Why the BMW i3-REx

No, just DC. There are some L2 (240 VAC) in a locked parking lot about a block away.

Bob Wilson

Second visit:

3:34 PM @37.8 kW
​

​

• 63.5% = (86.5% - 23.0% ) # Change in battery SOC
• 11.9 kWh = 18.8 kWh * 63.5% # Charge put into battery
• 12.9 kWh = 11.9 kWh / 92% # Estimated input power to DC charger
• $1.29 = 12.9 kWh * $0.10 / kWh # Estimated cost to charge
• ~2.1% SOC / minute # Measured charge rate, need full record!

No parking meter enforcement after hours and weekend.

​

Bob Wilson

I may have found a way to more than double the REX gas range by installing a second tank. Actually a BMW i3-REX tank from a salvage.

Bob Wilson

Finally got a good video capture of the two charging sessions. I combined them to generate the following charts:





Once the pack voltage reaches 394 VDC, it switches from a constant current limited by the charging station of a tapering current. The voltage remains constant (to the limit of the charging station read-out.) It may be that a fractional voltage increase in the 394 VDC range induces a current reduction.

There were two, 30 minute sessions, and with the second one started, the BMW i3-REX Battery Management System (BMS) started out soft. Then eventually the current resumed the tapering curve seen before. I suspect this is driven by the LiON battery versus a hard-coded function. Anything other than an active feedback mechanism would be nuts.

To run the SOC down, I started at 97% because of pre-conditioning and took a badly focused shot of the trip meter at ~5%. The tripmeter showed 61 miles @3.5 mi/kWh ~=2.45 kWh which suggests I included some REX charge. I'll have to repeat the tripmeter kWh test.

For those interested in the raw data: http://hiwaay.net/~bzwilson/BMW/CSS_200.mov

Bob Wilson

EXECUTIVE SUMMARY

With the BMW i3-REX, charging consists of two distinct phases: (1) fixed rate, bulk charge limited by the power source, and (2) exponentially decreasing rate, taper charge. The power source limit can be either the DC electronics or the on-board AC-to-DC charger. In either case, car will absorb that rate of charge until the 'taper limit' charge level is reached.

The taper region, charge adds an addition charge ranging from ~20% for a high-power DC charger to 3% for a low-power, L1 EVSE (12A @120 VAC). So we are faced with trade-offs:

1. Expensive fast DC charger can spend as much as half its time putting the taper charge in.
2. Cheap, L1 EVSE has very small taper time but the bulk charging is orders of magnitude longer.

MATH MODEL

As a first approximation, it looks as if the charger limited rate and battery taper, kW, are given by:

1. y_ kW = kW_limit # charger limit
2. y_kW = (kW_limit) * e ** (-0.096 * x_sec) # taper limit

One complication is the taper trigger is given by the exponentially decreasing rate of charge curve. So as the charge power increases, it has the effect of moving the taper charge limit to the left, shortening how long the peak charge can go in. This means the expensive, high-power charger spends more of its time tied up with the less efficient taper charge.

CONCLUSION

It makes little sense to increase the home charger capacity if the car is just going to 'hurry up and wait' on the taper charge. By the same token, the L1 EVSE that comes with the car is so over subscribed that the taper charge is vanishingly small. Worse, it becomes the daily range limit unless augmented by more powerful chargers.

There is probably an optimum power limit for the BMW i3-REX based upon one's requirements that probably exceeds the L1 EVSE. But it isn't clear that the 31 A @240 VAC, L2 EVSE (the 6.6 kw/7.2 kW) is the optimum either. What is clear is building out a 12 kW charger is such overkill that unless one has a significant hit in their ordinary driving, it makes no sense as the car will be spending as much time in taper charge as bulk load.

I know this is complex and I've probably not explained it well so let's use an analogy:

• High speed, high power DC charger - can look like someone who floors the accelerator to reach the next light and wait.
• Low speed, L1 EVSE charger - looks like someone creeping along watching multiple traffic lights change before they arrive.

