I really want to build a DC electric vehicle charger. Most people hear DC and think “DC fast charging,” because all the fast charging systems out there are DC, but that association is sort of backwards: fast charging must be DC, but DC charging need not be fast. In fact, I think there’s a very important use case for slow DC charging that’s totally unaddressed.
Some background: how does EV charging work? There are 3 components, really: the supply, the charge controller, and the battery itself. People tend to think of the box on the wall or the cable that came with your car (EVSE – electric vehicle supply equipment) as the “charger” but this is not so much the case: it’s basically just a plug adapter to attach household AC wiring on one side and the car on the other, with switches in the middle to make sure you can’t zap yourself at any point while handling the pieces. In reality, the charger is a bigger, more complicated box that serves two functions: rectification, and charge control. It rectifies the AC we pipe around our electric grid to high-voltage DC that can be consumed by the battery, and controls how much current is available for the battery to consume.
To support level 1 and level 2 charging (<2kW and <10kW AC), a charge controller is built into every EV on the market. A charge controller’s high-level job is to keep track of the power available on the supply side, the maximum safe power the load (battery) can handle, and pump the lower of those two numbers from input to output. Without too much detail, this charge controller box does rectification, power factor correction, DC-DC boost, and high-voltage DC switching. None of those steps is 100% efficient, and even 1% inefficiency at 10kW is 100W, so you can see how the physical size of the box would need to scale up to handle increased power beyond 10kW of charging.
That on-board AC charger is kinda big as it is at 10kW, so if we want to charge that same car at 100kW or even 300kW, it’s clear we can’t simply stuff a 10x or 30x larger charger in the car. A 30x larger box needs to go somewhere else not-on-wheels, and supply battery-ready high-voltage, high-current DC to the car instead, which is exactly how superchargers work. The charge controllers all live in that massive, loud gray cabinet off in the corner of the charging facility.
So now we know why DC charging is necessary if we want to charge fast. But why would we care if we want to charge slow?
Solar panels produce DC current
Firstly, solar panels generate DC current in the sun. Secondly, depending on how much sun hits them, they generate mostly the same VOLTAGE, but with extremely variable CURRENT available. For example, a 3kW solar array will be able to produce, say, 300V 10A in direct noon sun, but maybe only 5A a couple hours later. That’s fine if we attach a load to those panels that’s smart about detecting and consuming exactly as much power as they’re able to generate in-the-moment. Grid-tie inverters do just that – they pump into your AC mains exactly what the solar panels can generate at a given moment, and anything excess above what your household is using gets “exported” to the grid as current flowing out of your house. In other words, the grid is used like a big battery to make up the difference, positive or negative, and if you’re lucky your utility will charge or pay the same rate either direction.
What if it won’t, and you need to use the energy you generate or lose it? This is one use-case for Tesla’s powerwall and its various competitors: charge up during the daylight with your solar panels, then use that power later on when you need it. The powerwall has a charge controller that can keep track of what’s available from the solar panels and use it all.
But you already have a big battery that you’ve already paid for: your car. Why can’t your car do that same thing – take all the power your panels can generate, to use later [when you need to drive places]? The issue is that, if you do the traditional thing and use an inverter to make AC from the solar array, then the onboard charge controller of the car, there’s an information disconnect in the middle. The car’s charge controller doesn’t know how much power your panels are making: it just assumes a number for available power based on what kind of plug its attached to. Similarly, your (off grid, I’m assuming, since we don’t want to export) solar inverter doesn’t know anything about the load attached. If we consume less power than it can make, all is well, but we’re wasting some. If we consume more power than it can make, the system will quickly shut down to self-protect, and the car stops charging.
What we want is a charge controller WE have full design control over: one that can take DC from the panels, figure out exactly how much power is available, then push that exact amount of power into the car. As a bonus, high voltage is relatively easy to design for, where high current is what requires lots of physical space and cooling. We wouldn’t need to do any power factor correction (no AC involved), and DC-DC conversion could cover a very small voltage delta, making the circuitry compact and efficient as well. Simply stringing 10 or 12 panels together gets us close to the pack voltage at 400V, so our DC-DC converter really only has to control the current into the battery.
In other words, if we could design a box to “DC slow-charge” from a solar array, it would enable extremely compact, cost-effective, and energy-efficient solar harvesting straight into the car. Pretty neat, huh?!
I have had this exact same thought and I want to do it so much that I sent you a linkedin request.
1 Does having a solar array at home to charge a battery at home – then DC slow charge an EV overnight – avoid the inefficiencies of converting to AC and then back again when charging the EV battery?
