IoT (cell) uplink for GroWatt SPF 5000ES

[methods]

I have a Growatt SPF 5000 ES landing Monday... so we are just getting started here. I have a lot of experience with low-cost IoT and ... before I go off and reinvent the wheel I want to make sure nobody has already done it.

Inverter Manual

GROWATT 5kW 5000ES (USA Model) OFF GRID INVERTER & Growatt Transformer Kit

5000W LVM- ES Off-Grid 240V GROWATT Inverter with Solar Charge Controller + integrated GROWATT Split Phase Transformer for true 120V/240V Split Phase Electricity! The GROWATT SPF 5000W LVM-ES is the all-in-one inverter/charger you need to take control of your energy bill. The 5000ES can work...

www.signaturesolar.com

High Level Task
Link the Growatt to the cloud servers for the lowest price possible, without WiFi! When I ran an IoT company for a short period of time... this basically meant

* Intercept the signal with the correct protocol (maybe hardware, maybe WiFi, maybe BLE... we will emulate the phone interface if required)
* Push as little data up to the cloud as possible (you pay by the KB)

To show that I am not a spammer or wanker, here is my list of commonly used parts:

IoT Sim Cards ($1/mo - I bought 10, they do what they say they do)

IoT SIM Card | Hologram

A secure IoT SIM card that provides access to a global cellular network built for M2M data.

www.hologram.io

Arduino WiFi intercept modules (buy off Digikey, not Adafruit)

Search Results for 'WiFi' on Adafruit Industries

Adafruit Industries, Unique & fun DIY electronics and kits : Search Results - Tools Gift Certificates Arduino Cables Sensors LEDs Books Breakout Boards Power EL Wire/Tape/Panel Components & Parts LCDs & Displays Wearables Prototyping Raspberry Pi Wireless Young Engineers 3D printing NeoPixels...

www.adafruit.com

Arduino 3G (cell) intercept modules (buy off Digikey, not Adafruit)

Search Results for '3G' on Adafruit Industries

Adafruit Industries, Unique & fun DIY electronics and kits : Search Results - Tools Gift Certificates Arduino Cables Sensors LEDs Books Breakout Boards Power EL Wire/Tape/Panel Components & Parts LCDs & Displays Wearables Prototyping Raspberry Pi Wireless Young Engineers 3D printing NeoPixels...

www.adafruit.com

Adrduino, most powerful available (buy off Digikey directly)

Teensy® 3.6

www.pjrc.com

When I am programming for an Android Phone, I use MIT App Inventor 2

MIT App Inventor | Explore MIT App Inventor

That is a purely graphical programming language (like LabView) and it exposes EVERY function available on the Android devices. They are working on an Apple core (or they were) but apples sucks for DIY.

For micro processing I use strictly Arduino - but 32bit 600mhz versions that are basically small computers ported for Android

... Anyhow
This is my first post. (and it needs some pictures!)

* I ordered 10pcs of used 460W panels, those are paid for and ready to be picked up
* I ordered a 5KW off grid setup to experiment with
* I have in my possession 42KW of 1C LiFe in really good shape, broken up into 4S 40Ah chunks... heading toward 48V

Obviously...
The Scientific Method requires that we accurately record data so that others can duplicate our experiments. That is why I am here... to share and learn.... SO -> If anyone has done it, save me the time.

I presume the GroWatt (and I have not searched yet) attaches itself to a local WiFi and starts spewing to some web address with a login. I did buy the WiFi dongle, but we may intercept before that.

Options
* Emulate a WiFi hub, intercept and send to Cellular 3G
* Direct wire before the Wifi dongle, figure it out, translate what they want (we can electrically monitor the interface there)

So...
If anyone is into it... that is the sort of stuff I do on the side. I sell nothing, I advertise for no one, I have a lot of experience in strange and deep places. ...

