Thursday, August 30, 2007

Taking a Break

These last few days have packed with non-EV stuff, but I did take the DC-DC controller into the lab to see what was going on. First of all, there are no kick-back diodes on the inductors inside the converter. This kinda scares me since pulling on an inductor with a FET and then letting it go tends to create really high voltages (> 600V etc...).

I hacked together a 160V DC power source by simply tapping off the bulk capacitor inside one of our power supplies. It fired up the DC-DC converter, but the whole thing was moving up and down at 60 Hz, it was darn near impossible to look at things with an oscilloscope.

Next week I plan on taking five 32-volt isolated supplies (2.5 amp) and hooking them up in series to form a 2.5 amp 160V supply with a non oscillating ground so that I can probe all the internal nodes better. Fortunately, we use these supplies all the time for our inkjet printers, so there's piles of them around.

In the meantime, I'm e-mailing with Mike at Electro Auto to try and figure this out before he sends me another DC-DC converter to blow. I'm going to try and get by with charging up the 12V battery at the same time as the 144V pack and just running them both down at the same time. During the daytime, this isn't bad since the headlights consume most of the power. I hooked up a .01 ohm resistor to measure current draw from the battery and found that I can pull 30 amps with everything on, but this is rather atypical. Supposedly, the battery can supply 30 amps for an hour, so I should be okay with a 30 minute commute. This is still frustrating.

Remember, it's just a process...

Sunday, August 26, 2007

PakTrakr is Online! DC-DC blown again...

I spent most of yesterday and a few hours today wiring up the PakTrakr system. It's a bit of a pain to disconnect the most negative lead before attaching the display, but that's minor nuisance. It seems to measure pack voltage well; however, the voltage measurement on each battery seems to be slightly inaccurate. The Zivan charger is balancing the batteries well and each of them is within an 8.48V-8.50V range (20mv difference is pretty good!).

I chose to mount the display on the central console as shown above. It doesn't have a backlight but is easy to read during the daytime and the placement makes the buttons easy to access.

I had to greatly extend one of the PakTrakr remote module cables so that the rear battery pack could talk to the remote modules in the front of the car. That involved cutting one of the 3-wire cables on one remote module and splicing in an 8-foot extension of additional 3-wire cable. I added heat-shrink tubing to make the splices clean so they could be pulled through the right-hand heater duct under the passenger-side door.

The system seems to work fine. Having a backlight would be nice, but not necessary. My biggest gripe is that double sided tape seems to have a big problem sticking to the back of the PakTrakr modules, even though they tell you to use it to mount the modules.

On a downer note, the CCPower DC-DC converter blew again (fuse was vaporized and FETs shorted out). I actually caught the system in a working condition just before driving home and found out that it had blown when I got home. The only difference was that the DC-DC converter was under a heavy load on the drive home because the headlights were on.

Based on the circuit inside the DC-DC converter, I suspect that there is a significant amount of current (required from the headlights) going through 144V side. When the FET opens, the current rushing through the inductor causes a high voltage spike (300-400V is not unheard of) that gets added to the +144V side and just wipes out the 500V FET. I'll have to double check to see if the inductive kick-back diode is properly functioning. The CCPower website only has the input voltage of their DC-DC converters listed as 80 volts so, although these are designed for EVs, the underlying design may not be able to handle the higher voltage, especially with spikes from regenerative braking.

In the meantime, I'm using a 12V charger to charge the accessory battery as well. Maybe I can integrate that into the overall charging circuitry when I plug in the car.

Saturday, August 25, 2007

Talking with Mike at ElectroAuto

I just got off the phone with Mike at Electro Automotive. We chatted about the blown CCPower DC-DC converter and the engine vibrations. Since I fixed the DC-DC converter, we agreed that I would try to find out which conditions would cause it to blow again, if it does. It seems that Mike is more of a mechanical person than an electronics person. He graciously faxed me a circuit to generate a tach pulse from an ADC8" or 9" motor, but he didn't really understand the circuit. Unfortunately, the circuit relies on the fan inside the motor interrupting an optical emitter-receiver pair which doesn't exist on the AC24 motor in my vehicle.

He also agreed to provide a new motor to flywheel coupler if I still had engine vibrations. Mike commented that it was probably not the transmission, but probably had to do with the flywheel/clutch assembly. I'll have to do some more experiments to figure that out when I get the time and energy to take the motor/tranny back out.

I also said that I would look into the peak amps that the system draws if I could find a suitable ammeter. I still don't have the shunt or the console gauges since ElectroAuto is still waiting to see what peak amperage it has to deal with. When I told Mike that I easily pulled > 150 amps during acceleration, he seemed a bit surprised.

