Which ones?
DC Direct EV charger - DIY Build
- Thread starter jakepusateri
- Start date
jakepusateri
100K Ohm Resistor
Got any links? The only ones I’ve heard of are SolarEdge and Entelligent, neither of which are out yet.
jakepusateri
100K Ohm Resistor
Thanks for the link. I also know libresolar has some forums but they’re a little dead.
You're suggesting Vmp = 370V. With a gain of 1.5, you're up there to around 400V. But with a charge current of only 7A? The OP was thinking about 500V, 30A solar. That would be around 500/400 * 30A = 37 A battery charging current. What am I missing? How do you arrive at 52 ohm boost load?
jakepusateri
100K Ohm Resistor
What’s the advantage here? All the components would have to be rated for the max voltage, and presumably the amps would split equally with similar length traces, and they’d need to be in sync so use the same pwm signal. So now I have 6x the components. Is it a cost thing, as in I’m outside normal cheap components range and going 1/6 the power means 1/20th the price?
More granularity for panel string MPPT. You can do 8 panels in series with their own boost controller independent of another strings. Also as a beginner it may be easier to build smaller converter first before dealing with higher power. Depends what you want to do. If you are planning to build 100kW+ DC charger then going with higher voltage panel strings makes sense.
jakepusateri
100K Ohm Resistor
Agree about starting small, though it'd be a bit of a detour to do so. I have 12S, 12S, 11S strings on my roof currently, and a spare panel in the basement. I could start with 1 panel, but that's only 31Vmp at 13a, so we'd need a totally different transformer-based design to get to HV battery voltage.
The specs in the first post were guesses at design maxes to inform the process, and based on what other high voltage mppts (like on the 18kPV) are doing, not real requirement per se. Rewiring my existing panels would be very annoying at this point, so if I'm to personally test it I'll need to design to fit what I have.
What I could do though is use the design specced for full voltage, and only run the 31V through it, and boost to charge the 48v battery I have. Should be similar ratios, but less dangerous and lower power. Once that's proven out, move on to the car interface and full strength.
jakepusateri
100K Ohm Resistor
Yep that was the plan. Not too much more work, and makes it usable for others with different setups.
I like boost converter because it has output diode that points in the right direction to prevent current from flowing back from the battery in any circumstance which is great because you don't want battery to backfeed the panels at night. Looks like there is a more complex buck-boost circuit that keeps that output diode.
nickskethisnikske
New Member
- Joined
- Aug 27, 2020
- Messages
- 31
Ok, very tempted to help you. My bucket list is endless but this is for sure on it and also an inverter coupled to an EV.
Be aware of the safety risks involved, anyone that once experimented with 2 solar panels in series to make some nice arcs knows what I'm talking about. An EV battery has a lot more potential...
So how far are you with the idea? I like the buck/boost suggestion, do you need galvanic isolation? Nowadays there are little gti's with galvanic isolation either and galvanic isolation is not necessarily more safer. Since you have 3 strings I would aim for 3 modules of 5kW that can work in parallel, it won't hold you back to build them in the same enclosure.
Sic mosfets are nice but still expensive and used in this kind of applications were we work with high frequencies, although the price difference with IGBT's is getting smaller I would start with igbt's and work with a frequency of 20-50khz (on the edge to use igbt's). You might want to start with a "proof of concept" before going for higher efficiency. Also, higher frequencies will need a more careful pcb layout.
Are there things you already have laying around, because this can quickly get very expensive to source semiconductors, inductors (sendust, ferite, iron powder cores,..) capacitors, lems,... Broken gti's could be a good starting point, but it is advisable to use new semiconductors...
Allthough it is wisely to start with something like 48V to get a certain feeling, it surely is another world once you get in that 400V range. The buck and boost inductors need to be calculated for your working point, unless you start playing with the frequency. If you build in a current limit, you can start experimenting on low power levels.