Onward

-methods

 

[corporate.joe]

GitHub - johanmeijer/grott: Growatt inverter monitor

Growatt inverter monitor. Contribute to johanmeijer/grott development by creating an account on GitHub.

github.com

Written in python already. Can operate in proxy mode or traffic sniff for the data. Requires a growatt shine-wifi network device plugged into the inverter. I'm using this already with my SPF 3000tl lvm. It gets the data to my database but I'm still interested in something with higher update frequency. Currently limited to one "buffered" snapshot of data every minute. I'm looking into using the RS485 port on the inverter to network it with other charge controllers of different manufacture. Haven't gotten far in that regard due to complacency in current configuration.

About those SIM card plans that you have linked. I've seen those before. Never used nor priced any of those. What kind of cost is the data plan per MB? I'm looking at disk usage for my time series database right now at 450MB since it's install in late August this year(95 days). This includes data from shunt, two BMS's, and the inverter. Might also be some PI system data in that DB as well.

Growatt also offers a GPRS device that plugs into the inverter for cloud updates as well.

I have an older 2g, I think, seed studio GPRS modem board meant for arduino use laying around somewhere I've never tried to use. Bought it back when Radio Shack was closing stores.

I'd be interested in following any progress you make in eliminating the shine-wifi or shine-gprs devices and decreasing the data latency to any time series database.

Joe-

 

[methods]

Alright!
Awesome man! Let me parse some of that.

Written in python already. Can operate in proxy mode or traffic sniff for the data. Requires a growatt shine-wifi network device plugged into the inverter. I'm using this already with my SPF 3000tl lvm. It gets the data to my database but I'm still interested in something with higher update frequency. Currently limited to one "buffered" snapshot of data every minute. I'm looking into using the RS485 port on the inverter to network it with other charge controllers of different manufacture. Haven't gotten far in that regard due to complacency in current configuration.

Exactly! Data into our cloud

I have been on the free tier of AWS for a long time. Trouble there is that you have to write your own front end. Most recently I have started using a cloud NoSQL service. With that I can just dump JSON files all day without a hitch. HTTPS or otherwise secure transfer.

Slow data rates can be our friend*
Long-term storage is cheap, fast up dates are expensive, bandwidth limits are cliffs.

About those SIM card plans that you have linked. I've seen those before. Never used nor priced any of those. What kind of cost is the data plan per MB? I'm looking at disk usage for my time series database right now at 450MB since it's install in late August this year(95 days). This includes data from shunt, two BMS's, and the inverter. Might also be some PI system data in that DB as well.

Pricing changes and I did my test run over two years ago now. All of these guys are aiming at high traffic. . . so for what we are doing you can often slip in under the caps. Of course... you can cap your spend... so service will drop before you rack up an AWS size bill (eh hem)

IoT Data Plans & Pricing

Our flexible pricing is tailored around the unique needs of your devices. Choose from three plans with the features, services, and support that best meet your needs.

www.hologram.io

So... you read that as "A buck" for having the device online (registered in your list, so activated) and $0.34/MB

Gotta watch them
* MB?
* Mb?

Right? Bytes or Bits - we will see.

If you figure you can pack your data relatively well

16:20:30.2 PV420V PA42A BV84V BC100A GP4.2KW LP4.0KW
(Totally Arbitrary Nonsense or TAN)

So that is like 50 bytes. Lets say you want to do that 10 times a second for a 10hz update. 500Bytes a minute... see how quick that adds up?

* 500B/min
* 30KB/hour
* 210KB per day @ 7hrs (kick to a lower rate at night)

Lets round that up to ... 500KB because no doubt it will get there at those rates. That would be like $0.15 per day, so multiply by 30 to guess around $4.20 a month.

(I just woke up and those are SUPER rough numbers... but that is the order we are working on)

* Straight Talk Unlimited SIM in an Android with straight pass-thru
* IoT SIM in an Android or Arduino (or Pi) for modified or re-routed data

Growatt also offers a GPRS device that plugs into the inverter for cloud updates as well.

I have an older 2g, I think, seed studio GPRS modem board meant for arduino use laying around somewhere I've never tried to use. Bought it back when Radio Shack was closing stores.

