Plan: Off grid solar with a Model S battery pack at the heart

[wk057]

Not much effect from Matthew besides two days of no direct sun. The day before storage reached 100%, couple of days of Matthew clouds, and the day after was a zero-cloud day.

Two Matthew days:

And the day after (topping off stationary storage and two days of driving):

And some jokes with my brother, Matthew:

 

[MarcG]

Glad to hear! I was mostly concerned about some of your panels flying off your roof, or the ground mounts, with the crazy wind speeds

So are the SunPower panels still to your satisfaction I hope?

 

[pirpy]

Hello wk057

You are living my dream

I havent found ansver in long thread so I will ask.

Why didn't you consider UPS instead of 8 x 48 V outback inverters?
something like this:
Liebert NXC 10-200 kVA - Network Online UPS

and there are many others.

Best regards

B.

 

[Yoda101]

Water storage under the panels is actually something I'm considering, but not for rain collection. Heavily insulated water tanks for thermal storage.

If I remember correctly, you did not find a viable storage solution for your excess power , and you hate waste!

Have you considered the sunamp heat battery?
Broshure : http://sunamp.co.uk/wp-content/uploads/2016/05/Sunamp-Heat-Batteries-Brochure-web.pdf

Video :

 

[Nakinto]

I was wondering if you have any kind of radiator for the liquid cooling system? How did you set this up if you do? I am looking at putting a similar (much smaller) system in my Motorhome and I need about 8 of these packs.

Thank you!

 

[TheJMan]

Can you charge your Teslas DC now, straight from your MASSIVE battery?
Is this even technically possible somehow? like with a DC input CHAdeMO wallbox of some sort?
Would help massively with efficiency.

 

[miimura]

In order to charge the car from the stationary pack, you have to have a DC/DC converter that can raise the voltage from 50VDC nominal to the vehicle pack's 400VDC nominal. That converter would have its own losses. My understanding is that most of those devices are based on inverting to AC at a relatively higher frequency, going through a transformer, and then rectifying the AC back to DC. So, the only intrinsic benefit is being able to use a frequency other than 60Hz. He already has enough inverter capacity for the house to spare for car charging and the solar is producing excess for a large part of the year, reducing the benefit of improving efficiency. Those factors would discourage me from spending more money on a DC charging setup.

 

[Fat Bao]

2nd year update?

 

[arg]

In order to charge the car from the stationary pack, you have to have a DC/DC converter that can raise the voltage from 50VDC nominal to the vehicle pack's 400VDC nominal. That converter would have its own losses. My understanding is that most of those devices are based on inverting to AC at a relatively higher frequency, going through a transformer, and then rectifying the AC back to DC. So, the only intrinsic benefit is being able to use a frequency other than 60Hz.

It's certainly true that the existing battery->AC inverter, the AC->DC charger in the car, and the hypothetical battery(50V)->battery(400V) charger are all very similar devices with very similar efficiencies (depending in practice on how much money has been spent on them).

However, the AC charging passes the charge power through two of these conversion stages, while the direct charger only passes it through one conversion stage - so potentially saving some efficiency (in the range of 5%-10%).

He already has enough inverter capacity for the house to spare for car charging and the solar is producing excess for a large part of the year, reducing the benefit of improving efficiency. Those factors would discourage me from spending more money on a DC charging setup.

These are good reasons why it's probably not worth the (considerable) effort and expense.

 

[wk057]

2nd year update?

I lost a lot of my AC-side data for year two due to a failure in one of the TED sensor MTU units. So, some of the usage data is lower resolution (output from the inverters is only reported in 0.1kW increments vs 1W). I'll have to sit down and see what I can do with the data I have. Still some pretty good stats available.

Actually generated slightly less power this year: 42,181 kWh... mostly because I haven't concerned myself with utilizing the excess power generated as much as I did on year one. Hadn't had the time to complete my dump load setups, either. Also, my newer more efficient HVAC system had been online the full year instead of having my older less efficient system for 20% of the year like last year.

