The White Lightning 818e thread

[Speedy G]

Finally my first parts came in so I can say that my build started. This build is a pretty radical one since I'll essentially be putting a square peg into a round hole and I'm most probably biting more than I can chew, but what fun would it be if I didn't. The budget is unknown, and I'm not going to talk about it since I'm building my personal supercar so the dollar sign just went out the window.

So why White Lightning? It's my donor:

WL.jpg

I bought the drivetrain for the original white lightning and I'm putting it in an 818s. The car originally did 0-60mph in 8 seconds and had a top speed of over 300mph, although its bonneville official record was 245mph in 1999. The white lightning car weighed 2,550lbs and I'm hoping to brind that weight down to less than 2,400lbs in my build because of the use of different battery technology. Ed Dempsey's team used belt drive between the two motors, and hooked them up directly to a 9" Ford rear axle, no tranny.

What I'm doing:

1) Use the original belt drive setup for the dual motor setup (yes I know, not efficient and at 10,000 RPM it's way over the belt spec). The motors are AC Propulsion Generation 1, 150KW each (200hp). The motors were tested by AC Propulsion and yielded 165KW @ 350V or around 218 hp each. Ed's team modified the motors for better cooling, which suits me fine.

Motor+PEUonDyno.JPG

2) One of the original motor controllers has broken IGBTs so I'll need to replace them. This will mean redoing the power electronics since the original IGBTs cannot be bought anymore. I don't think it should be too difficult since IGBT technology has advanced considerably since the 90's, so an off the shelf IGBT should easily replace the original ones. The idea is to initially register the car with 1 of the controllers (so 218hp). The controllers are also so big that I may consider redoing the housings. They're that big because they're air cooled, so using a water cooled housing should make them significantly smaller. Also the original IGBTs are discrete which makes them at least twice as big as today's dual IGBT modules (hi +lo). All that means is there's potential for lowering the weight and size of the controllers.

** I decided to go ahead and start experimenting so I ordered some IGBTs. I went with the newest generation of Powerex IGBT MOD NX 600A 1200V DUAL and their VLA536-01R driver board. That pretty much means I can hook up a cpu directly to 3 of those and dump 600A @ 1200V (~960hp) into a resistive load (I can smell the burnt chip smell already!). I just don't want my electronics underdelivering. Btw, that's for 1 of the motors.

3) Batteries - Erik's build used 50 100Ah CALB batteries, I'm thinking 100-120 are required for my build. Where to put them??? We'll see, probably in the back, maybe some in front of the axle (ugh). That I have to do when the car gets in to see where they fit and where I can fit the electronics (currently the size of a medium fridge).

* Update - The batteries will be 110 CALB CAM72FI for the standard controller and 110 A123 20Ah pouches for the acceleration controller (added later). The initial pack fits 70 batteries in the center tunnel and the gas tank location and the rest on either side of the motors and tranny.

4) Gearbox - The main design issue. How to hook up the motor to a tranny? What tranny? The ideal EV tranny is very different from a regular tranny. First, my motors can go to 12,220 RPM (tested by AC Propulsion). Second,due to the high RPM and torque curve, the car can accelerate like an M5 from 0-62mph in first, while an M5 has to go through 1st and second, so I need less gears (3 is probably ideal). Third, since I can control the motor speed very precisely, I can shift into neutral from any gear, rev match and shift to the next gear. That means I don't need synchros or a clutch. There's companies that have built EV specific gearboxes that shift like this in less than 100ms, which makes them as fast or faster than ferrari or BMW dual clutch boxes. So... what to do... If I can get a cheap subaru donor I might try to use it, otherwise, a full sequential is my first pick (PBS S4 ).

** Update - I'm going with a subaru tranny just because it fits easily. I won't be running a clutch, just a straight shaft from the motor assembly to the tranny input shaft. Initially I'll run the car in 2nd gear only since it's close to what tesla runs in their Borg-Warner setup. With the 3.9 final drive 2nd has a total ratio of 8.04, which should be fine for city driving. We'll see if it's worth changing the final drive later. Later, I'll automate the shifting with a modified mastershift paddle shifter. Top speed in 2nd would be 150kph (94mph) at 10,000 RPM or 180kph (112mph) @ 12,000 RPM (if the tranny can handle it). Maybe I'll have to do something to the lubrication system of the tranny, we'll see.

5) Shopping list (bought so far)-

2015 STI steering Rack - Flatirons tuning
Aluminum LCA - Flatirons tuning
Wayne's Rear Trailing Arms - verycoolparts.com
Engine cover hinge - Michael Everson
Hood hinge - Mechie3
Speedometer/Odometer - Stack ST3802
Tachometer(13kRPM)/multi-display - Stack ST700SR-N
Fuel Gauge - Stack ST3315
TCI Paddle Shifters 301441
Wilwood hard brake lines Frnt & Back
Sparco L360 Steering wheel
Sparco Steering wheel hub adaptor -STRWL HUB SUBARU
Sparco Steering Wheel Ring - black

EV parts:
Gigavac GV200-MA EV Contactor
Gigavac Maintenance Switch HBD41AA
Brake Pressure Transducer for Regen - 5 Volt - 1000 psi
ZEVA Fuel Gauge Driver Plus
High Voltage Fuse 500 Volt 400 Amp A50QS400
High Voltage High Current PolyCarbonate Fuse Holder P243G G Type
Prius Gas pedal - EV West

Tools Bought:
Bosch PS10-2A driver/drill
Campbell Hausfeld Commercial CL153900AV Rivet Gun
Grinder - Dewalt 5"
Dremmel multimax MM40
Cobalt Drill Bits 19/64", 1/8" - Century Drill
Cobalt Drill Bits 25/64" - Bosch
Cobalt Drill Bits 5mm - Cle-line
Torque Wrench 3/8" 60-960 in lbs - Titan Tools 23147
Torque Wrench 1/2" 25-250 ft lbs - K-Tool 72102


Let the games begin!