That is Super good to know. I did accelerated research and had not taken notice. I will check out the GPRS offering -> Nothing like banging AT commands all day long.

I am an advocate for 2G... but it is falling away. I actually LIKE the limitations and coverage mapping it has... slows down data rates... but there is only one carrier left around here supporting it and 3G is becoming the new underdog. I would suggest going 3G or better if you want to have less headaches later.

There ARE legacy 2G signals bouncing around that are contractually obligated.... but those WILL fall away.

I'd be interested in following any progress you make in eliminating the shine-wifi or shine-gprs devices and decreasing the data latency to any time series database.

Ok - then we have a target.

That device has a USB1 / USB2.0 interface. For that we have 4 wires

* Ground
* Data high
* Data low
* 5V

You can tap right into that and monitor it, intercept it, pass it thru, etc. Very easy... no magic. It is not like USBC with its multiple lanes and coaxial wires. USB 2 is super slow and super simple. Three wires and you are in. You can bet they did one of two things

* Ran into an FTDI chip for SUPER SLOW old school data
* Ran right into the microcontroller (more likely these days)

We can sniff that

 

[methods]

STEP 1
Ol' methods over here has to actually have some hardware in hand... Right?

DISCLAIMER
I am NOT a solar expert or even a "Solar Guy" so you will hear errors, omissions, etc. I don't sell anything solar, this is the first solar I have ever purchased, and I am figuring it out. I learn by making small mistakes and correcting course... so... there you go.

Here is my Auto Transformer.

Here is my Growatt 5000 ES

Random Bits of wire and such

Combiner box - The only thing I think might end up as parts

... YEA
Getting right to it.

I picked up 4pcs of those 100Ah 12V

I wired them in series for 48V and dropped them in the Golf Cart for evaluation. I was pleasantly surprised ... those things are underrated rippers! So... Those are getting hard-cycled in on the Golf Cart. The bulk of the batteries we will be using are ... in... need... of QUITE a bit of work.

Been shopping BLE enabled BMS's in the 60V 250A range. I figure for one of these setups you want to double your DC ratings to not cook off... so... if we need about 120A then I am specifying about 250A.

(or we swap all the fets**)

In a previous life I used to take 36V 20A ebike controllers and hack them to 100V 100A burst. Rippin Hard

...

Anyhow
I have everything in hand except my panels which need to get picked up. I had not thought about the fact that they are 8'x 4' when I arranged pickup, so waiting until a good time to get those with a larger truck.

-Schindler

 

[methods]

Here is the 8awg for the 48V batteries, some splices, and the 1kV crimper.

I don't know how I feel about that cheapo crimper... I usually buy higher quality tools used. A used tool which is adjustable is far more valuable than a new tool which is not*

... I have a hydraulic crimper on order. In the EV industry we use tool-pack-powered crimpers! These dominate!

Just imagine an 18 volt Ryobi-backed crimper. Goes real fast, is real light, crimps like crazy. I was first introduced to them by an old dog at Tesla... or maybe ChargePoint.

I used to work at Tesla and ChargePoint - that's where I learned to respect high voltage - it is not a joking matter and it will kill you quickly.

Please be very conscientious when you work with these higher potentials. I can care less if you get electrocuted with 240 volts AC. I'll be very sad if you touch the 420 volt wires.

Me?
Prefer a 420V battery to a 42V battery.... But that's in a whole different League.

-methods

 

[methods]

HIGH VOLTAGE SAFETY

Somebody always dies. . . so. . . here is another reminder that we are dealing with HVDC here

* High
* Voltage
* Direct
* Current

I take it VERY SERIOUSLY. . . . -----> You should too. That is the only "Captain Safety" hard time that I will repeat over and over.

* DONT
* GET
* ELECTROCUTED! (to death)

-methods

 

[methods]

And FWIW -> Those are not a "legit" setup. That is a "poor mans" setup. The actual gloves that go over those rubber mitts hit a spec and are probably much safer. They are expensive tho... and I am a realist... so I ask that people AT LEAST buy the minimum.