I also used slightly more power from the grid... about 2.4 MWh this past year vs 0.7MWh the year before. This was mostly due to some rarer events that required power in times where I had a limited amount.... like charging four Model S/X, or returning from a long trip and needing a full charge when weather had been terrible for a while already and my other car was still being charged daily for a temporary 100 mile commute. Also had a few snow days in there where I preemptively switched to grid power to conserve battery power in case of a grid outage. I hadn't used any grid power since February, so all of my grid power usage for this past year was in the first half. (My year reset day is September 3rd)

So, even with my rare consumption spikes, over the course of two years I've averaged about 130 kWh/mo from the grid. 85% lower than the national average. But, keep in mind that I use ~3.6MWh per month on average (large home with multiple EVs), so that's actually still about a 97% reduction in grid power usage.

Overall the system is still going strong. No complaints.

 

[AntronX]

Overall the system is still going strong. No complaints.

Nice to see you still around. What are you working on these days? Did you complete the BMS for your Telsa modules?

 

[worldburger]

@wk057 (or anyone else) What are the limitations/reasons preventing a DC-DC home Supercharger/chademo system as a practical solution to charging a Tesla?

This project is remarkable on so many levels, but I didn't see anything addressing this specifically (thought I admit I did not read all 76 pages .

 

[worldburger]

Right... and 64kW @ 350v is <200A. Voltage loss is I²R so your wire size requirements go down A LOT; For the same size wire pushing 1500A generates >50x as much heat as 200A. My only point is that as high voltage options become available they're generally the better choice for loads >2kW. The other huge advantage is you can use EV packs as they come without rewiring them.

I have a 48v off-grid system too... that's all that was available. But if I had to do it again today I would either go with the Solaredge 'Storedge' solution or a Fronius Symo when they get released later this year. Smaller, lighter, cheaper and more efficient.

@nwdiver this post was helpful. has the landscape changed at all since your post for inverters you’d consider?

 

[spottyq]

So a little list of problems :
- high voltage DC power electronics are not really marketed/sold to single persons… so you've got to jump through hoops to even get a price, let alone buy something
- as a consequence of the above, it is only possible to get good prices when buying in bulk
- you'd have to reverse engineer the chademo or supercharger protocol (possible but time-consuming)
- if using a Tesla, the charging speed gets throttled when DC charging too much. (Though that probably wouldn't be a problem for wk ! )

None of them are show stoppers, but consider what you get in return :
- ~4-5x faster charging compared to normal AC charging (if you reverse-engineer the supercharger protocol, otherwise you are limited to 50kW with Tesla's chademo adapter.)
- 10% ? efficiency improvement ? (not if you want higher than AC charging speeds, because of the extra cooling needed.)
- You've got a good understanding of how the chademo/supercharger protocol works.
- ??

So basically, the only way it can make any sense (and that's still if you *really* need more than AC charging speeds) is if a company decides to make it so that they can get bulk prices, get a chademo license and sell the whole package.

To get an idea of the prices involved, check out EVTV's 50kW DC-DC chademo charger.

 

[miimura]

@wk057 (or anyone else) What are the limitations/reasons preventing a DC-DC home Supercharger/chademo system as a practical solution to charging a Tesla?

This project is remarkable on so many levels, but I didn't see anything addressing this specifically (thought I admit I did not read all 76 pages .

I addressed this for wk057's specific situation a few posts up, #1507.

In order to charge the car from the stationary pack, you have to have a DC/DC converter that can raise the voltage from 50VDC nominal to the vehicle pack's 400VDC nominal. That converter would have its own losses. My understanding is that most of those devices are based on inverting to AC at a relatively higher frequency, going through a transformer, and then rectifying the AC back to DC. So, the only intrinsic benefit is being able to use a frequency other than 60Hz. He already has enough inverter capacity for the house to spare for car charging and the solar is producing excess for a large part of the year, reducing the benefit of improving efficiency. Those factors would discourage me from spending more money on a DC charging setup.

I will add that there is a limited market for an expensive DC-DC charging setup, which would make the cost even higher due to limited volume.

The only practical way to do this is to start with a system that uses high voltage battery packs. EVTV has a bunch of equipment for sale that can be put together to do this with intact Tesla battery packs, HVDC inverters, and CHAdeMO chargers. However, you have to completely buy into Jack's total system approach, which is fundamentally different than Jason's. Jack's approach has merit, but Jason's is based on commercially proven equipment.

