Size matters..Growatt/LiFePO4 system build

[apctjb]

Diving in on building an off grid PV system for a vacation home in Baja. Building in a remote location so simplicity, redundancy, ability to monitor remotely, easy to " swap out"should there be problems are important criteria. Size/weight is also a consideration as need to drive the components down.

After reading lots of post on this forum decided to pursue an "all in one" inverter design with a LiFePO4 battery bank. Ordered 32 EVE 280ah cells on Alibaba from Amy at Xuba. Placed order, paid by wire to avoid the credit card hassles others have posted about and 3 days later the cells were shipped and on there way.

For the inverter decided to go with 2 Growatt SPF3000TL LVM 48 (3000W-48V) inverters. I had also considered the MPP Solar inverters but decided to go with Growatt because of brand (Growatt; top 10 PV inverter manufacturer), and with Growatt the parallel card/cables were included as was a wifi dongle and remote monitoring software. I purchased the inverters from Ian at https://watts247.com . Ian answered a ton of my questions, had the inverters in stock, shipped the same day and I had them 3 days later. Could not be easier.

2 x SPF3000 in parallel allows me to have up to 9kw of PV input (160A) and 6000Wac output at either 120 or 120/240Vac. I don't see needing 6000W output capacity often so intend to operate just one of the inverters , keeping the other in backup most of the time. I intend to have ~5.5kW peak of solar and will split the array output with 1/2 going to each inverter (My understanding is the inverters built in MPPT charger operates even if the inverter is turned off).

Below is a wiring schematic of how I am intending on wiring things up, and rendering of how I am thinking I will layout the components. I appreciate any comments/suggestions others may have to offer. This is my first time working with an "all in one" inverter or LiFePO4 battery. I still need to decide on which BMS to purchase, and welcome recommendations...

 

[Frank in Thailand]

Looking good!!

For BMS, at first I thought I was limited to use common port BMS types, as the MPPT charge and inverter discharge use the same wires. (Hybrid)

Later it hit me:
The BMS types that use contactors/relays can be used as well, and maybe even better!

To stop overcharge, you can set a relay/contactor at the solar cables going into the hybrid inverter.
No connection to the panels is no charge.

To stop too low discharge, either use the relay pulse to turn off the inverter (power button) or use relay/contactor to disconnect AC. (For me 230v)

For the rest, BMS doesn't do much.
It can do some balancing, depending on the type mili-amperes or real work.

I'm building the DIYBMS, it can burn about 10A top Balance per cell and have WiFi

(DIYBMS has a PCB per cell ($3,-), and one controller ($10,-) )

I do have the 1A active balancer from IC GOGOGO store.
I like it, it gives accurate cell voltage information via BT.
Most BMS balance only at top, this one continuous keeps the cells more close together.

There are many brand names and types.
After using the types that uses mosfets to turn on/off, I now see that they are NOT a good solution in the higher Amperage setups.
Below 50A, sure.
Good, simple, safe, cheap.
Higher Amperage, I can't advise.

 

[apctjb]

@forst; thanks for the recommendations. I have been struggling deciding on BMS; wanting to keep it simple but also able to monitor remotely.

Nice idea having the BMS turn off the inverter to prevent over discharge. As I have a main battery circuit breaker I was also considering using CB with remote trip capability. The BMS would signal the CB to open and the battery would be disconnected. No tare losses like a NO contactor, but circuit breaker would require a manual reset once tripped.

Interested in learning more about DIYBMS. Where can I get more info??

 

[Craig]

Looking good!!

For BMS, at first I thought I was limited to use common port BMS types, as the MPPT charge and inverter discharge use the same wires. (Hybrid)

Later it hit me:
The BMS types that use contactors/relays can be used as well, and maybe even better!

To stop overcharge, you can set a relay/contactor at the solar cables going into the hybrid inverter.
No connection to the panels is no charge.

To stop too low discharge, either use the relay pulse to turn off the inverter (power button) or use relay/contactor to disconnect AC. (For me 230v)

For the rest, BMS doesn't do much.
It can do some balancing, depending on the type mili-amperes or real work.

I'm building the DIYBMS, it can burn about 10A top Balance per cell and have WiFi

(DIYBMS has a PCB per cell ($3,-), and one controller ($10,-) )

I do have the 1A active balancer from IC GOGOGO store.
I like it, it gives accurate cell voltage information via BT.
Most BMS balance only at top, this one continuous keeps the cells more close together.

