Powerwall 2.0 Backup Runtime Extender

[pgrovetom1]

There is some work Jack Rickard has one at evtv.me regarding communicating with the BMS integrated into the Tesla Modules.

I have thought about the need to increase the battery capacity of a Tesla PowerWall. That was one reason I cancelled my Powerwall deposit because I wanted a system that I could increase the storage capacity. My choice of a hybrid inverter has not given me the programmatic choices that a PowerWall would have and therefore my need for additional battery capacity has not been needed. I will follow your project with interest.

Its not difficult to implement the BMS functions except balancing the cells by just controlling charging, discharge and a temperature monitor. It may not need balancing except initially and it would be easy to monitor the cells using the A/D and a 6:1 analog selector. Then it can be reported. If necessary, cell balancing modules for a high current 6s battery are available without messing around with the Tesla interface details. I would just remove it and bring the cell wires over to the controller.

I don't think adding 5KWhrs will make a huge difference plus the arbitrage is small stuff .... but it could be entertaining. The whole thing can be in the garage away from the outside environment. Only the radio based current monitors for the solar, PW2 and the home would be out at the service entrance. My service entrance ( and the PW2's and their stuff) is about 50 feet from my house.

The cost should be very reasonable at about $900 plus $1000 per 5.2KWhr battery pack - A wild guess!

1) 240VAC to 18V Power supply is about $100
2) The DC to DC 1800W constant current power supply is about $30
3) 2 good Micro inverters $500-$700
4) Raspberry Pi 4 Model B - 2GB RAMabout $40
5) DIY AC current monitors using Arduino and a LoRa transceiver plus misc $50-$100
6) DIY Variable current limiters $25
7) Tesla Model S Battery Module 24v 250ah 5.2kwh 444 Panasonic 18650 3200mah $900-$1200 each 5.2KWhrs

5.2KWhr - $2000
10.4KWhr - $3000
15.6KWhr - $4000 etc... very roughly

If there are no gotchas! Like the PW2's somehow can detect whats going on and complain! The trick is be invisible to the greatest extent possible.

Power injection - looks just like a reduction in home power load.. no more and no less
Charging - Looks like a home load.

By using a UL 1741 complaint micro-inverters, it looks just like the solar inverters. The power injection at transitions of home power ( an oven or high power appliance is turned on and then off) and the MI's must raise and lower the injected power from zero to maximum ( 500W in my example) without having the voltage spike. Since injection cannot be done while the PW2's are fully charged and the 60Hz has been raised, at all other times the PW2 can soak up any excess power being injected during on/off appliance transitions - I hope.

The radio link needs to be quick so the control loop doesn't have any problems. If the radio goes down, the system disconnects and the limiters are set to zero.

 

[miimura]

If you're going to use a standalone charger and you're going to use the single module at 6S, you should just use a hobby charger like the one linked below to control the charging. It will do charging current control and balancing, every time you charge. The only difference is that you need a power supply at a higher voltage than the module.

https://www.amazon.com/ISDT-T8-1000W-DC-Charger/dp/B0776PW3NW

 

[pgrovetom1]

If you're going to use a standalone charger and you're going to use the single module at 6S, you should just use a hobby charger like the one linked below to control the charging. It will do charging current control and balancing, every time you charge. The only difference is that you need a power supply at a higher voltage than the module.

https://www.amazon.com/ISDT-T8-1000W-DC-Charger/dp/B0776PW3NW

It would be nice to find a 6s commercial charger but this one has an inadequate discharge current at 5A and the balancing is limited to 2.2A per cell so it won't work in this application. Good idea though.

The Tesla Model S battery is a 74p6s configuration and the 74 cells in parallel would not balance very quickly at 2.2A. It would work but very slowly over time especially with well balanced Tesla batteries. The two micro inverters I'm proposing in my example ( minimum case) would require up to 1000W discharge ( 500W each) at 24.6V which would require over 40A of discharge current. This greatly exceeds this chargers 5A maximum. It also says its limited to 20W maximum discharge capacity which is inconsistent with the 5A. Not sure whats going on.

Finding a commercial charger with balancing would be nice. Do you know of one with much higher discharge and balance current. The constant current DC/DC converter is $28 and claims 1800W but using it at 75% or 30A would be wise. Its only an example that would work but at 30A @24V = 720W would charge a 5.2KWhr battery in about 7 hours. That's pretty close to the solar production time during a day.

