Open Source Electric Vehicle Program (OSEV)
State-of-the-Art Electric Vehicle Conversion Blog

Listed in reverse chronological order. Click on thumbnails to view full size pictures.

SPRING 2011
===========

June 1
--------
Posted a simplified view of the traction batteries at focus_batt_layout.pdf.

May 26
--------
Eight volunteers met 6-9 PM with the car on a lift: three electronic engineers, two mechanical engineers, one software engineer and physicist, and two JC auto/alternative fuel instructors. Three had worked on this EV, two worked on their own EVs, and the other three have no hands-on EV experience. Many pictures were taken and observations made. The rear pack, mid pack and radiator assembly were removed to get a closer look at the components, then placed inside the car. 100s of pictures were taken. A few components were removed from the car and kept for further analysis. Car will be hauled away in one week or more. Key observations include:
1. Traction fuse state: front pack F12 and F16 open, mid pack F13 and F15 closed, rear pack F14 TBD.
2. HV load fuse state: all 3 fuses open, F7 DCDC, F8 heater which went to open connection, and F9 charger.
3. LV load fuse state: all 6 EV fuses closed in junction box, F1-F6. 2 Fuses in motor controller accessible due to hole burned in aluminum were closed. Stock fuses TBD. 12V battery missing and DCDC believed to be off so unlikely any will be blown.
4. Lots of arc damage on drivers side of front battery pack hold down. No arcing damage seen in mid or rear packs. This may be due to the difference in orientation, with gravity pulling the front hold down down on to the battery plates as the plastic melted.
5. The front pack still had its aluminum end pressure plates in tact and plastic case material was still mostly in tact, inferring lower heat. The center of the pack right above the motor had all of its plastic, but it was melted together. The motor may have been deflecting or sinking heat.
6. SS compression straps eroded in the front center of the rear pack and the rear center of the mid pack, and the aluminum end caps were melted away there, and all of the plastic battery case material melted away, inferring high heat.
7. Lots of heat damage to aluminum on passenger side of motor controller and junction box, and aluminum passenger sway bar link melted away. The least damage is in the front drivers side corner, where there also was no tire attached. It is the only place with paint, brake lines, wire insulation and unrelaxed spring in tact. The large temperature gradient across the junction box can be clearly seen.
8. Of the three tires that were on the car, the front passenger one has the most tire remnants so it seems the coldest. The rear passenger one had no tire remnants, which seems the hottest.
9. W12 joining the mid and rear pack on the passenger side was severed by heat.
10. There is a half inch hole in the rear hold down.
11. Rodent problems have been reported at the north end of the building where the offices are, but not in the shop which is towards the south end, 100'+ away.
12. Floor where car was during the fire, one bay to the south of where we were inspecting it, was spalled -- the top was broken from what looks like burning fluid.
13. One cell in the rear (location TBD) looked like the plates had been ripped. Upon closer inspection it looks like something had burned out from the middle of the battery, inferring a possible site where the fire started, but since these batteries are supposed to not burn on their own, it may have occured after the fire started.

To do on-site within a week:
1. Measure rear traction pack F14 resistance.
2. Measure stock LV fuse resistances.
3. Take picture of floor where car was.
4. Determine slope of floor where car was by rolling a ball bearing on the floor, to help determine if gravity helped the fire spread via burning liquid electrolyte.

To do remotely:
5. Post and caption pictures.
6. Take information on this blog and other notes and emails and update report to include key pictures, observations, failure hypotheses (more have been proposed), analysis of hypotheses vs. observations, tentative conclusions and recommended actions (more have been proposed), then post on NBEAA website, post links to it on NBEAA and EVDL lists, and ask for feedback. Then finalize report and write Current EVents article and do a press release.

May 24
--------
Posted interim report for AAA at osev_sota_interim_report.pdf. Scheduled time to view the car with volunteers Thursday 2-10 PM.

