Monday, December 29, 2008

Add the Splash Guard

I'd like to get the car off the jack-stands and back into a position where I can drive it while still adding things like the guages. After installing the speed sensor, the last thing needed under the car is the splash guard to protect water and grit from getting into the Warp9.


After playing around a bit, I found that a piece of ABS plastic 15" x 18" and 3/16" thick will probably work. I've used a 9/32" drill bit to drill four holes in the plastic so I can tie-wrap it to the factory splash guard and the transmission.


Here's the plastic tie-wrapped to the original splash guard. For the lower-left tie-wrap in the picture, I had to drill through both the 3/16" plastic and the factory splash guard to get the tie-wrap through. The remaining hole in the upper right will be to attach that corner to the bottom of the transmission. The tranny mount hole is 1.5" in from the right edge.


Here's the splash guard in place. There was a spare bolt hole under the adapter plate with no bolt in it, so I found a bolt that fit and used that to attach the tie-wrap for that corner. Although the bolt sticks out, it's actually rather tight because it's bottomed out in the transmission case, so I don't expect any problems. I could have probably used a beefier tie-wrap, but it works for now.


Here's the underside of the installed splash guard. It completely covers the bottom of the Warp9 but gives it about 1" of clearance to breathe for cooling. This picture was taken from under the car looking towards the drivers side wheel hub. The front of the car is to the right.

The next task will involve hooking up the tachometer and the Link-10, which will be tricky since it involves a lot of under-dash work.

Good Night.

Mounting the Zolox Sensor

After futzing around awhile, I was able to get the spacer ring for the Zolox tachometer sensor to line up.


Here's the bottom of the mounting ring. The screws that came with the mounting kit stripped out the plastic, so I sunk some depressions in the backside of the ring where the bolt holes were and put in 8-32 x 1" bolts from the backside.

Notice that the large holes in the upper left and lower right are somewhat warped. I had to use a drill to hone out the holes in the proper direction so that the ring was actually centered on the motor tailshaft.


Here's the topside of the mounting ring with the Zolox sensor bolted to it with nylock nuts. There's actually a space between the black flange on the sensor and the mounting ring, so don't tighten these bolts too tight. The grey mount on the lower right holes the wiring so it doesn't flop around.


I drilled a hole in the center of the tailshaft and then used a 1/4" bolt to align the magnetic disk with the hole. I didn't have a tap to put threads in the center of the shaft, so I just epoxied the whole thing on, using the bolt to keep the magnetic disk centered.


Here's the sensor assembly bolted onto the motor over the tailshaft. The bolt holes had a little play to them, so I had to make sure thing were lined up. I put the Civic in third gear and blocked one wheel so that I could rotate the hub to spin the motor shaft. After spinning the motor shaft and verifying that the magnetic disk didn't rub up against the inside of the speed sensor, I tightened down the bolts.


The sensor comes with a shielded cable with three wires. Red is +12V, black is ground and white is an active pulldown. My test setup shown here includes a small 12V battery on the right (from an old UPS) and my multimeter. I added a 2.5K resistor to pull the white wire up to 12V when the sensor wasn't actively pulling down. Surprise! It actually works and generates high and low voltages as the motor shaft spins. The real acid test will be to see if it triggers the SpeedHut tachometer properly.

Frustrations with Zolox sensor mount

I tried installing the Zolox sensor mounting ring for the Warp9 motor from EVsource today. It looks like the holes on the mounting ring don't line up properly. The holes for the Warp9 casing don't match, so I had to drill them a bit larger so the bolts would go in.


Here I've placed the spinning magnet onto the end of the tailshaft with a short 1/4 bolt to properly center it. If you look closely, the spacing is off so that the shaft isn't properly centered in the mounting ring. It looks like the fabricator got the holes offset. This position is critical because the spinning magnet only clears the inside of the sensor by about 2 mm on all sides which is really close.

Furthermore, I tried using the supplied screws to attach the Zolox sensor and they immediately stripped out the plastic. I'm frustrated that a mounting ring that should just work wasn't really thought out that well.

With stripped out screw holes and a non-centered shaft, this spacer ring really needs some work. I intend to countersink some depressions on the bottom of the screw holes to allow for a true bolt to go through the ring with a nut. This shouldn't strip out and allow for better mounting. As far as getting centered, I'm not too sure the best way to do this. I'll either have to re-drill the holes in a different place or drill the existing ones larger and clamp the ring down hard in the proper position.