There are many brand names and types.
After using the types that uses mosfets to turn on/off, I now see that they are NOT a good solution in the higher Amperage setups.
Below 50A, sure.
Good, simple, safe, cheap.
Higher Amperage, I can't advise.

Glad you came around @fhorst the use of relays before SCC and after the inverter is exactly what I have been recommending for months now. It also has the advantage of cutting in the areas where the amps are the least.

 

[Craig]

There is nothing wrong with a DIY BMS just make sure it is bullet proof and test every scenario to make sure it doesn't fail when you are 1000 miles away. @apctjb where in Baja? I used to have a house in Loreto in another life.

 

[apctjb]

@apctjb where in Baja?

El Jalito; on the sea of cortez south east of La Paz and it is a 1000 miles away!. Loreto is gorgeous....

 

[Frank in Thailand]

There is nothing wrong with a DIY BMS just make sure it is bullet proof and test every scenario to make sure it doesn't fail when you are 1000 miles away. @apctjb where in Baja? I used to have a house in Loreto in another life.

That's exactly the reason why I can't use the Smart Daly BMS.
I need to pre-charge my 3 hybrid inverter to get it going.
If it does act, disconnect battery, I need to do the whole dance again..

If I'm not at home, the house and farm go without power = no water.
Crops don't like to be thirsty at +35 Celcius.

DIY BMS will be more reliable but takes a few days / weeks to figure out how to get all the things to work as I like.

One of the things is starting ventilation for the batteries.
That's a function I really like

It did take me a while to look beyond the Inverter - battery connection, as mostly that's where the BMS connection/disconnect is placed.

Separate port have its advantages, useless in hybrid inverter setup at that location.

Placing the contactors / relays before and after the hybrid inverter is "extension" of "separate port", no longer the direct cables but indirect.

Funny thing is that I bought long time ago 3* 50A contactor/relays for the solar and 400A
for the Inverters to make "emergency shutdown" (the big red button)...

Now it's used also for the BMS.
(And connected to my smoke /fire detector)

I won't be using the 400A contactor for this, but 63A AC contactor.


Size indeed matters.

It also does require different setup from "normal" as most products aren't build to handle higher Amperage.

My 43* solar generates max 14kw at peak hours, the 37.5 kWh (736Ah) potentially can be charged at +/- 250A

That is an amount of Ampere what most BMS won't handle.

Chargery and Electrodaucus can.
Probably a few more I haven't heard of.
Build safe!!

High Amperage can give big problems, if one connection isn't tight..

 

[RCinFLA]

Contactor coils take a lot of power when active.

If you have the know how, you can take a BMS with contactor, throw the contactor away, and make your own series switch with MOSFET's controlled by BMS contactor control voltage. At least you can be sure you have a good BMS switch that can handle your current without burning up.

A good way to solve the BMS switch re-engagement surge is to put an additional high Rds_on MOSFET's pair in parallel with high current MOSFET's that fire up a couple hundred microseconds before kicking on the main low R_on switching MOSFET's.

The lowest cost, lowest Rds_on 75v MOSFET's I have found are irfp4368. From Ali I have seen them for around $1 a piece.

 

[BiduleOhm]

Wow, be very careful about ordering components on aliexpress, especially power transistors, they are fake 90 % of the time (and it's not a legend, lots of people have decapped some components to see if they were fake or not). I love aliexpress for a lot of stuff but for electronics I always use reputable dealers like Mouser, Farnell, Digikey, ...

And I can tell you 1 $ for a 75 V 1.5 mOhm mosfet (plus non-SMD) is too good to be true, it's more like 3 $ or more in 10+ quantities.

Also when playing with mosfets and high power it's not that simple, you can have major problems really fast just because of a wrong value resistor or because you forgot about things like the inductive voltage spike at turn-off. Please be careful if you design your own SSR and test it extensively before using it

 

[apctjb]

Previously floated the idea of doing away with a high current contactor and using a shunt trip circuit breaker to serve two functions; protect circuit from overcurrent and disconnect the battery upon command from the BMS (just as a contractor or SSR would do). The drawing below illustrates the idea. The first battery bank is shown with a high current NO contactor which disconnects the battery, the second battery bank with a small NC relay that when the BMS signals (or fails) results in battery voltage applied to the shunt trip mechanism of the circuit breaker, opening the breaker and disconnecting the battery. The advantage of the shunt trip breaker it has no standby power loss. Disadvantage is once tripped it has to be reset manually.