May 19
--------
Here is the initial failure analysis, to be followed up with a group of volunteers with some expertise in electric vehicles and failure analysis to visit the car once the shop cleaned up is completed and we can see the car and put it up on a lift:

Less likely causes:

1. It could have been due to overcharging, which is a known way to ignite any battery and has been inferred to be the cause of other recent LiFePO4 battery fires including Niel Young's LincVolt. It could be due to either via a misconfigured charger, BMS or BMS to charger control interface. But it was not likely for this fire since:
a. The car was unplugged (verified the week of March 28) and had been instructed to remain unplugged until the BMS problems were solved.
b. The 12V battery was removed, which disables the charger from starting, and it was removed because it was at 8V (it was a spent battery being temporarily used and was dying because it was not being charged by the DCDC converter yet), and the 12V battery was not noted to be hot when removed, and the 12V ignition jumper was not connected (verified week of March 28), so it couldn't have been charging after the 12V battery was removed.
c. The charger was configured correctly, so had it been plugged in it only would have overcharged if one cell was unbalanced and very high which is opposite of what is likely when cells sit.

2. It could be an exploding battery, but not likely, because the LiFePO4 chemistry is supposed to be thermally stable. If these cells are shorted internally or externally, they only get up to ~200C, vs. 450C to ignite. Some lithion ion batteries have been tested to heat up to almost 600C when shorted, which explains why some laptops have burned by themselves.

3. It could have been lightning, since coicidentally there was a lightning storm at the exact time of the start of the fire, and a very rare tornado had touched down earlier that day and destroyed a shed less than 2 miles away, and the car was resting on a grounded conductive lift, but it is not likely since:
a. Lighting very rarely strikes the ground in Santa Rosa, it is usually sheet lightning between clouds, and no known strikes are known of that night.
b. The car was several feet inside the building, not near the edges where lightning is expected to travel.
c. There was no other evidence of lightning travel inside the building.

More likely causes:

4. It could have been due to a load fire such as the motor controller, charger or junction box, since there is no firewall between these loads and the front battery pack; no firewall between the batteries and their loads should be considered to be addressed on all EVs regardless of the cause of this fire. The motor controller, charger and DCDC converter seem to be in tact, and the DCDC converter should have been disconnected when off since there is an internal relay on its input. But the junction box negative high voltage terminal is >2V/mm from the positive side fuses that go to the charger, DCDC converter and heater, as well as the feed of the motor controller; this uninsulated spacing should also be considered to be addressed on all EVs regardless of the cause of this fire. The junction box lid is flared open as if there was high heat in it, so it looks like it could have been a source of fire. It was enclosed and had no potentially volatile components near it such as circuit boards, but it is possible to be the source due to the current design.

5. It could have been an arc across the front batteries that did not clear due to the high voltage being exposed and up to 10X more than the 1V/mm maximum recommended spacing; battery terminal spacing should be considered to be addressed on all EVs regardless of the cause of this fire. The arc could have been started by either a failing component on one of the BMS boards, a wire or tool that rolled down on to the battery pack, or a wet rodent or one carrying a conductive piece of material underneath the rubber sheets that were placed to cover the battery pack to protect passers by from electrocution hazard. The causes of these arcs are possible to be the source, given the current design.

As an aside, another potential failure mode could be a fire due to a collision where the batteries move, short to its rack or the chassis, and get the metal red hot that starts a fire, which also needs to be considered to be addressed on all EVs. This car had 30 cells behind the rear axle which were still more than a foot away from the edge of the bumper, but for maximum safety should be inside the cage like where the mid pack was, where the gas tank was. The front pack stuck out about a foot in front of the front axle which also could move during a collision. If the front and rear packs were removed, it would only have 1/3 the range and power, far too little. If all the batteries were put under the cab, like the Nissan Leaf and other OEM EVs, the body modification effort may be prohibitive on a conversion as opposed to a ground-up design, since they would need to be insured to be tucked up high enough such that they did not get damaged when running over something.