Speedy G

[Bob_n_Cincy]

Hey Speedy
Great to see another electric 818.
I am pretty familiar with the AC Propulsion controller and motor. I raced against them back in the late 90's.
Some evening I could have a long phone call information dump.

Take a look at these links. The tZero was built by the founder of ACP Allan Cocconi. It is similar size to the 818. It started as a kit car by Piontek engineering. Piontek used a motorcycle engine.

Just a bit of history. That motor and controller was the start of Tesla.
Bob

https://www.google.com/#newwindow=1&q=Tzero


http://piontekengineering.com/

[Speedy G]

Yeah, exactly. I was fortunate to talk to Paul Carosa about the motors. He was a bit skeptical due to their age, but he did some maintenance and sent me most of the literature. High RPM induction motors are it, the only problems are the gearbox and the high voltage.

Bob, I'm sure you could teach me a thing or two about batteries. What kind of batteries does the lightning bike have? How do they fit a 20kWh on a motorcycle? I'm trying to fit 30kWh into the 818! Any info on the battery front would be awesome. Btw, did the AC Propulsion guys beat you in the 90's?

Also, on the gearbox front, any info would be greatly appreciated. Here's a pic of what I have to hook the gearbox into:

DSC_0282.JPG

Any idea what kind of splined shaft hooks up to this and where I can get it, or how I could hook up a subaru tranny straight to this? Remember, I don't need a clutch or flywheel. I'll probably be using a mastershift paddle shifter, so shifting can be done without a clutch. The trick is to do the shift without power on the motor, and rev match to the new gear's RPMs.

Speedy G

[Bob_n_Cincy]

Bob, I'm sure you could teach me a thing or two about batteries. What kind of batteries does the lightning bike have? How do they fit a 20kWh on a motorcycle? I'm trying to fit 30kWh into the 818! Any info on the battery front would be awesome. Btw, did the AC Propulsion guys beat you in the 90's?

Hi Speedy,
Lithium-Ion Prismatic Cells. I don't know if we publish the brand name.
Very Expensive high power batteries. Only 1/4 of a tesla 85kw.

I had 4 cars running my controller and Allen had 3 cars.
Ohio State (ACP) and BGSU (EMS) were the 2 fastest cars. I remember (selective memory) that winners were about 50/50.
There was one race I did not attend and we got all three podium spots.
I need to get that picture for my scrapbook.

Here is a quick story about the Formula Lightning Series.

http://www.roadtripamerica.com/wheels/flc.htm

[Frank818]

Speedy!! Finally yes! Wow I didn't know you bought a rocket as a donor! The drivetrain, but you need a great drivetrain to build a rocket, right.

Much different than the VR6, but still similar level of challenges, more or so... Keep us posted, we want to see things moving!

I raced against them back in the late 90's.

And? Who won?

[Speedy G]

Since I'm still in the R part of R & D, I thought I'd write an update on alternative drivetrain solutions. Why? Because I looked at pretty much everything under the sun in order to optimize my setup. The short story is that I'll be using a regular 5MT tranny. We'll see if it can do 12,000 RPM though.

So an EV... What is the best tranny for a high rpm EV like mine or a Tesla? Single gear? Not really. First, at 12,000 RPM you're only going at 108mph. That's hardly sports car territory. Second, induction motors with sensorless vector controls are VERY inefficient at low RPM. Here's a Tesla range/speed graph, which can be thought of as an overall efficiency graph as well:

range-speed-model-s.jpg

Now if we want to see what "inefficiencies" are important at what speeds, here's another graph:

Whpermilevsspeed.jpg

Notice the drivetrain has some serious inefficiencies below 10mph.

So what would be the ideal gearbox for a high RPM motor like mine? Has anybody made one? Sure, and it's as good or better than BMW and Ferrari's dual clutch stuff, and simpler, and it's made for trucks!!!



Wrightspeed.com

Notice that it switches from first to second gear in 80ms (like the BMW dual clutch tranny). Btw, that's 80ms switching from over 20,000 RPM to 9000 RPM. However, it doesn't wear out because it doesn't have any clutches or synchros!! It's a dog box! What happens is the electric motor can be manipulated to exert 0 torque so the gear can be disengaged, then the motor recovers rotational energy from the motor by making it go from 20,000 RPM to 9,000. Once the motor and tranny's second gear are speed matched, the motor again exerts 0 torque, and the dog engagement shifts into second, gear. Once the gear is in, full torque is available again. Oh, also that's a clutchless shift.

So what would be the ideal ratios for an EV? Well, first is limited by traction. Remember I have 330 ft lbs from 0-6000 RPM, flat. So I chose first to be at the limit of traction ( ~13:1 - 14:1). Next, I'd chose third to give me my top speed, say 200mph at 12,000 RPM (so 4.5:1). Next, I'd chose second, although unnecessary, but it may give me better driveability at highway speeds and maybe improved efficiency. I chose 6.86 since that would make the shift points even (105 kph, 210 kph and top speed of 320kph). Why such wide ratios? I have full torque down low, so the lower the RPM that I can shift into, the better. Also comparing to an M5, my first gear will have better acceleration than the M5 in first AND second at peak torque.