In a super bad pinch
* Any insulating rubber under leather

Just remember that materials change behavior at higher voltages. Some things will stand off 200V and break down at 1,200V... so remember that. Perform a "HiPot" test on anything you are going to use. I have a 600V power supply that I test things against.

I bought that supply used for pennies on the dollar, replaced the fan, replaced the AC caps, replaced the output caps. Above is a video showing what happens when you over-volt an electrolytic capacitor.

More smoke than you thought - eh?

That was how we celebrated 4th of July under the oppression of the local governments ban on celebration.... and speaking of which... I am building OFF GRID because sometimes Government is lagging Engineering by a Decade or Two. Eh hem.

-methods

 

[methods]

Notes

Read that I don't need blocking diodes to parallel panel arrays. It's purpose was originally to block the battery from discharging into the panels.

I don't understand why one sunny panel won't discharge into a dark panel, but I accept it.

...

I read that all modern panels have bypass diodes Incorporated in a small black box in the back. I read that some people add additional diodes for extreme cases across the entire panel.

....

I realized that I never explicitly stated that a primary goal is to shuttle data to a private cloud database and not just to a third party database.

We can do both

If we wanted to use a third party... Then we can just use an Android hotspot. If we want to intercept, do we need a "man in the middle".

I use Arduino, 99% of other people use pi.

....

All the parts have landed except for the panels, I pooched that pickup.

....

Now to read the 5000 ES manual


-Schindler

 

[methods]

I was able to clear the TDC 531D (Cheapy DC Clamp Meter that I got from Signature Solar) for basic DC current measurements. I did not properly evaluate it but I ran it next to a CL800 and they tracked better than expected. You can see a longer integration period on the cheaper meter, but that is totally normal.



What you pay for in a really fancy meter is "faster settling and better noise rejection". You can observe this by taking a noisy resistance measurement with a Fluke 87V and an Auto-Parts-Store meter... higher end meters will settle faster and reject more noise.

Anyway
All we are doing is squaring away our tools, getting baselines, assessing materials, etc. I always confirm everything I have before I go and rig up a hundred amps or 420V. If I am going to rig up BOTH at the same time (lol - POWER, as opposed to HV or HC) then I really slow down and make sure I have a handle on things.

...

Current Meter - Field Check
We always do a "Sanity Check" with a meter before putting it into service. Usually this just means running the meter next to another meter to look for:

* Gross Accuracy
* Simple Precision
* Non-Linearity
* Ghost Readings

What this wont tell us is how the meter will perform at the edges of the envelope

* Cold Weather
* Hot Weather
* After getting wet
* Against high electrical noise
* Against fast moving signals (how does it integrate those?)

You may find that when you introduce a lot of noise that meters will integrate and display that differently, and ... that may not be linear. ... Anyhow, the meter works fine enough for Government Work. It NEEDS a white dot to indicate pointer direction.

...
...
In other news I scraped up a desktop system for doing basic checks.

* Whistler 1200W, 12V Power Inverter
* 4S Life 40Ah 1C pack
* 1500W house heater

This will suffice for spot checks, sanity checks, demonstrations, etc. I went ahead and got a head start on looking at some non-sinusoidal inverter signals on the scope and ... that is a part of a later test where we will look at the actual (measured) impact of square waves on full bridge rectifiers backed by advanced PFC. The readers should note that I have been running DC into "wall warts" and AC-DC power supplies for a long time. If you understand how a Full Bridge Rectum Fryer works then you understand that most AC-DC converters could not care less about Sinwave vs Squarewave vs DC input. It all gets settled out by the DC-DC in the PFC (for another time)

...
On the way there...

Example of Latent Failure
Here I am using a $42, 1.2KW cheapy-inverter as part of an investigation into Square Wave vs Sinusoidal output. During the test a kissing contact blew out, and that... is always good to review. Plenty of folks can wire something up reliably. Part of my job (in the real world) is inducing failure to assess response. THAT... can be a bit tricky at times.