April 10
--------
David Roden posted a link to the story about the EV fire caused by an accidental short across a 300V battery pack that was too close together (>1V/mm) that earned John Wayland his nickname "Plasma Boy": www.evdl.org/docs/plasmaboy.pdf.

March 23
--------
Here is a post from Bill Dube on the Electric Vehicle Discussion List that might explain what happened to our car:

"It is _really_ important for a number of safety reasons to maximize the distance between voltages in a battery pack. You want to put the highest positive terminal (and positive cell) as far away as possible from the lowest negative terminal (and positive cell.)

"Over the years we have discovered that you need to maintain a "one mm per volt" distance between terminals and other conducting surfaces inside a battery pack. Basically, you force an arc between conducting surfaces to travel a mm per volt of potential difference. This makes any accidental arc, perhaps initially caused by a metal shaving, loose bolt, wire "whisker", etc. travel one mm for every volt available. When plasma is fed by the metal vapor of the conducting surfaces and huge amps, it will self-extinguish _only_ if there is at least one mm per volt separation. Otherwise, like an arc struck by a welder, the plasma will sustain, and then likely spread to other "too close" spots inside the battery pack, leading to a chain reaction.

"You can place insulating barriers, (like G-10, and FR4) between surfaces that are unavoidably close. The barrier needs to be substantial enough to withstand the initial "fault" arc for a few moments while the accidental conducting bridge vaporizes. The barrier needs to extend beyond the "line of sight" between the conducting surfaces and present a total path length of greater than one mm per volt. These " arc revetments" also greatly reduce the consequences of >a dropped tool, nut, bolt, etc."

The spacing between the charger, heater and DCDC converter fuses and negative terminal inside the junction box exceeds 2V/mm, and the front battery pack, designed for symmetry to minimize electromagnetic radiation, exceeds this recommended level by an order of magnitude in several areas: .

March 19
--------
Informed at 9:30 AM that Focus burned up the night before. Estimated fire start time was 10 PM Friday night March 18th. Smoke detectors kept getting set off by gas engine exhaust so heat sensors were installed, but not connected to the sprinkler system. Alarm system went off, but was not connected to any communication system, so no one heard it. Finally someone reported smoke passing by and the fire department arrived about 11 PM, but the fire was almost out, it only took 50 gallons to extinguish the fire. Nearly all plastic, rubber and paint was burned, and the lift and garage door was damaged. The rest of the shop was filled with soot, but otherwise no other damage. Here are some pictures:
Front oblique: 
Front batteries: 
Front lower: 
Front junction box: 
Front junction box closeup:nbsp
Driver's fender: 
Interior: 
Rear oblique: 
Rear: 
Trunk: 
Trunk closeup: 

Note that there are no pictures of the mid battery pack, located where the gas tank was, since the lift was not operational.

March 16
--------
BMS tested with laptop with new 8th bank and still errors on bank 7, tested by plugging in charger temporarily to power BMS. Bank #7 is still not reporting. Traction pack is still wired up and reads 298V.

Determined that BMS was not shutting down charger when in an error state due to a junction box wiring diagram error: wires to K6 NO and NC are reversed and need to be corrected. W232B/C needs to be connected to NO and W245 needs to be connected to NC. Jbox wiring diagram page 2 needs to be corrected and jbox needs to be rewired. Planned to be done March 19.

March 12
--------
All batteries, fuses and interconnects installed. Pack now at full voltage of ~298V. Discovered inertia switch was wired wrong, the power feed was not hot because the ICE motor controller was not happy that there was no ICE so it needs to be fed from 12V direct, to be done later. So instead, hooked up 12V jumper wire to ignition input in junction box to activate controller. Pot box installed. Got wheels to turn. Ignition jumper was then disconnected.

Front battery pack 7 causing BMS errors, and BMS not shutting off charger despite this error. Agreed not to charge for any long period of time or unattended until these two issues were fixed. Long jumper in the middle of pack 7 suspected to be a noise problem causing BMS errors, so split in to two BMS banks 7 and 8. Reinstalled front pack hold down and tightened it. Hooked up front bank BMS wires to BMS control module, 1 7 and 8. Checked 12V accessory battery and it reads only 8 volts. Removed 12V battery and put on bench to charge. BMS did shut off motor controller during this error as designed.