Apparently I'm the only person with an operational AC kit currently, although several other people are hot on the trail to finishing theirs. Mike doesn't have an AC kit 914 back at the ranch.

Regarding the 8V golf-cart batteries, Mike also mentioned that they tended to either last 3-4 years or die right away (infant mortality). I hope I don't have a batch that demonstrates the "infant mortality" behavior...

Okay, I'm going on a bike trip next weekend and need to get biking gear, so the 914 will have to wait for a bit.

Friday, August 24, 2007

Fixing the DC-DC Converter

I cringe at the thought of having to send anything back to Electro Automotive for replacement or repair, so I took a chance and dissected the DC-DC converter to see if I could fix it. The company that makes the DC-DC converter is CC-Power in the UK, so I don't think sending it back to them would be timely either.

Here's the DC-DC converter taken out of its black aluminum case. The clips that held the FETs and diodes in place were a bit annoying to take off, but I finally got them with some needle-nose pliers. The circuit is basically a heavy-duty buck step-down converter with a bunch of inductors. The output stage is a transformer in push-pull configuration with two diodes to handle both sides of the coil. An opto-isolator takes the output and transfers a reference current back to the primary stage to regulate the voltage.

To the best of my knowledge, the inductor sitting on the table is directly in line with the +144V and is used to limit inrush current. The transformer coil on the left does the same to the +13.6V side, basically just a series inductor. The transformer coil on the right is the isolation transformer with +144V and +13.6V windings to transfer the current. The two TO-220 case parts on the top are the diodes to rectify the current coming out of the output side of the isolation transformer. The two filter capacitors on the left smooth out the +144V and +13.6V. The two FETs on the bottom edge of the board pulse current into the input stage of the isolation transformer. The rest of the circuit is just integrated circuits to drive the FETs and monitor the output voltage.

Here is one of the two IRFP350 FETs that clearly blew up. If you look closely at the burn marks, there is even a hole blown in the part package. Both FETs had blown and were shorted out. Since the FETs are rated at 16 amps and the fast-acting input fuse was rated at 5 amps, I suspect the FETs sustained damage from a high-voltage spike that went past their 400 volt rating.

The explosion from the FET shown above was so powerful, it even blew a hole in the steel clip that held it against the heat sink. (I'm glad I wasn't around when it blew!).

I went on a quest this evening back to the labstock at my work and found some IRFP460 FETs in the same package. They are rated for 500V and 18 amps instead of 16, so I thought they would be a good replacement. Heck, I might as well at least try since I probably voided the warranty when I opened up the case.

Here is the DC-DC circuit board with the new FET leads sticking out the bottom edge for resoldering. The middle lead is tied to the back of the FET, making it a good heatsink and difficult to solder without creating a cold solder joint.

Okay, I did a quick assembly without heatsink clips and attached the 144 volt pack to the input with a new 5-amp fast-acting fuse, and voila, the DC-DC started working again. So, I went back and did a full re-assembly with putting the heatsink clips back on and attaching the 12V accessory battery. The DC-DC seems to be working fine again. It creates a small whine unlike the blown unit before and it puts out 13.9 volts to charge up the accessory battery.

I have no idea why or how the DC-DC converter blew, but I suspect it will probably happen again given that I've only really driven the car for four days. As long as I check the fuse I'll know if it went bad or not. The red LED on top of the DC-DC converter lights up with a 12V accessory battery attached, even though the unit is blown, so I can't rely on that. Based on an idea from Ross Cunniff, I could rig up an op-amp comparator circuit that lights up the the generator warning light whenever the 12V accessory battery is below a certain threshold. That way I know the DC-DC converter isn't working and I can check it out. The other idea is just to add a green LED to the 144V inside the converter just past the fuse so that I can see at a glance if the fuse has blown. There are many other projects that I'd like to do first though, so for the meantime, I'll just periodically check the DC-DC fuse.

Good Night.

Getting the EV Back to the Shop and Fixing Electrical Stuff

Just before noon, I drove with the ICE vehicle over to the 914, charged the accessory battery with jumper cables for about 20 minutes, then quickly drove the 914 back home to the garage for repairs. Since the stranded EV was only 2 1/2 miles away, I walked back to the ICE car and drove it home too. Since the DC-DC converter was blown and I didn't want to deal with it, I focussed on some other electrical problems I'd been having.

The first issue is that I get major voltage drops when I turn on 12V accessories like the headlights and the turn-signals. Because the accessory battery is now in the front of the car and tied directly into the fuse block, the actual current path to the ignition switch (and the lights) goes from the fuse block all the way back to the original 12V battery position (see step VI.F.7 in the kit instructions), through a bolt connecting the large red wires, and then back to the ignition switch. The accessories were causing the voltage to drop at the main accessory battery, but I noticed an additional 1.5 volt drop from the wire resistance going to the back of the car and front again.