So that's it for ideals. What's really available? I looked at the following:

Powerglide - 2 gear, can be modified for manual shift, direct drive (no torque converter). Can be directly coupled to a Ford 9". The powerglide is light, but the ford 9" is really heavy. Also the setup ends up being way too long before the axle centers. The shifting action is also extremely harsh, so not really a street car tranny.

C5 transaxle - Heavy, Ratios kinda suck, scarce. C5 guys said I'd be better off with a Ford 9".

Porsche G50 - Original car was rear engined so you have to turn it upside down and do some minor changes to the diff so it drives the right way. Frank is trying to make it fit with his VR6. It's hard to find a good one.

Porsche G86.20 - Boxter tranny. This will probably fit, but it'll probably push the engine further to the front a bit. The tranny is strong and may be a good candidate, but I haven't found any reasons to use it to replace the standard subie 5mt, in stock form it's probably as strong as the subie tranny, and uprated gear sets are very expensive. This tranny will probably push the engine forward a bit, like 2 inches.

PBS, Medeola, Albins, Weddle or Vw Transaxles Some of the mentioned ones are sequential dog boxes. Only the sequentials seem to be stronger than the subie trannie. The sequentials provide no real benefit for an ev if rev matching is implemented in the control, like the wrightspeed box. These transaxles would move the engine forward quite a bit, probably up to 6 inches, which might be ok in an EV. These transaxles start at USD7k and got up to over USD15k. I'd only consider one if I were a hardcore racer, and for that, I like the 4 speed PBS sequential with helical gears.

Subie 5MT - Ratios are: first is 13.5, third is 5.85 and fourth is 4.24 - If I get a 4.11 final drive, then first is 14.2, third is 5.95 and fourth is 4.47. First and fourth are perfect for my ideal situation, fifth is useless, but I have second and third to choose from for highway driving. However, cooling the oil doesn't seem to be an option. The big advantage, the frame is built for it. Paddle shifting can be done using a modified mastershift system and rev matching using the drive and control system. The clutch can be a low MOI clutch. I'm currently looking at a tilton 5 inch tripple disc clutch. I need something that can hold at high rpm. Whatever it is, it'll have to be modified. Eventually, if the control system is good enough, maybe I can do without a clutch altogether.

So my only question is, is someone that's pushing 300whp having serious problems with traction or is it manageable? What tires? What ratios?

[Frank818]

Porsche G50 - Original car was rear engined so you have to turn it upside down and do some minor changes to the diff so it drives the right way. Frank is trying to make it fit with his VR6. It's hard to find a good one.

True. And I am a couple of days of seeing it fit or not. I just received the adapter plate and stuff. Gotta fab a shifter bracket kit as well and having in mind now that shifting also works upside down. But it's strong enough no issues and has the possibility to use a lot of different ratios per gear. My mechanic has a second one, but maybe not in such an excellent condition mine is.

I wish I could tell you how it runs with 400whp but I can't until late next spring or early summer.

[Speedy G]

Yeah, but your torque is gonna be way more than 330 ft lbs. You're gonna be burning rubber in second. I have a traction advantage stemming from around 400 lbs worth of batteries in the back plus hopefully less than 320 from the motors and controllers. By modding the controllers I should be able to shave at least 70 lbs off the drivetrain weight. That's a target of 650 lbs vs 270 from the subie motor. For comparison's sake, your VR6 is around 420lbs.

Btw, for those that aren't building from donors, it seems like the 2015 STI steering rack fits. The advantage is a ratio of 13:1 vs 15:1 for regular wrx or 18:1 (I think) from a forester. In my case, that means leaving the power steering, essentially for low speeds. I'll be using an electic motor as the power steering pump. That also means I could control to a point how much power assist I have at any speed with an additional simple control system and a driver (mosfet of sorts).

[Frank818]

Speedy, where are you in the build?

[Speedy G]

There's updates up on the first post btw. I'm still in the design stage since the kit hasn't arrived. My date was the 27 of Dec, but in this holiday season it's hard to arrange everything. Banks and my shipping company as well as FFR have days off, etc. I'm thinking I'll get the kit in mid-february at this pace. I've solved several issues though. I now know that I can register the car here in Colombia as long as I register my company as an automotive assembly plant, and therefore I can say what model year the car is. The process will be long, but I can start it once the car drives but there's a ton of post initial drive projects that I can be doing while the whole legal stuff gets sorted out. Here's a short list:

1) Replace the monstrous power stage of the AC propulsion drive with my modern day design - Powerex IGBT MOD NX 600A 1200V DUAL + driver board + snubber capacitors + water cooling. This alone should shed maybe 15 lbs and use less power than the air cooling system the original power modules had.

2) Battery management system - I'm thinking of using a battery management system just so I an keep track of the battery voltages. I plan to abuse the batteries to 10C (1C would be a discharge rate of 30kW/h), so keeping track of failing batteries is probably a good idea.

3) Main control system - I'll be using a Cubietruck (ARM based computer) as the main entertainment and control system of the car. It'll have 500G of solid state hard drive, and run some flavour of linux.
3a) It'll be connected to a DAC which in turn will be connected to a power amp using balanced connections. My cell phone will also connect to the cubie so I can use music on my phone over the car stereo.
3b) It'll control the power steering pump. Less power steering as speed goes up. The power steering pump is electric btw.
3c) It'll control the vacuum pump for the brakes. I'll be using computer controlled power brakes. Another electric pump will provide the vacuum to power assist the brakes. You gotta remember this thing will weigh in at around 2300lbs when I'm done.
3d) I'm thinking maybe I'll have some power ventilation in the car. If I do, the cubie would also control that.
3e) The cubie will have either a 7" or a 9" touchscreen built into the center console. So I imagine I'll have to develop some software to display the control options as well as various sensors (water temp, motor temps, battery temps, total current used, kwh/km, battery voltage, etc). I'd also like to have google maps and a wifi connection through my phone.

