High voltage MPPT Charge Controllers preferably with flexible battery output?

[JohnGalt1717]

So I've fallen victim to the Growatt scandal... and what I really always wanted was separate of everything from MPPT to inverters to AC/DC charging.

I have 8x605Watt panels in series (and I can't parallel because of major losses going to 4S2P because of shading etc.) They're 48VDC max open voltage with 14A max (theoretical). I have 4 of these series. And plan on adding another 3-4.

Ideally, I'm looking for an MPPT that will scale with ENNOID-BMS: https://www.ennoid.me/bms/gen-1 as I'm slowly building up to a 384VDC pack, but I can live with 48VDC MPPT output for now.

Does anyone know of any 450VDC or higher MPPTs on the market that aren't integrated hybrid units that are any good? Any that do more than 48VDC?

And if not, since I can, would anyone be interested in a design like ENNOID-BMS for doing a scalable MPPT design? If there was interest I might build them and open source the design for DIY. (and I'd license it for companies that want to sell it as a product for profit.)

 

[Mr-Sandman]

I guess I misunderstood your question in the other thread. I thought you was looking for a high voltage PV MPPT stand alone controller, that could push above the standard 48v DC limits. Missed the high voltage battery part.

 

[JohnGalt1717]

I guess I misunderstood your question in the other thread. I thought you was looking for a high voltage PV MPPT stand alone controller, that could push above the standard 48v DC limits. Missed the high voltage battery part.

That's optional for the moment. I'm trying to control the $450/month power bill first, then I'll worry about optimizing later if I have to, since this is green field stuff apparently.

 

[Hedges]

Here are a couple 600V SCC

Schneider Electric 865-1032 Conext MPPT 80 Amp 600VDC Solar Charge Controller

Schneider Electric 865-1032 Conext XW MPPT 80 Amp 600VDC Solar Charge Controller

www.solar-electric.com

Morningstar TS-MPPT-60-600V-48 TriStar 600 Volt 60 Amp MPPT Solar Charge Controller

Morningstar TS-MPPT-60-600V-48 TriStar 600 Volt 60 Amp MPPT Solar Charge Controller

www.solar-electric.com

There is an efficiency loss, 600V (or about 400 Vmp) to 48V, back to 240VAC.
Higher voltage battery (or rail) can help that. Some like SolArk have a high voltage DC bus; MPPT, battery, inverter all talk to it.

I use AC coupling, so 600V (max) PV inverters deliver 240VAC. Loads use that, or battery inverter pulls it down to 48V (lower efficiency round trip.)
There are "400V" battery inverters. But small home units actually contain a few banks of 48V batteries and boost converter.
SMA has commercial battery inverters which would use large high-voltage DC batteries. Need 75kW to 4 MW?

 

[JohnGalt1717]

Both of these designs I'd have to spend $5-6k just on MPPT to get the same as $2400 for Growatts that also have the inverter.

Which, considering these are just 2 half bridge ZVS converters with a small feedback circuit and a $2 micro-controller is a hard pill to swallow. The parts for the whole thing is about $60!

 

[Hedges]

Is that 15kW to 18kW of PV to battery charge capability, from $2400 worth of GroWatt?

To me, a 5kW to 10kW (or larger, 3-phase) GT PV inverter is better way to go.
If much of the power is used during daytime, it never cycles through battery.
In my system, when grid is down, surplus power is discarded. My battery is AGM and undersized, so about 5kW max charge rate.

For Alaskan Noob, GT PV doesn't help so much because goal was to put almost 100% of power into battery, for use during the other 20 hours of little to no sun. It can work but requires twice as much battery inverter just to get double charge current.

 

[JohnGalt1717]

Is that 15kW to 18kW of PV to battery charge capability, from $2400 worth of GroWatt?

To me, a 5kW to 10kW (or larger, 3-phase) GT PV inverter is better way to go.
If much of the power is used during daytime, it never cycles through battery.
In my system, when grid is down, surplus power is discarded. My battery is AGM and undersized, so about 5kW max charge rate.

For Alaskan Noob, GT PV doesn't help so much because goal was to put almost 100% of power into battery, for use during the other 20 hours of little to no sun. It can work but requires twice as much battery inverter just to get double charge current.

Ya, sadly that's why a lot of us went with the growatts. They drive 4500W per inverter. You can buy 4 of them for $2400 give or take and they handle 450V open circuit on each of those.