To fix this, I found the two thickest red wires under the fuse block going to the back of the car. I cut them both near the firewall and used an ohm-meter to figure out which one went to the ignition switch. I then added a yellow 1/4" quick-connect connector on the end of the ignition switch input routed it along the wiring loom and plugged it directly into the fuseblock where the other thick red wire came from. This eliminated the 1.5 volt drop from the long wiring and the 12V system is much happier.

The second issue I wanted to address was the lack of a heater/cooler fan blowing out the ducts. To get at the contacts which turn on the blower motor and fan light, you simply remove the ashtray in the center of the dash all the way (press down on the center clip).

Here's a look into the ashtray hole down onto the electrical contacts (just behind the black/blue and brown wires near the center of the picture). These didn't seem to work at all, or only intermittently. I globbed a bunch of dielectric grease onto the contacts and used a long flathead screwdriver to scrape away all the corrosion and carbon buildup. There was a tremendous amount of dust and other dirt in there that I cleaned out with an old worn-out toothbrush.

After all that scraping (the dielectric grease acted somewhat as a solvent too), the fan operated at all three speeds and the fan light lights when the fan goes on. The next step for the climate control system will be to add hair dryers to heat the cabin. Unless I come up with a better idea, I'll probably tie the hair-dryers directly to the 144 volt system through relays controlled by the rear-defroster pull-knob.

DC/DC converter blown

Well, I ran out to Radio Shack this morning and got some more 5 amp fuses for the DC/DC converter. After plugging a fast-blow fuse in, powering up the car blows the fuse. I tried a slow-blow fuse and that went too. I then disconnected the 12V battery to see if that was causing a problem and it still blew a fuse.

What's sort of interesting is that these glass fuses didn't just melt and break, they really vaporized and completely coated the inside of the fuse glass with metal bits. It looks like there's a pretty serious short in the DC/DC converter.

The car is still stranded at Krista's place. I brought over jumper cables and tried to charge the battery with my ICE vehicle. The voltage drop across the 300amp jumper cables was 0.8 volts so I'm guessing the 12V battery was mostly dead. Since this is my day off, I might take a risk, charge it up some more and try to crawl home for repairs. The distance is only 2.2 miles and the main 144V pack is happy up around 151V. Grrr...

EV Toolbox

Before the accessory battery discharge yesterday, I wrote down a list of things to keep in the EV in case there was a breakdown out on the road. I came up with the following items. Any comments for additions or modifications would be welcome.

Emergency Equipment
  • cable cutters
  • small fire extinguisher (for electrical fires)
Repair Tools
  • Haynes manual for electrical diagrams
  • digital multimeter
  • wire crimper and extra pink,blue,yellow quick-connect crimps
  • two battery wrenches wrapped in electrical tape
  • electrical tape
  • extra cable lugs and hammer-based crimp tool
  • medium hammer
  • LED headlamp
  • extra fuses (DC-DC converter anyone?)
  • pre-made 24" battery cable with lugs (from NAPA)
  • #2 philips/standard screwdrivers
  • pliers
  • small utility knife to cut insulation
  • small 12V SLA battery to bootstrap the system if 12V battery dies
Again, what would you add/subtract/modify in this kit?

Next up: fixing the stranded car, looking at the PakTrakr and getting the heater fan to work.

Thursday, August 23, 2007

12V Auxilliary Battery Dies!

While cruising over to my girlfriend's house this evening, the whole system just shut down on the road. That was a rather scary experience. The hazard signals started flickering and nothing would turn on. After pulling over and checking the voltage on everything, the pack was nicely up at 150V but the accessory battery was down to 10.5 volts. With the key on, the DC-DC converter light was on; however, when I checked the DC-DC fuse, it was blown. Apparently, I've been running down my accessory battery for awhile now.

The car is stranded for the moment until I try replacing the fuse to see what the issue is. Worst case, I can charge the accessory battery with a long extension cord just to get the car home.

Ugh, goodnight.


Despite the issues with the car, I've decided to start commuting 15 miles each way to work to see how the car does. By not revving the motor above 5000RPM, I can minimize the vibration problem.

The batteries are starting to break in. I'm getting slightly more power on the freeway now and starting from a dead stop is acceptable. I've taken the 914 up to 70mph. It doesn't have the pep that the original 2.0 liter engine had, but I still haven't really cranked up the AC controller and done a parameter analysis.

The car wobbles at 60mph with a low-frequency side-to-side shaking, so I took it into Les Schwab this morning and they found one of the back tires to be significantly out of balance. I'll take it back up to 60mph on the trip home.