4) Automate the tranny with a modified mastershift shifting system since I'm running a directly coupled motor to tranny. The mastershift workflow needs to be altered to shift, like this: run-> depower motor-> neutral-> rev match-> shift-> repower motor.

5) Redo all the fibreglass parts in carbon fiber. There's a local guy that works on racecars that makes the carbon fiber molds. I should be able to shed quite a bit of weight with this mod. Locally it's not as expensive as it would be in the US.

But then again this is all still vaporware since I haven't turned the first screw yet...

[Frank818]

Oh it's not there yet, sorry I thought it was. It'll miss New Year's, but it gives time to design.

Very cool stuff, lots of non-FFR stuff! Actually, your build will be way more complex than mine. You have will, the 2nd strongest power in the Universe. How long you estimate to build everything?

[ninjanick]

SpeedyG,
Have you laid out your traction pack yet? I'm still in the middle of figuring out the best layout for my build.

[Gary Livingston]

Hey Speedy!
Wow! a very ambitious project. I thought the White Lightning was in a museum somewhere. I'm building electric also so I will be following your thread closely as well.

Regarding C rate mentioned below......

2) Battery management system - I'm thinking of using a battery management system just so I an keep track of the battery voltages. I plan to abuse the batteries to 10C (1C would be a discharge rate of 30kWh), so keeping track of failing batteries is probably a good idea.

C rate is actually a discharge rate relative to cell Capacity. So, 1C for 100Ah cells would be 100 amps. 10C would be 1000 amps.
Ah and kWh are both measures of energy. Ah usually used for individual cells and kWh usually used for packs. WH (watt-hours) are a more accurate measure of energy than Ah (amp hours) because they take into consideration the voltage and current, not just the current.
So, a pack consisting of 100 pcs of 100Ah cells (if LiFePo4 chemistry) would have an energy capacity of 3.2 (cell nominal voltage) x 100Ah x 100 units = 32,000wh (watt-hours) or 32kWh.

To determine how much horse power you could get from this pack, you would need to know the voltage level at the specific current level. Lets say at 1000 amps, cell voltage would sag to 2.5vpc (volts per cell) x 100 cells x 1000amps or10C. That would be 250,000 watts or 250kw.
250,000/746 = 335HP
If your drive train was 90% efficient, you would have 300hp at the wheels.

Sorry, got a bit carried away there. Was just going to correct the C rate. . . but thought some folks may benefit from the explanation.

[xxguitarist]

Speedy,
On the BMS front:

Will you be running active, passive, or manual charge balancing? It's a big thing with lithium cells to maintain good balance. Depending on the chemistry, it varies from detrimental to performance.. up to dangerous, to have mis-balanced cells. It can lead to over discharge and over-charge events.

What is your plan for battery cooling? At high C rates, the cells generate a bit of heat, again depends on chemistry some. On the other side, what about heating, or will the car not be driven/charged if it's below 40F?

Charging? Only at home, or perhaps with J1772 Level 2 or the more recent DC fast charge CHAdeMO etc infrastructure support? The former isn't too difficult to implement, it's a relatively simple handshake process.

[Gary Livingston]

Speedy,
On the BMS front:

Will you be running active, passive, or manual charge balancing? It's a big thing with lithium cells to maintain good balance. Depending on the chemistry, it varies from detrimental to performance.. up to dangerous, to have mis-balanced cells. It can lead to over discharge and over-charge events.

What is your plan for battery cooling? At high C rates, the cells generate a bit of heat, again depends on chemistry some. On the other side, what about heating, or will the car not be driven/charged if it's below 40F?

Charging? Only at home, or perhaps with J1772 Level 2 or the more recent DC fast charge CHAdeMO etc infrastructure support? The former isn't too difficult to implement, it's a relatively simple handshake process.

The topic of BMS or not has been a hot one on several forums for quite some time. Enough years have passed now with enough experience and data to prove that LiFePo4 does not need to be actively balanced. I subscribe to manually balancing, preferably bottom balancing, and not exceeding a charge voltage of 3.6vpc. I actually charge to only 3.45vpc. The capacity above this point is not much and the life is prolonged with this approach. As long as you have a well balanced pack, this approach works well. I have been running this way for years on the older CALB SE cells, with no balance issues. I've also run a booster pack of A123 32157's the same way.
There are some caveats, such as having cells in the same pack exposed to different temps, as well as significant differences in capacities.
It is widely believed that most catastrophic fires are caused by BMS's failing to stop the charging process and overcharging leading to thermal runaway.
Other lithium chemistries seem to require a little more balancing care . . .but I'm not sure i would charge most of them at night in my garage... lol
I think a BMonS (monitoring) is not such a bad idea. . . BManS (managing, meaning top balancing) not so much required. Some folks have coined these systems Battery Murdering Systems... lol

Discharging is ok down to -20C typically for LiFePo4 although performance suffers greatly at colder temps. The real thing to avoid is charging below 0C. This will cause lithium plating and shorten battery life significantly.

cheers!

[erlihemi]

Speedy, Awesome project! If you can program for zero torque crossing at shift points, then any circle track 2 speed manual should be lighter than the powerglide. Why not just direct couple a subie CVT? then you have programmable torque/speed control potentially... You have a heavier 818 than design so contact patch can be increased and traction is less of an issue. Assuming regenerative braking?? Good Luck!