 

[nwdiver]

@nwdiver this post was helpful. has the landscape changed at all since your post for inverters you’d consider?

SMA has a new HV inverter that's supposed to be released this year. If it's priced similarly to their grid-tie inverters you should be able to get a 6kW off-grid inverter for ~$1500...

@wk057 (or anyone else) What are the limitations/reasons preventing a DC-DC home Supercharger/chademo system as a practical solution to charging a Tesla?

I don't think the cost-benefit weighs out... even in places where theres tons (MWs) of ~400vDC like at the new 40 bay Supercharger-Lounge with PV covered parking it's just easier to do ~400vDCPV => 480vAC => => ~400vDC Supercharger.

 

[anttinnnn]

I'm in the process of prototyping my custom BMS.

Short question regarding the BMS, what is the reasoning behind only connecting 48VDC twin (6s/24VDC) module packs in parallel, but leaving the middle (+24 VDC) potential floating and not connected together? Surely this would equalize and balance the packs more, and save on the BMS-modules?

Also I guess you could also connect the bricks inside the modules in parallel, then the whole system would only have 12 serial "cells" to monitor and balance? Much simpler and more balanced setup, what is it that I am obviously missing?

 

[wk057]

Short question regarding the BMS, what is the reasoning behind only connecting 48VDC twin (6s/24VDC) module packs in parallel, but leaving the middle (+24 VDC) potential floating and not connected together? Surely this would equalize and balance the packs more, and save on the BMS-modules?

Also I guess you could also connect the bricks inside the modules in parallel, then the whole system would only have 12 serial "cells" to monitor and balance? Much simpler and more balanced setup, what is it that I am obviously missing?

It is a good point.

First, to make such a connection make sense, it would require paralleling them in a way where resistance was equal. This would likely require a third large bus bar. Alternatively, the center of the sets of 4 modules could be connected in those sets without too much trouble.

Overall, I did consider doing this. An issue with this is actually a safety concern. There are some failure modes that could result in this center connection causing gross overcharge of a module.

 

[wk057]

Not sure if anyone really follows this thread anymore, but, I made some updates to wk057.solar to make a graph that shows net usage, battery usage, solar input, etc etc. I ran the script to regenerate all of the previous graphs as well in the updated format, although some pockets of time don't have all of the info required.

Here's a cool example from a couple weeks ago when I had my Model S running a script to keep the charger power in line with available power.

I've been doing this for a while now (years?) when convenient, but never really graphed the effect. You can see the SoC remains roughly flat despite ~25kW's flying around in there. The load line is a little under the solar curve because the load line is AC power and doesn't compensate for DC losses... which add up at high power usage like charging a car at 20kW.

Overall, pretty cool to see my system graphed out like this finally.

I originally didn't show anything load related because of privacy concerns. Essentially, the internet can know when I'm charging a car now. But there's no real pattern to my usage, overall. I'm home and not home pretty randomly, so, until that changes I'll leave it public. If I need to I'll just put the usage related stuff on a time delay (couple of days) or something.

Anyway. System is running great still.

Just about 128 MWh generated in total now.

My grid usage has gone up, but that's mainly because I run a set of servers in-house now that are about a 1.2 kW continuous load... which adds up. Also, I have three to four EVs here now, depending on who's here. My system was really only sized for two.

In the summer all of this isn't as much of a problem, but it shows in the very limited usefulness of my storage now, which is at best a couple of days worth of power. The Chevy Volt charges from 120V, and the Tesla's between 40 and 80A. Tesla usage is pretty obvious in the graphs, but the Volt tends to hide in there. The server rack usage stays pretty solid, too. Also using more HVAC with more of my house occupied these days.

Surprisingly, taking all of this into account, solar input is probably still pretty close to being sufficient, on average. I'd just need another... maybe 500-600 kWh of storage to make it completely usable year round..... which doesn't make a lot of sense considering even at the lowest prices that would cost somewhere around $100k... and I don't think I'll spend $100k on occasional grid usage in multiple lifetimes.

 

[brkaus]

Of course we still watch this thread