diy solar

diy solar

New System Design - 40kw / 80kwh / 20kw

HighTechLab

AKA Dexter - CTO of Current Connected, LLC
Joined
Sep 23, 2019
Messages
1,637
I came up with a concept for a massive system, that I'm considering having pre-designed for customers. Based around LF quattro inverters, the goal on this is for someone off-grid that is unable to reduce their system demand and is wanting to run everything electric. Supports adding a generator quite easily also.

Having 20kw of solar for an 80kwh battery bank, the thought is this should be sufficient to recharge the bank within 4 hours of sun, which is pretty average across the board.

40kw of inverter in split-phase +parallel to give some major output capability.


1653397813522.png

16 SOK 48v server rack batteries = ~80kwh battery
4x Victron Quattro 10k's = 40kw (120kw surge)
4x Victron MPPT RS 100a = 20kw solar
3x Lynx Distributors with a bunch of fuses
1x Cerbo GX

A few notes:

  • Victron requires all cables to the Quattros be equal length when used in a multi-inverter cluster. We dispersed the battery interconnections within the inverter connections so that we could reduce the load on the busbar and avoid having 1200a through a single point.
  • The SOK's have internal shunts/monitoring, so that would handle the SOC% and current measurement calculations
  • Solar panels are TBD, I just pulled in a standard block from a different diagram I made
  • Equal length battery cables for even current sharing. 4/0 to Lynx (400a max of batteries), 1/0 from group 1-2 to group 3-4 (200a max), 2awg from 1 to 2 and 2awg from 3 to 4 (100a max)
With solar panels included this has a pricetag in the range of $75,000 for everything included.

What stands out to you?
 
What stands out to you?
The battery circuits span the left and center lynx lynx distributors.
Since the batteries span 2 distributors the resistance in the joinery between the distributors may make the resistance to the batteries asymmetric.
I would connect the batteries to the left lynx then add a lynx shunt.

lynx_distributor(battery)<->lynx_shunt<->lynx_distributor(inverters)<->lynx_distributor(SCCs)
 
What do you see the market being for a 40kW system? I have a pretty substantial house (albeit in a moderate climate), and my demand only peaks around 8kW... so 12kW surge capacity is justifiable. I can see a bit extra for additional HVAC and a pool, but that seems quite big. Being off-grid at that size makes it even more odd.
 
What stands out to you?
The fact that on the face of it, it sounds like a DIY install for $75,000 and there is no way I'm dropping that kind of money on a DIY system with no support to come fix it when it goes sideways. (vs other $75,000 turn key commercial installed systems with full customer support and warranty)

Any $75,000 system I'd spec would also come with a auto-controlled diesel/propane generator in the 10-15kw size with install and 200 gal tank.
 
The battery circuits span the left and center lynx lynx distributors.
Since the batteries span 2 distributors the resistance in the joinery between the distributors may make the resistance to the batteries asymmetric.
I would connect the batteries to the left lynx then add a lynx shunt.

lynx_distributor(battery)<->lynx_shunt<->lynx_distributor(inverters)<->lynx_distributor(SCCs)
I was considering adding a 4th Lynx and putting half the solar controllers on the other side, but reality is all SCCs may output differing amounts of solar based on panel placement and potential shading. I also considered interdispersing the SCCs as well, but didn't like how it ruined the symmetry. Not sure there is a perfect way to solve this problem.

No lynx shunt is needed. The batteries report SOC to the Cerbo directly. Also the lynx shunt is underrated. The quattros can easily exceed 1000a in peak scenarios.

I also agree on the 400a fuse. It's not a Victron thing, but rather LittelFuse who supplies Victron. Victron's manual for the Quattro's specifically states to use a 400a fuse that they do not have available.

What do you see the market being for a 40kW system?
We just provided a customer a similar system but with a 128 cell CALB battery bank and REC BMS. It's for an all-electric house. Heat pump HVAC, heat pump water heater, electric stove, electric dryer. The demand adds up QUICK in that type of build.
 