[Speedy G]

Thanks! Yeah, I saw those gearboxes, I forgot to list them. Colemanracing.com sells them:

http://www.colemanracing.com/Lightwe...ion-P6179.aspx

2 x 2 speed gearboxes would be better than 1 gearbox with a differential since I'm using 2 motors anyway, but my problem was finding something that could couple each gearbox to the axles. Also, what drove me away from those gearboxes was the fact that most of them say that they're not meant for hard starts, but rather just to get the car moving. They're also straight cut and require rebuilds often. Then again, each gearbox would only be pushing 170 ft lbs so they should be able to hold up.

Regarding CVTs, they usually use torque converters for initial slip and to ease the shifting. That makes them less desirable due to inefficiency. Also, they're really not a good idea for electric motors. My wife's car has a CVT and when you step on it, it goes all the way up to 4000RPM and stays between 4000RPM and 5000RPM between 30 and over 100mph. It does that to optimize for a gas engine's max torque. My motors have max torque between 0 and 7000RPM so a CVT would actually be counter productive.

As for regen, yeah, I have regen up to 200hp worth. Honestly, I don't know how it'll be to drive a rear wheel drive car with regen entering curve. Some of the Formula E guys talked about it being difficult. The big question is how to bias the rear brakes.

[rtz]

Like what Jack was saying during the show this week in regards to using a standard 5 or 6 speed transmission. Examples: You're going up a steep driveway; put it in first. Driving around town; pick 2nd or 3rd. Out on the highway; put it in 5th or 6th. Essentially single speed when in use; but allows you the choice of the best gear for the job. The Subaru trans will work just fine.

[Bob_n_Cincy]

Like what Jack was saying during the show this week in regards to using a standard 5 or 6 speed transmission. Examples: You're going up a steep driveway; put it in first. Driving around town; pick 2nd or 3rd. Out on the highway; put it in 5th or 6th. Essentially single speed when in use; but allows you the choice of the best gear for the job. The Subaru trans will work just fine.

I built my first EV in 1998. It was AC system tied a 5 speed manual in a minivan. I drove it a lot like you describe above.
That was back in the Lead Acid days and only 50 miles range with at 1300 lbs battery pack.
Bob
377e.jpg ASSYBLY1.JPG BOB_BAT.JPG BOB_VAN.JPG VAN_BAT2.JPG VAN_SIDE.JPG

[Speedy G]

I agree about the subaru tranny being obviously fine for driving around town, but the build idea is to optimize for acceleration with decent range and top speed, and I like paddle shifters. That's why I considered all the other trannies. Also weight is a big factor in my build because of the batteries, so something lighter could've been advantageous which is why I looked at the powerglide and the 2 gear circle track trannies. The truth is that I wasn't able to find anything better than the subie tranny, but hopefully someone develops something cool soon. I like what these guys are developing, although I don't agree with in-wheel motors and don't think rare earth motors are a good choice, even if they're slightly more efficient (not in this case though), and slightly lighter than induction motors.

Bob, did everyone think you were crazy back in '86 building an EV? Talk about early adopter...

[Speedy G]

SpeedyG,
Have you laid out your traction pack yet? I'm still in the middle of figuring out the best layout for my build.

Nah, I need to wait for the kit to get here. You gotta remember I'll be trying to reuse as much as I can from the white lightning drivetrain. I have an aluminum structure where the 2 motors are bolted onto, and 2 pulleys that tie the motors together. On the lower pulley, in the back, I have a female 16 spline pinion coupler. I'd like to have one motor over the other if there's enough room, but we'll see when the chassis gets here. After that I need to figure out where to fit between 100 and 120 batteries, that's the hard part.

[Speedy G]

Hey Gary, I got ya. The thing is in this case 300kW is the equivalent of 10C because the estimated current is 1000 A @ 300V. Also about the 100 batteries delivering only 335hp, that's why I'll be trying to fit 120 instead, but honestly, I'm not sure where yet. I don't want 80% of the weight on the rear axle. Also battery to tranny input shaft efficiency is 91% at it's best, but since I have a synchronous belt coupling the motors together, I'll only have high efficiency at low rpm.

[Speedy G]

The topic of BMS or not has been a hot one on several forums for quite some time. Enough years have passed now with enough experience and data to prove that LiFePo4 does not need to be actively balanced. I subscribe to manually balancing, preferably bottom balancing, and not exceeding a charge voltage of 3.6vpc. I actually charge to only 3.45vpc. The capacity above this point is not much and the life is prolonged with this approach. As long as you have a well balanced pack, this approach works well. I have been running this way for years on the older CALB SE cells, with no balance issues. I've also run a booster pack of A123 32157's the same way.
There are some caveats, such as having cells in the same pack exposed to different temps, as well as significant differences in capacities.
It is widely believed that most catastrophic fires are caused by BMS's failing to stop the charging process and overcharging leading to thermal runaway.
Other lithium chemistries seem to require a little more balancing care . . .but I'm not sure i would charge most of them at night in my garage... lol
I think a BMonS (monitoring) is not such a bad idea. . . BManS (managing, meaning top balancing) not so much required. Some folks have coined these systems Battery Murdering Systems... lol

Discharging is ok down to -20C typically for LiFePo4 although performance suffers greatly at colder temps. The real thing to avoid is charging below 0C. This will cause lithium plating and shorten battery life significantly.

cheers!

Gary, it sound like you're going to be able to help me out when the time comes... I'm actually thinking of getting the newer CALB cells. I was thinking of doing bottom balancing and monitoring voltages as you stated. The monitoring is more of a neat-o than a real need in my opinion. I just like having all the variables in any given system. In other words, if my cousin's M5 beats me in the 1/4, I want to know what to blame ;-) Also, I'm on the equator, well, 14 degrees North but close enough, and although I live at 10,000 ft of altitude, the temperature here barely ever gets below 40F, but if that's not recommended, the on-board computer will have an environmental temp sensor to shut off the charging.