Is there a place were I can get more information about "how to acces the battery itself"? I supose there is a a relay and other stuff before you can realy acces the potential of the battery and is there a difference between drawing current or pushing current?
Be aware of the safety risks involved, anyone that once experimented with 2 solar panels in series to make some nice arcs knows what I'm talking about. An EV battery has a lot more potential...
So how far are you with the idea? I like the buck/boost suggestion, do you need galvanic isolation? Nowadays there are little gti's with galvanic isolation either and galvanic isolation is not necessarily more safer. Since you have 3 strings I would aim for 3 modules of 5kW that can work in parallel, it won't hold you back to build them in the same enclosure.
Sic mosfets are nice but still expensive and used in this kind of applications were we work with high frequencies, although the price difference with IGBT's is getting smaller I would start with igbt's and work with a frequency of 20-50khz (on the edge to use igbt's). You might want to start with a "proof of concept" before going for higher efficiency. Also, higher frequencies will need a more careful pcb layout.
Are there things you already have laying around, because this can quickly get very expensive to source semiconductors, inductors (sendust, ferite, iron powder cores,..) capacitors, lems,... Broken gti's could be a good starting point, but it is advisable to use new semiconductors...
Allthough it is wisely to start with something like 48V to get a certain feeling, it surely is another world once you get in that 400V range. The buck and boost inductors need to be calculated for your working point, unless you start playing with the frequency. If you build in a current limit, you can start experimenting on low power levels.
Is there a place were I can get more information about "how to acces the battery itself"? I supose there is a a relay and other stuff before you can realy acces the potential of the battery and is there a difference between drawing current or pushing current?
Last edited:
jakepusateri
100K Ohm Resistor
For accessing the battery, the NACS spec has a simplified view of the charging circuit on page 4.
https://digitalassets.tesla.com/tes...g-Standard-Technical-Specification-TS-0023666
As for how far I am, it's mostly schematic phase. I'm taking the schematic of https://libre.solar/hardware/mppt-2420-hc.html as a starting point and scaling up where I need, but it's going to be a completely new pcb. I was probably going to start with a stm32 dev board and prototype the rest on a perf board. Once I get it built, I can take that hardware and run it lower voltage and power and turn it up once it doesn't melt.
On mostfet cost, I'm not sure what people are talking about. I don't have anything lying around, but https://www.lcsc.com/product-detail/MOSFETs_Alpha-Omega-Semicon-AOTF190A60L_C454182.html seems to be a 600V 20A mosfet for $1.37. I need like 2 of them. Is there some spec sheet detail I'm missing or will these be fine? Nothing else seemed that expensive either on a cursory glance of the BOM of that mppt.
https://digitalassets.tesla.com/tes...g-Standard-Technical-Specification-TS-0023666
As for how far I am, it's mostly schematic phase. I'm taking the schematic of https://libre.solar/hardware/mppt-2420-hc.html as a starting point and scaling up where I need, but it's going to be a completely new pcb. I was probably going to start with a stm32 dev board and prototype the rest on a perf board. Once I get it built, I can take that hardware and run it lower voltage and power and turn it up once it doesn't melt.
On mostfet cost, I'm not sure what people are talking about. I don't have anything lying around, but https://www.lcsc.com/product-detail/MOSFETs_Alpha-Omega-Semicon-AOTF190A60L_C454182.html seems to be a 600V 20A mosfet for $1.37. I need like 2 of them. Is there some spec sheet detail I'm missing or will these be fine? Nothing else seemed that expensive either on a cursory glance of the BOM of that mppt.
Watch out for body diode reverse recovery. Silicon MOSFETs rated above 200V have terrible diode reverse recovery specs. You will need to use SiC MOSFETs or IGBTs to reduce it. Make sure you understand issue explained in this video before you get into high voltage stuff.
jakepusateri
100K Ohm Resistor
Wow, the differences are pretty dramatic for Qrr. The test conditions don't match exactly, but 6.8uC vs 73nC for https://www.mouser.com/ProductDetail/Qorvo/UF3C065080K3S?qs=B6kkDfuK7/Aa0CxrobLSgw== is almost 100x difference.
