I'm scared. Post-reconditioning the HVB overcharged and warbled.

Good catch RE: outliers.

I do have bad modules and weak modules.

To be clear, I'm only pack charging and module discharging, as of now.

If the recommendation is to test the discharger (it only discharges) on bad or weak modules first, I can do that.

Though it would be ideal if there is a recommended max discharge rate for a module in a pack down to 6V. Or even 6.3V.

My goal is to measure capacity, even if it's not entirely accurate, in a safe but fast way. I'm looking for data to inform pairing modules.

I'll keep digging through the forum and online for consensus.

If it helps to make this less confusing, let's separate the capacity testing process of 1 prolong charge & 1 (2A/3A/5A/8A/10A) discharge per module from a 3 cycle reconditioning process.

If this new process for capacity testing works (for the purpose of pairing modules in an otherwise healthy pack), and is fast, and is safe, it could be valuable information for future readers and I'll make a new thread documenting it clearly.

Edit: PriusCamper discharges with a 50w 12V halogen, down to what appears as low as 1V. That's ~4.2A. Sometimes they add a 20w bulb to the mix for a total of 5.84A. "which is super fast when modules are at full charge but potentially damaging below 5v"

I'm just going down to 6V or 6.3V if I want to be extra safe. Yes, I know that the capacity measurement going down to 6.3V will be less than actual. My thinking is that I'll still get a good enough idea of which modules to pair. Better than having no data at all and pairing them haphazardly. Some modules, I would think, even then, would show higher capacities, with some showing lower.

 

Half the pack is ebay modules at this point. I have no insight into the capacity of any of the modules as I bought the car used with a dead pack. I have 11 spares in condition ranging from poor to weak but potentially ok after reconditioning.

So I know (or at lease think I know) this much: the blocks aren't paired and failures are likely happening from poor pairing. I'm trying to address this and gather data. Most people are fine chucking a module in, and I could do that too as I've done before, but where's the challenge in that?

Do I have a few 3400mAh modules in there? Are they all 3xxxmAh? I'd like to know ballpark where I stand, because no amount of ebay modules will help the situation if I just keep making bad pairs.

Ideally, I can target 5500mAh across all the blocks. Identify which modules need more attention, more cycles, deeper cycles individually on the side to get them into parity.

But it could also be as simple of an optimization as knowing which modules to pair from the current pack, swapping only the bad module in block 5. No more reconditioning, for now, because I can get a smaller spread on the blocks with proper pairing.

 

Are you inferring that there is no benefit to having blocks with the same capacity in a pack? I was under the impression that this could extend the time between whack a mole events. And that a large part of why whack a mole events happen is because this intelligent pairing isn't taking place.

 

Noted.

I think what I'm going to try first is compare "usable" capacity of the pack modules going from 8.4V to 7.1V. (after a prolong topping charge, cool down, and trickle charging/discharging individual modules to 8.4V before the test)

Before doing that however, I'll see if the bad module in block 5 will even reach 8.4V on an individual trickle and establish baseline bad with the measurements for it. Then I'll see what the questionable other module in block 5 does. Then I'll see what a probably good module does.

Edit: DrPrius tested at 33.74%. I've gone up from 30% to 60% before, but that's still pretty bad.

 

Wouldn't DrPrius only show the capacity of the weakest block?

 

From what I understand, it pulse charges at high amps (up to more than 60A!, but mostly 30A) (hit the gas, wait, hit the gas, wait, etc) until the car won't let you do that anymore. Typically 75%-80% SoC, but less than that if the pack is in poor condition. It'll stop letting you pulse charge when the pack hits its maximum SoC. This is determined by when the first module (or maybe block) hits it's max SoC.

The test then discharges from that point at ~6.5A and times how long it takes to drop pack SoC to 40% (but really 39.4-39.9%).

Then the time is converted into a rough estimate of capacity by comparing to the measurement taken on a new pack.

So, if measuring usable capacity is good enough a metric for DrPrius on a pack level, I think it's good enough a metric to measure myself on a module level. Though, yes, as has been stated, I don't have a comparative new module measurement using my method to infer a true capacity %. But I will have relative measurements from the other modules which should be sufficient for pairing.

Actually, Jack would have that number and I should ask him what it is.

