diy solar

diy solar

Solar house generator I started DIY back in 2000 - My path from Trace to Xantrex (on FLA battery) to XW Pro inverters on Tesla Model S batteries

mike sent me the following:

Have been building my retirement home which is offgrid, boat access only for the past 9 years and need to get my electrical system
planned and into my home. The problem is me sitting on the fence on the batteries and inverter.
So many people are telling me that the Tesla batteries aren't safe. I think that using EVTV controller would look after any issues like this.
I love the modular ability to add more batteries should I need to.
Looking at eBay as a source for batteries do you know if all of the Model S batteries are compatible ?
Anything that I should look for ?
Seeing that the Singineer makes inverters specifically for Tesla's I'm leaning this way but..
No real user feedback on their gear, probably minimal support, therefore Victron would be a better choice due to support ?
The EVTV inverter can be used with other inverters like your Trace/Schneider, SMA. Any issues with the interface with your Trace and EVTV ?
I'm not computer savvy by any stretch, does the EVTV controller/display work well ?
Thanks very much for your time, really appreciate it.
 
After I composed my thoughts here are my answers

My Phase 2.0 of my 20 year ago system began in 2019. I wanted a very large battery bank but not break my bank. When I found EVTV and his solution I spent months watching all Jack's videos (at 1.5 or 1.75 speed). Being a software geek I loved the ESP32 and whole model.

People do dismiss the NMC Tesla battery (nickel-maganese-cobalt) because of its low thermal event level (some 185 degrees C). Most of that is because people do not respect those batteries nor respect what the Tesla engineers did to protect the batteries. What Jack did was to reverse engineer the software protocols of the battery modules BMS and then monitor and control the interfaces to the batteries.

Effectively Jack uses the exact protocols that Tesla uses in protecting the batteries in the vehicles. To me that was brilliant. Others quickly dismiss the Tesla battery as a viable storage because they don't understand. Once bias, always bias.

I found the cost per energy density of tesla to be a great deal: Around $300 per kwh, once one had 4 modules AND the JITB (Jack-in-the-box: nick name for the ESP32).

Verses $700 per kwhr for Fortress, Simplify and other commercial batteries on the market.

So for me, the JITB and the modules became:

1) Safe
2) cheaper per kwhr
3) well protected from dangerous events

Luckily, I had 22 year old inverter that could fully handle the tesla battery charge profile.
A Trace SW4024, would do a CUTOUT at 16 volts, this allowed me to set actual cutout at 18.5 volts on the teslas.

It allowed me to use the full spectrum of the battery's range.

Later I upgraded to 48v inverter a Xantrex SW5548 and it too allowed for the low cut out (I now use 38.5 volts for cutout and 49.5 volts for bulk charge - thus cycling 80% of battery capacity in order to get more life cycles).

I understand, an inverter is the major component in a system - got to get that right

I had first thought I would end up with the schneider inverter - that is why I did the schneider charge controllers. But upon reflection I could see that schneider was like the GM of the industry. Good system, complete functionality, but a bit aged. Originally based on the Trace/Xantrex models. Big company solutions.

SO I expanded my options. Kept my old inverter for awhile, and used the teslas in 24v mode. As mentioned I moved to Xantrex for 48v because it could handle the charge profile of the teslas. The Schneider could not. Neither can all the other inverters. Growatt, MPPT, Victron, etc all cutout at 42 volts, or maybe 40 volts. They can NOT utilize the full range of the teslas

But the Sigineer could. It is massive, heavy and $3500 (with Shipping) and 15kw

I am still aiming to go with that - I am invested in the Teslas, I have half of my planned battery capacity. (I am preparing for grid down state in the coming months/years


So many people are telling me that the Tesla batteries aren't safe.

True if not properly handled
Gas cars are not safe either if not properly handled

they seem to work fine in the vehicles using tesla's BMS and controller


I think that using EVTV controller would look after any issues like this.

That is exactly how I view the ESP32 (JITB)
it is my monitor, protector, and barrier to "bad events"
It is amazing.