Ugh. Happy Monday.

Christmas Presents


Santa was good to me this year. I received a Link-10 battery monitor from my parents. I don't think they understand what it does and why it costs so much, but okay, they got it for me anyway. The items on the right came from Belktronix. The wires on the upper left are the Link-10 voltage prescaler. The module on the upper right is the DC-DC to supply an isolated 12V to the link-10. The grey cable on the lower left is the tachometer interface which activates the LVP motor shutdown when the tach shift light turns on. The pink bag on the lower right is the fixed BatMon board with a zener diode.

I have a week left of vacation to try and get this stuff working. We'll see how it goes.

Happy New Year to all.

Friday, December 19, 2008

Tach Mount, Splash Guard and POR-15

It's still pretty cold in the garage, but I motivated myself to get out to TAP plastics (10 blocks away) and pick up some ABS plastic for the improved splash guard.


Here's the 3/16" thick piece of ABS lying on top of the existing splash guard. I got a piece 36" x 18" which is way too big, but I didn't know what would be needed. In the final directions, I'll probably list a smaller piece to acquire. This one cost me $22.25. I'm going to figure out how to cover under the motor vents to prevent grit from getting in there while still allowing adequate airflow for cooling.


Since I cut off the tailshaft on the Warp9, the Warp9 mounting ring for the Zolox tach sensor was too thick. I took that to TAP plastics as well and had them cut it to the proper thickness while keeping all the proper mounting holes for the sensor. The original donut was 1 1/2" thick. After cutting the tailshaft, I had 3/4" left over. If I cut the mounting donut down to 1/2" thick, it would leave 1/4" plus the spinning magnet (another 1/4") to protrude inside the tach sensor sleeve.


I wanted to make sure the spinning magnet sat in the center of the tailshaft. With the full tail shaft, there would be a 1/4" threaded hole to mount it into. Without the hole, I chose to drill a 1/4" hole in the center of the shaft about 3/8" deep to match the 1/4" hole on the spinning magnet. This will allow a short 1/4" bolt to keep the spinning magnet centered while I epoxy it on. Ideally I would have used a tap tap to make a threaded hole, but I didn't have one and this was quick and easy. Note that I used masking tape to cover up the tail bearing to prevent metal shavings from getting into the motor.

Since the mounting hardware that came with the donut above was meant for a 1 1/2" thick donut, I'll get the proper hardware tomorrow to install the donut and Zolox sensor.



Another nice surprise! The POR-15 starter kit arrived today as well. There were three battery hold-down pieces I didn't finish before sending all the metal pieces off for powder-coating, so I'll have to coat them myself. POR-15 is really tough stuff and this $20 starter kit will have everything I need. (no, this isn't an ad for POR-15)


For those who are curious, the starter kit contains metal cleaner, surface preparer, the POR-15 paint itself, rubber gloves and application brushes. I'll have to find a slightly warmer day to apply this stuff since I don't want to inhale fumes in my closed garage.

I'll be going snow-shoeing tomorrow, so maybe I'll get to installing the sensor on Sunday.

Thursday, December 18, 2008

Tachometer is Here!

This is a cool day(And not just because it's freezing outside). The package with the custom tachometer from SpeedHut arrived.


Although I received the 2-hole pillar pod last Friday, I wanted to show it here since I ordered both items from Speedhut. The wiring harness is nice and long. The small module in the lower left of the photo is the dash light interface which allows a variable dash light brightness to control the tach brightness as well.


This is just too awesome for words. I hooked the tach up to a small 12V battery to light up the dial. I'm psyched that I got the "Civic EV" lettering for free.


On the backside of the tach is what looks like a standard 1/8" stereo audio jack to drive an optional shift light LED. This is what I'm going to drive the LVP signals with to disable the motor controller PWM if the RPMs get too high. Bryan from Belktronix is working on an interface now (tentatively for $10-15) that will tie this output into his controller system as a rev-limiter.

Progress is slow these days due to the cold weather. The garage is freezing and I'm not motivated to work too much with frozen fingers. My employer is forcing everyone to take off the next two weeks to make the financials look better, so I'll have time to wire up the tach and possibly finish the splash guard under the car. Since the roads have been covered with ice, I'm leery about taking the car out and sliding it into a ditch.

A small can of POR-15 should be arriving tomorrow that I can paint the remaining battery hold-downs with and there's one Batmon board left with a bad zener diode on it that's back at Belktronix for repairs.