...

I see that things are pretty quiet so I am going to carry on. I usually pick one thread and dump everything there. Eventually this will build up into my stated topic. Sometimes I get criticism for going a bit off topic - BUT - I.... believe all of this is on topic. Before we can leapfrog into hacking WIFI we have to get a lot of little things verified. I always cover that long run-up... as... we dont want to confuse the new guys into thinking this is a weekend job.

Setting up a 6KW system that throws around 420V and pumps 100A into and out of large battery banks is... a non-trivial exercise.

-methods


P.S. Regarding the Review
So far, so good.... Folks get bombarded these days with over-produced and technically-underwhelming reviews... so when I dig in and do an in-depth review it can throw folks off (expectations). I don't perform my testing (or hacking) in a traditional way. I always do it down in some far corner of a forum (out of sight) and I take my sweet time. In the end... I will detail my failings... so that if you follow the same or similar path... then you should have a pretty good idea of how things will roll out.

 

[methods]

I picked up the 10pcs of damaged Solar Panels (pennies on the dollar) and now it is time to set up my "Solar Dyno".


(Displaying in Watts per Meter Squared)


From some LED work I did a while back I remember that meters like this can be fussy. If you really want to get a good measure you need to repeat the test on some kind of grid - 4 measurements - 42 measurements - 420 measurements... you want to make sure that you dont have hot spots and cold spots in your Solar Dyno.*

I am going to start building mine using "Grow Lights". There are plenty of failed entrepreneurs who offload very high power lights on CraigsList. As everyone moved over to LED, the traditional (especially transformer driven) lights have fallen out of favor.

PERFECT!

* Set up the table
* Set up the light with adjustable height
* Set up a proper diffuser <-------
* Set up the blackout curtain
* Prove the setup in

Like any Dyno -
This wont be an "Absolute Measure" where you and I can compare numbers. What it WILL be is a "Relative Measure" (for me) so I can start to compare the impacts of changes I make to the panels.

* Panel as delivered
* Panel with a sheet of plastic over it
* Panel with a coat of Silicone over it
* Panel with a coat of RainX over it
* Panel with a coat of Epoxy over it
* Panel with a piece of Lexan over it
* Panel painted by roller, brush, sprayer, etc....


... I have 10 panels to assess and experiment on ->
AND
This is where we are going to need that WiFI data coming out of the Growatt. Of course I would prefer to test panels 1 at a time, but the tool I have right now in front of me is a 5000ES -> So that is the tool I will adapt to my needs.... while I look for a single panel MPPT used.

STEP 1
* Put 4 of the panels up on the roof
* Wire them in Series, bring the two lines down
* Jimmy-Hack a "First Proof" on the 5000ES

Before I go building a Dyno, I need to at least make sure I am working with SOMETHING that is worth fooling with. So...



I have 10 of those in damaged condition for Experimentation. In hand, at this moment.

... Now we are going to start getting serious.

PMAX (MPPT)
* I see an AbsMax of 460W
* I pre-accept a ~10% loss to call that 420W (in perfect conditions)

Ok, so if somehow we have a perfect sun day here, my target would be on the order of 1,680W. Notice that I pre-calculate the 10% loss in right off the top so I dont have to keep re-calculating.

VMAX (Open Circuit)
* I see an AbsMax of 51.7V +/-3% = 50.15V (assuming worst case of the 3%)
* I pre-accept a value as low as XX.XX (TBD)

IMAX (Short Circuit)
* I see an AbsMax 11.5 +/-4% = 11A (assuming worst case of the 4%)

Ok -
So what are our working numbers for a passive load?