So could this work??

 

[Frank in Thailand]

Contactor coils take a lot of power when active.

If you have the know how, you can take a BMS with contactor, throw the contactor away, and make your own series switch with MOSFET's controlled by BMS contactor control voltage. At least you can be sure you have a good BMS switch that can handle your current without burning up.

A good way to solve the BMS switch re-engagement surge is to put an additional high Rds_on MOSFET's pair in parallel with high current MOSFET's that fire up a couple hundred microseconds before kicking on the main low R_on switching MOSFET's.

The lowest cost, lowest Rds_on 75v MOSFET's I have found are irfp4368. From Ali I have seen them for around $1 a piece.

LOL

you are joking, yes??

Contactors do NOT us a lot of power.

That is just PLAIN BS!

500A medium quality (Jennings) DC contactor at 24V use 1.1A to make contact (that is a few milliseconds 26.4 watt)
After contact it uses 0.16A, or 3.85 Watt per hour.

You can't possibly say that is "a lot of power'' !!

My cheap Aliexpress 400A use 6 watt in contact.
More, yes, still not a lot of power.
I don't have a full test bench to be able to measure the wattage used during the millisecond of contact

Spec sheet of a contactor (=proof) :
(yes, this is a direct invitation to you, to proof your BS statement "a lot of power'')



FIY:
A "BMS with contactor" doesn't come with contactor, it is something you order separate.
Clearly you have no idea what you are talking about!!

 

[Frank in Thailand]

Wow, be very careful about ordering components on aliexpress, especially power transistors, they are fake 90 % of the time (and it's not a legend, lots of people have decapped some components to see if they were fake or not). I love aliexpress for a lot of stuff but for electronics I always use reputable dealers like Mouser, Farnell, Digikey, ...

And I can tell you 1 $ for a 75 V 1.5 mOhm mosfet (plus non-SMD) is too good to be true, it's more like 3 $ or more in 10+ quantities.

Also when playing with mosfets and high power it's not that simple, you can have major problems really fast just because of a wrong value resistor or because you forgot about things like the inductive voltage spike at turn-off. Please be careful if you design your own SSR and test it extensively before using it

I totally agree.
Aliexpress products need "China safety marge", as you seldom get what is advertised.
not long ago i found windturbine advertisement, rated 6000watt, with free (750watt) controller ?
Maybe during a tornado the generator can give 6kw spike.

Mosfets and other key components better can be bought from respectable source like Mouser or Digi-key.
Yes, you pay a few cents more, but can choose between quality, and are sure you get what you ordered!!

Playing with mosfets and high amperage is not that simple, at all.
The problems that occur when you make a mistake.....
One 100Ah LiFePO4 gives 1300A @ 3.2v on short.... (source: YouTube video where an 100A LiFePO4 cell was abused)
My 736Ah powerbank would give about 10.000A @ 52.4v Yikes!! ☠ ?
That is not something to joke around with.

If you are educated in electronics and know what you are doing, sure, you can make your own mosfet based on/off switch. (SSR)
Solid State Relays are not always the best solution, especially not to use for high ampere BMS.

Their "power" is the fact that they are small and can be on the PCB, what makes it more customer friendly to use.
Really useful at lower amperage.

SSR basic purpose is to switch a lot, for shorter time. (mostly lower amperage)
Contactors basic purpose is to switch not so often, for longer time (mostly higher amperage like turn on electric forklift)

SSR tends to get really hot, and need additional cooling. (heat = power used???)



Recommended heatsink size for 200A SSR
=100x100x80mm




Contactors run a lot cooler, at the correct voltage.
(I've done several tests, at little under voltage (40v for 48v) the contactor is warm and can be touched without problems.
Al little over voltage (54v for 48v) the contactor gets a lot warmer, almost too hot to touch.
Below 30v the 48v contactor doesn't make fast contact.