Even so I would have liked to have placed my power house in a separate building like Erin and Jost did - I love their YT channel





I love the modular ability to add more batteries should I need to.
THAT also is a major attractant to this for me as well


Looking at eBay as a source for batteries do you know if all of the Model S batteries are compatible ?
Yes, get the same capacity. I use the 5.2 kwh modules, there are 6.2 kwhr modules but there are fewer of them

2013 to 2018 at least I know of
I have 2017 - 4 of them
and some earlier years, which I don't know which ones I have
older years use 6 individual BMS leads, that you can see through the plater cover

new years use a flat ribbon cable fr the BMS leads


Anything that I should look for ?
MAKE sure the modules comes with the great Tesla BMS card as that is required for the JITB to work

I am sure you read my experience when I failed to ensure that
Once you get a good seller, reuse them
I now buy from the seller in New York


https://www.ebay.com/itm/174654844336?epid=25021027276&hash=item28aa3d6db0:g:CZAAAOSwLsxbNp4B

www.ebay.com



1x Tesla Model S 85 battery module block , 24V, 250Ah 5.2kWh | eBay


For sale is an Clean good Condition Tesla S 85 1x modules with pigtail and bms and wiring plug if asked for. All listed amp hours, volts, KWh, are Tesla ra tings and will slightly be different depending on actual battery usage and conditions, load, ect .

www.ebay.com







Seeing that the Singineer makes inverters specifically for Tesla's I'm leaning this way but..
No real user feedback on their gear, probably minimal support,
Since Will and others on this site avoid teslas one won't find much support info hee

Sigineer did flash their firmware with the charging profile for teslas - that indicates they understand that market

NOTE: The sigineer has an idle consumption of 200 WATTS all the time. You need enough solar to cover that as well as battery
It i intended for BIG system, just keep that in mind


therefore Victron would be a better choice due to support ?
Again the cutoff limit of 40 to 42 volts reduces use of the battery's capacity between 38.5v and 42 volts

Read EACH inverter's cutoff limits and bulk charging numbers


The EVTV inverter can be used with other inverters like your Trace/Schneider, SMA.
Inverters?

maybe, but I will be swapping my Xantrex inverter out for the sigineer

or are you referring to the charge controllers?

if so, they work find. The Charge Controllers (CC) charge the battery, the inverter drain it. well at low charge the inverters have a battery charger from Grid or GenSet


Any issues with the interface with your Trace and EVTV ?
No, or can you explain your thoughts


I'm not computer savvy by any stretch, does the EVTV controller/display work well ?
I LOVE the Raspberry Pi Display
great information

I use a computer (linux - maybe available under windows) call VNC where I can remotely view the 1 Display from anywhere



Thanks very much for your time, really appreciate it.

Mike Ruegamer
N Vancouver , BC


In the past few years I see that choices of good inverters and cheaper batteries have grown
Growatt, MPPT
and other inverters have gone down in price and up in features, even though they do not support the Tesla charging profile
Can stack multiple Growatts to get more wattage
video from Will or DavidPoz show the capabilites
Batteries like the EG4 from signature Solar look great
The patent on the LiFePO4 batteries expired in Sept 2021 and I think we will find many new batteries coming in from China (well at least until we get a war) like the EG4
EG4
48v 5.12kwh 100amphr
very similar energy density to the tesla (24v, 5.2 kwh, 232 amphr)
$1500 vs $1250 for tesla
Nice stacking box, with disconnect, etc
Today, I would consider a Growatt 6k/8k/12k inverter, 48v
EG4 batteries, will support upto 15 modules, 75kwhr
Things are much better with options than just 3 years ago
You have more options now.
I suggest write your requirements, and develop 1 or 2 solutions for those requirements
look at price
flexibility
future growth
understand ability


kinda fun though, isn't it?
 
Last edited:
Yet another incredibly helpful and time intensive reply by Doug, you're a true champion for the use of Tesla with Jack's BMS, thank you for being here.... You truly are a god send to everyone on this adventure.

73
Jen
KB6JEN/VE7OTH
 
For those are using the Jack-In-The-Box battery controller for Tesla batteries, here is note.

In December (15th) I discovered that the ESP32 JITB was failing to proper harvest battery modules from the pack. I communicated with the sw engineer about and I provided him some good logs. I was running the following ESP32 firmware:

EVTV Tesla Battery Module Monitor Version ESP32-2.07

He provided a link and info for newer firmware. Now I am running:

EVTV Tesla Battery Module Monitor Version ESP32-2.19

2.19


Well today I found the same frozen voltage report. The JITB was reporting 46.3 while my charge controllers were reporting 49.5v (full battery for me). I harvested the data and the logs and sent them to him.