January 7
---------
Modified potbox from 2 wire to 3 wire to work with Azure: rear view  
Front view  
Installed radiator assembly:  
Picture of junction box with cover off:  
Installed mid and rear battery boxes:  

January 3
---------
Built heater input cable assembly, junction box status circuit board and a subset of the console, missing the fuel and temperature gage amplifier and speedometer spoofing circuit board.
Building the status circuit board:  
Testing the status circuit board and console:  
Top view of status circuit board:  
Bottom view of status circuit board:  

FALL 2010
=========

December 19
----------
Made radiator assembly -- side view:
Bottom view:
Made low voltage chassis EV wire loom:

December 3
----------
Added more detail to front rack for BMS controller and wire loom mounts: more front rack

November 28
-----------
Updated the following wiring diagrams:
Chassis:
upper left
upper right
lower left
lower right
Details A-C, batteries:
Detail A: Fused front battery box
Detail B: Fused mid battery box
Detail C: Fused rear battery box
Detail D, junction box:
upper left
upper right
lower left
lower right
Parts list, wiring list and design issues spreadsheet
Detail E: Console
Detail F, low Voltage EV chassis Wire looms:
upper left
upper right
lower left
lower right

October 23
----------
Designed radiator tray assembly design:
- assembly
- custom part detail
-  welded part detail

October 16
----------
Designed traction battery interconnects: sketch

SPRING 2010
==========

May 15
------
Updates for the entire semester are in this entry.

Procured EV parts:
Spreadsheet with EV component details
Used from Make Mine Electric:
Link to website: www.makemineelectric.com
- Motor: Azure Dynamics AC24

- Motor controller: Azure Dynamics DMOC445

Link to drive system datasheet: www.azuredynamics.com/products/force-drive/documents/AC24_DMOC445ProductSheet.pdf
Link to user manual: www.azuredynamics.com/products/force-drive/documents/MAN-080001-001_DMOC445_and_DMOC645_User_Manual.pdf
Link to pedal control manual: www.azuredynamics.com/products/force-drive/documents/MAN-080002-001_DMOC_Pedal_Controlled_Application_User_Manual.pdf
Link to CAN control manual: www.azuredynamics.com/products/force-drive/documents/MAN-080003-001_DMOC_CAN_Controlled_Application_User_Manual.pdf (not planned for the first revision but may be used in the future)
From EV Components:
Link to website: www.evcomponents.com
- Traction Batteries: 90 Thunder Sky LFP-60AHA LiFePO4 cells
Link to datasheet: www.thunder-sky.com/pdf/TS-LFP60.pdf
Picture of single cell:

Picture of all cells:

- Battery management system: Elithion Lithiumate
Link to web site: liionbms.com/php/liionbms.php

From Metric Mind:
Link to website: www.metricmind.com
- Charger and cables:Brusa NLG513-SC
Link to manual: www.brusa.biz/assets/downloads/datasheets/NLG5_0503.pdf

- DCDC converter: MES-DEA 400-1000
Link to manual: www.metricmind.com/data/mes_dcdc.pdf

- Cabin heater: MES-DEA RM-4

From EV Source:
Link to website: www.evsource.com
- Power brake vacuum pump kit:
Link to web page: www.evsource.com/tls_braking_system.php

- Power steering pump kit:
Link to web page: www.evsource.com/tls_steering.php

- Traction fuses: 5 Ferraz-Shawmut A50QS300-4s
Datasheet: us.ferrazshawmut.com/oem/media/pdf/A50QS.pdf

Had motor to transaxle adapter kit fabricated by Make Mine Electric.