2 Does the slow charge preserve your EV battery by not creating much heat?
3 Isn’t this much more convenient because I park my EV at work during the day so I need to charge it at home overnight?
1) kind of, but it’s probably about as inefficient to invert+rectify as it is to DC-DC>battery>DC-DC>battery. Converting costs efficiency, not necessarily JUST inverting. Plus, then you’d need a huge battery, comparable in size to your car. If you want to do this, for sure the easiest option is to just buy six power walls. I will not be buying six power walls.
2) This isn’t really a relevant concern. The relevant “battery preservation” metric is charge rate as a ratio to pack capacity. If you’re charging at “less than 100%/hour,” the pack is going to be just fine. That would be 80kW for i.e. a Tesla model 3 or Y long-range. We’re not talking about making an 80kW solar array, and that would be “fast charging” if we did, anyway, albeit very safe, pretty slow “fast charging”
3) If you can’t park your EV next to solar panels during the day, my whole concept here is moot. The idea is you’d have a system to directly solar harvest either at home (I work at home) or in the work parking lot.
I agree with this concept entirely.
I have a small stand-alone PV battery system at home to play with. .So simple. > Panels > MPPT > Battery.
The MPPT charge controller does not care how much power is coming from the PV, it just pumps the avaialble amps into the battery until the boost and float voltage specs for that battery is reached. Now if I disconect my 24V battery and plug that MPPT directly into the DC port of the EV battery, it will work exactly the same. All we need is an MPPT controller that works to the Lithium battery specs for the EV, whether that is 100V, 200V or whatever.
Additionally I don’t think the PV array needs to be high voltage either.
There is lithium specific MPPT controllers out there that will buck-boost the output voltage when the input PV voltage is lower than the battery voltage. For example this mob has a 56.8V MPPT that can boost, https://genasun.eu/collections/genasun-lithium-mppts
I think EVs though are using higher voltage battery packs than 56.8V, (up to 300V?), In theory all we need is an MPPT controller for the required voltage and it should all work. It will trickle charge whenever the sun is out and at whatever rate the PV array can deliver. I think it could be safer for the batteries too than a 1000A fast charge.
Note, I would rather stick with a safe low voltage DC PV array than play with 300V DC which is somewhat deadly.. All we need is the right MPPT controller and it’s job done.
Much higher voltage, yes. For reference, the mode Y has a 96S pack configuration for 350V nominal, 403V full charge. Newer EVs like the Ioniq 5 and Taycan have a reconfigurable pack arrangement that can put two halves of the pack in series for charging, up to 850V max for current gen CCS.
I almost agree about boosting, but I bet jumping such a large gap is pretty inefficient, and 400V anywhere is almost as bad as 400V everywhere.
That said, at a reasonable array power, boosting is almost mandatory. It’d take 10 full-size rooftop panels to hit 400V, and that’s in noon sun, every cell in series, sourcing 2kW.
ON THE OTHER HAND, charging with less gets increasingly silly. 3 Panels at 600W feels like a decent amount of power to continuously, slowly harvest, but in a Tesla at least, simply running the computer (mandatory for charging) consumes half that.
Really, at the end of the day, this kind of connector would be good for maximizing use of a large home-scale solar installation, and in that setting it’s not uncommon to run a 400V string anyway.
I was thinking along the lines of really simple DIY stand-alone systems built solely for the DC EV charger, similar to any remote off-grid PV system and lower voltage.
And yes, assuming 3x200W panels at approx 96V DC it would be a pretty average 600W starting point, but we can add more 3x200W strings in parrallel to get more multiples of 600W to scale up to something more useful whilst still maintaining a relatively safe DC voltage. Panels are very cheap.
Hmmm, Imagine building a carport with a solar panel array as the roof. All you need is that mppt charge controller device of the right specs and an EV cable and Chamedo plug. Simple setup.
But I believe you are suggesting using an existing large home grid-connected inverter type PV system and feeding the PV DC power direct to the DC EV charger (with the required charging regulation).
It should be do-able if there was a smart energy management system that could direct power to where you want. i.e either to the DC charger, or the grid inverter (or hot water system etc etc whatever you have)
Good suggestion. But I reckon your grid inverter would not be doing much when you plug the car battery in, and thats OK if you get little or no feed-in tarrif.
Using the EV battery in reverse direction to power your home at night is a very nice idea too.
Bi-directional DC EV charging is already thing. I see there is a 7.6kw (level2) chamedo DC charger such as Wallbox Quasar which allow energy to be both charged or discharged from an EV.
That makes the EV a home battery in effect. Sadly though the DC part is only charger to EV and EV to charger. House to charger is AC 240V @6A-32A.
Powering the home from the EV obviously requires DC to AC conversion, so the inverter is just going to be there anyway, so may as well use it, as bi-directional EV charging is pretty damn attractive.
