People have suggested that the Model 3 is very smart about the 12V battery and flags errors when you add drain from there. You will have better results if you tap the 12V directly from the top of the "Penthouse" under the rear seat. Search threads about stereo installs. At least one guy did a bunch of research on the Model 3's 12V system and that was his conclusion.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Powerwall 2.0 Backup Runtime Extender
- Thread starter miimura
- Start date
mckemie
Member
Thanks! I had concluded that the Tesla was monitoring 12v demand somewhere down stream from my battery tap. I was intending to look for that CT or monitoring point today but not beyond the area of the battery. With your information, I will not waste that time and may seek the right point in the rear.
I wonder if S/X are the same. I have installed trailer lighting directly from the battery on Ss. Of course they are tiny draws and not likely troublesome.
"12V Battery Need Service" error and aftermarket sound system
leads to
Tesla Model 3 Stereo - Part 9: Summary and Lessons Learned
Last edited:
I don't know if it's been done for a Volt, but a Prius traction battery connected to a large UPS has been successfully used as an alternate power source during power outages. It's not grid-tie additive, but I assume that during extended power outages, the Prius UPS could be your main source of power and the Powerwall would sync/recharge to it.
Details of Prius UPS is here: PriUPS-getting electricity FROM your hybrid vehicle
Details of Prius UPS is here: PriUPS-getting electricity FROM your hybrid vehicle
Since I opted to do the non-grid tie route with my Volt (just got a quality 1200w pure sine inverter) I've been considering looking in to what it would take to feed my critical loads panel properly. The Gen 2 Volt's supposed maximum traction DC to 12v DC wattage is fairly low (I've read anywhere from 1-2kw) which is limiting.
That said, as part of my Powerwall install my 240v circuits (oven, dryer, a/c, ev charger) are already sub paneled out which just leaves the 120v circuits on the main (aka critical loads) panel. They moved those 240v breakers out which leaves room at the top of my panel for a proper interlock or if I wanted to go all in a manual transfer switch would be even better. Volt -> Inverter -> Garage input -> Interlock/MTS -> Critical loads panel
The challenge with either option is that I'm looking at a few thousand dollars to properly permit and install the correct hardware and the reality is that outages that will exceed my solar + Powerwall capability will be exceedingly rare. For now a large extension cord will suffice.
As a final note I'd caution against trying to use a non grid-tie inverter to sync and charge your Powerwall with - I'm fairly certain that will either not work as expected (or worse) and the DC-DC-DC losses will stack quickly. My plan would be to physically plug the fridge, furnace, tankless gas hot water heater, and freezer in to the inverter and let the solar charge the Powerwall.
I agree with all this. Even though all your loads are 120V, you probably still need to provide split phase 240V power if you're going to connect to a sub-panel through a transfer switch. An autotransformer can do this at a reasonable price even if your inverter is 120VAC.
Example: https://smile.amazon.com/Victron-Energy-Autotransformer-120-240VAC-32A/dp/B01MXI87GZ
The main difference with the Prius solution is that it is taking energy from the hybrid battery, not the 12V system, so the power capability is much higher than taking 12V from a PHEV or EV. Vehicles like the Outlander PHEV also have an optional 120V inverter built in that takes energy from the traction battery, but it's still only rated at 1,500W.
Yes, the Prius is capable of supplying about 6kw via it's traction battery and running engine. The system I have uses a 6kw UPS which is fed into my load panel via a regular generator transfer switch. It's only running when the grid power is off so it really functions like a backup generator type system. I don't really know how powerwall interacts with solar and grid tie, so can't speak to that, but when it is in backup generator mode, the UPS is supplying extremely good quality power to the circuits that are connected to it. I haven't tried it yet, but it should be possible to charge a Tesla at about 4-5kw via the Prius, in a pinch. The UPS is "portable", so in essence, you could make a Prius portable Tesla charger that can drive to an empty Tesla, and charge it (slowly) enough to get it to a supercharger...