Ideally this is all done in a DC bus at 340V internally so that you don't have any of the losses and can power everything directly on solar and only the excess goes into the batteries.

 

[Hedges]

This would be a nice way to go, but 5x to 10x the power homeowners need.
I have no idea what the batteries cost.

https://files.sma.de/downloads/STPS60-DS-en-30.pdf

https://files.sma.de/downloads/STP50-41-DS-en-15.pdf

 

[JohnGalt1717]

This would be a nice way to go, but 5x to 10x the power homeowners need.
I have no idea what the batteries cost.

https://files.sma.de/downloads/STPS60-DS-en-30.pdf

https://files.sma.de/downloads/STP50-41-DS-en-15.pdf

Definitely on the right track. The thing about all of this is, that this stuff doesn't have to be expensive and can easily scale on battery voltage. Both of those units are 3 phase, and require massive battery packs. There's just no reason for these fixed battery voltage designs.

If your BMS is communicating properly, the MPPT and inverters can know exactly what the voltage of the battery is, and the inverters have to use 340VDC internally to power 240VAC output so in a simple half bridge boost converter with $1 microcontroller you need 6 fets (About $100 right now) to drive 50A output from any battery from 12-300VDC assuming the battery has enough amps to do it.

And with 8 fets you can use any size up to 600VDC battery @ 50A. Double those and you double the output effectively.

So I could drive 200A 240VAC from about $350 in parts with a board from PCBWay. I've been trying to avoid building it, but damn, it's getting hard to justify not just designing all of this stuff myself.

 

[Zwy]

So I've fallen victim to the Growatt scandal... and what I really always wanted was separate of everything from MPPT to inverters to AC/DC charging.

I have 8x605Watt panels in series (and I can't parallel because of major losses going to 4S2P because of shading etc.) They're 48VDC max open voltage with 14A max (theoretical). I have 4 of these series. And plan on adding another 3-4.

Ideally, I'm looking for an MPPT that will scale with ENNOID-BMS: https://www.ennoid.me/bms/gen-1 as I'm slowly building up to a 384VDC pack, but I can live with 48VDC MPPT output for now.

Does anyone know of any 450VDC or higher MPPTs on the market that aren't integrated hybrid units that are any good? Any that do more than 48VDC?

And if not, since I can, would anyone be interested in a design like ENNOID-BMS for doing a scalable MPPT design? If there was interest I might build them and open source the design for DIY. (and I'd license it for companies that want to sell it as a product for profit.)

Just remember everything you install has to be rated for the much higher DC voltage, including breakers. That gets expensive and not easily found, that is why the prices are higher, low production runs. It can be done, but I have to wonder at what cost.

If you have shading, putting those panels in series will lead to low PV output. You would be better served by more parallel strings.

You could just buy the 5000ES and use those as charge controllers, but really you should just look at the LVX6048 if you desire the 450V PV arrangement. https://watts247.com/product/lvx-6048-hybrid-solar-inverter-split-phase-120v-240v/

More inverter output until you build your high voltage system, plus larger capacity on the MPPT than the 5000ES. Yes, it costs more but autotransformer is built in, no rigging for neutral or ground.

 

[Hedges]

You could probably build a multi-kW inverter for $350 in parts (cost in China may be 1/10th what we spend on prototype quantities here).
But you'll have something about as robust and capable as a $350 no-name inverter.

It isn't likely to do what a multi-$thousand SMA or Schneider inverter can do.

Having battery voltage higher than peak AC voltage does make economy inverter more feasible. I paid $200 15 years ago for a Hitachi VFD that runs a 3-phase motor. But that doesn't need inductors beyond EMI filtering.

Sunny Boy Storage for $2500 is a 6kW 240V inverter (9kW surge). Unfortunately, the "400V" battery it uses costs as much or more, and is really a boost converter. Wired straight to a high voltage battery (and faking BMS communication), could be the ticket. BOM cost is probably fairly low.

 

[JohnGalt1717]

Just remember everything you install has to be rated for the much higher DC voltage, including breakers. That gets expensive and not easily found, that is why the prices are higher, low production runs. It can be done, but I have to wonder at what cost.

If you have shading, putting those panels in series will lead to low PV output. You would be better served by more parallel strings.

You could just buy the 5000ES and use those as charge controllers, but really you should just look at the LVX6048 if you desire the 450V PV arrangement. https://watts247.com/product/lvx-6048-hybrid-solar-inverter-split-phase-120v-240v/

More inverter output until you build your high voltage system, plus larger capacity on the MPPT than the 5000ES. Yes, it costs more but autotransformer is built in, no rigging for neutral or ground.