Sunday, December 14, 2008

Charging parts back and more miles

I received the fixed charger parts back from Belktronix on Friday and installed them with no problems. After going through a full charge cycle, I did find another Batmon board that was activating the shunt resistor at 13.5 volts instead of the expected 14.6 volts. I'll look into that early this week or send it back for repairs.

After fixing up the charging system, I took a quick test drive with my friend Jonathan for about 7 miles and the car behaved well. The suspension is still a bit off (rear is 2" high while the front is 1" low), so I'll look into fixing that.

Stuff still to do:
  • Fix the zener diode on the failing Batmon board
  • Fix the suspension ride heights
  • Figure out how to encase the BatMon boards while allowing FET cooling
  • Paint remaining hold-downs (not powdercoated) with POR-15
  • Install Link-10 system after Christmas
  • Install tachometer with shift light going back to controller for over-RPM protection
Last evening, I drove the Civic to a party at FreeGeek, a local computer recycler and showed off the car. There's a significant hill between FreeGeek and my house; the Civic did quite well starting on the hill. This is a far cry from the white-knuckle hill-starts I've been living with in the 914. What a relief.

We just got about three inches of snow on the ground today, so I don't think I'll be joyriding or testing the full acceleration of the Civic for fear of sliding off the road. I just updated the EVAlbum entry with some mileage so I'm officially out of the "under construction" stage.

Thursday, December 11, 2008

Plastic Boxes and Speed Sensor

I got some fun stuff today. The plastic boxes for the BatMons and the Zolox speed sensor with Warp9 mounting hardware came in.


Here's a pile of 2x2" plastic boxes 1/2" high that fit the BatMon boards quite well. If I cut off one side, I should have at least some protection from splash and other grit that enters the engine compartment.


Here's the Zolox speed sensor from EVSource. I don't know if it will work correctly with the SpeedHut tachometer that's coming, but I hope so. It took me awhile to figure out what the wires were. Red is +12V, Black is GND and Green is an active pulldown (Otmar uses a 2.6K pullup) every time a magnet (4 per revolution) in the small wheel passes the sensor.


Here's the spacer ring that allows easy connection of the Zolox sensor to the tailshaft of the Warp9 motor, also from EVSource. Since I cut my tailshaft off and there's very little clearance, I might not use this, or slice it thinner. With the cut tailshaft, I also lost the threaded hole in the middle of the shaft. I'm thinking about epoxying on the magnetic wheel instead.

Monday, December 8, 2008

Assembling a Price List

I always wondered how much this kit really cost me, aside from tools, so I assembled a total parts price list and uploaded it to the Civic-EV Google Group here:

http://groups.google.com/group/civic-ev-kit/web/SystemPricelist.xls


I'm a bit surprised I came out below $10K (barely).
For those of you who have done conversions, please check my spreadsheet and let me know what I missed. One of the high-wants for the kit was that it would cost < $10,000 and I might have actually done it.

Cheers,
Tim

A Different Tachometer

After a suggestion I received from Bryan Belk (of Belktronix), I decided to go with a different tachometer than the cheap Shucks one:

http://www.speedhut.com/custom_gauge_description-gauge_type-Tachometer-auto_number-700.htm

This tach is only $96 and comes with a customizable face. I was even able to add the text "Civic EV" for free! The other cool thing is that it has an LED driver on the back for a particular shift point. This could easily be set and tied into the LVP (low-voltage-protection) circuit on the controller to prevent over-revving the motor. Initially, I was going to create a someone complex circuit that measured the spaces between tach sensor pulses and triggered the LVP circuit when they got too quick. This gauge takes care of that function and allows non-circuit designers with the open-source kit to have over-rev protection as well as a really cool tachometer that fits easily into the 2" pillar pod.

With all this stuff ordered, I'm just enjoying a little down time until things come it. Cheers.

Ordering Gauges

While I wait for the charging components to get fixed, I did some research on gauges to monitor the state of the system. My goal is to use as much of the existing dashboard as possible and then add low-cost gauges in order to prevent any motor problems and detect faults.