In general, I'm pleased with the freeway performance. Although, if I put the car under heavy acceleration for a long period of time, the batteries "brown-out" down to 128V or so and the controller cuts back the power significantly as the voltage drops. Letting up on the accelerator allows the battery voltage to come back up and power comes back. I could adjust the parameter in the AC controller to set the voltage cutoff lower, but I fear damaging the batteries.

I looked at installing the PakTrakr battery monitoring system last weekend; however, the wiring distance between the front and back of the car is longer than the provided cables. Ken Hall (makes PakTrakr) said that I could extend the cables as far as necessary, so I grabbed some 4-wire cable and will be extending things soon.

I've given several people at work rides to show them the technology and everyone has been very supportive.

Next up: More commuting and checking out PakTrakr

Friday, August 17, 2007

NEDRA Pictures

Going to NEDRA at Portland International Raceway was my first major excursion in the 914 since both the flywheel was balanced and the clutch was properly adjusted. The car handled very well and cruised on the freeway without much problem. I did get up to 65 but kept my speed mostly at 50 to not pull too much current while breaking in the batteries.

Here's the wait in line for the ticket office at Portland International Raceway. I felt proud to not have to waste gas while idling in line.

I spent most of the time answering questions about the 914 so I didn't take too many pictures or see too many races. Maybe next year I'll try it out.

Here's a more serious electric drag racer. As you can see, the original engine compartment up front is fairly barren. The Zilla2K controller sits in the bottom.

Here's looking in the rear window where all the A123 batteries are. There are probably more pounds of high-current Anderson connectors than batteries.

Here's the "purist" control cockpit.

While not a racer, this entry was rather interesting since its electric motor didn't power the wheels...

but it powered a huge fan on the back of the car.

Here's Miles' antique vehicle that he brings to most of the OEVA monthly gatherings.

Miles uses Hawker Genesis batteries to power the antique vehicle. Much lower maintenance than the originals!

This is a Honda Civic Delsol conversion done by Chris Brune who helped me understand how to drive my electric car better using lower gears and higher RPMs.

That's all the photos I took. There were many more vehicles and many fun EV drag races. Just as I left, one EV drag racer had just spun out into the wall. Exciting times indeed.

Grommets Anyone?

In my frustration to take care of the vibrating of the motor/tranny, I unbolted the motor from the front engine mount and raised it up with a jack. The clearance above the motor is tight so you can't raise it too much. After revving up the motor, it still resonates at 5000 RPM which leads me to believe that there's an issue with the motor to transmission coupling. If the transmission shaft is slightly off center then the whole thing could wobble.

To try and fix this, I put two 1" rubber grommets around the bolts between the front engine mount and the black motor mount. After re-tightening all the bolts and revving up the engine, it still resonates, but significantly less. I can probably deal with this for awhile. We'll see...

Off to NEDRA...

Fixing the Flywheel - Still Vibrations!

This has been one heck of a week. I took the flywheel into an automotive machine shop to balance along with the attached clutch assembly. In the meantime, I tried to figure out if the real problem was the motor to flywheel adapter from Electro Automotive.

I borrowed a precision dial gauge from a coworker and measured the wobble on the motor adapter. Initially, it had a full 7 mils (thousandths of an inch) of difference in depth. After trying to alter the tilt by tightening down the two screws on the high side first, I got that down to 4 mils.

Turning the outer hub of the adapter had absolutely no effect, so it must be a problem with the inner mating cone. I filed down the motor shaft key and that had no effect. Only using a semi-round file on the inside of the cone (that touches the motor shaft) under where the key rests started to move things out of tilt. After much filing on the inside of the cone, I got the face of the adapter to tilt less than 1 mil, so I was happy with that.

In order to adjust things and remeasure, I purchased a 7-ton gear puller (with posi-lock) from Sears for $70. It worked remarkably well at popping off the adapter from the motor shaft with minimal effort and no possibility of damage to the threads on the inner cone of the adapter. I had to pull this thing off about twenty times to finally get the adapter face flat, so the $70 was well worth the effort.

One person suggested I send the adapter back to ElectroAuto since the cone was flawed. Quite frankly, they aren't responding right now and with NEDRA happening tonight, I was just too impatient to wait (possibly for months) to get a new adapter...

Here's the modified flywheel back from the shop. The backside was slimmed down to reduce weight and the starter motor teeth were removed to save weight as well. The markings in the middle are my own when figuring out the high and low spots.

With the adapter face now within 1 mil of depth, I put the flywheel on in several positions and it's face now moves in and out only 4 mils. That's acceptable for now.