If all the circuits are 120V two-wire (separate neutrals for each ungrounded conductor), there is nothing wrong physics-wise with installing a transfer mechanism which disconnects from all other sources, jumpers together both busses in the panel, and feeds them from a 120V source. I'm not aware of it violating anything in the NEC or the UL listing of the panel, either.
Cheers, Wayne
NAE (not an electrician) but this is my understanding as well - nothing technically against it just really strange. I have read that some AFCI/GFCI breakers and outlets may not like it depending on the ground wiring.
I suppose you're right even though it's a little unconventional. I wonder what building inspectors would say about it.
mckemie
Member
Progress.
I received:
Hot DC12V to 36V 10A 20A 30A Step-up Boost Converter Power Supply Module Car | eBay
and tested overnight using two M215 inverters, getting just under 500 watts of PowerWall agumentation. I was unable to get the inverters to start with only the boost but adding a 36v battery in parallel got them started. The 12v supply was an imiev. The overnight test pulled an estimated 12kwh from the imiev and supplied an estimated 5kwh at 240vac. The boost is the 30amp/1kw model. I have a 10amp model on order. The 1kw should supply either 3 or 4 ~250w inverters. I will eventually test the 10amp/350w model with a single inverter. My 36v "inverter starter" battery was a golf cart module containing 8 surplus 4 ah hoverboard batteries but I expect a single hoverboard battery to serve. Will test a single 36v battery.
So, I have several configurations that work to agument a PowerWall with 240vac outputs:
1) Banks of the above mentioned golf cart / hoverboard batteries suppling M215 inverters. I have tested as many as four M215s in parallel giving near 1kw. This is the easiest and probably most efficient configuration.
2) Low quality "500w" grid tie inverters doing 12vdc -> 240vdc. I've run two in parallel and found them to be both inefficient and not reaching rated power. The power is highly dependent on input voltage; 14v gives about 200w and 12v gives about 125w. This set up works but not very satisfactorily or efficiently.
3) The above described boosted 12vdc -> 36vdc. That setup shows good promise.
It seems the supply of cheap old stock M215 inverters is drying up. There is hope that a low cost supply of the similar NEP BDM-250 inverters will come on the market. The BDM-250 appears to be identical to:
https://www.walmart.com/ip/Micro-Re...RfixFwCVG8Ou0JGRo1smnOzg7DwmEIThoCZeQQAvD_BwE
I have two of the Replus in hand for testing.
My 12vdc sources are imievs with connections directly to the 12v battery and my Model 3 with connections to 12v from a point under the rear seat. The inability of the Model 3 12v battery to supply suitable power has been discussed previously here. Here are some photos of my Model 3 12v connection:
https://photos.app.goo.gl/1zCvjxzJXnwHjHu98
I received:
Hot DC12V to 36V 10A 20A 30A Step-up Boost Converter Power Supply Module Car | eBay
and tested overnight using two M215 inverters, getting just under 500 watts of PowerWall agumentation. I was unable to get the inverters to start with only the boost but adding a 36v battery in parallel got them started. The 12v supply was an imiev. The overnight test pulled an estimated 12kwh from the imiev and supplied an estimated 5kwh at 240vac. The boost is the 30amp/1kw model. I have a 10amp model on order. The 1kw should supply either 3 or 4 ~250w inverters. I will eventually test the 10amp/350w model with a single inverter. My 36v "inverter starter" battery was a golf cart module containing 8 surplus 4 ah hoverboard batteries but I expect a single hoverboard battery to serve. Will test a single 36v battery.
So, I have several configurations that work to agument a PowerWall with 240vac outputs:
1) Banks of the above mentioned golf cart / hoverboard batteries suppling M215 inverters. I have tested as many as four M215s in parallel giving near 1kw. This is the easiest and probably most efficient configuration.
2) Low quality "500w" grid tie inverters doing 12vdc -> 240vdc. I've run two in parallel and found them to be both inefficient and not reaching rated power. The power is highly dependent on input voltage; 14v gives about 200w and 12v gives about 125w. This set up works but not very satisfactorily or efficiently.
3) The above described boosted 12vdc -> 36vdc. That setup shows good promise.