He responded later that he believes he knows what the issue is and will work on an upgrade to fix it.

The bottom line is that every 49 days a 32-bit value wraps and triggers this condition.
While this voltage data lockup occurs, the system is unprotected from too high a voltage condition - so this is a BAD condition

The work around until new firmware comes, is to reboot the JITB within a 49 day window.

To determine the firmware, one used the USB comm port like CoolTerm, to read the output from the JITB.

Here is a snippet from the USB output"

================================================================================================================
********** EVTV Tesla Battery Module Monitor Version ESP32-2.19 Tuesday, February 08, 2022 20:06:27 ******
================================================================================================================
******* WIFI STATUS..connected to local Access Point: dougbert-solar as IP: 192.168.8.142 signal strength: -53 dBm ****
================================================================================================================
Module 1: 24.114V 19.0/18.2C Cell01:3.998V Cell02:4.033V Cell03:4.032V Cell04:4.034V Cell05:4.034V Cell06:4.005V [3c08/3eab/4751] (36/0/0)
Module 2: 24.170V 19.2/17.5C Cell07:4.000V Cell08:4.036V Cell09:4.033V Cell10:4.035V Cell11:4.035V Cell12:4.008V [3c16/3e85/4717] (36/0/0)
Module 3: 24.170V 19.5/18.7C Cell13:4.006V Cell14:4.035V Cell15:4.034V Cell16:4.035V Cell17:4.035V Cell18:4.009V [3bf6/3e97/474f] (36/0/0)
Module 4: 24.192V 20.0/18.4C Cell19:4.004V Cell20:4.033V Cell21:4.033V Cell22:4.034V Cell23:4.033V Cell24:4.010V [3c17/3e8b/4715] (36/0/0)
Module 5: 24.235V 21.4/20.0C Cell25:3.990V Cell26:4.037V Cell27:4.035V Cell28:4.035V Cell29:4.036V Cell30:4.019V [3c17/3e90/46ca] (36/0/0)
Module 6: 24.241V 21.4/18.9C Cell31:3.995V Cell32:4.038V Cell33:4.037V Cell34:4.038V Cell35:4.038V Cell36:4.021V [3bd6/3e7b/5711] (36/0/0)
Module 7: 24.180V 22.4/20.8C Cell37:3.985V Cell38:4.030V Cell39:4.029V Cell40:4.030V Cell41:4.032V Cell42:4.016V [3bab/3e97/4717] (36/0/0)
Module 8: 24.211V 22.0/20.0C Cell43:3.990V Cell44:4.038V Cell45:4.037V Cell46:4.038V Cell47:4.038V Cell48:4.024V [3c00/3e8e/5751] (36/0/0)
Module 9: 24.180V 21.5/20.6C Cell49:3.994V Cell50:4.036V Cell51:4.037V Cell52:4.035V Cell53:4.037V Cell54:4.018V [3bf7/3e89/4770] (36/0/0)
Module 10: 24.182V 22.4/19.8C Cell55:3.984V Cell56:4.030V Cell57:4.030V Cell58:4.029V Cell59:4.029V Cell60:4.006V [3bbb/3e8a/560b] (36/0/0)

PACK STATUS:No Faults Modules:10 Voltage:48.375v Avg Cell:4.024v Avg Temp:20.1C SOC:96.12%

CURRENT:-32.547A Voltage:48.37v Ampere Hours:-45.43AH Power:-1574 Watts

Max System Discharge Current: -68.54A Max System Charge Current: 90.54A
Max Pack Voltage: 49.42vdc Min Pack Voltage: 48.36vdc
Current High Cell Voltage: 4.040V Low Cell Voltage: 3.984V
Battery Lifetime Charging: 3635.897 kWh Discharging: -3196.248 kWh
Negative Contactor:ON Reported ON
Positive Contactor:ON Reported ON
Temperature Alarm:OFF 0 Voltage Alarm:OFF 0
Charge Enable:ON Heat Enable:OFF 12v Reported ON


Enter ? for Configuration Menu......


See the following video piece on how to connect to the COMM port by utilizing a small WiFi router:


On my system I have another method of using the Raspberry Pi display itself, and use one of it's USB ports and directly connect to the USB port on the ESP32 without wifi. Now I use kermit under linux on the Pi to do this, so that is an advance method.

To update the firmware requires using this COMM port to upload the binary.
 