 

[Frank in Thailand]

Previously floated the idea of doing away with a high current contactor and using a shunt trip circuit breaker to serve two functions; protect circuit from overcurrent and disconnect the battery upon command from the BMS (just as a contractor or SSR would do). The drawing below illustrates the idea. The first battery bank is shown with a high current NO contactor which disconnects the battery, the second battery bank with a small NC relay that when the BMS signals (or fails) results in battery voltage applied to the shunt trip mechanism of the circuit breaker, opening the breaker and disconnecting the battery. The advantage of the shunt trip breaker it has no standby power loss. Disadvantage is once tripped it has to be reset manually.

So could this work??

View attachment 17351

Thanks for sharing the idea!! I did not yet think about this way of solution.

In short:
Yes, shunt trip will work perfectly.

It's "disadvantage" also can be seen advantage
When it needs to trip, something is "wrong" and needs your attention.
Upside is, its really easy to turn back on, everyone who can reset a normal AC circuit breaker can do this also.

Other disadvantage is its price.
400A quickly is + $ 600 up to $ 4500 for higher quality.
that is huge disadvantage.

DC circuit breaker needs to have an Arc defuse, and works a lot different from AC circuit breaker.
Standard a DC circuit breaker is about 10x the costs of AC circuit breaker, and not even a really good quality.
Shunt trip DC circuit breaker, at High ampere.....
That will cost a lot of money.

 

[BiduleOhm]

SSR tends to get really hot, and need additional cooling.

That's because they use marginal design as mosfets are expensive. For some reason the BMS tend to use more mosfets to have a lower Rdson and dissipate less power, I guess it's because the higher power ones followed the same designs as the lower power ones who didn't have heatsinking (or maybe just an aluminium plate).

 

[Frank in Thailand]

Makes sense.

Probably also why the higher Ampere rated ones are high in price, more quality mosfets.

Basic principle for 10A , 100A or 1.000A is the same.

One or more cells out of range, and stop.

 

[BiduleOhm]

Yep, for a nice 48 VDC SSR who only needs a small heatsink you have a base cost of about 30 $ and then about 30-35 $ per 100 A you want to handle, but that's for the materials only, you need to add assembly, shipping, manufacturer margin, vendor margin, etc...

 

[DelChamp]

Hey, How is this project going? I'm looking to build this same system with a buddy that wants to power his Auto shop in the back of his property. The electric Company wants $2,500 just to run electricity out to it.

 

[Michael B Caro]

El Jalito; on the sea of cortez south east of La Paz and it is a 1000 miles away!. Loreto is gorgeous....

I have been meaning to get to both of these areas to do some diving......Hopefully it won't be too awful long now before I can get there.

 

[apctjb]

Hey, How is this project going?

A learning experience for sure. Below is a picture of the assembled bank.

The M6 terminal screws that came with the battery were zinc and to short for double busbars . I ordered these on Amazon, they worked great


I torqued them to 5nm using a T handle torque wrench (used by bike mechanics). Plastic handle , not likely to short busbars if dropped...

After tightening I use a IR temp gun to check to see if there is heating at any of the connections. (Also bought on Amazon)

 

[DelChamp]

Hey apctjb,
Have you thought of using one of the Solar Surge Suppressors in the combiner boxes? I am going to have just one Growatt 3K unit initially and was thinking this would be perfect -Midnite Solar Solar Surge Protection Device - 115V MNSPD-115. It is actually rated to 150V max which seems to be a perfect match for these Growatts which are rated to 145V.

Also, I will have a 3p3s array for mine as well. My panels are rated at 9.71 Amps each and the manufacturer recommends using 15A breakers/fuses for each string. Also, I was actually going to use two 80A main breakers before going to the inputs on the Growatt to match it's max Amps input, just in case I wanted to add a couple of more strings.

Your thoughts?

 

[apctjb]

Hey@ apctjb,
Have you thought of using one of the Solar Surge Suppressors in the combiner boxes? I am going to have just one Growatt 3K unit initially and was thinking this would be perfect -Midnite Solar Solar Surge Protection Device - 115V MNSPD-115. It is actually rated to 150V max which seems to be a perfect match for these Growatts which are rated to 145V.

Also, I will have a 3p3s array for mine as well. My panels are rated at 9.71 Amps each and the manufacturer recommends using 15A breakers/fuses for each string. Also, I was actually going to use two 80A main breakers before going to the inputs on the Growatt to match it's max Amps input, just in case I wanted to add a couple of more strings

Surge suppressor for sure if you live in a lightning prone area. 15A string fuses, and 80A breaker on the PV input are all good. The breaker is really more a disconnect than circuit protection as PV modules are by their nature current limited so don't worry about it being oversized. Your string fuses have you covered.

