ReInVolt PHEV Plan

INTRODUCTION

I recently bought a ReInVolt traction battery to replace a perfectly working, OEM NHW11 traction battery:

This photo shows a spare, traction battery cover on top of the original NHW11 pack. The red, ReInvolt pack is in the trunk.

Goals and Objectives:

1. Benchmark existing NHW11 traction battery after normalization drive: list 19 module pair rest voltages, all resistances, and temperature characteristics.
2. Replace with ReInvolt traction battery after normalization drive: list 19 module pair rest voltages, all resistances, and temperature characteristics.
3. Design PHEV system as mileage improver (the NHW11 had no native EV mode, yet.)

Background:

The growing domestic market in China has started to limit rare earth exports needed to make NiMH and other critical hybrid parts. In response, Toyota announced a serious battery recycling program in Japan in contrast to the 'eye wash' $200 bounty in the USA. I've helped with two incidents of owners whose local Toyota service center failed to honor the bounty. So I'm expecting a shortage of NHW20 modules and used NHW11 modules which will impact prices.

Based upon my earlier studies and hands-on, I've often recommended ReInvolt to replace failed traction batteries. But my NHW11 battery is no where near needing replacement. Now I can compare and contrast a working NHW11 battery versus the ReInvolt pack but the existing, perfectly healthy traction battery still has service life remaining.

My NHW11 traction battery is a prime candidate for a Plug-in Hybrid Electric Vehicle assist pack. All 19 modules are operable and the pack has no know defects. Retirement of the NHW11 pack to extra power storage is feasible without having to rebuild it. The original pack will provide additional power to reduce fuel consumption and the lower duty cycle will extend its life.

Mounting

I have a spare, traction battery pack cover that fits in the space above the existing traction battery:


To handle sudden stop or frontal impacts, I'll add a reenforcement bar in the open space between the seat back and the traction battery:

A plywood arch will provide additional support over the existing traction battery. Covered with a fiber-glass, epoxy, the part will be strong, light, and keep the upper pack from overloading the ReInvolt pack cover.

A new traction battery management circuit will fit in the space of the unusable OEM electronics. But one safety challenge is to provide short protection for the power wires leaving the case. The most likely location for the short circuit protection will be the raised areas under the shell . . . if there is space:

Note that protecting a 270 V, DC wire is tricky because unlike AC, there is no zero-crossing. The risk is trying to open the circuit will lead to a non-extinguishing arc. There are snubber and solid-state solutions but these are not trivial and have to be tested . . . without a runaway arc.

There is plenty of space for a battery management system operated by an MSP430, microcontroller:


This is another view showing the space available for the top-mounted, PHEV battery array:


Battery Management Specifications

• 30 A - maximum charge or discharge rate (L14-30 socket and plug compatible)
• 19 pair balance - the 19 module pairs will be shut regulated
• ~8,160 W, ~11 hp
• 45 C maximum charge temperature
• provisions for battery heater to maintain 5C minimum, may extend to ReInvolt battery (N. Alabama won't see much use)

Interface Electronics

The interface electronics will go through several versions:

A ) Sustainer - using diodes and current limiting resistors, the old NHW11 banks will be kept at the peak voltage of the operational, ReInvolt banks. This lets the NHW11 pack remain 'charged' while various designs are built and tested.

B ) AC charge - with an interlock, the NHW11 battery will be charged and the 19 module-pairs will be voltage balanced. It will include temperature monitoring. There will be a separate feed for an engine block heater and plug for emergency house power.

C ) Traction boost - under software control, send "n" amps with voltage limits to the ReInvolt pack. One approach to investigate is to shadow the discharge current from the ReInvolt pack. Since the NHW11 pack is independent of the ReInvolt pack, the charge and discharge limits can exceed the Toyota limits of 40-80%. Careful management allows a wider range of charge-discharge.

D ) Replace 12 V battery - helps offset the extra traction battery weight.

TESTING

I have a 10 mile commute each morning and on the weekends, I often take longer trips. So I'll use Auto Enginuity and Garmin to record the routes, speeds and fuel consumption. Then I'll take a highway loop from Huntsville along I-565, I-65 over the Tennessee River, Hartselle exit, and back to Huntsville. Finally, I'll do some hill climb tests that should warm the pack enough to force the fan to run:

a) baseline existing NHW11 pack with ReInvolt pack in trunk
b) Swap NHW11 pack and ReInvolt pack and repeat.

Thoughts? Suggestions?

Bob Wilson

 

Just adding another, independent set of data points to what we already know. I'm into 'facts and data,' not speculations.

