non producing’ hours is closer to 19c/kWh which is 10.5c/kWh return. Considering the ~90% round trip efficiency that goes to more like 9.5c/kWh. Adding a 10kWh battery that I dead-flattened every single day would give me $1 a day return. So even ignoring the environmental concerns, I’ve never been able to make the numbers add up for my personal situation…

I know a few people who have them and are completely sold on them, but there’s a weird thing where out of all the people I’ve talked to who rave about how much money they’re saving, very few of them have any way of actually tracking that or put any thought into it beyond ā€˜well, my power bill used to be $300 and now it’s $100’, despite that being not much better than the return they’d have gotten from an index fund…

Environmentally speaking, I’m a HUGE EV proponent. House batteries just time-shift energy, an EV battery is removing a combustion engine from the equation. Not to mention all the other benefits like air quality (in NZ asthma and respiratory illnesses cost the country billions a year edit: Study reveals health impacts and social costs of air pollution | Ministry for the Environment claims that human-generated air pollution causes $15.6b in social costs with motor vehicles being the largest contributor by far), noise pollution (the reason we left our previous house), trade deficit (the cost of petroleum imports for the transit sector is more than half the value of our total trade deficit, which is wild… $5.4b imports in 2022 vs a trade deficit of $8.5b in 2023) etc. etc. etc.

The next thing on my to-do list is to spec out a good solar diversion capable charging point. I’ve got a pretty specific set of requirements, though, so it’s proving to be a struggle to find one that isn’t something like a $2,500 ABB unit or otherwise. EVNEX in NZ make a fantastic quality one but sadly it needs another power meter in the main board to do solar diversion, doesn’t support OCPP, and they can’t confirm it’ll function correctly if I don’t give it open internet access, which is a real bummer.

V2H is still a bit early to be used that way, though, in my opinion. I don’t know of any systems that are actually any good and ā€˜doing it now’. There’s the Wallbox Quasar which has been in the wings for a long time, I think Enphase have one but it needs the entire Enphase setup which is spendy, plus I don’t know if it’s ā€˜actually’ available. Some of the vehicle manufacturers have dedicated units (Nissan did one in Japan with the Leaf, Ford does one for the F-150 lightning, I believe there might be a GM one, too…) but they’re not really ā€˜general availability’ and I think are more along the lines of tech demonstrators that aren’t widely deployable due to standardisation issues.

If you know of any then I’d love to hear of them. This time last year I was working on making a V2x capable charging system and that ended up getting canned due to being too complicated to do at the moment due to the regulatory landscape and lack of support from OEMs.

2 Likes

Here we’re at:

  • A8c per Khw from 10am to 3pm
  • A54c per Khw from 3pm to 9pm
  • A24c per Khw from 9pm to 10am

So, you can see why just having a battery for 3pm to 9pm would be great thing, even with no solar and charging it from the grid at the cheap rate it still pays for itself well inside the 10 year guarantee period. I’m hoping sodium batteries come on-line soon which’ll make the math a no brainer.

Our grid was coal and gas, all that is to be gone by 2030, there’s massive grid scale battery projects in the works such as these:

Here come the batteries: Six new big battery projects in line for capacity payments in W.A. grid | RenewEconomy

The excess we send to the grid, we get A2c Kwh for so essentially nothing. But I look at it as help to make a headache for the authorities who really need to push on with the big batteries to soak it all up. And of course, it’s being used somewhere by someone and not coming from the power station.

I’m not that far into the EV research V2H was approved by the authorities recently, I see cars that do it are generally limited to a 3Kw export rate, which is enough 90% of the time. But I’ve not got into the weeds yet.

2 Likes

Yeah, that’s such a nutty inversion of the power pricing compared to here.

There’s potentially the option within that timeframe to still do ā€˜diversion’ type things as well, I guess. I have smart switches on the dehumidifiers now and want to add one to our chest freezer. There’s potentially a bunch of stuff that could handle being turned off for a few hours to avoid that expensive timeframe. Also, presumably you’ll have a decent chunk of solar beyond 3pm? At this time of year that’d be importing for max 2 hours with our usage. Maybe more like 4 hours during winter.

Yeah, some of the battery options that are being floated could be huge for home storage. I was a big fan of the Redflow redox flow batteries but sadly that company collapsed mid last year.