Last edited:
I am beyond PISS*D about this.... they've stopping doing any reasonable video's on Jack's channel anymore and now this, NO notification of a safety issue... the ONLY reason we overpaid for this stuff was for the safety factor and to see they've SH&T the bed on letting us all know we could burn our houses down and failure to support this product is exactly what I feared when Jack passed. I have yet to install my system, and at this point, depending on the outcome of this, may just sell it all, go to rack mount modules and spend the rest of my life bad mouthing their lack of follow up support on each and every forum I'm on. They've ruined that company since Jack passed. Why the hell are they not telling us ALL who spend thousands on this stuff when new firmware comes out, not just you. Why has it taken TWO F'ing months for us to find out about this risk and not from them, from you!!!!!! (rhetorical)

Sent them an email demanding an explanation as to why they've failed to let us know their bug (they've kept secret) could burn us all to the ground and they have done NOTHING!!!! If I do not have a proper response by the end of the day (or a full refund on every item I've purchased from them) I'll be notifying appropriate regulators about their failure to do a proper recall as required by law and I'll be consulting my attorney. I mave have some Tesla modules for sale Doug, at this point I have ZERO faith in that company or that product.. I do not plan to use it till they earn back the trust they've lost, not by the issue, but by their backyard hack BS failure to to deal with it by notifying us and resolving. This is exactly why we're all forced to get inspections and have laws, to protect us from this stuff, their keeping this secret and risking all our lives sets a very bad example for the
DIY industry and is what leads to further regulation. Stay tuned....

Richard (info@evtv.me), pull all my invoices and start drafting a full refund to me!!!!!

Thanks for bringing this to our attention Doug, at this point I do not recommend ANYONE buy this product or anything else from them until this is resolved.... at the end of the day if I've not been responded to... I'll deal with it my own way, I don't take lightly the risk of a company killing my family by not releasing information about an issue they know is dangerous.


whoa there, we just found this issue,

It has JUST come to attention (this issue) since Dec (event 1) and 2 days ago (event 2) - they have not kept it secret, EVTV didn't know about it yet - with both events code inspection and test cases have been occurring to locate the problem.
It is just over 49 days from Dec 16th (when I first reported first event) until now with the second event.
We found 2 cases of the event, which aids in locating the issue, and it did


I am also a software engineer and it takes debugging time when finding an issue.
There are always updated firmware on products, some are needed fast and others are simple bug fixes

Not all of product firmware updates need to be updated to all products in the field.
Some firmware updates are done to add DEBUGGING output information.

I do agree with you that a mechanism needs to be put in place by EVTV for firmware updates
along with release notes indicating fixes and impacts, etc.


The reason I received new firmware was after I first reported the issue, to get to the latest so as to verify if the issue I
found was still failing. Since it was, one goes to the next step to cause the failure to
occur faster than 49 days, as a unit test. Once that test is in place, then fixes can be
applied and the test run again. That does take some time. I was watching for
the issue to occur, and it did, I reported it and now solution is pending

I also do miss videos on the products, that I believe should return

Yes this one issue is a serious problem,
yet we/they have found it, the problem, and the engineer is working on the fix.
I don't know when/how they will send out info but EVTV probably doesn't know about it fully yet
as the engineer is now working on it

I suppose I jumped here too fast with too early of a customer notice, considering your reaction.
I was wanting to give a heads up to others of a concern. Inspections would not have found this issue.
It is an edge case.

Again the work around is to cycle power on the ESP32 for now, with period of time less than
49 days.


I reported the issue, research was done on that issue, and a solution will be forth coming
My intent was NOT to cause a fire storm here.
Rather as a place where discussions of a product can be share with other users of the product over time
and quickly notify users of it.

I still fully believe in the product - it does a good job at that - as it is improved.
Product support and notifications can be improved, fully agree

dougbert
 
Last edited:
Again the work around is to cycle power on the ESP32 for now, every period of time less than
48 days
.
good to know

handling the timer counter wrap around from INT_MAX back to 0 with firmware for microcontroller, just one of those things that is easy to omit or not handle properly. i've surely done it before ? the counter can only count so high, so after enough time it loops back to 0 and it can look like time went backwards if the code isn't checking properly

thanks for documenting the bug and assisting them in zeroing in on it ?
 
good to know

handling the timer counter wrap around from INT_MAX back to 0 with firmware for microcontroller, just one of those things that is easy to omit or not handle properly. i've surely done it before ? the counter can only count so high, so after enough time it loops back to 0 and it can look like time went backwards if the code isn't checking properly

thanks for documenting the bug and assisting them in zeroing in on it ?

for a wrapping 32 bit Time stamp integer, instead of using a comparison being done, a 2s complement subtraction seems to work better.