We want the home charger to be the crafty driver who moderates their speed so every time they arrive at the light, it turns green.

Bob Wilson

 

Earlier today I realized solving the problem in reverse simplifies everything. In effect, the 'taper' is reversed to become an exponentially increasing value until it reaches the supply limit. Then it flat-lines until the area under the curve, the integral, matches the battery capacity. This simplifies making a family of charging curves. So simple and obvious.

Oh well, something to do Thursday.

Bob Wilson

Ok, I used the model to develop this table:

	kW	time	description
1	1.2	15:17	L1 12A @120VAC
2	2.6	08:06	L1 12A @240VAC
3	4.4	04:46	L1.5 20A @240VAC
4	6.8	03:19	L2 31A @240VAC
5	12.0	01:59	DC 12kW
6	40.0	00:58	DC 40kW

So this pretty well defines my end game. The high-speed DC charger is nice but does not support any reasonable cross-country drive.

The L2 EVSE at home means the car is pretty much ready whenever I want. The L1.5 EVSE looks to do a good job too and is portable, replacing the L1 EVSE that came with the car.

The L1 EVSE is pretty weak. On 120 VAC, it takes a long time, longer than a sleep over. On 240 VAC (not tested, yet) it is barely a 'sleep-over' rate. It is more of a 'topper' than a useful charger.

Bob Wilson

 

I'm thinking the fast DC chargers work best co-located with taxi and delivery services:

• airport
• entertainment district
• medical lab
• delivery service
  • post office
  • UPS
  • Fedex

Higher power means shorter charge times with rapid turnover. Properly located, these can charge concurrent with loading, delivery, or waiting times.

At one time, I was thinking perhaps one located midway between Huntsville and Nashville. But the block-times, the net mph do not make sense. Better to just fire up the engine and keep on making miles.

Bob Wilson

 

My thinking is multiple charging cables with a simple switching mechanism between the electronics and the cables. In an ideal configuration, charger in the middle with four parking places and four charging cables. You find an empty parking place; plug-in the cable, and; an estimate for when charging begins. With four parking places and cables, 15 minutes each and then cycle to the next. If no car is connected, continue to poll each position until a car arrives and plugs in.

Bob Wilson

 

Improved version: http://hiwaay.net/~bzwilson/BMW/CSS_200.mov

Bob Wilson

FYI, the local, fast DC charger can bring the car up to 97% +/- 0.5%. The built-in AC-to-DC chargers can bring the car to 100%. Given 1% ~= 1 mile, this can be significant, especially if one wants to set a personal best, EV range record:

The 103 km (64 miles) would have come from one of the maximum discharge tests of the local fast DC charger. The App claims that range is 196 km (122 mi) but my model reports significantly further, ~180 mi (~288 km).

Bob Wilson

 

How would we tell?

Bob Wilson

 

Actually I like to do it in cool weather to avoid battery cooling overhead:

• discharge test - record the change of SOC times the expected kWh capacity compared to the tripmeter total miles divided by the 'mi/kWh'. Need to use the widest SOC % to handle the lack of precision.
• charge test DC - what I've done with the fast DC charger. Record the current * voltage for each minute and convert to War. Sum the results and compare to the rated capacity adjusted for the change in SOC %
• charge test AC - same as the DC charger but with a 92% efficiency estimate for the in-car AC-to-DC chargers.

I am happy to report all three methods give the similar results greater than 18.8 kWh. Not bad for a 2014 BMW i3-REX now with 17,000 miles (bought with 6,440 miles.)