I'm thinking about it.

I just wish this stuff was separate. These integrated units are great in theory but a bad idea in the real world....

As for the breakers, 600VDC breakers are readily available now for about the same as AC breakers (especially the required combo breakers now in the US). And Gan/SiCFETs can easily switch 600VDC themselves (and will fail open) for all internal stuff, you only need breakers for circuits that are supplying devices outside the box.

 

[JohnGalt1717]

You could probably build a multi-kW inverter for $350 in parts (cost in China may be 1/10th what we spend on prototype quantities here).
But you'll have something about as robust and capable as a $350 no-name inverter.

It isn't likely to do what a multi-$thousand SMA or Schneider inverter can do.

Having battery voltage higher than peak AC voltage does make economy inverter more feasible. I paid $200 15 years ago for a Hitachi VFD that runs a 3-phase motor. But that doesn't need inductors beyond EMI filtering.

Sunny Boy Storage for $2500 is a 6kW 240V inverter (9kW surge). Unfortunately, the "400V" battery it uses costs as much or more, and is really a boost converter. Wired straight to a high voltage battery (and faking BMS communication), could be the ticket. BOM cost is probably fairly low.

12-400V LiFePo4 batteries puts everything within the bottom 80% of GanFETs. As long as you spec the PCB trace thicknesses and distances properly, you'll have a very robust design. But we're talking home stuff here, not 75kVA, so it doesn't have to be that level of design.

But the dirty secret, is that when you look at the boards, they're all doing exactly the same thing. A transformerless inverter is nothing more than:

1. Boost from battery to DC output voltage (AC output * 1.41 for RMS). So if 2x120V legs at 180 degree phase, then you're boosting to 170VDC at 2x amperage. If you're doing EU 240 then 340VDC. This is 2 caps, an inductor, and 2 fets.
2. Run through H-Bridge using SPWM which is a very straightforward technique with a micro controller. This is 3 caps, 4 fets and an inductor per leg.
3. Control it all with a single micro controller. STM32 most of the time but ESP32 will also work. In Either case $1. There's a few resistors to do voltage dividers to monitor the output and adjust as necessary with feedback, but those are pennies.

That's it. Literally, MPPT controllers are more complex because they're 2 full bridges (or a full bridge and a half bridge) so 8 fets, 3 caps, and 2 inductors + a few resistors and a microcontroller.

And the AC/DC UL rated power supply is by far the most complex because high power (> 400W) power supplies have to cause resonance in the transformer (which is required for UL) for efficiency, and the AC side has to be completely galvanically isolated from the DC side which means optocouplers or X/Y rated capacitors.

This is first year EE stuff with the exception of the AC/DC converter and even that I don't think is beyond your average first year EE person.

If you take apart a Schnieder Electric or similar, you'll find exactly the above. There's NOTHING special to them at all. They aren't some dark magic. The only difference is that they'll use Ti, Analog Devices or similar Taiwanese fets etc. whereas the Growatt uses Chinese knockoffs. And all of this stuff became easily doable because of high speed microcontrollers and GanFETs that don't have small miller effect switching losses

 

[Zwy]

I'm thinking about it.

I just wish this stuff was separate. These integrated units are great in theory but a bad idea in the real world....

You can replace boards and other parts, so these are serviceable. Anything one off kills so much time, I've done plenty of custom fab with metalworking and other projects before simply because no one made what I wanted. The time spent sourcing items alone is such a time killer. Look at how much time you have spent already researching a solar project.

As for the breakers, 600VDC breakers are readily available now for about the same as AC breakers (especially the required combo breakers now in the US). And Gan/SiCFETs can easily switch 600VDC themselves (and will fail open) for all internal stuff, you only need breakers for circuits that are supplying devices outside the box.

Beware of some of the cheaper 600VDC breakers out there, the arc chute isn't there. Higher DC volts, you want the good stuff and it will be pricey.

I was surprised they were running up to 30a on the GaN in the video below. and claim 100a. I have to wonder longevity before FET failure running higher amps. Heat kills and long term it is failure prone.

I do have to ask why such a large system?

 

[Zwy]

You could do like Ian did here with four 5000ES units feeding a 12Kw 12000T unit. It's at the 8:00 min mark where he mentions it in print on the screen.

I looked for a video where he showcases his house setup but didn't find one.