Designed motor mount adapter:
- 3-D sketch
- top view
- front view
- side view
- power steering pump mount
- material list
Designed front battery hold down:
- 3-D sketch
- top view
- front cross section and material list
Designed rest of front assembly:
- Front allthread
- Front assembly material list
Power steering pump to be attached to motor mount, behind motor. The motor side equal length half shaft support is less than 1 inch away from the motor casing. It would be convenient to mount it to the motor supplimental attachment points but Azure Dynamics confirmed not to add any forces to the casing. They suggested to lengthen the motor side half shaft by removing a section from the part that comes out of the motor that includes the mount bearing, reattaching it to the tip, and adding that much to the length of the motor side. Torque steer should not be a problem with such a low power motor. The plan is to bring it to the local Drive Line Shop to have them do it. Began fabricating motor mount. See motor picture above for progress made.

Designed and fabricated mid battery box:
- 3-D sketch
- rack
- hold-down
- material list
Picture of gas tank space where mid battery box will go:

Picture of mid battery rack and hold down:

Determined that only 30 cells fit under hood instead of 40 as feared. Decided on no tunnel box and to cut out spare tire well to cut down on potentially two extra battery boxes. Needed to reduce rear axle weight, so changed to 30 rear cell design, reusing mid battery box design for rear. 90 total cells, 288 volts, 17.3 kWh, 495 pounds. 60 mile range assuming new batteries at 60 Ah * 93% of 288 volts being drawn under load with heat loss in batteries at 269 Watt-hours per mile. 48 mile range expected after a minimum of 5 years down to 80% capacity, but vehicle may be able to be driven up to 10 years down to 36 mile range and noticable power reduction.
updated weight analysis spreadsheet

Decided to make the rear battery box the same as the mid battery box, with spare tire well removal and trunk floor reconstruction required. Picture of rear space where rear battery box will go:

Did high level design for location of motor controller, DCDC converter, charger and junction box where radiator was located. Fabrication drawings forthcoming. Picture of radiator space where power electronics will go:

Picture of DCDC, charger and motor controller below where they will be mounted:

Designed heater mount to firewall. Picture of where heater will go:

cabin heater to firewall mount

Accessory battery needs to be big to run power steering and brakes for an entire slow long trip if DCDC quits, so similar sized battery to starter battery will be used. Will mount vacuum brake system in front of it. Picture of where accessory battery and vacuum brake system will go:

FALL 2009
=========

November 21
-----------
Measured "magic number", distance from rear vertical surface of motor to rearmost precision surface of the flywheel, to use to make hub to mount flywheel on electric motor the same distance from the profile plate. Used digital calipers with straight aluminum block on flywheel to provide measurement surface:

Picture of motor face; adapter plate needs to look like this:

Picture of transaxle face; adapter plate needs to mate to this. One of the precision guide pin holes is being pointed to on the right side:

Designed motor to transaxle adapter:
- assembly cross sectional view
- motor ring
- adapter plate (outline and flywheel housing bolt dimensions to be added)
- hub

November 14
-----------
Completed preliminary weight analysis.
preliminary weight analysis spreadsheet
Removed 585 lbs, 499 front, 86 rear.
Will add approximately 839 lbs, 518 front, 320 rear. Final curb weight will be approximately 2985 lb, 320 lb more than wet ICE curb weight.

Consumption rate estimated in the following manner:
- Start with 375 Wh/mi for the following 2975 lb 375 Wh/mi 1966 Mustang DC EV conversion: link to web page
- Assume 12% reduction due to AC efficiency being 85% vs. DC being 75%: 331 Wh/mi
- Assume 10% improvement due to regenerative braking: 298 Wh/mi
- Assume 5% reduction due to front wheel drive being more efficient: 283 Wh/mi
- Assume 5% reduction due to modern car being more aerodynamic: 269 Wh/mi
- Assume 0% improvement due to no weight reduction

Plan is for 40 batteries in front, 12 in tunnel, 36 in gas tank area and 8 in rear, 96 cells, 307 volts, 18.4 kWh, 528 pounds. 64 mile range assuming new batteries at 60 Ah * 93% of 307 volts being drawn under load with heat loss in batteries at 269 Watt-hours per mile. 51 mile range expected after a minimum of 5 years down to 80% capacity, but vehicle may be able to be driven up to 10 years down to 38 mile range and noticable power reduction. Concern is that only 24 cells may fit comfortably under hood once motor and electronics are installed. This may result in a shift in cells towards the rear and a reduction of rear seat and trunk payload, or a reduction of number of cells and reduction of range and power to meet safety requirement to not exceed GVWR and GAWRs. Final layout and updated weight analysis to be performed once motor is in and front battery box is designed.

