Quantitative Results of Grid Charge/Discharge

NEW TEST OF A 2006 W/130K MILES

Car: 2008, 143K miles. Purchased and lived in the Phoenix, AZ area.
Original owner is a long time friend.

He complained of deteriorated mileage. He was accustomed to his wife getting worse mileage them him because of her driving style; however, her mileage had dipped from the low-40s to the mid-35s. When he made a concerted effort to get good mileage, about 42 was the best he could get.

I offered to take a look at his battery for him.

Techstream reported very consistent IR of .023-.024 and a typical min/max block voltage difference of 0.2V. Driving charge/discharge tests didn't show anything alarming. The only outward signs of an issue were the battery meter dropping pretty quickly and the reduced mileage.

I conducted a baseline test as follows:

Force charge to max SoC.
Drive to test location (about a mile)
Conduct discharge test via reverse and slipping the brake

Chart of output:

The Capacity was computed by the average current between each time point and represents the area under the red curve. The average current of all readings was 15.3A. The total capacity used was 625mAh. The range of SoC was then used to extrapolate the total capacity of the pack. In this case the SoC went from 73 to 38.5 or 34.5% total. So 625/.345 = 1813mAh extrapolated total capacity of the pack.. about 27.9% of the rated 6500mAh.

Following the Grid charge/discharge, I conducted another discharge test that I was unable to export for some reason, here is the Techstream chart:

It looked improved, but with a lot of tedious effort in transcribing the data, I couldn't quantify the results in the same way. Additionally, the engine restarted at just under 50% SoC, which led me to believe that the ECU did not yet have an accurate representation of SoC.

Drive 70 miles over two days

Repeat initial discharge test, except I do it much better. I'm able to maintain higher currents more consistently. Here's the chart:



As you can see after the first minute or so, the current climbs, and I'm able to maintain 35-45A discharge throughout most of the test. The average current for this test is 25.5A - significantly higher than the original.

Computed Capacity: 1,679mAh
Extrapolated Capacity: 5,895mAh or 90.7% of rated.

DeltaV also dropped from a high of 0.58V before to 0.38V after.

6 week update:

Repeated test today. Results added to table. Captured charge as well:

Discharge:


NOTE: the step in the SoC and drop in current was the result of the owner inadvertently terminating the recording. I immediately stopped the test and we resumed recording. When building the chart, I chopped out extraneous data for chart continuity. For those concerned, the step drop in SoC demonstrates that if anything, I have unrecorded used capacity and the test results represent values lower than actual (conservative).

Summary:

	Parameter	Initial	Final	6 wk 1800mi
1	Capacity used during test	625 mAh	1679 mAh	1462 mAh
2	Extrapolated capacity	1813 mAh	5895 mAh	3440 mAh
3	State of Health	27.9%	90.7%	52.9%
4	Max Block DeltaV under load	0.58V	0.38V	0.53V
5	Average Test Current	15.3A (38A Peak)	25.5A (50A Peak)	34.9A (18 sec 60-103A Peak)
6	Peak Test Current near end	38A	48A	44A
7	Block IR during test	27 - 28 mΩ	27 - 28 mΩ	23 - 24 mΩ

In addition, the performance feels great. I was able to easily cruise my neighborhood at speeds up to 30mph purely on electrics alone before the ICE kicked in (1.6mi). Mileage of 48mpg+ is attainable with only minimal efforts to "drive it like a hybrid." Getting less than 43 is tough even when I'm accelerating at full throttle to freeway speeds and deliberately avoiding regen (putting it in "N" when off-throttle and coming to a stop).

I will be returning the car to its owner tomorrow. I plan to repeat the discharge test in a few months for comparison.

Steve

CONCLUSION:

The benefits resulting from grid charge/discharge remain in effect. Extrapolated capacity and state of health are much lower than the test immediately after the grid charge. Two factors influence this are likely the inconsistent and increasingly aggressive discharge test (34.9A average vs. 25.5A average and significant peaks in excess of 100A vs. no peaks) and the car having a better established SoC over 1800 miles vs. computed SoC over 70 miles.

Real-world driving remains at the improved level with only minor deterioration likely more associated with a recently licensed 16 year old girl as the primary driver.