KIDS: DON'T TRY THIS AT HOME. I attached the motor with flywheel up to the AC controller to see if it would vibrate when I got it up to 5000 RPM. By tuning back the torque and maximum RPM in the parameters of the controller, I safely spun up the wheel to 5500 RPM with nowobbling and very minimal vibration. In the picture above, the flash of the camera makes the wheel look still, but the clutch alignment pins on the outer edge are a bit of a blur.

With things balanced, I finally mated the motor/flywheel/clutch assembly back into the transmission. Notice the flywheel has no teeth and has just enough diameter to hold the clutch pressure plate. Also notice the two drill divots in the side of the flywheel that the shop drilled to balance it.

Here's the assembly on the furniture dolly. When I removed this from the car, there was only one 2x4 under the front motor mount. I realized that there is enough clearance to use two 2x4s and it puts the whole assembly closer to the underside of the car when I jack the motor/tranny into place.

Here it is again, ready for insertion. I taped up the shielding on the 3-phase AC cable to keep the end from fraying.

With the assembly rolled under the car, I simply put a jack under each end of the front motor mount and jacked it into place so I could insert the motor mount bolts again. Since the whole thing has to tilt while I attach the rear transmission mounts, I left the nuts on the front motor mount bolts somewhat loose.

Here's jacking up the rear end of the transmission so I can bolt the transmission mounts back into place. Don't forget to tighten the front motor mount bolts when you're done!

After re-attaching all the other wires and mechanical cables, I fired up the AC motor again and spun it up to 5500 RPM. What!!??? The car still vibrates!!

I think I've been barking up the wrong tree for the past few weeks. Since the electric motor is directly bolted to the front motor mount without any rubber engine mounts, I highly suspect that the 914 chassis has a resonant frequency around 5000 Hz and the light vibrations from the electric motor are stimulating it.

I'm taking the car to show off at NEDRA tonight anyway. I never get up to 5000 RPM if I stay in 2nd gear in neighborhoods and 3rd gear on the freeway.

Next up: Finding rubber mounts to add to the system to see if that helps things out. Grrrrr.

Sunday, August 12, 2007

Taking a Day off to see the EV Show

I took the day off to walk around Portland to the farmers market and EV awareness day with Krista. There were probably twenty electric vehicles at EV awareness day. Here are just a few of the more notable ones.

The electrically augmented bicycle. This was originally just a human powered vehicle made in Belgium, but the owner added an electric motor to power the rear third wheel to get up the hills in Portland (of which there aren't many in Belgium).

Here are some cool looking Neighborhood Electric Vehicles (NEVs). They run on 48 volts and the company sells them as a kit for about $3500. With batteries and additional items, the final cost is around $5000. Not bad for something that has a top speed of 50mph.

Here is my (tongue in cheek) arch nemesis John Benson. He built the DC version of the Electro Automotive kit awhile back and challenged me to a drag duel at Nedra next week. Based on my experience with acceleration, he'd toast my AC 914 in the drag race, but I'd win in an endurance race.

Next Up: Tweaking the adapter hub and trying out the PakTrakr

Friday, August 10, 2007

Balancing the Flywheel/Clutch

Okay, since I take Fridays off, this gives me a chance to take the flywheel into a machine shop for balancing. I took the flywheel and clutch pressure plate to Portland Engine Rebuilders. They basically said that they couldn't balance the flywheel because they couldn't attach their balancing machine to the small center hole. Many thanks to Alan at A&P Specialties (Audi and Porsche repair) for lending me a 1.7Liter 914 crankshaft so I could provide something for P.E.R. to balance.

Here's the crankshaft from a 1.7L 914 that Alan graciously let me borrow.

Here's the crankshaft with the connecting-rods removed.

Here is the crankshaft with the bolted on flywheel. I delivered this to Portland Engine Rebuilders to 1) remove the outer 1/2" of the flywheel (basically the starter motor teeth) and 2) balance the whole assembly along with the clutch pressure plate so everything is stable. Ted, the owner of P.E.R., was very kind and explained the whole process as well as making some suggestions for how to lighten the flywheel without losing strength up at 6000 RPM. Machining the flywheel and balancing should take about a week.

Back at home, I purchased new 10-24 x 1" allen head bolts to re-attach the adapter hub to the AC motor shaft. Since the face of the adapter seemed shallow on the side where the motor key sat, I tightened up the bolts on the opposite side first to try and bring that side deeper to flatten out the face.

Here I'm using a piece of plastic (yes that's a filler plate for the 5.25 inch drive on a computer case) to determine if the face of the adapter is flat. By holding the piece of plastic in the right place and bracing it against non-moving parts, I can turn the motor shaft by hand and hear where the face scrapes the plastic. This second time around I think the adapter is flatter, but still slightly off. I smeared some red loctite on the deeper surface to act as a liquid shim to try and flatten it out more. We'll see what happens when the flywheel comes back.