GaN, which doesn't even have a body diode also seems similarly priced at around $8 (https://www.mouser.com/ProductDetail/Transphorm/TP65H070G4PS?qs=ulEaXIWI0c94%2BPmd6%2Bn24w==). According to https://www.ti.com/lit/an/slyt801/slyt801.pdf, GaN and SiC are both good choices at my design specs, but GaN seem to be more efficient.
GaN, which doesn't even have a body diode also seems similarly priced at around $8 (https://www.mouser.com/ProductDetail/Transphorm/TP65H070G4PS?qs=ulEaXIWI0c94%2BPmd6%2Bn24w==). According to https://www.ti.com/lit/an/slyt801/slyt801.pdf, GaN and SiC are both good choices at my design specs, but GaN seem to be more efficient.
nickskethisnikske
New Member
- Joined
- Aug 27, 2020
- Messages
- 31
Yes, you can say individual components are cheap but it adds up eventually, you probably want to have some spares as well or different kinds of it.
Do you have some basic testing equipment as well? You will need an oscilloscope, power supplies, volt/current meters, etc...
That are nice links you provided!
I don't know if I'm allowed to put a link of an other forum, but on the backshed forum we have build some basic mppt controllers, you can search for "150V 45A mppt - roll your own" thread. It has a "basic" but bullet proof pcb (buck converter), it's also really nice to work on during development phase, you can make your own controller board for it that you can keep external, it is well documented and there is code written for an arduino nano. It will take the amount of power you want but the pcb might need some adjustments to make it safe for 400V.
Do you have some basic testing equipment as well? You will need an oscilloscope, power supplies, volt/current meters, etc...
That are nice links you provided!
I don't know if I'm allowed to put a link of an other forum, but on the backshed forum we have build some basic mppt controllers, you can search for "150V 45A mppt - roll your own" thread. It has a "basic" but bullet proof pcb (buck converter), it's also really nice to work on during development phase, you can make your own controller board for it that you can keep external, it is well documented and there is code written for an arduino nano. It will take the amount of power you want but the pcb might need some adjustments to make it safe for 400V.
Last edited:
400bird
Solar Wizard
Maybe I missed it, but why discuss the boost converter? You could run 10s or whatever it takes to get your panel voltage to 400 volts at VMP and skip the boost section. Evening going higher means you can buck down to pack voltage.
You'd loose some output, but you could skip the MPPT all together (added back into the design later) and just connect the panels straight to the pack, once comms are established and the charge contractors are closed. Then all you need is a dual pole relay/contactor.
Connecting a 400 VDC array to a (dead)300 volt 60kWh pack, tada! Now the array is at 300 volts.
I assumed that communication/getting through the handshake with the car was the difficult part, tackling that first would be my starting point. Once that is figured out you can add in the buck boost MPPT to increase solar harvest/efficiency.
You'd loose some output, but you could skip the MPPT all together (added back into the design later) and just connect the panels straight to the pack, once comms are established and the charge contractors are closed. Then all you need is a dual pole relay/contactor.
Connecting a 400 VDC array to a (dead)300 volt 60kWh pack, tada! Now the array is at 300 volts.
I assumed that communication/getting through the handshake with the car was the difficult part, tackling that first would be my starting point. Once that is figured out you can add in the buck boost MPPT to increase solar harvest/efficiency.
fromport
Solar Addict
I also think that is the hard part. I was involved with a Chademo diy charger and it really meant sending information multiple times per second to keep the connection open. CCS uses over PLC communication and is imo even harder to do.
Article with dated info but a nice overview (imo)
EV DC Fast Charging standards – CHAdeMO, CCS, SAE Combo, Tesla Supercharger, etc
greentransportation.info