In terms of charger, all I need is a plug, 220V, 110V, 308V or even 420V. My system has AC Propulsion drives which use the on-board IGBTs as the charger. They patented the idea which is why everyone else needs a charger. I can charge at a rate up to 20kwh. Too bad it's not a Gen II, since I could use the battery pack as my home's UPS (probably for a couple of days).

Regarding cooling anything, it'll all be water cooled if need be, although the LiPo cells don't get that warm, usually. Temp sensors in the battery packs sound like a good idea.

[Gary Livingston]

Hey Gary, I got ya. The thing is in this case 300kWh is the equivalent of 10C because the estimated current is 1000 A @ 300V. Also about the 100 batteries delivering only 335hp, that's why I'll be trying to fit 120 instead, but honestly, I'm not sure where yet. I don't want 80% of the weight on the rear axle. Also battery to tranny input shaft efficiency is 91% at it's best, but since I have a synchronous belt coupling the motors together, I'll only have high efficiency at low rpm.

You mean 300kW right? Not kWh.
Have you looked at fitting 120 pieces of 100 Ah cells in that car with two motors/controllers? I don't want to sound pessimistic and I haven't laid that out but. . . wholly crap I would eat my hat if you could fit those all in there.

You mentioned wanting to build a super car. CA cells are awesome, but they are not "super car-ish" cells. If you want a super car cell, go for the A123 AHR18700F1 cells. . . . but ur gonna need very deep pockets. lol

Seriously though, what are your range and performance goals?

[Speedy G]

Yeah, the space issue... If we start out with Eric's 50, maybe add another 9-12 since my motors are shorter, and another 12 can probably fit in the center channel, even if we have to remodel it a bit. I'll probably have to fit another 18 in the back on top of the other ones and the rest are a bit ???. We'll see how everything fits once I get the chassis and I'm able to digitize it into solid works or something similar. However, CALB now has the CAM cells that could be useful if they make 120Ah cells. Currently only 70Ah cells are out. We'll see. If I can't get them to fit between the axles then they'll have to go in front of the front ones (most probably). More compact cells would definitely be usefull, but I've been thinking of staying away from other chemistries due to combustability.

Regarding the controllers... yeah, especially the AC Propulsion ones are just huge. They'll have to go on top somehow. For the first stage of the project, I only need to fit 1 since the other one is broken. To get both motors working I'll redesign the power stage with modern stuff. That should save a lot of space, especially if I use water cooling.

The range isn't all that important. I'm more interested in performance in the 1/4 and daily driving short distances (<50 miles), and keeping a reasonable budget (not more than my beamer). Beamers cost 2x as much here btw.

[ehansen007]

Hey speedy. I can say that there is a bunch of room left for me to add batteries including in front of the nose but running them down the middle might be tough but doable. I found that dropping the shifter made it harder to fit cables along with the small harness for the controllers. One thing to rember for sure is to keep those batteries accessible for balancing! Are you going to run a BMS? Also, if I had to do it over, or if this second tranny goes like the first one, I would go with STi tranny for strength. 1st gear is almost useless so with the torque you'll be making you'll still have a nice 5sp. I'll see what can do about the solid works file.

[Gary Livingston]

Hey speedy. I can say that there is a bunch of room left for me to add batteries including in front of the nose but running them down the middle might be tough but doable. I found that dropping the shifter made it harder to fit cables along with the small harness for the controllers. One thing to rember for sure is to keep those batteries accessible for balancing! Are you going to run a BMS? Also, if I had to do it over, or if this second tranny goes like the first one, I would go with STi tranny for strength. 1st gear is almost useless so with the torque you'll be making you'll still have a nice 5sp. I'll see what can do about the solid works file.

Hey Eric. How do you like the car now that you have had if for a while? Have you put many miles on it? Drive it regularly?
The solid works file would be awesome. I heard EVWest modeled it. It's probably worth something to them though. I still may do that. . . but for one car, not sure it's worth it. . . maybe if I was building more.. . Did you look at the EVIC display for Curtis?

[ehansen007]

The car is awesome. Just plug it in and go. I'd always like more range but that's just a function of cost for me. I've got about 400 miles on it and the only thing that bugs me is the rubbing on the front fenders (I'm running 18's with spacers so it's right at the edge) and how low the car is in general. I love the way it looks as it sits but it's a bit scary going in and out of driveways with the splitter on. And, (and this is something that everyone should know) this car is so low I can't even get it on a flatbed trailer to have it towed home. I'll carry some spacers for the springs with me and hope that if I do need a tow, that he'll have a low profile jack so I can at least get the car high enough to get the spacers in. I hadn't looked at the EVIC as it was not available but it looks pretty good. I like having all that data in one place. The problem is that they don't list whether or not you can convert it to Fahrenheit for the motor and controller temps. The pro would be the way to go though since it's got MPH too so you can make it the instrument cluster. As long as you could design the interface to fit within the subie cutout it would be awesome. Square just doesn't look OEM. Finally someone has done it though which is great.

[Speedy G]

Hey Eric, no worries about the solidworks file. It's EV West intellectual property and I respect that. I talked to them, but they didn't seem to really get into the project. They gave me some recommendations for which I'm grateful, but they're more focused on lower cost systems like yours where you have to put together a bunch of systems to be able to have a complete system. The truth is, the only parts I could buy from them are the batteries, and there's better prices out there, not by much, but still.