It seems the supply of cheap old stock M215 inverters is drying up. There is hope that a low cost supply of the similar NEP BDM-250 inverters will come on the market. The BDM-250 appears to be identical to:
https://www.walmart.com/ip/Micro-Re...RfixFwCVG8Ou0JGRo1smnOzg7DwmEIThoCZeQQAvD_BwE
I have two of the Replus in hand for testing.
My 12vdc sources are imievs with connections directly to the 12v battery and my Model 3 with connections to 12v from a point under the rear seat. The inability of the Model 3 12v battery to supply suitable power has been discussed previously here. Here are some photos of my Model 3 12v connection:
https://photos.app.goo.gl/1zCvjxzJXnwHjHu98
Last edited:
mckemie
Member
I think I've reached a stopping point on this project. I can augment from two power sources in three or four different ways.
Willie: PowerWall agumentation, new layout
Willie: PowerWall agumentation, new layout
OppositeLock
Member
This thread is brilliant! I think I could come up with something backed by a gasoline generator instead of an EV. You need a 24V battery to buffer, but generator charges the battery via charger, while inverter phase matches micro grid. If I wire this upstream of my PV ammeters, the PW will think it's the solar producing. Yes, not efficient, but a way to extend runtime in long outage. Cheap generators are terrible at running continuously at low load, so running it at high load to charge the PW and turning it off would be awesome.
Timing is good as I just installed the battery harness with quick disconnect plug on my Chevy Volt over the weekend. Actually connected it directly to the 2kw step down transformer that charges the 12v battery as some others online have suggested. Was fairly simple to install and has been working great in my light testing. Pure sine wave inverter was $180 and cables were $70 so all in I have a 1200w generator that can run for about 80 hours on 9 gallons of gas for $250. Just a quick note that this isn't a grid-tie setup but rather a load shedding approach.
I don't know if it's in this thread or elsewhere, but a guy in Australia took a surplus solar string inverter and wired the generator AC output through a bridge rectifier into the DC side of the grid tied inverter. That is a brilliant setup if you absolutely need to put generator power into a Powerwall system.
Had my first real test of my own Backup Runtime Extender due to a 2+ hour power outage after substantial snow and high winds yesterday. Not that I actually needed the extender as my Powerwall should last about 24 hours by itself without any PV (the snow blocked all production) if we're conservative - this was just a great opportunity to run the test 
The goal of this setup is not to physically charge or augment the Powerwall directly in any way but rather to shed some of the house load on to my EV (2018 Chevy Volt) in the case of a prolonged outage that would exceed the backup capabilities of my Powerwall + PV. With a standard, non grid-tie inverter like the one used here backfeeding the house or attaching it to the Powerwall is extremely dangerous.
As I mentioned in my previous posts I attached a quick disconnect power cable directly to my Volt's traction battery to 12v battery step down inverter. If you're not familiar with the Volt the step down inverter is under the trunk mat and is conveniently located in a way that with a long enough cable you can snake it under the side panel and in to a storage compartment for super easy stowage when not in use.
With the power out I started the Volt, turned off the headlights, any climate settings, and the radio. Next I set up my inverter, ran a heavy duty extension cord to the critical appliances (freezer, fridge, Xbox, TV), and then plugged it in to the Volt. To keep the car "running" so that the traction battery kept the 12v topped off I used a rubber band to keep the gearshift button pushed in. This is enough to convince the car that it needs to stay on even if not moving.
For about 2 hours I was drawing anywhere from 150-300w from the car, as indicated on the inverter, and the Volt registered a 0.5kWh total consumption from the traction battery over those two hours. With the 18.4kWh battery and a full tank of gas I estimate I could get approximately 110kWh from the car alone with this setup, which is easily over a week of totally off grid operation for the whole house. Of course with the Powerwall and PV that duration will be longer.
Overall I'm super happy with how it worked. I'll need to get a longer extension cord and a better way of setting up the inverter - having it lay on my ladder to keep it off the ground is not the best. The components total were under $250 and I think this was a great ROI.
