EV Attack project

[fastco]

Jeff, I can't tell you how informative and helpful your posts have been. I purchased an Attack last year with the intention of going the EV route. Fortunately, I have been able to avoid costly mistakes and pitfalls by letting you do it first! After many problems having nothing to do with cars, I am finally beginning work on the car. I am planning to incorporate the same AC150 motor but am interested in having a transaxle with more than one gear (if for no other reason than having the luxury of going over 80 mph and not hearing a motor scream at 12000 rpm while doing it) Wondering if anyone has used a Porsche 911 transaxle?? Should be beefy enough to hold up to the torque? Also, don't know if this is prying, but what did the cells from A123 cost? I know that they don't just sell to any Bob off the street so I am thinking of using a friend at a university who happens to run the electrical engineering dept. to (a) purchase the batteries (b) use the project as a tool for his class. What would be the possibility of the two of us flying out to have a look at your car?

[Jeff]

Originally Posted by fastco

Also, don't know if this is prying, but what did the cells from A123 cost? What would be the possibility of the two of us flying out to have a look at your car?

No Problem. The A123 M1 series cells, in quantity will set you back a little over $10 each in many thousands quantity. So the math is pretty simple, my cost for the cells alone are $36k And that excludes all of the interconnect and packaging necessary to turn them into battery banks. Expect another $10K for that.

Great to hear someone has gleaned some value from my ramblings about the Attack EV project. At last check, I reached unity with the cost of a Tesla roadster, and it's still not finished. I'm in so deep now that I can't stop.
My wife suspects I've developed a gambling problem. I wish it were that easy

You're certainly welcome to come and take pictures and notes. It's an open-source project that all are welcome to visit and to learn from.
It is complex by design, and requires an in-house machine shop when you take on a project of this scale. Additionally, you'll need the help of a competent EE with ambidextrous software kungfu.

Two Speed Transmission:
Well, on a "do over" I would opt for a three speed auto transmission and omit the torque converter. This is the most sensible approach in an AC motor EV. The transmission would be reworked and valved for two speeds; say 1st and 2rd gear, and shift up-downn would be commanded manually.
The advantages of this are: (1) no clutch (2) shift on the fly.

The disadvantage is that you will have to provide an external electric hydraulic pump for transmission band clamping, and modestly for cooling and lubrication. A perfect example of a compact and very high pressure hydraulic pump: the 1990 Range Rover hydraulic assist brake pump made by Wabco ~$900 http://www.britishparts.co.uk/.sc/ms...664/ABS%20PUMP

And I have no idea why those rocket scientists at Tesla Motors did not use this approach. Their two-speed trans is a modified manual design, and it eats gears, spitting them out like hillbilly teeth. They are shipping the initial production as single speed until the replacement transmission design is integrated into the production vehicles.

Regards, Jeff

[fastco]

Yikes...46k for the sparky end of things huh? That's a little discouraging. And I suspect AC systems won't be knocking at my door tomorrow to give away a "surplus" 150kw either! (And I assured the Mrs that I could do it for less than a Tesla)

Okay, for simplicity sake, why couldn't I just use a stout 5-speed transaxle, (Porsche?) and leave the gears in? Sure I would sacrifice a few lbs, but at the end of the day still have my hair and sanity.

As for the machine shop, we got that licked...it's the EE end of it that I'm shy on, thus me enlisting the talents of the local university. However, I've just learned that my partner in this project has the financial stability of Enron, so I fear that this project may be awhile in the works.

[Jeff]

Originally Posted by fastco

Okay, for simplicity sake, why couldn't I just use a stout 5-speed transaxle, (Porsche?) and leave the gears in? Sure I would sacrifice a few lbs, but at the end of the day still have my hair and sanity.

The transaxle with clutch approach is standard for a DC motor drive package.
I should have elaborated a bit on why I recommended the modified auto trans approach with an AC motor. I'll explain:

A series DC motor is typically limited to less than 8K RPM. Anything higher would result in destruction from centrifugal force (flying apart), or plasma arc-over on the commutator. Additionally, the power band is relatively narrow, and so gearing is useful for operation at highway speeds.

A variable frequency AC motor can easily spin to 13K RPM and enjoys a very broad power band by comparison.

If you attach a manual shift transmission to an AC motor, a gear change can only be safely accomplished when parked. Remember that AC/DC motor torque is nearly 100% at zero RPM. If you start out in 3rd gear, you stand a strong chance of spitting teeth off the final in an ICE transmission since the wheels might not spin to free up the torque. So, they start in 1st or 2nd gear.

When you add a clutch to an AC motor, you risk a very serious chance of splodin' the clutch, and seriously over-revving the pressure plate simply because of the high-revving nature of these low mass variable frequency AC motors. Bad things can happen.