AC propulsion motors I think start at 25k for each with the controller, but I can hook up a regular fuel gauge, a regular odometer (to 13000 rpm though), and even a regular speedometer (2.5V/12,000rpm). All I have to do is hook up the motor to the controller, the batteries, the gauges, and the gas pedal, and I'll be ready to roll, charge, etc. I was lucky to get a second hand ACP system so the cost was much lower, but still a little higher than yours. The second motor/controller, and the second 200hp will take more time. Also, the ACP advantage is lower weight (50kg/110 lbs for the motor) due to a higher rpm design. Higher RPM 400 Hz designs need less back-iron.

In terms of the chassis schematic, I'm sure there's been some changes to the chassis as well. I do plan on doing some structural simulations since I'll be using quite a bit more weight, which is why I prefer redoing the digitizing. It might be useful for others since there's the question of the brace bars in front of the seats, etc, so I'll probably post it if I get it done properly. I'm not sure how FFR would feel about me posting it though. Maybe I'll distribute it on a case by case basis.

Btw, since I'm thinking of only running in second gear initially I won't be needing a shifter (well, reverse is a switch, and I will have a start button). Eventually, I'll be getting the mastershift system so my shifting cables can be really short and shifter motor can be right next to the tranny, behind the batteries. The only issue, then would be the e-brake cables. There is an e-e-brake conversion where it becomes a switch, but I'm thinking it might be more fun to have a mechanical one. If I'm going to have a heavy butt, maybe I can use it to drift a bit? Also, there's the issue of how to route wiring. I'll be using the isis stuff, so few wires will be needed. I'm thinking of hanging them out the side like the water ducts that go to the front. Btw, one reason for having the batteries in the tunnel is that fewer long power cables can be used. Only a return cable from the front to the back would be needed (battery ground sounds like a good idea).

In terms of BMS, I love the display of the LIGOO system. I'd love to have an automatic balancing system, but I'm not paying $3500 for battery monitoring and questionable balancing. I'm not sure what I'll do, but what I'm thinking is monitoring the voltage of each group of 9-12 batteries. 9-12 seem to be the most convenient grouping when looking at Eric's build. The software is easy enough, It's just reading from some a/d converters and compare them to historical and common sense numbers. The issue might be the voltage conditioning circuits (I'll need isolated differential reading and conversion of 28.8V -> 5v or 38.8V ->5v). I'll probably use some of the design of the ligoo displays on the display computer.

[Speedy G]

At this point I'm realizing this project is getting so big that I'm going to split it into sub-projects. That way I can focus on just what's to be done for the current stage and this thread can be more focused. Stage I is the stage after which I register the car, and the following are what I see as landmarks or stages:

1) Stage I: The most basic system needed to get registered. Drivetrain consists of 1 ACP controller, 1 200hp motor, 1x 5mt tranny stuck in 2nd gear, no clutch (direct coupling), 100 batteries, basic instruments (gas gauge, speedometer, odometer, basic gear control (N, D, R, but not sure how yet). Obviously this includes a finished 818 in every other aspect. Wiring will be Isis, and at this point, no bells and whistles. The toughest problem at this stage is the coupling of the electric motor to the 5mt tranny. It's going to have to be custom made. Fitting the batteries will be another issue.

2) Stage II: Although some of the work will be done in parallel to stage I, this stage includes a display computer with a touchscreen mounted on the center column. There's tons of data available, so the inital display computer will just show data (rpm, speed, motor temp, controller temp, battery voltages, battery current, battery capacity left, etc.). The display computer will also include the entertainment system. The cubietruck computer will be hooked up to a DAC that drives a power amp directly. Media options will just be bluetooth, local hard drive and USB drive. I'll post everything regarding this project in the body & interior forum.

3) Stage III: This stage is about getting the second 200hp motor online. This calls for redesigning the power stage of the ACP controller. Hopefully the ACP logic circuits can be used from both PEUs and be hooked up to a much smaller footprint power stage. If I have time I'll start doing the power stage design in parallel since all I need is my laptop.

4) Stage IV: This stage is about control systems. Speed dependent power assisted steering, brake vacuum asist, customizable traction control, different drive by wire programs and regen. And the mother of all control systems, the shifting control system. I'm thinking I'll need a second onboard computer for this.

... Hopefully I have more than 2 years worth of work ahead of me. I'm more of a grease monkey than a pilot, so it should be fun.

[Bob_n_Cincy]

He speedy,
Now take stage one and split it into a 15 to 20 weekend task.
I thought the ev part would take most of the time and the kit would be erector set. I WAS WRONG.

I don't know this guy(Steve Cornett), but he made and adapter and flywheel for a Subaru MT5. If you contact him, let me know what you get.

This is one of the better conversions I've seen.

Looks like he's get 40 mile range with a 16 kWh pack. The lite 818 should get 20% more.

Here is a link to his motor adaptor. http://www.electricsubaru.com/adapter.html
Bob

http://www.electricsubaru.com/


http://www.evalbum.com/4229

[Speedy G]

VERY helpful, thanks! Erector set... yeah, I've been reading. If it were easy would we be doing it? Actually the EV part for my stage I just looks time consuming. The rest looks like it needs some real skills.

[ehansen007]

EV West makes the adapter plate andd coupler as well since they did model it for my car. They are just busy as hell over there right now. I had one on order from Rebirth however the integrated coupler/bearing was having wear issues with the key on the shaft of the motor. At least on the HPEV AC shaft anyway. I remember talking to Steve Cornett as well. I was talking to EVERYONE at the time as it was hard to find info and make the right choices! Again, be wary of the stock WRX transmission even in 2nd gear only. If the tires are hooked up the stress is going to go straight to that tranny and I'm also not sure how long those bearings will last going over 7K RPM constantly. Might want to look at a rear diff or purpose built gear box for single speed.