Next Up: Taking a break for a few days and possibly installing the PakTrakr battery monitor.

Thursday, August 9, 2007

PakTrakr Arrives and Adapter Misaligned

Well, this day definitely had its ups and downs. The PakTrakr battery monitoring system I was expecting early next week actually showed up today.

Here's the PakTrakr still in all its bags. The main display and one battery monitor unit is in the upper-left bag. Three extension battery monitors are in the other three bags. The serial cable in the lower left can hold 2MBytes of log data before dumping it down to a laptop. Each monitor can handle six batteries. My vehicle has 18 batteries, but I chose to purchase a total of four monitors due to the distribution of batteries in the car. This means more cost, but far less wiring and installation hassle.

Okay, on to the gritty part. I dropped a bunch of leftover parts at camp914 today and picked up the 12-point CV joint bolt removal tool from Craig. After removing the CV joints, I rolled the motor/tranny out from under the car and wiggled off the transmission.

After close inspection of the flywheel, it looks like it wobbles about 30 mils (thousands of an inch) which could be causing the problem. I unbolted the flywheel and used a piece of plastic rubbing against the hand-spun adapter hub to see if it was unbalanced. Unfortunately, the adapter hub from ElectroAuto was tilted to one side about 6 mils. This may not seem like much, but this tiny distance gets amplified through the radius of the flywheel up to 30 mils. Just to make sure the flywheel wasn't tilted on its own, I put the flywheel back on rotated approximately 180 degrees. The tilt was still 30 mils and it followed the tilt in the adapter hub, not the flywheel rotation, so now I know it's a problem with the adapter or the AC motor shaft.

After removing the adapter plate and large heavy spacer ring around the adapter hub, I found that one edge of the wedge inside the hub was slightly more depressed than the opposite edge. This is a bit of a relief because it points away from the shaft on the AC motor being bent. I had to use a 7-ton gear puller from Sears (model 104 Posi-Lock) to pull off the adapter hub.

The instructions say that you can remove the adapter by backing out the small allen head bolts and rapping them with a brass hammer. After slightly backing out the bolts, only a few threads were still in the adapter hub, so hitting them with a hammer only broke off the threads. That's why I had to resort to the gear puller. I'm quite surprised that ElectroAuto suggests using allenhead bolts for the hub that are only 3/4" long. The threads are much deeper. Since I damaged the shallow threads, I'm going to mount this again using 1" bolts instead.

Make sure you do NOT put Red Loctite on these small allen-head bolts like it says in the instructions. The flywheel prevents them from coming out and the loctite simply glues the wedge inside the adapter hub tight so you cannot remove it! (my bad)

Here's the mess after I took everything apart.

The object of my frustration: the adapter hub. The left edge of the inner wedge seemed to pull in 6 mils further than the right edge. I'm going to screw in the bolts on the right edge (near the slot) first to see if I can compensate for this tilt.

Next Up: Having the flywheel balanced and lightened at a machine shop.

Wednesday, August 8, 2007

Biting the Bullet - Book 2

In my eagerness to balance the flywheel, I unhooked all the mechanical and electrical cables from the motor/transmission assembly and lowered it onto a furniture dolly. Note that I still need to borrow the CV joint bolt remover from Craig at Camp914, so the assembly is just sitting on the floor right now. I'll try to describe the process in a little detail since Joe posted a comment a few hours ago asking for more pictures.

First up was removing all the mechanical cables. The speedo cable came off easily. I put the end in a plastic bag to avoid dirt and grit. The clutch cable (which I had just adjusted) took a bit more time and I saved the nuts and stop-wedge in a plastic bag. I had to thread this through the engine mount so it wouldn't interfere with the motor coming down. The shift linkage came off much more easily since I had greased it up and the throttle cable came off too since it was wrapped around the motor.

For the electrical stuff, I first popped off the reverse light wiring and tie-wrapped it out of the way. Next came the grounding strap (don't forget that!). The hardest wiring was the speed sensor wiring from the motor.

The picture above shows the speed sensor box opened up. I had to carefully unscrew the wires on the bottom and feed them through the small entry hole and then through the two-inch hole in the wall between the engine compartment and the rear luggage compartment. I also tie-wrapped these in a plastic bag to protect the ends.

After disconnecting a copper bar on the battery pack and pulling off the plug on the Zivan charger (caused sparks before), I carefully unbolted the three thick cables coming from the AC controller. After feeding this thick cable down through the hole in the luggage compartment floor, the assembly was free to lower.