Bob Wilson

 

Had a little fun tonight:


Bob Wilson

 

December 9:

I was stopped at a traffic light, Four Mile Post Road and 6500 Whitesburg Dr, Huntsville when it changed to green. I moderately pushed the accelerator; there was a 'clunk' and; no motion. Power cycling the car, same thing. I set the emergency flashers and shifted into "N". A slight grade let me turn into a parking lot and coast to a parking place. I put it in "P" and tried again and it cycled through a series of error messages, stayed in "P", and shut itself off. So I called BMW Connect.

It was 4:00 PM Friday afternoon and after an hour and a half, no luck getting a tow truck with the "tire skate", a plastic part that fit under the rear wheels so the car can be wrenched on a flat bed tow truck and taken to the dealer. By now the sun was down and it was headed to freezing temperatures. So we agreed to start again Saturday morning.

Various error messages seen:

• Drivetrain: Stop carefully
• Pedestrian Warning - Malfunction. Consult service center.
• Vehicle battery charge low - Charge by driving for longer periods use external charger. Function requiring battery will be switched off.
• iRemote reported 6i2% traction battery SOC

I have a video of each display during the aborted starts but it will take a while to reduce them to small, downloadable, mp4 files.

All of my tools and meters were at home but the symptoms are consistent with a 12V battery failure. Can't be sure until tomorrow when I go back to the car and take a peek. I'll also put a Scangauge on and see if there are any error codes to read out. But the primary goal on Saturday will be finding a tow truck company that can safely take the car to the dealership.

Good thing we're weren't mid-way to Arizona when this happened. Stuff happens so we deal with it and move on.

December 10:

Finding the 12V dead when I arrived this morning suggests a controller failed and effectively put a heavy load on the 12V so normal functions soon failed. I can't call it a failed 12V as another part may have been the excessive load.

The Saturday morning crew was much more clued and actually scheduled the work and sending an SMS message with the ticket:

The Friday evening the BMW rep told me she couldn't find a tow truck with 'tire skates' which are large plastic pieces that under the tires, allowing the car to be pulled onto a flat-bed. In reality, all that was needed is a tow truck with a tire, lift gate. Since it was 4:00 PM, it would not have been seen by the techs in time to diagnose, much less, repair. Still, it did mean a second trip to the car Saturday morning.

So I called the tow company and let them know the 12V is dead and the rear wheels are locked and can not be unlocked because of the dead battery. He showed up 20 minutes later and quickly got the car up and on the way to the dealer:



After securing the rear tires, the tow truck operator used the seat belt to 'tie' the steering wheel.



Calling the tow company was important because it let them know the actual condition of the car. So they sent the truck with the wheel lift gate. Unfortunately the magnetic, rear light bar would not attach to our carbon fiber hood so the operator put it against the base of the windshield ... and during the tow it fell down (there is a $500 fine if the bar is not there.) I probably should have laid a sweater or something under it to keep it from falling below the windshield.

No technicians on the weekend so I left the key and the VIN 7 digits with a sales clerk. Since I had a tire check that showed all but one at 5 mm with one at 6 mm and the right front tire alignment off, I'll have them do an alignment and won't worry if they also rotate them side-to-side. The tires have 10k service miles.

I'll share what they find but it looks like 'a week at the spa' for our BMW i3-REx.

LATE THOUGHT: It is important to test every aspect of a car during the warranty. The addition of a 100A, 40kW, charger two weeks ago allowed testing the fast DC charger. Other than the row of 'programming buttons,' I think everything has been tested. But there are two non-standard mods: (1) coding for 75% REx operation and (2) the 2" receiver hitch for my wife's wheel chair carrier. In theory, these pose a warranty risk . . . we'll see.

Bob Wilson

 

Not my first rodeo, a first generation vehicle like our 2003 Prius, last year manufacture. Also, Gen-3 first year which had the early brake pause. The 2014, December manufacture, BMW i3-REx is both first generation and first year, perhaps not my best choice but it was affordable.

So I'm sanguine about what happens and just fell back to our 2010 Prius, a first year of the Gen-3 Prius. Today I am glad I kept it.

Bob Wilson