From a fellow engineer:

wow, that's classic. I coded a time wrap around bug many years ago. To test my fix, I loaded the time counter with an initial value that would guarantee it would hit the wrap around in a few seconds instead of months

to measure delta-T, if you always do a 2's complement subtraction, and not a comparison, it usually works

(even when the time counter wraps)
 
Last edited:
for anyone interested in more info about time wraparound

Description​


Returns the number of milliseconds passed since the Arduino board began running the current program. This number will overflow (go back to zero), after approximately 50 days.

Description​


Unsigned long variables are extended size variables for number storage, and store 32 bits (4 bytes). Unlike standard longs unsigned longs won’t store negative numbers, making their range from 0 to 4,294,967,295 (2^32 - 1).

4,294,967,295 milliseconds in days = 49.7103009259 days
 
if they were aware of the bug which potentially compromises over voltage protection disconnect, and did not immediately send an email out to everyone affected (to whatever extent possible), that is a bad look for the org. the delay in communicating the bug seems kind of not good imho. bugs happen, good to proactively communicate about :)

the way it manifests is mundane, but yeah... it seems to affect a rather important aspect of the system operation.

anyways, hoping everyone can have a safe machine happy batteries happy day ?

oof!
 
I read your comment as you notified them on Dec 2nd..... over two months to let people know about an issue that could burn their house down is what set me off, that said, unless they only learned of it within the last 24 hours.... I'd expect to know about it for such a serious issue.. I'm on their mailing list for a reason... I love the product, I love what Jack did with the company and I trusted Jack.... this has blown all that.... not the issue itself... I use to work at Microsoft... I know about release schedules.... what is unacceptable to me is the lack of letting us know the system we trusted could fail and burn us to the ground... and kill our kids on the way down...

UNACCEPTABLE!!!!

But I'm sure they appreciate your "whoa there"... I don't "whoa there" when it comes to fires and death. We're already walking on grenades with this stuff, if they're going to tell me the pin is made of noodles that melts at a certain temperature unless replaced/cycled regularly... oh wait... they haven't told me.... if not for your hard work the pin would be on the ground...

We paid for a sense of security... I've lost that.
Correction from the saved email: Dec 15th

I understand the concern and I expressed that as well. That began the search for the bug.

The manifestation was the failure to update module voltages based on packets received from the modules. that is what I reported in Dec and sparked the search for the cause.
This week's failure event gave info leading to the Time Delta calculation and to the time wrap operation. Hence: good news leading to the fix
 
if they were aware of the bug which potentially compromises over voltage protection disconnect, and did not send an email out to everyone they could, that is a bad look for the org.

the way it manifests is mundane, but yeah... it seems to affect a rather important aspect of the system operation.

anyways, hoping everyone can have a safe machine happy batteries happy day ?

oof!
yes, that does point to the support processes failing of EVTV.
I will suggest to Richard that needs to be updated

I don't know if the engineer communicated the issue to EVTV that he was researching it. He contracts for EVTV
 
Just a reminder of the risk this creates that we all paid PREMIUM money to avoid...... it's easy to forget...... "whoa there"?????

 
Just a reminder of the risk this creates that we all paid PREMIUM money to avoid...... it's easy to forget...... "whoa there"?????

watched this again.. worth the rewatch. it's such a good video. they made the best of a big oops.

they put their mistake out in the open, documented and went over the postmortem analysis. rich did the community a great service by sharing the lessons learned.. truly some epic "please use my mistake to learn" up in smoke energy going on.

in the end, in that video, it was a lack of over voltage protection disconnect during charging that precipitated the fire.

rich straight up acknowledges the error in allowing an overvoltage condition, leading to fire.

the best part, nobody was injured, and LOTS of footage to get a real idea of what working around it means.

thanks for sharing
 
Yup, and if I'm not mistaken, that's exactly what this issue is they've not told us is an issue...

How many people have these and are NOT on this forum and have no clue about this and we can assume based on current lack of comms, never would know......

Whew......
Agree with those points

Notification to customers is needed for sure for this product.
I asked about that as well
 
Back
Top