I didn't want to go through the hassle of lowering the whole car to lower the motor/tranny, so I took a different approach. First, I put the rear of the car up on six-ton jackstands. Then, I removed the rear tires to give myself more space to work.

Since I have two floor jacks, I put one under each end of the front motor mounting bar. With the jacks pressed against the bar, I could remove the mounting bar bolts and slide a furniture dolly under the motor. I had to slide the dolly forward quite a bit and put a 2x4 across the end so that the other end of the dolly didn't crush the clutch cable bracket when it was lowered. I then slowly lowered each of the floor jacks on each side little-by-little until the motor mount sat on the 2x4. I had to remove the white cups from the floor jacks so that the raising surface was low enough to get the motor mount arms on to the furniture dolly.

With no more weight on the floor jacks, I moved one around to the rear of the car under the transmission and put slight upward pressure on the tranny. With this slight upward pressure, I could remove the bolts from the rear transmission mounts without much trouble. I then added two more 2x4s to the furniture dolly and lowered the floor jack until the bottom of the transmission just forward of the clutch pulley bracket dropped onto the 2x4s.

Here's the lowered motor/transmission. Note the single 2x4 on the front of the dolly holding up the motor mount. The rear 2x4s are just under the transmission. I still have to get a CV joint bolt remover tomorrow from camp914 to take off the CV joints. Note the ground strap dangling off the top rear of the transmission. After I remove the CV joints, I'll jack up the car higher on the jackstands and roll the motor/transmission out the back.

Note the blue reverse switch wire in the upper left of the picture, the grey speed sensor cable off to the left and the thick black AC controller cable on the left. (Click on the picture for a larger version)

I performed all the above work in about two hours. This is far faster than the 1 week it took me to lower the engine/transmission in the first place. Things go much faster when you do them the second time. I'm a little surprised at how much I've learned in this process.

Next up: finding a CV joint bolt tool, removing the flywheel and balancing the whole thing.

Tuesday, August 7, 2007

Swallowing my Ego

Over the past few days, I've gone through the four stages of acceptance regarding the rather noisy vibration that the car undergoes when I rev the motor above 3500 RPM.

Denial: "My new electric car is wonderful and it doesn't really rattle above 3500 RPM."

Anger: "Why the hell does my car vibrate and rattle about 3500 RPM???!! It's not supposed to do that! Can't the Porsche folks make a good flywheel? Can't ElectroAuto make a good adapter coupling?"

Negotiation: "This would be sooooo cool to show everyone at EV awareness day and to show my friends. Maybe I can just drive around and keep the RPMs under 3000 for a week or two until I figure out the best path forward.

Acceptance: "Okay, I remember when I installed the flywheel that it didn't go on completely straight (only a few hundredths of an inch off...) and this could really screw up the electric motor bearings and the transmission. I better bite the bullet now and deal with this so I can get on the road more quickly."

I'd like to thank Mark the mechanical guy I met through an OEVA member and my co-worker Steve (retired mechanical engineer and AutoCross enthusiast) for beating some sense into me about this. I'm an electrical guy, after all. (Can't we just reboot it?)

Next up: A character building experience involving removal of the transmission and repositioning the flywheel, if it's possible.

Monday, August 6, 2007

Shifting Fixed and Charger Outlet Installed

Things just keep getting better and better with this car. I did some experiments this morning with operating the transmission and found that my clutch wasn't fully disengaging. After adjusting the clutch to fully release this evening, the car shifts MUCH better and I don't grind gears anymore! What a relief!

I also took this evening to finish up the wiring for the external charging outlet on the front fender of the car. This means I don't have to open the hood to charge the car.

The instructions call for cutting a 1 7/8" hole. Sears only had a 1 3/4" hole saw, so I cut the smaller hole and then used a grinding wheel to grind off the plastic threads on the outlet insert down to 1 3/4". I believe the fit to be much more solid.

I also didn't have a 9/16" drill for the power cord hole, so I used a 1/2" drill and greased up the cable with silicone spray so that it would easily slip through the slightly smaller grommet. The silicone spray also helped get the cord through the weather seal on the backside of the outlet insert.

After rummaging around in my leftover parts from ElectroAuto, I also found a temperature sensor for the Zivan NG3 charger, plugged it in and tie wrapped it to the inside of the front battery box.

I just sent an e-mail to a local guy who has helped a few folks out with mechanical issues on their electric vehicles. I'm hoping he will be able to help me with my flywheel balancing issue. Craig from camp914 suggested I take the flywheel/clutch assembly to a shop for balancing too.

In preparation for EV awareness day in Portland this coming weekend, I plan to label all the components so I can better explain things to interested people.

Several folks have asked about 0-60 time, range and top-speed specifications for this car. I haven't really pushed it yet, and probably don't intend to, but I'll capture more data over the next few weeks.