I’ll be looking into it when I buy an EV and my juicy solar feed-in tarrif expires in a few years.
“Imagine building a carport with a solar panel array as the roof.” This is exactly what I have in my mind, yeah. It makes more sense for me than most, since I work from home and have complete flexibility to co-locate my solar and vehicle during the heat of the day. I can imagine this being a “less regulatory burden” way for employers to provide at-work charging, but that also might not work well (many employees who charge at work probably require more reliability in that regard, vs getting stranded at work because it’s cloudy).
I AM also suggesting a DC-side energy management scheme where, in a grid-tie scenario, the car can get first priority for higher efficiency. But that’s not a requirement in my mind, more a “most interested parties would probably want that.”
If vehicle-to-load to power your house REALLY jazzes you, I’m trying to pick up an Ioniq 5, which has that built-in: it runs its charger in reverse as an inverter to pump OUT up to 3kW mains AC. In theory, you could leverage that to charge from solar, and discharge to mains in an outage, using your car as a full-blown “powerwall whenever it’s plugged in.” But there are so many regulatory and practical reasons that won’t work well that I think it’s not worth pursuing. Namely, you NEED a non-export relay in that case to disconnect your house from the grid if you’re locally powering it during an outage, or else you must do that step manually which means you won’t know whether power’s back without regularly checking it yourself, etc etc. And assuming grid-tie solar, you couldn’t simultaneously power the house from solar and the car if the outage were during the day, since the car is definitely not designed to phase-match external AC.
V2L is definitely cool and I’m definitely going to use it if I can get one of these cars this decade, but from a home-backup standpoint I think it’s more of a “run a long extension cord to the fridge until the power’s back” scenario.
I’ve been scouring the internet for information on just this. Any luck with making it work? It makes no sense to me to convert DC-AC-DC.
Sadly, no – I haven’t attempted such a project yet and may or may not ever have time to. The overhead of keeping the car electronics enabled (2-400W) kind of kills the allure for me. If you could go direct-to-pack without any of the overhead, as if you had the solar roof version of the car, I’d be a lot more tempted since even relatively small solar installations would produce useful output.
This is finally getting easier! This link shows examples of the software and hardware requirements of enabling the DC contactors in the car. They won’t enable unless there is DC voltage they requested on the HV pins but that is expected..
https://github.com/uhi22/pyPLC/blob/master/doc/EvseMode.md
Chester – that’s a very cool link, thank you for sharing it here!
Thank you for this. I was surprised at how little information there seems to be about this concept.
As you said, you will need quite a few panels to get the 400 volts you need to charge. So I’m not sure why the ~300 W in charging losses is a turn off for you. So maybe 1 of your 10 panels is enough to cover those losses? Plus, to me, the ~300 watt in losses is a sunk cost — it is consumed no matter what charging technique you are using.
My understanding of the voltage-current characteristics of solar panels is the voltage is fairly constant through the sunlight hours of the day. What varies is the current. This should allow enough voltage to charge for many hours a day (albeit at a reduced level earlier and later in the day), not just at solar noon.
@bruce The 300W isn’t sunk cost because it’s power, not energy. Imagine the two extreme cases where we could charge with either 300W only, or infinite power. If we have 300W available, there’s no sense even plugging in, since we’ll be sitting there forever while the computers burn the power and the battery sees nothing. The energy lost in kWh is unbounded. The cost to charge is infinite.
On the other hand, if we charge instantly with infinite power, the car consumes 300W*0s to keep the computers on – no energy whatsoever, everything goes into the battery. This case is of course absurd, but illustrates the point that the higher power you can charge with, not only the higher the system efficiency (given that one particular loss) but the lower the absolute cost of the waste.
Now, to your points about voltage: the charge curve of the battery varies in voltage quite dramatically, from 3.2V to 4.2V per cell. And my car has 192 of those in series, so at dead-empty it’s got a pack voltage of 614V, at full 797V. Even if I built a solar array with 797V output (which would be, like, a lot), that array would burn 33% of its output capacity as heat in the panels themselves if I was trying to charge a dead battery at noon. That’d be quite bad for the panels. Nevermind the car’s likely issue with requesting a given charge voltage and getting something entirely different – I’m pretty sure it has safety checks for that sort of thing.
The only practical way to do this at all is with a DC-DC converter that boosts or bucks the array voltage to the current pack voltage, in which case you can pick an array of any size (voltage) you please.
Taking the counterpoint, if I needed a 19-panel array to make this work at all, that’d be 3.8 to 5.7kW depending on the panel size. That’s so far in excess of the 720W J1772 minimum and the 300W computer overhead that you might as well just do the J772 MPPT thing, which would be quite a lot easier (mostly off-the-shelf components) than building the DCDC charge controller anyway.