There’s also the option for re-using end-of-life EV batteries for home storage that bypasses a lot of those considerations. Here in NZ you can pick up a Leaf battery with a ton of life left for a couple $2k which bypasses any environmental concerns (2nd life) and cost concerns ($2k for >10kWh usable is nuts). It’s just that it’s something that gets a bit sticky in terms of compliance/electrical safety, sadly, so I’m hoping some company will pick it up and run with it by turning them into a fully off-the-shelf BESS.

Those vehicles are all V2L or vehicle-to-load. They don’t inherently support being hooked up to a house. With some of them it’s possible to do it in the same way you do a generator inlet, i.e. switch the entire house over to a secondary supply inlet and then hook the car up, but that’s not really feasible for situations other than grid outages. It’s a truly monstrous regulatory hurdle to go from V2L to V2H/V2G and likely impossible for a purely car-based solution.

Most of the V2h/V2g systems that are in the works are using DC from the vehicle to an bidirectional inverter on the wall so should have a much higher power limit than 3kW. I think the F-150 lightning is the only one I’ve seen that does V2H via an on-board inverter and even then I think that’s a bit of a technicality…

Regardless, definitely give the EV thing some thought. We got our first EV back in 2014 and we’ve been an EV only household for 4 years now (Leaf and Tesla model 3). The landscape has changed SO much over that time and there are so many great options out there. Admittedly I’m very much on the ā€˜EV Enthusiast’ side of the equation given that I’ve been designing EV charging systems for the past decade and a half of my life but more and more people are seeing the writing on the wall for combustion tech and it’s just such an all round nicer experience!

@vicious1 I see you lurking there :smiley: Can you tell it’s a slow day at work? :stuck_out_tongue:

1 Like

I needed a specific charger and my electrical utility paid for it. We ended up getting a charge point home flex for $509 and the utility gave me a rebate for $500. It works fine. But we aren’t doing any on demand scheduling or rate limiting. We just charge overnight off peak. If I had solar, I would probably make it charge from 10-3pm everyday or something. That would probably be plenty and we are usually home during those hours.

We have a chevy bolt. The equinox ev has a V2H system available but it is in the $5-7k range. Some bolt owners use a 1800W inverter on the 12V system and convince the car to keep the 12V charged from the HV system. That would be enough for an emergency here. Although so would the power ports on the back of an F150 lightning. It makes no sense to me to have a battery in my garage for the house. But some places with lots of solar and high prices it might make sense (California for example). Having a battery that you drive around makes more sense. The use case is that it charges during the day and then discharges at night. But I want it for driving in the morning. IDK if I would like that either. At any rate, solar usage in Denver is not that high yet. So I would be doing fine by the pocket book and the environment to just run the meter backwards during the day. My neighbors would get green power then, at least.

I’m well aware of the OpenEVSE :slight_smile:

It’s not applicable to NZ/Australia because of the way we handle electrical safety here. Charging equipment needs specific documentation and testing that is basically impossible with something like the OpenEVSE. I spend a LOT of time dissuading people from using it, partly because I think it’s kinda ā€˜meh’ as a design but also because it’s specifically illegal to sell/supply here in NZ and very questionable to use in terms of insurance/liability etc.

For that same reason, pretty much anything that gets recommended from someone in the US is going to be irrelevant for us, outside the bigger names like ABB. It’s kinda irritating in some circumstances but it also protects us from a LOT of garbage.

Yeah, the Equinox V2H system looks like one of the ones I mention above where it’s specific to the vehicle and not ā€˜really’ standards compliant so unlikely to ever be usable outside the US and potentially even only usable in specific US jurisdictions depending on supply agreements. Not that it matters because you can’t get an Equinox outside the US, anyway!

I agree, I’m kinda lukewarm on the V2H idea anyway because I think most of the better use cases are charge off solar, back supply overnight which doesn’t fit a lot of use cases particularly well unless you’ve got an extra vehicle that doesn’t get driven much during the day or work from home. Still good and would work well for us, just it’s more of a ā€˜suits some of the people some of the time’ type of solution.

As for running the house off the 12V inverter, yep, that’s exactly what we’ve been doing since 2014. During the big storms a couple of years ago we spent an entire week running the house off our Leaf via a 12V inverter. It was enough to run the water pump to give us running water as well as running the networking setup, charge phones/laptops and even run both my main linux PC and my gaming PC.

2 Likes