Sunday, August 5, 2007

Test Drive to Work

Since traffic is really light on Sunday morning, I drove to work (15 miles away) and back to see how the 914 would handle.

Shifting was a real pain on city streets since the electric motor changes RPM very quickly during acceleration or regeneration. The RPM stays constant when I "coast" unlike a gasoline engine which tends to slowly drift back to "idle." This makes synchronization of the gears rather difficult.

As expected, power was rather lacking on the freeway. It took me 30 seconds to get up to 60 mph. After analyzing the AC controller logs, it turns out that I only took the motor up to 3000 RPM and it is spec'ed at 12000, so maybe I'll try a lower gear next time to get more torque at the wheels. I just have to get used to the concept of driving at higher RPM, since I got better gas mileage by keeping the RPMs low on my old car (which is 19 years newer than the 914 :).

If my ammeter from West Marine is correct, then just cruising takes about 50 amps and every time I accelerate even slightly, the current goes up over 200 amps. I'm not sure how much current these golf cart batteries are designed for. The battery voltage sunk from 152 volts at rest down to 132 volts after awhile during heavy acceleration. The controller cuts current to the motor if the voltage drops, so I definitely lost power on the uphill exit ramp off the freeway. I think people experience cognitive dissonance when they see a little red Porsche sports car being the slowest car on the road.

I'll be happy to e-mail anyone logs from the AC controller if you'd like to parse the data with Excel to see how the controller behaves. If I can capture a screen, I might even post a picture here.

Friday, August 3, 2007

My First EV Video

My friend Pete dropped by for lunch today with his video camera and we captured a four-minute video of the electric car in action as well as a description of the various parts. This is rather unrefined, so enjoy...


Wednesday, August 1, 2007

Hacking the Regen System and Fixing Acceleration

Well, today was a good day. I got word back from Azure Dynamics after they analyzed my acceleration logs and they pointed me to the EE2NoAccelBat variable. Mine was set to 136 volts. It turns out that if the voltage coming from my batteries fell into the 136 to 149 volt range (EE2NoAccelBat + EE2AccelRamp = 149), then the controller would linearly limit the power going into the motor down to zero when the voltage reached 136.

With a 144V nominal pack, at moderate current draws, I was getting major power loss. Azure recommended setting EE2NoAccelBat to 127, so that this "scale-back" range was a more reasonable 127-140V. This made a HUGE difference in acceleration performance as my battery voltage rarely drops below about 138, even drawing 120 amps.

In other words: READ THE FRIGGIN MANUAL (RTFM) and understand all the graphs

The car was still rather difficult to drive with the regenerative braking happening when I released the accelerator. In a comment to a prior blog entry, Paul J suggested that I try using the brake light signal from the brake pedal to activate the regenerative system. With my driving frustration, this seemed like a good idea.

This change actually reverses the function of the regen relay. In the original kit instructions, the regen relay turns on the brake lights when the AC controller is in regen mode. With the PaulJ brake mod (I might just call it that...), I remove the regen relay and use a similar (normally closed) relay to activate the regen function when the brake lights come on.

Again, this involves purchasing a normally closed relay so that when the brake lights come on, it opens up and lets the regen circuit activate on the AC controller.

This worked like a charm! The car is soooo much easier to drive because it behaves much more like a gas powered vehicle. That is, you push the acclerator to go and press the brake to stop. The relatively weak brakes are augmented strongly by the regen function of the AC controller, so the brakes are actually much more responsive, despite the heavier car. I recommend this alteration to anyone using this kit!

A final task was to mount the ammeter that I had been using for experimental purposes. Until I get the cool looking dials from ElectroAuto, I just shoved the ammeter in the place where the engine temperature gauge was. I had to break off two of the long mounting pins so it would fit in the hole, but this will work until I get something better.

On another really bright note, I contacted the facilities manager at HP about having a plug-in at work for charging my vehicle. I had contacted him last October and he said he would get around to it. Much to my delight, the work actually happened without him telling me and there's a plug-in (with an Electric Vehicles Only sign!!!!) just outside my building!!! Yay!!! I'm sure this is all part of HP's plan to look more "green," but I'm still ecstatic that I don't have to walk halfway around the site just to plug in my vehicle. Here's a picture:

I've pushed really hard these past few days to get the car going. The Oregon Electric Vehicle meeting is tomorrow and EV awareness day is August 11th. Also, the NEDRA races are on August 17-18, so if I'm really lucky, I'll try to get the car to these events. The test drive with the improved acceleration and "PaulJ Brake Mod" makes the car feel much more road worthy.

Off for a quick test drive and then to collapse. Cheers!