[Speedy G]

Yeah, the thing is I'm running without a flywheel and without a clutch. Since I inherited the white lightning rig, the tranny will hook onto a converter that couples directly to a spline socket. For this reason, adding a flywheel would be more difficult. So it was either do without the flywheel and clutch, or totally redo the rig that couples the two motors, so I opted for the easy way out, even if it's not the most efficient. All I have to have custom made is the coupling between the spline socket and the tranny splines, with 1 additional issue, the coupling should be insulated. Also, aligning the tranny with the spline socket is very difficult. Just a little off and your bearings are kaput. I might use some kind of forgiving coupling, we'll see. I'm still not sure how much room I have.

Btw if it's the bearings that die from going over 8kRPM then I'll be happy. Those I can replace with ceramic bearings. I'm more worried about the gears themselves. The higher the RPM the higher the stress on the gear teeth, and I don't think the relationship is linear... I figure I won't be driving much over 100 kph often (that's 7kRPM speed). At least not initially. Once I get it shifting, there's really not much advantage shifting at 10,000 RPM (optimum shift point) vs 6000 RPM except that the motors might heat up more shifting at lower RPM. I've been wondering if oil cooling would benefit the durability of the tranny. It would have to be a DIY kit though.

[John Rubicon]

Hi Speedy G,

I just read this thread and may I say, you are an intrepid builder. I've used the ACP drive and would be delighted if someone replaces the air-cooled IGBTs with modern silicon cooled by water. Keep us posted on that aspect, please.

On your discussion regarding autoshifting, I have some experiences to share. Getting custom firmware from the Mastershift guys is, in my experience, not worth the wait. Been there, done that. The device uses a cable pull, but the potentiometer-encoders are at the motor end of the cable, so cable stretch makes for less than precise shifting. This may not manifest as a problem in transmissions with loose gates without lockouts. We ended up just driving the Mastershift hardware with our own H-bridge and a separate controller. Worked, but not as reliably as we wanted. I'm going to look at using a servo-hydraulic x-y shifter that VW made. I have it and it's OEM quality.

Another potential problem is that the ACP motor has a speed encoder, not a position resolver, so its "RPM" output from the controller is a low precision variable that gives a poor state estimator for matching RPM targets during shifting.

I'm using PMAC motors and also plan an autoshifting setup.

As for 12000RPM motors and transmission life, I have to be honest and say that my experience is that the bearings go south quickly above 8-9000RPM. My PMACs are 6-7000 RPM and 530 ft-lbs.

I look forward to your thread and in collaborating with you.

JR

[John Rubicon]

Speedy G,

one other thing to consider is that the ACP motor output shaft should be galvanically isolated from your transmission to prevent the large induced currents in the rotor from driving current through the rotor bearings and causing them to gradually fail. Heck it may even throw a ground fault if you don't. The belt drive provided this function in the White Lightening prototype. ACP used a Honda(?) clutch hub mounted to a phenolic disk to provide that isolation in Gen 1.

Oh, and if you do replace the controller phase boards, please don't throw them away! I and others need spares ;-)

JR

[John Rubicon]

Sorry, double post...

[Speedy G]

Awesome to hear someone' s further along my madness. Good info on the bearings and the mastershift stuff. At least now I'm not flying as blind. Yeah, I figured I'd have to drive the mastershift with my own h-bridge. Let me know how that vw x-y shifter works. Is it easily available or is it a 1-off? It actually sound like a cheaper option? I still like mastershift's paddle and bump shifters though.

Hey John, do you have any info on the igbt outputs? I'm needing switching frequency and maybe how much they dampened the switching. An o-scope image would be great with scale info. I can' t start working on that since I don't yet have the voltage to turn the suckers on. I already have the IGBTs, the conditioning circuits and the cooling plate. All I need is the snubber caps and I can drive one of the motors with a microcontroller. For that I need the frequency and a sample voltage/time graph.

[John Rubicon]

The x-y from VW was used in a couple of european models over a decade ago. So OEM, but rare. I could make it a paddle shifter with the right code and a paddle shifter hardware set that mounts to our GTMs.

Typical switching frequencies are 4-12kHz. Higher than that and the IGBTs get lossy. I have some experience with SeimKron IGBTs - great products. What IGBTs are you using?

Switching, of course is modulated to try to follow the back-EMF of the motor At higher speeds, the more irregular the back-EMF, the high frequency modulation one has to do to control the motor efficiently. At some point you run out of frequency headroom trying to match the ripples in the back-EMF.

The above should be tempered by the fact that I have experienced only PMAC control and not induction motor control. The PMACs can have a "lumpy" V vs. time graph and so switching speed limits effect how well you can "follow" such a curve. Induction motor back-emf may be much "smoother" and therefore easier to control.

In either case, a gate driver will help you protect your expensive IGBTs as they have safety timing features built in. noise, stray capacitance and other effects can alter the signal coming from your micro-controller before it gets to the IGBT gate terminals. So, typically, a gate driver circuit is located very close or on top of the IGBTs. IF your not careful, however briefly, you may inadvertantly command both the upper and lower IGBT on a leg to be "on" at the same time. That's a direct short and a very expensive fireball. The gate driver circuit takes the signals from your microcontroller and then triggers it's own moderately powerful transistors to trigger the IGBTs.

[Gary Livingston]

Hey John, do you have a build thread? What brand/model motors/controllers are you using?

Cheers