Here are threads local to PriusChat:

• http://priuschat.com/forums/prius-h...ays-china-assures-us-rare-earth-elements.html
• http://priuschat.com/forums/prius-hybrid-news/86139-toyota-start-recycling-hybrid-batteries.html

That would waste resources better applied to improved products.

The best numbers I have are ~3% of the USA vehicle fleet are lost per year. It isn't totally linear as older cars, ~10-15 years, are more likely to fail from major mechanical but given our USA driving patterns, I suspect accidents are the chief risk. As for NHW11s, just under 54,000 were sold in the USA:

Prius Fatalities 2001-2007

The discredited Sudbury report led to analysis of nickel use in our Prius, at most, about 1% or less of the total output. Projected across all specialty metals, hybrid electric cars are a fraction of the materials market. In fact, between 2-3% of the USA market but compared to an ordinary car, a rare combination of these metals. When metal thieves begin stealing Prius to sell to recyclers, we will have 'arrived' at a bad place.

Curiosity.

Bob Wilson

 

Initial survey of NHW11 pack after driving about 8 miles and parking with the car in "N":

	Column 1	Column 2	Column 3
0	module pair	Volts	Resistance
1	1	15.23	0.022
2	2	15.13	0.022
3	3	15.14	0.023
4	4	15.12	0.022
5	5	15.11	0.022
6	6	15.2	0.022
7	7	15.08	0.021
8	8	15.16	0.023
9	9	15.15	0.023
10	10	15.12	0.023
11	11	15.16	0.024
12	12	15.16	0.022
13	13	15.14	0.022
14	14	15.13	0.023
15	15	15.25	0.023
16	16	15.17	0.022
17	17	15.17	0.021
18	18	15.22	0.021
19	19	15.21	0.022

Bob Wilson

 

This is for accident mitigation. The extra pack will not have a connection to the air bag system so in the event of an accident, it won't know to 'safe' the external power cables. I also like to know the cables are safe by seeing the switch is "OFF" and even a safety LED that shows no voltage on the wires. I'm a "belt-and-suspenders" guy when dealing with power electronics.

Bob Wilson

 

Funny you should mention this as that is on my agenda this evening. For the record, ~134,000 miles of service with the past 85,000 being in my hands (not always so gentle when testing!)

TEST PROTOCOL

1. Drive normally to area near Brindley Mountain 'hill climb'
2. Connect Auto Enginuity and record everything except resistances similar to earlier spreadsheet: all 19 module-pair voltages, current in/out, and four temperature probes. I will include the fan status.
3. Climb Brindley at 55 mph - a speed that the engine can handle without drawing significant charge from the traction battery.
4. Drive ~ 6 miles to make sure everything is normalized.
5. Descend Brindley at 55 mph on cruise control and in "D" using brake as needed if speed exceeds 65 mph (wicked curve at bottom.) This same profile fully charges my wife's ZVW30 so the 'pink' terminals show up.
6. Park nearby and shift into "N" to monitor voltages and temperatures for a 10-15 minute 'cool down.'
7. I may choose to do a series of hill climb test to 'pump' the traction battery temperature up high enough to get the fan to come on. This will let us measure how quickly the battery cools off in stock configuration as a function of fan speed.

Now my expectation is the original NHW11 traction battery with 38, 6.5 Ahr modules will have lower total energy storage than the new 38, 6.5 Ahr batteries. However, we won't really know until I swap the traction batteries.

Can you think of anything else we need to measure as baseline for the NHW11 traction battery performance?

Bob Wilson

 

Here is my initial data:

The chart shows:

• heat pumping - driving up at 55 mph and down the 525 ft. (160 m) Brindley Mountain charges and discharges the traction battery and raises the temperature. There were four descents followed by a short, forced charge to make sure the traction battery was at peak charge. The temperature change: 19 C -> 37 C, 18 C increase!
• Used "N" to put an ~1.6 A. discharge on the battery. However, the fan did not come on until the ambient air in the traction battery reached 26 C.
• Latent heat, the battery continued to warm-up, plateau, and finally starts to cool off. I measured the vent ~18 C while the metal around it was ~15 C.
• For a several minutes 3:14 to 3:20, I put the car in "R" to accelerate the discharge. Then back into "N" until it got too boring.

This isn't perfect data but we do have a temperature slope with the fan running at the lowest setting. Sad to say, this was done on Sunday, not Saturday. I'll have to swap the traction battery later this week.

Any questions, concerns or comments?