The fact that on the face of it, it sounds like a DIY install for $75,000 and there is no way I'm dropping that kind of money on a DIY system with no support to come fix it when it goes sideways. (vs other $75,000 turn key commercial installed systems with full customer support and warranty)

Any $75,000 system I'd spec would also come with a auto-controlled diesel/propane generator in the 10-15kw size with install and 200 gal tank.
We are providing systems to homeowners working directly with qualified electricians, where the electrician is responsible for the install. These are not systems to be mickey-moused together by a "DYIer".

Generator is optional.
 
Here is my idea.
Notice the inverter_chargers get a 250 amp fuse to increase the possibility of isolating a fault to a single branch circuit.
 

Attachments

  • HighTechLab.001.txt
    3 KB · Views: 5
FWIW, I would guess you are better off skid-mounting it as a single system. In about a 3' x 6' x 6'h rack/skid/cabinet I think you could have everything integrated. You could also use vertical busbars and marine rated terminal fuses rather than struggling with matching cable lengths. It would add about $10-20k to the equipment cost, but easily make up for it in installation cost.

You would need front/back or front and sides access depending on how you configured it

Also, the system is really 32kW/40kVA, but that is being pedantic.
 
We just provided a customer a similar system but with a 128 cell CALB battery bank and REC BMS. It's for an all-electric house. Heat pump HVAC, heat pump water heater, electric stove, electric dryer. The demand adds up QUICK in that type of build.
You should encourage them to document everything! I love seeing large all electric builds like that!
 
I’m not sure how you would charge the battery bank in 4 hours? I see that the SOK batteries have a maximum charge rate of 100A. 100A X 48V = 4800 watts. 80k WH battery / 4800 watts = 16.7 hours. Sorry if I am missing something. I am new to off grid solar planning and am not up to speed yet. I am also trying to build a large system like this, but am struggling on how to charge this large of a battery bank in under 5 hours of peak sun that I would get in CO during the winter.
 
I’m not sure how you would charge the battery bank in 4 hours? I see that the SOK batteries have a maximum charge rate of 100A. 100A X 48V = 4800 watts. 80k WH battery / 4800 watts = 16.7 hours. Sorry if I am missing something. I am new to off grid solar planning and am not up to speed yet. I am also trying to build a large system like this, but am struggling on how to charge this large of a battery bank in under 5 hours of peak sun that I would get in CO during the winter.
100A per battery
 
100A per battery
If I am understanding the diagram correctly then, each charge controller is connected to a bank of 4 batteries totaling 20kwh. The charge controller can deliver 100A of charging at 48V. 100 X 48 = 4,800 watts. So 20k/4.8k = 4.16 hours. Then each of the 4 battery banks is connected to its own inverter?
 
I just saw the system you made for Ambition strikes https://www.youtube.com/c/AmbitionStrikes
I'm guessing this thread was about that topology.
You mention in the video that the inverter leads have to be the same length.
Why is that?

I think the inverters would function properly in parallel on the ac side even if the inverters were completely separate systems on the dc side.
 
Last edited:
Here is my idea.
Notice the inverter_chargers get a 250 amp fuse to increase the possibility of isolating a fault to a single branch circuit.

Just to prove that I read your "diagram", the solar charge controller section lists positive for in legs. Is that right?

scc.1 {
in {
positive<-dpst<-6s_panels
positive->dpst->6s_panels
 
Just to prove that I read your "diagram", the solar charge controller section lists positive for in legs. Is that right?

scc.1 {
in {
positive<-dpst<-6s_panels
positive->dpst->6s_panels
You may be the first. :)

Thanks for the eyeballs.
fixed.
 
So the wiring of your batteries is interesting. Is that the recommended way to do it if we are planning on daisy chaining? Would you recommend a bus bar instead?
 
Having 20kw of solar for an 80kwh battery bank, the thought is this should be sufficient to recharge the bank within 4 hours of sun, which is pretty average across the board.

Dexter, could you speak as to how important using this wiring is method us versus the package you supply on your website? This is from the Victron book if I'm recalling correctly. I just a bought a few of the 5 battery stacks and had wondered about it doing this way but with everything else I never circled back to it. Obviously I want the best performance and life from my battery bank and now is the time to change things up.

Thanks!


1672406073329.png
 
Back
Top