Oh, and the Powerwall made us the envy of the neighborhood as we continued about our business during the outage as the only house lit up on the block
Copy of the equipment list if you're curious
Rough Country 7 FT Quick Disconnect Winch Power Cable Compatible w/Any Standard Size Winch RS107 -https://www.amazon.com/gp/product/B07G79K5SF
Power Inverter Pure Sine Wave 1200Watt 12V DC to 110V 120V with Remote Control Dual AC Outlets and USB Port for CPAP RV Car Solar System Emergency - https://www.amazon.com/gp/product/B07G36Z2TH
The goal of this setup is not to physically charge or augment the Powerwall directly in any way but rather to shed some of the house load on to my EV (2018 Chevy Volt) in the case of a prolonged outage that would exceed the backup capabilities of my Powerwall + PV. With a standard, non grid-tie inverter like the one used here backfeeding the house or attaching it to the Powerwall is extremely dangerous.
As I mentioned in my previous posts I attached a quick disconnect power cable directly to my Volt's traction battery to 12v battery step down inverter. If you're not familiar with the Volt the step down inverter is under the trunk mat and is conveniently located in a way that with a long enough cable you can snake it under the side panel and in to a storage compartment for super easy stowage when not in use.
With the power out I started the Volt, turned off the headlights, any climate settings, and the radio. Next I set up my inverter, ran a heavy duty extension cord to the critical appliances (freezer, fridge, Xbox, TV), and then plugged it in to the Volt. To keep the car "running" so that the traction battery kept the 12v topped off I used a rubber band to keep the gearshift button pushed in. This is enough to convince the car that it needs to stay on even if not moving.
For about 2 hours I was drawing anywhere from 150-300w from the car, as indicated on the inverter, and the Volt registered a 0.5kWh total consumption from the traction battery over those two hours. With the 18.4kWh battery and a full tank of gas I estimate I could get approximately 110kWh from the car alone with this setup, which is easily over a week of totally off grid operation for the whole house. Of course with the Powerwall and PV that duration will be longer.
Overall I'm super happy with how it worked. I'll need to get a longer extension cord and a better way of setting up the inverter - having it lay on my ladder to keep it off the ground is not the best. The components total were under $250 and I think this was a great ROI.
Oh, and the Powerwall made us the envy of the neighborhood as we continued about our business during the outage as the only house lit up on the block
Copy of the equipment list if you're curious
Rough Country 7 FT Quick Disconnect Winch Power Cable Compatible w/Any Standard Size Winch RS107 -https://www.amazon.com/gp/product/B07G79K5SF
Power Inverter Pure Sine Wave 1200Watt 12V DC to 110V 120V with Remote Control Dual AC Outlets and USB Port for CPAP RV Car Solar System Emergency - https://www.amazon.com/gp/product/B07G36Z2TH
Attachments
pgrovetom1
Member
I posted this in the thread I began but since this thread was similar, I thought I would post the same thing here to get any ideas from the people here.
I wanted to share a diagram of the concept example I was investigating to see if I could get ideas and feedback.
The main purpose is to add additional backup on top of the PW2 to help ride out cloudy days while in backup and perform time shifting of solar production. My understanding is the PW2 cannot perform time shifting of solar if the system used the Federal solar credits.I received a Federal Tax credit.
An example would be to utilize two 500W UL1741 120V AC Micro inverters, one per leg ( since they use different amounts of power) ( L1/L2) to inject power from the battery. I chose in my example a 5.2KWhr Tesla battery given its about the right minimum size. Additional Tesla batteries could be used up to the charger limit. It could supply 10 hours of 500W as one example. My home never uses less than about 800W at night baseline so one strategy would be to monitor the home L1/L2 power and adjust the current limiters to match that power up to 500W per leg. The charger would be enabled during solar production hours if there was adequate power being generated as measured at the solar inverters.