 

[JohnGalt1717]

You can replace boards and other parts, so these are serviceable. Anything one off kills so much time, I've done plenty of custom fab with metalworking and other projects before simply because no one made what I wanted. The time spent sourcing items alone is such a time killer. Look at how much time you have spent already researching a solar project.


Beware of some of the cheaper 600VDC breakers out there, the arc chute isn't there. Higher DC volts, you want the good stuff and it will be pricey.

I was surprised they were running up to 30a on the GaN in the video below. and claim 100a. I have to wonder longevity before FET failure running higher amps. Heat kills and long term it is failure prone.

I do have to ask why such a large system?

GanFets are about 6.5x more efficient than standard Silicon Fets and use about half the quisent current to charge the gate than SicFets. They also have a miller plataeu of about 1/3rd that of others so they're in resistence mode for a MUCH shorter period of time between on and off and visa-versa.

It's straight forward to find 80a at 125C GanFets. And that's constant current. At pulsed they're rated for 150-200A continuous pulsed as you'll find in DC/DC converters and SPWM inverters.

A standard offgrid and carbon free house:

1. 30A @ 240V Geo/Air source heat pump: 7200W + 10A @ 120V for circulating pump + 7A @ 120V for HRV = 9240W
2. 20A @ 240V Geo/Air source hot water heating heat pump: 4800W (resistence would be 7200W)
3. 60A @ 240V Induction Cooktop: 14,400W
4. 13A @ 120V Microwave: 1500W
5. 30A @ 240V Oven: 7200W
6. 30A @ 240V Dryer: 7200W (20A heat pump model possible for small savings)
7. 13A @ 120V Dish washer: 1500W
8. Various 120V stuff: 5000W


#1, #2 come on by demand. You have to assume they're on in your calculations at all times because you can go way over easily. So you start with 15,000W.

Then using standard derating at 40% of total, you're around another 15-18,000W of power.

32kWh is about what the average American house uses and that's with propane/nat gas/oil for heating, so this is entirely in line with systems that are burning fossil fuels.

Thus, I have to have 30ish kW of inverter unless I can get some of those on to DC only, which is doable with the first 2, which halves the size of the inverters I need effectively which is major $$$$. Unfortunately, my geo thermal beat the DC inverter tech by 2 years, so I'm stuck with an AC unit and separate hot water heating for the moment, so I have to size the inverter accordingly because you can't hack a non-inverting heat pump to run off of DC power.

And in my house I also have a garage, with another air source heat pump and hot water heater, a welder and an air compressor. I can convert all of that eventually to DC because they're all switched mode DC devices internally, but initially before I take them apart and void warrantees, I still need to cover the inverter for the moment.

And that doesn't mention the Tesla that is about to be parked in the garage at 80A AC (which I'm going to probably build my own DC level 3 charger for to run directly off the batteries) The good news is that the only competition an AC charger has (other than being 15-25% inefficient when being powered from a battery bank) is the hot water heater and the heat pump because it charges over night when usage otherwise is low.

I.e. you need 30kW to power a standard home that is off grid and not using fossil fuels, and that's minimum.

I already have a 61kW battery setup, that I'll be expanding to 100kW once the rest of this settles down.

 

[Zwy]

You're like the super power user, the power company must love you. ?

I myself prefer to know what my system required before install. I spent $99 on one of these https://www.amazon.com/gp/product/B07FN3X9GX/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1

I always thought my house used around 20 Kwh per day, the shop is with the same meter. What I've found is 18 Kwh average per day, with 33 Kwh max for a day, Christmas dinner the wife really sucked the power.

The highest demand at any one time is 5Kw per leg so a total of 10Kw.

Heat is LP gas as is hot water, clothes dryer is electric, well pump is biggest Kw draw.

I think spending the money for a monitor (several makes out there) and determine just how many Kw you will need at any one time might be a wise investment. Spending money for capacity that will never be utilized is not economical.

 

[JohnGalt1717]

You're like the super power user, the power company must love you. ?

I myself prefer to know what my system required before install. I spent $99 on one of these https://www.amazon.com/gp/product/B07FN3X9GX/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1

I always thought my house used around 20 Kwh per day, the shop is with the same meter. What I've found is 18 Kwh average per day, with 33 Kwh max for a day, Christmas dinner the wife really sucked the power.

The highest demand at any one time is 5Kw per leg so a total of 10Kw.