Thanks,
Bob Wilson

 

The black line is the current into and out of the traction battery pack. A positive value is current drawn from the pack, a discharge.

The "Ah capacity" normally means measuring the integrated amps (dot) time of the charge or discharge from peak charge, 100% SOC, to ~1 V/cell, the 0% charge. But our battery management ECU only allows 80% to 40%. To properly, and accurately measure the Ah capacity, I would need to have a separate "super" MRC 989 (aka., the Dept. of Energy approach) or test each module with an MRC 989, the "oldnoah" and my preferred approach. I did not try to do that in the chart in #9.

I am more interested in the dV and amps (dot) time on the four hill descents. These will be compared with the new ReInVolt pack to see if my expectation of higher capacity proves out . . . the ability to absorb more energy with the ReInVolt pack.

I will look closely at the dV vs current drain to see if it is possible to project Ah to 1 V/cell. If we assume the battery controller at peak the end of the forced charge event, ~02:51 PM, in the last major charge, we can divide that Ahr/.8 ~= pack Ah capacity. This is not perfect but based upon an assumption of linear discharge to 1 V . . . it should be close enough for our work.

Bob Wilson

 

Hummm:

15.25
15.08 -
-----
0.17 V

170 mV.

As I mentioned, there is nothing wrong with this pack and it is not even marginal. So it looks like a good candidate for a 'Plug-In' battery to the ReInVolt, the new primary pack.

Bob Wilson

 

Hi Patrick,

Based upon the measurements, I can give a range:

• 4.4-5.5 Ahr @1.3 A. :: the higher value assumes the battery controller accurately keeps the peak charge at 80% of the actual capacity.

After taking out the initial voltage drop when charging stops, I had 131 samples over 20 minutes and 31 seconds. The voltage per cell, 228 cells, dropped from 1.395 to 1.351 V with an average current of 1.28 A. This projects to 1309 samples to reach 1V.

If we use the starting sample as 100% charge, the linear projection shows 4.4 Ahr to 1.00 V/cell. If we call the starting sample as 80% charge, the capacity comes in at 5.5 Ahr. This is the most accurate projection I can come up with based up this set of data on Sunday at ~35 C.

I changed the focus to each of the 19 pairs:

	Column 1	Column 2	Column 3
0	Module	100% SOC	80% SOC
1	1	4.12	5.16
2	2	4.15	5.18
3	3	3.83	4.79
4	4	3.78	4.73
5	5	3.93	4.91
6	6	3.99	4.98
7	7	3.57	4.47
8	8	3.91	4.89
9	9	3.92	4.90
10	10	3.93	4.91
11	11	3.88	4.85
12	12	3.97	4.96
13	13	3.79	4.74
14	14	3.93	4.91
15	15	4.40	5.50
16	16	3.66	4.57
17	17	3.97	4.96
18	18	4.08	5.10
19	19	4.08	5.10

As expected, pair #7 is the weakest at 3.57 Ahr. Since it would reach 1 V before the others, the rest of the pack energy is unavailable. This becomes the usable limit and consistent with "oldnoah" reports.

The next question is the rate of temperature decline. The data shows:

• 24.9 minutes
• 19.6 min - fan low on
• 24.5 C - Inside Air Temp
• 3.7 C = (36.7 C - 33.0) average of probe 1, 2, and 3, starting and ending temperatures
• -0.15 C / minute, probes 1, 2, and 3
• 2 C = (32 C - 30 C) probe 4 starting and ending temperatures
• -0.08 C / minute, probe 4

I need to swap the battery and Thursday is a holiday.

Bob Wilson

 

Yippie!!!

Re-InVolt pack is in and the car started right up. It is lunch time and after a bite and a drive around, I'll be headed to Brindley Mountain for some data gathering.

YIPPIEEE!!!

I hired a local, independent shop to swap the battery:

LBJ's Automotive
12403 South Memorial Parkway
Huntsville, AL 35803
(256)880-4956
Cost: $187.50​

You' all need to understand I've already helped swap one battery . . . I've also replaced one clutch . . . one muffler . . . rebuilt an engine (VW 1500 cc) . . . shingled a roof . . . put up a fence . . . plowed a field . . . hauled hay . . . Marine bootcamp . . . There are somethings that after the first time, the 'newness' wears off and there is no need to repeat. Plus it is good to support your local businesses.

So if anyone lives in the North Alabama, South Tennessee area, I would throughly recommend LBJ's and a Re-InVolt pack.