So, if you want to use an AC motor, and need another gear, do it with a modified auto transmission, functioning as manual shift. The torque converter must be removed to reduce centrifugal mass (and wasted energy). The "shifting" is accomplished using friction bands, and is tolerant to high RPM operation. An external pump is required to power the bands, and for lubrication. The wiki below might help understand the concept a little better.
http://en.wikipedia.org/wiki/Semi-au...c_transmission

Also, I failed to mention earlier that much of my success in assembling this disaster-of-a-kit into a vehicle is due entirely to Flyinfej and his www.attackbuild.com website. Much of his work serves as the primer for what any builder would encounter along the way.

Regards, Jeff

[Jeff]

I've provided a few recent images.
The first represents the installation of the S2k electric rack, and the vacuum pump/cylinder for assisted brakes. Brake lines are still not installed. Vacuum storage is left and below the rack, made from a recycled propane canister. Diaphragm switch on the end controls the pump motor mounted to the right.
The original fiberglas firewall had been modified for heating/cooling by the first owner, and so a replacement was fabricated from sheet aluminum (not shown).

The second is another view of the rack, and how the tie rods did not require the extensions used on the Accord rack.

The third is a firewall inside view. Steering linkage and pedals were installed to locate the firewall panel. Side note, I moved (heated and bent) the brake pedal arm 40mm to the left to improve the positions. They were too crowded before.

The fourth shows the gaping hole in the engine space, without the power electronics bay installed. All of the original supports and truss pieces for the F22 were surgically removed from the framework. Extensive rework was performed on the underside of the rear-most stamped accord subframe to accommodate moving the transmission back two inches, otherwise the CV shafts would not have been parallel. The front engine support has been attached to the frame underneath (hidden) to limit deflection from the motor torque.

The fifth image shows the lateral support arm. In the lower right of the image you can see how the transmission extends beneath the rear accord subframe.

It's obvious there have been extensive modifications to the Attack to accommodate an EV design. Next week I'll post a few newer images showing the assembled motor and electronics bay.

Regards, jeff

[fastco]

Anybody have any idea what's going on at AC Propulsion? We called yesterday to source out our motor, and got the ole' "sorry, we are unable to fill any orders at this time, can't get parts...bla, bla, bla" routine. They suggested contacting UQM out of Colorado. ???
Jeff, thanks for the 'splaination of the transmission dynamics, it makes total sense, I just never thought about it long enough. Right now I'm so frustrated with this project that I'm considering scrapping the whole thing and selling the car.

[Jeff]

Call me when you have a few extra minutes. I want to discuss your application and needs on your project, and then conference into ACP. ACP may have simply "kicked you to the curb", as they have to frequently do because their intent is to sell to OEM's and not individuals.

Expect to spend $35k, and perhaps a 25-40 week ARO to delivery. This is completely normal for this industry.

Sadly, UQM likely won't be any improvement on materials or delivery. They've just now announced a competing BLDC motor/controller package rated at 150kW. At worst case -it's vaporware, best case - prototype hardware and software.
http://www.uqm.com/products/motor.html

Also, the motor and controller are only half of the fun. Have you sourced a supplier for batteries? There's a couple new large format LiFeP04 battery supplier offering impedance numbers close to what A123 is delivering on.

Regards, Jeff

[Jeff]

Following completion of the pedals and other firewall items, I moved onto building the brake lines.
These were already run by the previous owner, though it really needed a do-over. I reviewed other car build images, and attackbuild.com. I was surprised to find all had used long flexible lines. Mine were feeding to compression tee's. What the heck were they thinking And what was the idea of the goofy remote brake fluid reservoir I read that Jeff (flyinfej) fabricated an extension block to extend the master cylinder assembly beyond the shock/spring for clearance on his car, and wisely used AN flare fittings. Nice. Most other build images were too ambiguous to identify what they had done with the brakes.

The previous owner of my kit had attached a barbed cap over the top of the master cylinder assembly and secured it with what looks like JB weld. I could not determine if it was a threaded, or pressed on cap with epoxy as a gap filler. I tried heating it with a heat gun to soften the epoxy, but it would not turn or move at all. Crap. I resolved to fabricating a support plate for the original remote reservoir, and linking the cylinder and reservoir with right angle hose pieces and barb fittings with spring clamps. It's not the cleanest, but it will do. My posting from a week ago has a firewall image that shows the reservoir and master cylinder mounted.