 

I emailed Jack, but I also asked ChatGPT (AI) to calculate usable capacity with the knowledge that the DrPrius discharge test ideally pulls 6.5A for 900 seconds on a new pack to reach a 100% capacity result.

The AI says...

The usable capacity between 80% and 40% on a new Prius NiMH battery discharging at 6.5A can be calculated as follows:

First, we need to find the discharge rate in Ampere-hours (Ah). The discharge rate can be found using the formula:

Discharge rate (Ah) = Current (A) x Time (h)

In this case, the current is 6.5A, and the time is 0.25 hours (15 minutes), so the discharge rate is:

Discharge rate = 6.5A x 0.25h = 1.625 Ah

Next, we need to find the total capacity of the battery. For this calculation, the total capacity of the battery is 6.5Ah.

Finally, we can calculate the usable capacity between 80% and 40% as follows:

Usable capacity = Total capacity x (80% - 40%) = 6.5Ah x (80% - 40%) = 6.5Ah x 40% = 2.6Ah

So, in this scenario, the usable capacity between 80% and 40% on a new Prius NiMH battery discharging at 6.5A is 2.6Ah.

 

I'll need to think about why DrPrius doesn't discharge 2.6Ah but rather 1.625Ah for the life expectancy test. I'll also look around for a DrPrius video where a new pack is tested. I have seen a few where packs test >110%, exceeding the 900 second timer.

 

Ok I tried again but asked the AI to factor in the operating range of voltages. This feels more directionally correct.

The usable capacity between 80% and 40% on a new Prius NiMH battery can be calculated as follows:

First, we need to find the energy that has been taken out of the battery in 15 minutes, given a discharge current of 6.5A. The energy in watt-hours (Wh) can be found using the formula:

Energy (Wh) = Voltage (V) x Current (A) x Time (h)

We can assume an average voltage between 8.4V and 7.2V, which is 7.8V, during the discharge period. Therefore, the energy taken out of the battery in 15 minutes is:

Energy = 7.8V x 6.5A x 0.25h = 12.57 Wh

Next, we can calculate the usable capacity between 80% and 40% as follows:

Usable capacity = Energy / Total capacity = 12.57Wh / 6.5Ah = 1.9Ah

So, in this scenario, the usable capacity between 80% and 40% on a new Prius NiMH battery discharging at 6.5A is 1.9Ah.

Well, dividing watt hours by amp hours to get amp hours seems kind of weird, and wrong. Disregard this. It's NOT directionally correct.

1.625Ah is the closest answer to being right thus far.

Update: Jack said "40% usable", but I'm following up to clarify if it's 1625mAh or 2600mAh.

 

I'm an extremely persistent startup founder with a tech generalist skillset. If I don't know how to do something, I learn how to do it. I take things from concept to functional product in the market, sometimes by myself if I don't have venture capital to hire others to work with me.

I wouldn't say I'm an engineer by trade, mechanical or software, but I think like an engineer.

When a relative breaks a fridge and mentions they are going to throw it out, I fix it for them so they don't. I very much like fixing things as well as data analysis. Systems optimization.

 

This is from a new cylindrical pack. DrPrius went past the 15 minute discharge timer by another 5 minutes. Complicating the measurement slightly is the fact that the amperage wasn't maintained at 6.5A near the end of the test.

 

Hehe.

My situation is that I bought the car for 1/2 Kelley Blue Book and 1/6th of MSRP. 8 years old and with 139k miles plus a dead pack at the time. The used car place sold it to me at cost without taking a profit because they didn't think anyone else would walk in wanting to buy a dead hybrid vehicle.

I've had good fun reconditioning the pack with prolong charger and discharger, both of which I bought used on ebay.

I've had good luck with the modules from ebay.

I bought some tools I didn't need to, like an impact wrench.

Altogether and including my most recent purchases of a hobby charger and a discharger with cables, I've still spent far less than the cost of a new pack. (To say nothing of the cost of the car, which I love. It's a CT200h.)

KISS in this case would be buying a new pack. Which is great if you can afford it. I develop free and open source software that might make money eventually from complimentary products. So Keep It Complicated Korey is what I must practice until then.

...And then I'm upgrading to the NexCells muahahaha.