There is some nuance in how he explained that. Basically you have to go back and watch two prior videos that he linked via picture in picture, but the short and simple is that he buys dead laptop batteries and usually it is only one cell that failed - the rest are still viable. He has a lot of labor and equipment tied up in testing and separating the wheat from the chaff. That never factors in to his final. But in the end he gets to $300 by what he spent and how many viable cells he ended up with.
What he doesn't outline is that these are all heavily used cells and may have less than 50% of their life left. That means that almost immediately your 4.4kw battery is going to start degrading as cells die off. The odds of this wall making 3 kw in 4 years is slim. So you will fairly quickly find yourself in perpetual maintenance mode.
What he needs to do, or what we need to do, is make that wall with:
1. Replaceable fuses (the ones that he talked about that would pop when a battery failed).
2. An LED that indicates which battery / fuse had failed.
3. Individual replaceable batteries.
That way you don't have to unsolder the whole damn thing to replace three or four and maintaining your power wall is as easy as maintaining the TV Remote in your living room. Think about it. You get five or six that died off over last month and instead of shelling out thousands of dollars, you go down to Costco and buy a pack of 10 or 20 batteries for $60 and just pop them in. Making this an end-user serviceable product will open it up to a LOT of people and reduce the overall amount of cash that it takes to keep it operational. Because remember - all of the profitability and cost offset models assume the device is working at 100% - and we all know that no solar set up and no battery set up operates at 100% indefinitely. As a matter of fact degradation starts immediately and is measurable in the first year. Then five years down the road batteries are really starting to show it. Go ten years and you have to start considering large replacement. But if you had done small replacements all along, then you kept close to the 100% mark and you won't have another $3500 payment in 10 years.
I think Tesla tried to sue him, He referenced one video that he had to take down and I couldn't figure out why he spent so much money on the glass panel that said telsa. My guess is that they settled and basically he could use the name Tesla because it is already out there, but something in at least one of his videos went to far.
I also think that his fuses are the leftover leads from resistors that he used on another project.
I thought that the failure rate would be high and wanted to make a battery tray of sorts so they would be easily replaceable. The LED thing is brilliant though. What I haven't figure out is if you can rig an LED that tests one cell when it is wired in series or parallel. Also, even dead, these things might light up an led so I'm not sure how to determine if the battery is indeed not taking a charge. If you wired the led so that it only lights as current is going into the battery, it seems that might work. Your LED's would only show cells not taking a charge when they were being charged. That might work very well, because you wouldn't be testing the battery, but it's acceptance of current. That way you might find LEDs that were rated at the charging capacity of the batteries and therefore not need a resistor or worry of LEDs burning up. From what I can tell these cells push to much current for a LED to last very long.
This solution presents two problems, I think. First is that you might find that the stronger (or weaker) batteries charge first and suck more juice, therefore your LEDs might not all light up at the same time. I.E 12 batteries in a bank, apply charge, the ones that take a charge well (or are more discharged, therefore not as AHish, IDK) might light up the LED strongest at first and then taper off as they top off, resulting in other LEDs turning on, getting brighter as they start to charge. So you might have to observe which LEDs shine first or which shine last to find your dead cells. I'm not an electrical engineer so I have no idea if I'm barking up the right tree, yet.
The second problem is that LED are one-way. So if you charged them through an LED, you couldn't discharge them (I.E. use them for juice) through the same circuit (LED). So if this would work you'd have to have a charge circuit going to the battery and a load circuit coming out. This doesn't seem like a big problem, just more soldering, until you start charging while you are using them. Might not matter, but I don't know.
I definitely need to study that DC Electronics book I've had since I was 19 and been meaning to read.
I'm charging my test batteries and they are so far taking a charge. By that I mean the voltage is increasing and holding, at least so far. I'm using a hobby charger I bought for charging NiCad and NiMH batteries and the instructions were simply to charge at 10% of the AH rating. Back in the day, I'd charge a 500 mAH battery at 50mA until warm. This charger is putting out 18 volts and I'm using the 500mA setting, I've only had one battery get slightly warm. Mostly I've stopped charging once they get over 4 volts and some are getting there in an hour. I left one overnight and it reached 4.3 volts, has held it all day and wasn't appreciably hotter than body temp when I checked it in the morning.
I suspect that a lot of laptop batteries go bad because of the thermal sensor. Batteries are a safety liability and I think that the manufacturers are very conservative when it comes to overheating. I also think that most thermal sensors are crap.
If anybody has any ideas about putting these thing under a load, I'd love to hear it.
As for removability, so far all I've come up with is a tray with old-school copper/brass weatherstripping lining each side. Problem is how you you manage the groups in parallel into a series. Again, I need to read/draw it out...