The charger would not be connected if the battery was fully charged or the temperature was low. Each UL1741 micro inverter would be fed by the 24V battery through a variable current limiter. The current would be varied to manage the power produced and injected by the micro inverters. This would make the micro inverters power vary as the MPPT adjusts to the maximum current at 24V. At 24.6V, the limiter would be adjusted up about 20 A to allow the production of the full 500W. This allows the injected power to be any amount from zero to the maximum of 500W in this example. Each AC leg would be using a different amount of power which could tracked. The current would be set to zero if the temperature was too high or the battery fully discharged.
The current monitors on the solar, PW2 and home would communicate the current power to the Raspberry Pi controller via a radio link using LoRa as one example. By utilizing a real time clock and monitoring the solar power, PW2 power, home power and injection power, the injected current would be adapted to the circumstances and time of day. So for example, charging would be restricted to when there was sufficient solar power. Power injection would be based on an effort to provide the current homes power use up to the 1000W in this example.In non-backup mode, it could be programmed to charge during maximum solar production and discharge during the evening highest tariff time. This would allow power arbitrage even if the PW2's won't do it due to the Federal tax credits.
One issue, that needs to be investigated is when the PW2's raise the 60Hz to shut off the solar inverters. Do they raise the 60Hz only when they are fully charged? If they are powering a home at night and not fully charged, do they lower the 60Hz allowing the solar inverters to come on in the morning. Do they only raise the 60Hz when they are fully charged and cannot accept charge or do they know the time of day and solar production hours etc.. The concern is these UL1741 micro inverters will comply with the raised 60Hz shutoff. If the PW2s only raise the 60Hz when they cannot accept charge irrespective of time of day, then this should work ok.Otherwise the micro inverters would not be allowed to discharge if the PW2's raised the 60Hz to shut off the solar inverters.
Thoughts?
Here is a very rough diagram of one implementation of the concept:
I wanted to share a diagram of the concept example I was investigating to see if I could get ideas and feedback.
The main purpose is to add additional backup on top of the PW2 to help ride out cloudy days while in backup and perform time shifting of solar production. My understanding is the PW2 cannot perform time shifting of solar if the system used the Federal solar credits.I received a Federal Tax credit.
An example would be to utilize two 500W UL1741 120V AC Micro inverters, one per leg ( since they use different amounts of power) ( L1/L2) to inject power from the battery. I chose in my example a 5.2KWhr Tesla battery given its about the right minimum size. Additional Tesla batteries could be used up to the charger limit. It could supply 10 hours of 500W as one example. My home never uses less than about 800W at night baseline so one strategy would be to monitor the home L1/L2 power and adjust the current limiters to match that power up to 500W per leg. The charger would be enabled during solar production hours if there was adequate power being generated as measured at the solar inverters.
The charger would not be connected if the battery was fully charged or the temperature was low. Each UL1741 micro inverter would be fed by the 24V battery through a variable current limiter. The current would be varied to manage the power produced and injected by the micro inverters. This would make the micro inverters power vary as the MPPT adjusts to the maximum current at 24V. At 24.6V, the limiter would be adjusted up about 20 A to allow the production of the full 500W. This allows the injected power to be any amount from zero to the maximum of 500W in this example. Each AC leg would be using a different amount of power which could tracked. The current would be set to zero if the temperature was too high or the battery fully discharged.
The current monitors on the solar, PW2 and home would communicate the current power to the Raspberry Pi controller via a radio link using LoRa as one example. By utilizing a real time clock and monitoring the solar power, PW2 power, home power and injection power, the injected current would be adapted to the circumstances and time of day. So for example, charging would be restricted to when there was sufficient solar power. Power injection would be based on an effort to provide the current homes power use up to the 1000W in this example.In non-backup mode, it could be programmed to charge during maximum solar production and discharge during the evening highest tariff time. This would allow power arbitrage even if the PW2's won't do it due to the Federal tax credits.
One issue, that needs to be investigated is when the PW2's raise the 60Hz to shut off the solar inverters. Do they raise the 60Hz only when they are fully charged? If they are powering a home at night and not fully charged, do they lower the 60Hz allowing the solar inverters to come on in the morning. Do they only raise the 60Hz when they are fully charged and cannot accept charge or do they know the time of day and solar production hours etc.. The concern is these UL1741 micro inverters will comply with the raised 60Hz shutoff. If the PW2s only raise the 60Hz when they cannot accept charge irrespective of time of day, then this should work ok.Otherwise the micro inverters would not be allowed to discharge if the PW2's raised the 60Hz to shut off the solar inverters.