43.08V * 10.68A = 460W

Ok -
Lets turn those into numbers I can remember

42V * 10A = 420W <---------------------------

420 is a magic number that I (and others) have worked into many products which are all around you. If you are looking for a keystone number, try 420. Try it in Octal, Hexidecimal, Binary, ... in nibbles, bytes, BCD, ASCII etc. (eh hem)

So...
In order to build a SUPER JENKY "quick test" load for panels that will give me more than just Open Circuit Voltage and Short Circuit Current then I would need something about like this:

Calculation for a Passive Load

W = V*I
W = I^2*R
W = V^2/R

Three formulas... two that include R (resistance in ohms)

R = 42^2 / 420
R = 420 / 10^2

Right? Basic rules of the equation.

(42 * 42) / 420 = 4.2
420 / (10 * 10) = 4.2

So we are looking for a 4.2ohm resistor capable of dumping 420W. THAT... is a big-daddy of a resistor and THAT>... is why it is easier to just just an MPPT and an arbitrary load. For Arbitrary Load we want

CONSTANT POWER

* Constant Current
* Constant Voltage
* Constant Power

Remember that... CONSTANT POWER for our Test Load... and one of the best ways to do it is thru a conversion. I plug in a standard appliance like a Hot Air Gun, Heater, etc. I do this via an inverter (1.2KW) and that inverter takes in 12ish volts. I create that 12ish volts with a small MPPT (or I am about to) and THAT... is where we are at.

* 1 panel we need at least 500W
* 2 panel we need at least 1KW
* 3 panel we need at least 1.5KW

OK - moving right along. We will omit any tests involving a 4.2ohm resistor that is 420 watts because it is "tarded". I may be able to find just the right thing if I dug thru the ol junk pile... but... using an MPPT into an inverter into a fixed load sure sounds better. Resistors change temperature with heat unless they are Nichrome or otherwise build of temperature stable materials. This wont create constant power.

OH
That is what we have - right?
We have a big MPPT box that will bring in arbitrary panels and directly convert them to AC 240V


Ok
Let me get to it. The coffee is starting to kick in and braincells are starting to fire.

(I just wake up and start writing)

-methods

 

[methods]

First light, first test.

Now all I have to do is perform accelerated aging for the ingress requirement*

Get ready to see pressure washers, submersion, extreme flexing....

... Now we're starting to do some test engineering...

-methods

 

[methods]

Half way thru testing.
Doing load tests today on Battery, Inverter, Charger.

First V-limits then I-limits then thermal

 

[shavermcspud]

I have written a node red flow that does this automatically and works with growattt and solis , i have a disk image for raspberry pi that uses an rs485 to usb, stores the data locally in influxDB and outputs to MQQT and grafana, you could easily use this to send small packet of important information via GPRS or 3G with a data sim

 

[methods]

MPPT Validation - Basic Power Maximizing

MPPT Validation - Hysteresis and Settling Time

Getting on with it now!

Power Supply was only $420 broken. Bad Primary caps, no output caps, no fan. It did have the GPIB card... Which I have if anyone is looking for one*

It cost me about a hundred bucks (and 10 hours) to fully restore the supply to functioning condition. It was sold as... "For Parts" IIRC.

... WORKS PERFECT FOR THIS SORT OF VALIDATION

Sorensen DCS series
* 60V 18A
* 600V 1.7A

Takes 120V or 240V input. Provides a kilowatt rail to rail. Produces 600V @ 1.88A off of a NEMA 5-15 outlet


-methods

 

[methods]

(Sorry to go so far off topic... I'm literally doing this for the first time and I have to get everything set up. Figured I'd lay some tracks in the snow)

Today is the big day.

20 (no 30!) Years in the making.

-methods

 

[methods]

Frozen Over... You definitely don't want ice in the cracks! Ice expands!

So far I've been able to push one kilowatt hour into the damaged pack. LiFe really is a far more forgiving chemistry than LiPo.

-methods

 

[methods]

GOT IT

My MPPT current could not converge on my short circuit current because 168V is too close to the 160V lower limit.

Short circuit was 7A (suboptimal angle) but MPPT never took more than 6A.

Given margin for error, limits, etc... I was basically banging off the lower limit. That was 4x50 volt panels that put out max power around 42 volts.

For the setup I need five panels!