I decided to go with solid brake lines. I bought a set of four aftermarket stainless lines for a Prelude, along with a roll of 3/16 tubing, two tee's and nuts. I've fabricated four flange mounts to secure the short flexible brake lines. These are welded to the frame, with a spring clip holding the fitting to the welded flange. The Prelude flexible lines used 10mm metric nuts rather than the SAE nuts used for the tee's and master cylinder. I know that Flyinfej used and recommended the AN flare connections, but I went ahead with the standard hardware and was careful with flaring. I had a high quality flare tool.

The CV shafts were a quiet do-over.
Months ago, a CA racing equipment fabricator evaluated the motor performance numbers and recommended I use hollow shafts. They'd twist less, and would be lighter than the solid shafts. They are unequal length, unlike the f22/h22 that use an intermediate shaft.

Well, an armchair-grade review from an engineer who came to visit last Oct explained some important dynamics of the torsion flexure in the solid shafts. He noted the Tripod joints (inboard CV joints) would snap during a wheel hop, as none of the peaked energy could be accommodated by the hollow shaft (no twist).
I had that same feeling you get after eating bad chinese food.

Last Nov I called a highly recommended CV joint/shaft supplier for the race industry, and asked for an independent review.
http://www.markwilliams.com/custom.aspx

Brett at Mark Williams Ent. in CO was extremely helpful, and after reviewing the design, he offered a couple of choices: (A) new spools and shafts, custom fabricated to the car at $5k (bulletproof), or (B) fabricate new CV shafts to fit my original stock joints, sized and tempered to allow for additional twist, and hopefully not lay waste to the Tripod joints at $1500.
I went with plan B. The new CV shafts arrived early January, but the Tripod joints did not fit the CV shafts. It turned out a CAD mistake carried through the order, and they splined it to 31 tooth. An oops on their part. New shafts have been fabricated and tempered, and should be here this week.

Regards, Jeff

[flyinfej]

lol, you mean non-standard fittings into compression fittings followed by ten feet of flex line didn't seem the best choice for brakes?

I mainly went with AN because I was having a hell of a time sourcing some metric fittings. AN is technically 'better', but I would have been plenty happy with double flared... anything but that crap in the kit.

[cbulen]

subscribing

[Jeff]

I received the two replacement CV shafts from Mark Williams on Monday.
These were the new shafts that replace the two I received the first week of Jan that were rejected for a spline error on their part (see earlier post).

To ensure there were no more mistakes; I shipped the CV joints and the Tripods back with the rejected shafts to use for test fitting at Mark Williams.
Long story shortened, on arrival to me, the Tripod spline was oversize on one shaft and required a 12 ton press to install the Tripod. I wondered how they tested them?


The saga goes on and on like my aunt Margie's ass.

So, I sent the one shaft back to Mark Williams on Tuesday for rework along with a few choice words to describe my disgust. It arrived back to me on Friday, and appeared they had broached the spline to allow for a slip fit on the Tripod. It wasn't the cleanest since they originally cut the splines with an indexing horizontal mill, and the broached reworking looked like crap, though it does fit. Thankfully it appears I have closure on this.

I'm now ready to complete the drivetrain installation work.
No compliments to the folks at Mark Williams Enterprises...

Regards, Jeff

[Jeff]

I've been busy with work and had little time for the Attack lately. The brakes were completed a few months ago, as was the new firewall.

The brake lines were assembled from 3/16 steel tubing and standard flare to tee's. I used aftermarket flexible steel lines between the flange mounts and the calipers. The flanges were fabricated and welded to the tube frame.

The original fiberglas firewall was full of holes, and a complete mess. I decided to fabricate a new firewall from sheet stock rather than trying to trying to salvage it. It's made of .040" sheet aluminum (two pieces), and coated with a plastic spray-in truck bedliner coating.

Jeff

[Jeff]

I welded a support bracket for the Bias Valve to the back side of the front shock support. It's close to the master cylinder, and easy to adjust if needed.

The brake system is vacuum assist by a compact electric diaphragm pump. A vacuum switch is attached to the storage cylinder. The storage cylinder delivers enough volume for two complete pedal cycles before the pump activates. The switch activates at 14" Hg, and stops at 21"Hg (7"Hg hysteresis).

Jeff

[Jeff]

I have some images to post showing the Attack before the body was installed.
I forgot how dirty the shop was from all of the fiberglas dust. Anyway, there are images showing the PEU (Power Electronics Unit) installed to the rear frame and view of the new outriggers that needed to be fabricated. The outriggers carry a battery payload on each side, and accessed from beneath for service. The zig-zag rows of rivnuts around the outriggers for the side pod battery packs dictated the 1.5" square stock material dimensions I used. I'll follow up with images of what the side pod packs appear like later.

Regards, Jeff

[Jeff]

One images shows the flexible brake lines coupled to the hard lines. And another image of the motor and controller.

The next images will be of the body installed, and my troubles with the door hinges explained.

Regards, jeff