Thoughts?
Here is a very rough diagram of one implementation of the concept:
I saw that thread. What kind of ideas did you get from the people on that thread?
Your concept appears to be good. Since the additional micros would only be used in grid-down situations, you would want them to do proportional curtailment as the frequency increases, or you should measure the frequency and use your variable current control to do the curtailment yourself.
I am curious how much you have thought through the controls and the details of the circuits to do the variable current control on the DC side of the micros.
I am curious how much you have thought through the controls and the details of the circuits to do the variable current control on the DC side of the micros.
pgrovetom1
Member
The system can perform in two modes: "Grid Power On " - power arbitrage and "Grid Power Off" PW2 extension.
Arbitrage simply charges during high solar time and discharges in the evening when the new tariffs at PG&E will be after solar time. So the system will move the abundant non-peak solar power to the evening peak time to arbitrage the rate difference. Its not much but its a side effect of the design.
The tricky part in both modes is to properly coordinate with the PW2. My thinking is to try and create the illusion to the PW2 that the system is "erasing" the home vampire power and low end power ( < 500W per leg - lights etc..). by measuring the home power on each leg, the current limit can be used to track the home load in near real time. It quietly charges while there is good solar and the PW2 has not shut off the solar inverters with high 60Hz because its fully charged. The UL1741 micro inverters would be compliant so its no different than my 18 M215s on my solar. When the PW2 raises the 60Hz, the micro inverters ( MI) would shut down. If I buy MI's that meet the new curtailment spec, they would also be curtailed if the PW2 slowly raised the 60Hz - just like the solar inverters.
Charging would "look like" a home load on the 240V with a BMS function which stops on battery fully charged or high temperature alarm ( or any other alarm that makes sense). I'm hoping the PW2's only raise the 60Hz when they are fully charged and see excess solar power to shut down the solar until they discharge to 95%. When the PW2's are cycling from 100% down to 95% and back, these MI's would only produce power when 60Hz was normal. It should be ok as long as the PW2's only raise the 60Hz when they are fully charged.
I have a design engineering background so the variable current limit should not be a problem. I would either use an Aurdino or a Raspberry Pi or maybe both. They would send an 8 bit code to an eight bit DAC which would produce a voltage between 0-5V which would control the pass transistor. Non of the circuit controls should be a problem but my largest concern is whether there are any "gotchas" where coordinating with the PW2 becomes a problem like the 60Hz issue I've mentioned. I'm also a bit concerned with the delay in the current/power monitors and the radio delay to the main controller. If someone turns on an oven in the house, both 500W MIs would be turned on in less than a second and when the oven turns off, that one second delay could cause a spike in the MIs output voltage if the PW2 doesn't soak it up.
I've located a low cost power supply and constant current which should be less than $150
Sign In | AliExpress Mobile
Sign In | AliExpress Mobile
The power supply would be DC 18V, AC 185-265VAC and 1500W. The combined efficiency should >90% even with the Tandem power supply and Boost constant current converter.
The Tesla battery is a 74p6s configuration with a fully charged ( 4.1v per cell) of 24.6V
Another issue would be adding a real BMS should balancing be needed. If the batteries in the Tesla Module track well once balanced once, than it might be reasonable to leave balancing off and just manage over-charging, low battery, and temperatures. The controller could have an 6:1 analog selector and ADC so it can monitor the balance and report any problem.
I have a Raspberry Pi and Aurdino and a LoRa transceiver. Its mostly software with a few circuits and many off the shelf from the Aurdino world.
Once my PW2's are installed, I can monitor when they raise the 60Hz to see if thats ok. I could also build a simple version using a much smaller battery before buying $1000 Tesla pack.
I just don't know if the PW2 software can detect odd things going on. Only way to know for sure is try it.
Last edited:
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.
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.