My example DC-DC charger
US $26.92 10% OFF|1800W 40A CC CV Boost Converter DC DC Step Up Power Supply Adjustable Board DC DC 10V 60V to 12V 90V DIY Electronic Kit Module|Integrated Circuits| | - AliExpress

 

[miimura]

I think you're overestimating the balancing requirement for Tesla modules. Either that or you're expecting it to balance quickly. That's just not how it works. 2.2A balance current per cell group is relatively high, even for a 74P pack.

When you discharge the module you're taking energy from the main (+) and (-) terminals. The 6 cells in series are discharging equally. They will only go out of balance if they are not well matched in characteristics. Tesla is very good at assembling modules with cells that match. So, comparing the discharge current to the balance current makes absolutely no sense. Also, the 5A discharge spec is only for the discharge test mode of the charger. That doesn't come into the picture at all when you're charging.

I have a boost converter almost exactly like the one you linked. Your math is correct. You would only get about 720W through it for charging. I would recommend a buck converter and higher voltage power supply because it will be more efficient regulating the voltage down than boosting and regulating the voltage up. Also, the amperage is on the input side. If you are starting with 12V and boosting to 24V, you will only get 360W through it. Regarding charging time, I would hope it wouldn't take that long because you should not be discharging 100% of the capacity every day. That will be a hard life for that battery pack. Using it 90% to 20% would give a much longer life.

 

[pgrovetom1]

It appears as you say there are off the shelf commercial chargers and balance modules that should work fine. No point re-inventing the wheel. It appears I should be able to build a simplified prototype after my PW2's are installed and see if they complain. I'll focus on the Raspberry Pi controller, current limiters and radio link current measurement remotes. If the PW2's are ok with it, I'll refine it and add the non-backup mode.

My current E6 tariff which ends in 2022, charges $.396 per KWhr in the peak first tier and $.203 in the off peak first tier. So it would be possible to arbitrage the difference of $.193 per KWhr. So its only 20 cents per KWhr. So even if the whole 5KWhrs was charged and dumped each day, thats only 5KWhr x 20 cents or $1 per day. That may not be worth much effort.

 

[miimura]

I pulled out the Runtime Extender hardware last night due to the dearth of solar production with the smoky Bay Area air yesterday. It allowed me to make it through the Peak hours before hitting the Reserve. After running for 4 hours the DC converter got pretty warm. I should get a 12V fan to cool these things. Total AC energy out was 80% of DC energy from the car's 12V system. Today looks just as bad for solar generation. I think I will just leave the 25% in the Powerwalls for actual power outage and accept whatever Peak power I have to buy from PG&E this evening.

 

[gpez]

I pulled out the Runtime Extender hardware last night due to the dearth of solar production with the smoky Bay Area air yesterday. It allowed me to make it through the Peak hours before hitting the Reserve. After running for 4 hours the DC converter got pretty warm. I should get a 12V fan to cool these things. Total AC energy out was 80% of DC energy from the car's 12V system. Today looks just as bad for solar generation. I think I will just leave the 25% in the Powerwalls for actual power outage and accept whatever Peak power I have to buy from PG&E this evening.

I was fascinated by this as a PoC, cool to see that it was put to use.

Would you mind sharing the app charts for the time you ran it? Curious to see the data.

 

[miimura]

I was fascinated by this as a PoC, cool to see that it was put to use.

Would you mind sharing the app charts for the time you ran it? Curious to see the data.

It's a little unremarkable. The area under the red line was supplied by the Runtime Extender Grid-Tied Inverter. It starts out at about 650W, then drops to about 500W as it heats up and starts running its fan.

 

[gpez]

Still cool to see, thanks for sharing.

I wish the Tesla app could also show the PW SoC.

 

[gpez]

Still cool to see, thanks for sharing.

I wish the Tesla app could also show the PW SoC.

...on the graph

 

[n2mb_racing]

Stumbled on this thread... Someone here might know, are there good small'ish solar inverters that will work with 60-84V DC battery inputs? And can they be started up without a grid connection?

 

[miimura]

Stumbled on this thread... Someone here might know, are there good small'ish solar inverters that will work with 60-84V DC battery inputs? And can they be started up without a grid connection?

If you don't want to push the energy into the grid or Powerwalls, you can use any off-grid inverter that is compatible with your battery voltage. That said, battery inverters over 60VDC are uncommon because they require more rigorous isolation and protection from touching live contacts due to the hazard of the higher voltages.