Tomorrow we're testing with six and possibly seven if I get one on the roof of the truck

 

[methods]

Complete (enough)

6 Panels running out in the bush. Polish to be documented over the holiday.

-methods

 

[methods]

Over 6Mb/day for a $1/day

Pretty good.

(200Mb/30 days)/24hrs = 278KB/hr
Note that bites not bits!
So multiply it by 10 if you're comparing....

4.6kB/min ought to be enough. You don't get the iot services, but who needs them. You pay up front and have no overages.

You'll have to parse your texts to set a flag for data overage. It's super easy to text and read texts from iot.

-methods

 

[methods]

We are trying to assess the cost and viability of distributed solar.

For practical reasons most people go Monolithic with their install. Yea there is copper/conduit-cost... but far more back-breaking is the practical action of trenching your electrical connections.

* Every root (even 1/4" thick)
* Every rock (even 3/4" wide)
* Every soil type, full of lessons

Where we live there are a lot of trees, roads, obstacles, and fire hazards. Since we just got burned out last year by the CZU fires. . . we are especially interested in installing modular solar "stations" that can work Stand Alone or Networked.... that are far-clear of combustibles.

To connect these modules means trenching hard-packed rock. The first foot down into the earth is really tough on the back, but after that it is basically soft sand. Using a tool like the one shown above makes an impossible job possible! The dude operating that (robot) has decades of experience. He was doing this when I graduated from Highschool in the 90's.

* Super Cool guy
* Disciplined Artist
* Through and thoughtful. . .

You really have to have a vision when you start in on a project like this. Even more than, that you have to drive HARD to get MVP (Minimum Viable Product) in place to start racking your learnings. More Will Be Revealed . . . is my motto in the Startup World.

I am a fan of the Bob Ross approach...



Just one brush stroke at a time. . . easy as that. If you dont start you will never finish. If you dont like it... let it dry... and paint right over it.

...

* Got to get into tight spaces
* Got to get in and out
* Got to get done fast

Sure - Maybe some people can bring in a 40,000lb tractor and rip 3' into the earth in a single pass while auto-feeding ground burial wire. What we are showing here is a practical method that can be rented right down at the Heavy Rental shop. You can rent anything from a 40' pole truck, to a 3/4" mainline auto-feed, to a trailer, to a tractor.

Gotta be practical with your time!

... so
I gotta say...

I did all the Electrical but as far as "Getting the Job Done" (our Prototype) the lions share of the credit goes to another man. He knocked it out of the park by quickly knocking together:

* A pallet-mobile base station
* A tractor-mobile array stand

Over a 2 day period we went from nothing but me fiddling on the desk and the roof of my house... to an install that is collecting data. Super awesome. Most importantly he met our core requirements.

* Must knock together over a weekend
* Must knock back apart if required
* Must be 4-hr mobile

What this means is that we can pack all this crap up on a flatbed trailer between the end of the day and dark. . . or we can move it around as needed. THIS -> Is fundamental to what I am trying to prove. Mobile Stationary Storage.

* We have a requirement to produce 7KW for 6hrs a day
* That is 42KWH of "work"
* That must be covered sun or shade...

So a Minimum of a 42KWH battery here. MINIMUM...

Yes
We can at any time split the load between the battery and the sun... but at the end of the day... we have to be able to push 7KW for 6hrs straight.

* I figure that will require 42KW of Solar
* On a 10% day we will be pulling down 4.2KW
* 7KW - 4.2KW = 2.8KW
* 42KWH / 2.8KW = 15hrs = 2 days

So with a 42KW array of panels backed by a 42KWh battery we should be able to run 2 days straight starting from a full battery and 10% light hitting the panels.



Just did that math - does not really matter. More will be revealed ... as we forge ahead.

* Golf Carts hold 4.8KWH (4pcs of 12V 100Ah lithium)
* Broken down EV holds 48KWH

Lots of batteries to fill. Eventually we will get to 420KWh - and - at that point we will have met my goals.

-methods