Bob Wilson

 

Initial, Re-InVolt data:

	Column 1	Column 2	Column 3
0	Batt Blk	Volts	Ohms
1	1	15.85	0.021
2	2	15.80	0.021
3	3	15.78	0.021
4	4	15.81	0.020
5	5	15.79	0.021
6	6	15.81	0.022
7	7	15.82	0.022
8	8	15.78	0.023
9	9	15.79	0.022
10	10	15.79	0.022
11	11	15.85	0.022
12	12	15.79	0.023
13	13	15.80	0.023
14	14	15.82	0.022
15	15	15.83	0.023
16	16	15.80	0.023
17	17	15.77	0.021
18	18	15.81	0.022
19	19	15.81	0.023
20	min	15.77	0.020
21	max	15.85	0.023
22	diff	0.08	0.003

Re-InVolt pack, average of seven samples.

As for as the min/max voltages are concerned, I'm very happy. This is not a trivial problem as user 'oldnoah' shared. Still, it is half the difference we'd seen before. One note, the Re-InVolt pack was not installed as soon as delivered. This could have been nothing more than different rates of 'self-discharge.'

We know the end modules appear to be in healthier state than the middle modules. Re-InVolt may have surveyed failed packs that came in and could provide a better statistical analysis than my few samples. My few samples suggest a better rebuild strategy would cluster the strongest modules in the middle and the weakest towards the ends. The goal was best expressed in the poem, "The Bishop's One Hose Shey."

I am curious that the battery controller 'reported' internal resistance does not match the expected lower resistance of the NHW20 modules. However, this pack has yet to face serious stress tests, the limits of low and high SOC as well as charge and discharge currents.

In the next hour, I will be repeating my earlier 'hill climb' tests but they did not include worst-case, maximum current drain. However, they were very good about peak charging the traction battery, the most stressful test because it induces endothermic, charging.

NOTE: I have no intention of maximum rate, hill climb in the next set of data with the NHW11 pack back in. There are some tests that make no engineering sense.

Bob Wilson

 

Here is the raw data and a summary chart:

• four descents at 55 mph
• three ascents at 55 mph
• one maximum ascent - starting from dead stop, brake on, maximum acceleration, the indicated speed at the top was 78 mph (83 mph corrected.)
• some data loss events due to running a computer game during the battery cooling phase

More analysis is coming but folks are welcome to use this data.

Bob Wilson

 

The first comparison is 'hill descent' which imposes exothermic charging heat as well as internal resistance heating. Since this data includes both the original, NHW11 pack measured on November 7, and the ReInvolt pack measured on November 20, I'v appended "-07" for the November 7th readings and "-20" for the November 20th readings:

Because the runs were at different times of the day, I converted them into minutes and then offset segments of the ReInvolt run on November 20 to coincide with the NHW11 pack tests of November 7.

It looks like the NHW11 module battery (aka., "-07") heats at a faster rate and shows higher voltage during the charging from hill descents. This is consistent with a higher internal resistance of these 2003 Prius battery modules. The temperature was ~72F during the ReInvolt test and ~50F during the NHW11 test.

Bob Wilson

 

Two other aspects are capacity and cooling:



TEMPERATURE AND COOLING

The NHW11 pack was subject to four hill descents with four, 55 mph hill ascents. At 55 mph the NHW11 engine handles the climbing energy without a problem. Most of the heating comes from the exothermic, NiMH heating and current.

The ReInVolt pack had four hill descents and three, 55 mph hill ascents. But giving into temptation, I did a standing start, engine running, sprint to the crest peaking at 78 mph indicated, 84 mph true. That segment is not in this data but will be analyzed in the future to measure the kWhr. But this also put a serious I**R heat load on the battery and a higher temperature.

As for the rate of pack cooling, during the cooling sections of both packs under the same fan voltage, ~10 V., the cooling rates are very similar. What is curious is the lower fan voltage at the early part of the ReInvolt pack cool down seems to cool the pack faster. However, this could be another latent defect of the Auto Enginuity package. I'll need to repeat this test using a ScanGauge XGAUGE.

CAPACITY

I still have to calculate the Ahr using the slope of the discharge curve in the linear section. However, it is obvious that ReInVolt pack has a shallower slope, more capacity under the ~1.2 A. load (read from graphs: )

• -0.58 V/min - ReInvolt pack voltage loss per minute
• -0.80 V/min - NHW11 pack voltage loss per minute

The NHW11 data also includes a shift into "R" and back to "N". These transitions provide the raw data needed to calculate the internal resistance of the original pack. I have similar data for the ReInVolt pack.

Bob Wilson