Ann: Redflow enters Voluntary Administration, page-111

What the actual issue is? No. All I can do is provide details of what I saw: the battery was incapable of sustaining the rated power output at the required voltage.

To explain a bit of technical detail: power is equal to voltage, times the current (in amps). So you can have 50 volts and 20 amps, to give 1 kilowatt of power (50*20 = 1000). Or 100 volts and 10 amps is also 1 kilowatt.

Now, any battery has what's known as internal resistance: the current has to go through internal circuitry before it reaches the terminal and presents itself to the outside world. Another relevant equation here is that voltage equals current times resistance. Meaning that as the current goes through that internal resistance, you will see a drop in voltage on the outside. If there's no current, you'll not see that voltage drop; that's called the "open circuit" voltage. As soon as a current starts to flow, you'll see that open circuit voltage drop to a lower value. How much of a drop depends on both the internal resistance, and how much current is flowing. The higher the internal resistance, the more the voltage drops. The more current that's flowing, the more the voltage drops.

The internal resistance of the ZBM is pretty high, so as the current ramps up, the external voltage drops, quite a bit. If it drops below some critical value, the outside power circuitry assumes that it can't handle the load, and drops it on the floor. Which brings the external voltage back up above the threshold, and the dance repeats: supplying power, not supplying power, supplying power, not supplying power.

This is what I saw with my unit. I deliberately set up code to direct the maintenance discharge into the car (by turning on the EV charger when maintenance started), and I saw the battery supplying 1.5 kW ... nothing ... 1.5 kW ... nothing ... 1.5 kW ... To put that into perspective, 1.5 kW at a nominal 50 volts is 30 amps. If the internal resistance is one tenth of an ohm, that's a three volt drop, which pulls the open circuit voltage of 50 volts down to a "real world" voltage of 47 volts, and the power to 1.4 kW - so the current needs to go up to 32 amps to compensate. (I'm pulling figures out of my backside here; I don't know what the internal resistance of a ZBM is offhand. I'm just trying to illustrate the maths.)

Now, the nominal cell voltage of a zinc bromide cell is 1.8 volts. To get the 48 volts that the ZBM is supposed to (nominally) be at, you need to stack several cells in series: 48 / 1.8 = 26.67 cells. Call it 27 cells, for 48.6 volts. Maybe 28 cells. Don't know. If several of those cells fail and just pass current through without adding any voltage, you'll see a drop in the voltage on the output, and that - combined with the internal resistance - would be sufficient to explain the whole thing going down the way mine did.

Maybe that was what was going on: internal cells failing at a faster rate than expected, causing the whole unit to go pear shaped. We don't know. Redflow wasn't talking, and the fact that they weren't talking - and that what they claimed to be the issue didn't align with the symptoms I was seeing - was one of the factors that pushed me to get out. Internally, Redflow knew far more than we do. I suspect they didn't know root causes - if they knew root causes, they would have been targeting a fix for it, and they would have been making it much more clear to shareholders than they did. ("Yes, we're seeing a lot of failures, but we know why - here's the details - and we have a fix coming for it." That sort of messaging would have been reason to give them a better chance at the CRs than they actually got. But without that messaging? Nope. Not putting more money into a black pit.)

Ultimately, we only have guesses. There's nowhere near enough technical information to make an informed comment. Installers might know, but I suspect that they were also mushrooms. (Fed on BS and kept in the dark.) Which, for a startup that's struggling to put together a working product, is a death knell.