October 31
----------
Curb weight measured after all ICE components and transaxle removed. Picture of car on scale pads:

Picture of scale display:

Left front 482
Right front 474
Left rear 503
Right rear 499
Total 1958 lb
The transaxle was weighed at 120 lb, so adding that back in the glider weight is 2078 lb.

Weight measured with 4 passengers totalling 803 lb after all ICE components and transaxle removed. This is 100 lb more than the payload target, but gives us an idea of how the passenger weight gets distributed for weight analysis reference. Picture of scale display:

Left front 640
Right front 637
Left rear 739
Right rear 745
Total 2761 lb

Per GVWR sticker on door sill:
Front GAWR 1975
Rear GAWR 1745
GVWR 3715 lb, = front + rear GAWRs - 5
5 passengers max, 3 front, 2 rear; maximum payload with luggage 827 lb
Assuming wet curb weight was originally 2659 lb per September 12 entry, payload should be 3715 - 2659 = 1056 lb, 229 lb more. Picture of GVWR sticker in door sill:

October 17
----------
Separated the transmission from the engine:
- Disconnected the K-member from the bell housing.
- Removed the starter.
- Removed the bolts connecting the engine to the transmission.
- Removed the drive axles and the bolts that supported the axle on the engine.
- Removed Power steering pump and tubing.
- Removed K-member and wheels from transaxle.
- Detached the transmission from the engine.
Picture of engine hanging from hoist:

Picture of transaxle:

October 10
----------
Removed radiator. Removed entire front end suspension assembly as a unit, which includes the engine, transmission, steering column, struts, drive-axles, wheels, and the K-member:

Picture of front end assembly removed:

October 3
---------
Performed 0-60 test. Average was 11.06 seconds.

Removed fuel system -- fuel tank, fuel filter, filler hoses and charcoal canister:

Removed exhaust system -- tailpipe, resonator, muffler, exhaust pipe, and catalytic converter, as well as most of the heat shielding and supporting hangers and brackets. Drained engine oil, coolant and transmission fluid. Picture removing exhaust pipe while draining oil:

September 12
------------
2005 Focus sedan with 77K miles in very good condition purchased by SRJC staff from Nader's Preowned Cars and Trucks in Santa Rosa and delivered to class.

Ride height measured with 2 gallons of gas, no passengers, tires at 30 PSI, using tape measure from floor to top of inside of wheel well hole:
Left front:    right front:    left rear:    right rear:
Left front 26-1/16"
Right front 26-1/8"
Left rear 26-11/16"
Right rear 26-3/4"
The left side is 1/16" lower than the right when empty. The back is 5/8" higher than the front.

Curb weight measured: no passengers, 2 gallons of gas, motor oil and coolant, with 4 wheel scale:

Left front 770
Right front 796
Left rear 527
Right rear 494
Total 2587 lb
Left/Right 50.1%/49.9%
Front/Rear 60.5%/39.5%.
12 gallons of gas or 72 pounds below estimated wet curb weight, all on rear wheels. Estimated weight 2659 lbs. wet.

Weight measured with 164 lb driver added:

Left front 837
Right front 815
Left rear 587
Right rear 512
Total 2751 lb
77%/23% added to left/right. Note less than 100% on the left; anti-sway bar is partially levelling car.
52%/48% added to front/rear. Note that the front seats are between the axles. Rear seats and trunk are on the rear axle. 50/50 front/rear expected at GVWR.