Friendship allowing, I will test again at the 6 month mark. I will make a greater effort to replicate the post-grid charge/discharge testing profile for a more useful comparison.

 

Thank you.

I have four Honda hybrids, so I've been mucking about with packs and sticks (modules) for a little over a year now. Hondas do not have anything as powerful as Techstream or even Torque PIDs available, though Peter Perkins' OBDIIC&C system is a very respectable solution (available only on G1 Insight & HCH1). Honda IMA does not have a reliable way of operating in electric-only mode, so even if better data collection could be had, comparative testing would be difficult.

I've seen real-world and seat-of-the-pants improvements from both grid charging/discharging and module-level work. The only way I have shown quantitative benefit (about 20-25% capacity improvement) was to compare before and after discharges to 1V/cell at about 4.5A (500W halogen bulb). Unfortunately, 4.5A is nothing compared to car-induced loads, so I can't relate that to in-car performance.

I was giddy with the prospect of being able to quantitatively demonstrate a before/after difference.

There is still variability to this as a human is required to make this test work, and it's technique dependant. The computed capacity and particularly the extrapolation are less than certain as 70 miles may be insufficient for the car to properly establish the full SoC range across various currents. That's why I plan to repeat the test in about 3 months. I believe normal driving over that period by my friend's wife should get the SoC range well flushed out (she just drives it with no consideration for its mileage potential).

I'll post the results here.

Steve

 

No affiliations of any kind. I am employed in the aviation industry, and it is my sole source of income.

While not affiliated in any way, I readily endorse Hybrid Automotive due to my familiarity with them from their long-time support of the Honda Insight community, but I receive no benefits of any kind. I do not personally know, nor have I met Jeff, but I have witnessed numerous cases of positive outcome with his customers at Insight Central: Honda Insight Forum over the past year.

 

is this version of charger is the same fro prius and honda if not do you have a Prius version?[/QUOTE]

My DIY version is Mike Dabrowski's version with the addition of a single APC-35-350 in series.

 

I forgot I had photos of the action:

Charging:



Discharging w/200W:

The LED voltmeter is something I added originally for Honda use. It doesn't include the voltage from the APC-35-350 because it's limited to 199V. I don't recommend it. As you can see, I rely on cheap ($5) multimeters for voltage and current readings.

Steve

 

The charger in the picture is an evolution/adaptation.

 

Concerning the results, my friend is thrilled to report his wife has seen the mileage climb to 48 with her normal driving. His wife was pleased as well, but was was most pleased with the light detailing job I did on the interior...

Steve

 

I absolutely do. Coming from the Honda IMA world where cell imbalance is more common, people are doing "preventative maintenance" on their packs. Frequency varies, but I have homed in on a grid charge every oil change, which is typically 2-4 times per year with accompanying discharging as required. However, that recommendation is made without an evaluation tool like Techstream. Honda IMA systems engage in very noticeable positive and negative recalibrations that increase in frequency as a pack deteriorates, so we have obvious warnings that we may need to do a grid charge/discharge sooner but no way to quantify pack status.

With Techstream one can assess the health of the pack. If one does not want to just go according to an arbitrary schedule, one could use Techstream to asses the extrapolated pack capacity against a known 6500mAh available capacity for a perfect battery. It probably makes sense to identify the "acceptable threshold" of 80% rated as a benchmark since we need to account for use. This is 5200mAh.

So, every oil change, assess your pack's health. Export techstream to excel/google sheets, compute mAh, and extrapolate pack capacity. If it's less than say 3120mAh (60% of 80% rating), conduct a charge/discharge. Note that this is an arbitrary number based on ratings and what I observed with the deteriorated pack.

EDIT: To heck with extrapolation. I think it's grossly inaccurate. I think it's more important to go off just the measured capacity alone. Given the results and the retention of capacity after 1800 miles, I would say below 1300mAh measured.

Operating the packs in a state of imbalance is what damages cells as it drives them outside the 40-80% SoC range where cycle deterioration accelerates. If we can reduce this imbalance, we can slow the deterioration and increase the useful life of the pack. This isn't a certainty, but it's definitely an optimization.

Steve