Have fun.

 

[RCinFLA]

Wow, be very careful about ordering components on aliexpress, especially power transistors, they are fake 90 % of the time (and it's not a legend, lots of people have decapped some components to see if they were fake or not). I love aliexpress for a lot of stuff but for electronics I always use reputable dealers like Mouser, Farnell, Digikey, ...

And I can tell you 1 $ for a 75 V 1.5 mOhm mosfet (plus non-SMD) is too good to be true, it's more like 3 $ or more in 10+ quantities.

Also when playing with mosfets and high power it's not that simple, you can have major problems really fast just because of a wrong value resistor or because you forgot about things like the inductive voltage spike at turn-off. Please be careful if you design your own SSR and test it extensively before using it

Do not necessarily disagree. What do you think you are getting in Chinese inverters and BMS's? I am 100% sure they will not be International Rectifier parts. Chinese close clone is HY5608. It is a little cheaper and at least you have Chinese local Hooyi Semiconductor that can locally defend their reputation against bad copies.

 

[BiduleOhm]

What do you think you are getting in Chinese inverters and BMS's?

Well, I don't plan on buying chinese inverters and it's also a reason why I wasn't satisfied with the chinese BMS and making my own instead.

Now that you said that, I wonder if there's any sense buying a chinese inverter and replacing the mosfets with better ones (and the caps too...).

 

[Cancerfood]

Diving in on building an off grid PV system for a vacation home in Baja. Building in a remote location so simplicity, redundancy, ability to monitor remotely, easy to " swap out"should there be problems are important criteria. Size/weight is also a consideration as need to drive the components down.

After reading lots of post on this forum decided to pursue an "all in one" inverter design with a LiFePO4 battery bank. Ordered 32 EVE 280ah cells on Alibaba from Amy at Xuba. Placed order, paid by wire to avoid the credit card hassles others have posted about and 3 days later the cells were shipped and on there way.

For the inverter decided to go with 2 Growatt SPF3000TL LVM 48 (3000W-48V) inverters. I had also considered the MPP Solar inverters but decided to go with Growatt because of brand (Growatt; top 10 PV inverter manufacturer), and with Growatt the parallel card/cables were included as was a wifi dongle and remote monitoring software. I purchased the inverters from Ian at https://watts247.com . Ian answered a ton of my questions, had the inverters in stock, shipped the same day and I had them 3 days later. Could not be easier.

2 x SPF3000 in parallel allows me to have up to 9kw of PV input (160A) and 6000Wac output at either 120 or 120/240Vac. I don't see needing 6000W output capacity often so intend to operate just one of the inverters , keeping the other in backup most of the time. I intend to have ~5.5kW peak of solar and will split the array output with 1/2 going to each inverter (My understanding is the inverters built in MPPT charger operates even if the inverter is turned off).

Below is a wiring schematic of how I am intending on wiring things up, and rendering of how I am thinking I will layout the components. I appreciate any comments/suggestions others may have to offer. This is my first time working with an "all in one" inverter or LiFePO4 battery. I still need to decide on which BMS to purchase, and welcome recommendations...

View attachment 17150
View attachment 17151

Thanks for the level of detail in your schematic. It has really clarified the components needed and wiring. I am wondering why you chose to go with two separate spf3000 growatts instead of like a single 12000watt(or smaller) growatt unit? I am looking at the growatt for my off-build and appreciated your thread!

 

[schroederjd]

I was considering the shunt-trip option for discharge disconnect as well. I noticed that this is the approach that Batrium seems to suggest (https://support.batrium.com/article/239-circuit-breaker-and-shunt-trip-why). Note also that in their designs, they use a low current shunt-trip (
$39 Schneider Electric A9N2694) which physically trips an adjacent larger breaker. If you're planning to add a large breaker to the system anyway, seems like a cheap way to add discharge protection.

Midnight also has remote trip breakers. 125A for $43, 175A or 250A for $102 (from Alt-E store). Considering that BMS dischage disconnect would be 'last-resort' protection method, having to reset the breaker doesn't seem like too much of a down-side. I was thinking about using this method for the discharge circuit, and a 100A contactor for the Charge side. This way, a low-temp disconnect event (which will be common in my setup) would be reversible, and would not impact discharging.