Heat is LP gas as is hot water, clothes dryer is electric, well pump is biggest Kw draw.

I think spending the money for a monitor (several makes out there) and determine just how many Kw you will need at any one time might be a wise investment. Spending money for capacity that will never be utilized is not economical.

I know what my usage is because I have hung amp meters on both legs. I know that I have to just assume that the hot water heater and geo thermal is on, and then there rest is a risk analysis for which my metrics (my meters plug in for recording data) show 40% is safe.

I have 3x 200a panels, and 2x 200a meters. At peak I use about 300A with the induction cook top going, a load of drying in, the heat pump on and the hot water heater going. (and I didn't mention the well pump you you have to assume is on in your load calcs, which is one of my smaller 240V draws)

My electric bill last month because rates went up and I was crypto mining was $450. At the cost of this system, I'll be able to pay for the entire solar setup after federal tax credits in under 2 years. (When I built the house 6 years ago that bill was $240 / month!)

My current plan for the moment is going to be to use 4 growatts for the 240V stuff and my garage with an auto-transformer because there isn't much out there other than light bulbs that would blow. In the house I'm just going to shift the 3 240V breakers that are on the one panel to the other panel and use the growatts to power those. And then I'll use 1 or 2 of these: https://www.ebay.com/itm/LVX-6000w-...ar-inverter-PV-450vdc-100A-mppt-/143999827478 for the 120V stuff separately fed is my plan. Means more wiring though, so I might sell the Growatts entirely and just buy those or build my own since it doesn't seem like anyone is building separate equipment sadly.

PS: My heat pump costs about half as much to run per month as your LP heat and hot water and in the summer my hot water is free from AC and I get AC

 

[Zwy]

Air source heat pump wouldn't work too good here, last night it was -20F. Again tonight and next night. Right now it's 0F at 2 pm.

Air conditioning is another matter, air source mini split would work fine here, high humidity in summer.

Ground source here maybe but can't see the payback with LP gas price here. I buy in summer under $1.50 and don't use 800 gal per winter. At over $35K for a ground source heat pump installed, it would take $5/gallon LP gas to break even at 10 years, $2.50 for 20 years. LP gas goes in spurts for high prices, the cure for high prices is usually high prices as production increases and demand keeps a lid on it.

I'm thinking more along the lines of a solar dump load combined with hydronic heat instead for kitchen and living room. I have it in the shop, just love it. It would require more PV but have plenty of room for more ground mount. I would consider it supplemental and not total heating usage.

I can always burn wood. Grew up with a house entirely heated by wood and will always have wood available. Takes work though, I can find more productive uses of my time.

If you're really intending on 4 5000ES's, might just be better to install 4 LV6548's such as DMI did.

 

[JohnGalt1717]

Air source heat pump wouldn't work too good here, last night it was -20F. Again tonight and next night. Right now it's 0F at 2 pm.

Air conditioning is another matter, air source mini split would work fine here, high humidity in summer.

Ground source here maybe but can't see the payback with LP gas price here. I buy in summer under $1.50 and don't use 800 gal per winter. At over $35K for a ground source heat pump installed, it would take $5/gallon LP gas to break even at 10 years, $2.50 for 20 years. LP gas goes in spurts for high prices, the cure for high prices is usually high prices as production increases and demand keeps a lid on it.

I'm thinking more along the lines of a solar dump load combined with hydronic heat instead for kitchen and living room. I have it in the shop, just love it. It would require more PV but have plenty of room for more ground mount. I would consider it supplemental and not total heating usage.

I can always burn wood. Grew up with a house entirely heated by wood and will always have wood available. Takes work though, I can find more productive uses of my time.

If you're really intending on 4 5000ES's, might just be better to install 4 LV6548's such as DMI did.

So Air source heat pumps are good down to -13F without having to defrost most of the time now and are about equivalent efficiency at -3F. Above that they're more efficient (and cheaper) than propane.

And if you use a 2 phase thermostat with an external temperature sensor like Venmar makes, you can easily have it switch to Propane for heating whenever it drops below your outdoor setpoint, and then you get AC for 8 of 12 months and propane as backup.

And, you can get air source heat pump hot water heaters too which are super efficient and they use CO2 as the refridgerant and will go down to -40F without trouble, or you can hack your own using a Mr. Cool and a $100 heat transfer plate like you'd see in a geo thermal super deheater and you're good. (which is what I did)

We have wood backup as well, but it's sure nice having a house that is 72F and sunny year round at 50% humidity.