The list I came up with:
  • Mount a Link-10 E-meter in a pillar pod to monitor current and state-of-charge
  • Mount a tachometer in a pillar pod to monitor RPM and use the sensor as an over-rev protection device
  • Make a circuit to drive the "low oil" light when the 12V car battery is less than 13V to detect DC-DC failures
  • Tie the "battery" light to the low-voltage signals on the Belktronix system to point out undervoltage on the main pack
  • Make a circuit to drive the "check engine" light when the motor overheat contacts open on the Warp9
  • Future Option: find a temperature sensor that attaches to the Warp9 that will drive the temperature gauge on the existing dashboard
To that end, I went ahead and ordered the following
I still have to find out the best pillar pod for these gauges since the Link-10 monitor is physically deep and will probably stick out quite a bit. I also have to design the schematics for the oil light and overtemp light.

In order to protect the BatMon boards from water splash (at least in the engine compartment), I ordered a bunch of 2x2" plastic boxes. I don't know if these will work since I don't know the heat dissipation requirements, but I'll give it a shot.

Have a great week everyone,
Tim

Saturday, December 6, 2008

Cleaning up Loose Ends

After working with Bryan from Belktronix yesterday, I found out that I had blown some components in the IsoBatMon, the Charge Detector and one of the BatMon boards. Those went off in the mail this morning back for repairs. I'll try the system again when the parts come back.

In the meantime, it looks like the speedomter is working again. I also purchased a new headlamp to fix a burned out right hi-beam. It turns out the problem wasn't the bulb but a *missing* fuse in the fusebox. I suspect the prior owner sacrificed the fuse to fix another circuit but didn't replace it. Oh well, I now have a spare bulb...

The Belktronix system calls for a 60A fuse between the DC-DC and the main battery. I utilized the existing alternator input for the DC-DC output and replaced the factory 80A fuse with 60A equivalent.

Theoretically, I could drive the vehicle now, but charging would get very tricky as I don't have a charge monitoring system. I'd have to watch the battery voltage like a hawk or risk hurting the batteries and shunt resistors.

There's still much work to do in the way of deciding which gauges I should use and writing open-source instructions. At this point, I'm a little burned out, so I'm going to take this weekend and just clean up the house (there are EV parts *everywhere*) and get my dining room table and kitchen back.

Many thanks to all the congrats and well-wishers out there who have supported me in this effort. I hope this open-source project proves fruitful, especially if gas prices go back up next year.

Cheers,
Tim

Friday, December 5, 2008

Wiring up and Testing Batmon Boards

Okay, it's time to get your wiring spaghetti on. I spent part of last evening (after fixing the speedo) and all of this morning wiring up the Batmon boards. In the process I blew the OVP channel on the IsoBatMon unit.


Here I'm crimping yellow 3/8" ring terminals on all the blade fuse holders that will hold the 7.5A fuses. I tried to do as much as I could on a workbench instead of in the cramped quarters of the car.


Okay, I've finished wiring up the front set of batteries. If you look closely (click picture to zoom in), there are the two grey LVP/OVP cables on the right side of the picture. The two thinner red/black paired wires in the upper left of the photo are the LVP/OVP signals going to the two front batteries.


Here is the finished Batmon wiring in the rear trunk, showing the speaker wire going to the shunt resistors and the twisted, colored computer wires connecting the LVP/OVP signals between the boards.


I came up with this little device to test all the Batmon boards. It's basically a 3V supply with 2 AA batteries and spade terminals on the wires. The spade terminals are smaller and much thinner than your standard 1/4" fast-on terminals. This enables them to slide easily into the fuse socket without much entry or removal force.

How does it work? Instead of plugging in the fuse, you plug this 3V supply in to the fuse sockets in both directions. In one direction, it adds 3V to the 12V battery supply, making 15V which triggers the OVP circuit. In the other direction, it subtracts 3V from the 12V supply, making 9V which triggers the LVP circuit. With this, you can verify that both Batmon circuits are working AND you can check the red and green LEDs on the IsoBatMon board to make sure this board is properly talking to it.


Here's the 3V tester in action. I have the red(+) side plugged into the left side of the fuse holder which is tied to the +12V of the battery. The black(-) side is plugged into the right side of the fuse holder, giving the BatMon board 9V. It's hard to see due the camera flash, but the red LED on the BatMon board is lit up, showing that the circuit works. The red LED on the IsoBatMon board has also lit up, indicating a proper connection back to the main system.

While working in the trunk, I briefly touched one of the OVP signals to a battery terminal and fried the OVP side of the IsoBatMon. The BatMon board itself works, but when it's opto-isolator shorts out the master OVP pair running through the car, the green LED on the IsoBatMon fails to light.

Since the IsoBatMon is powered from a battery "low" in the pack (around 12V), touching the OVP signals (or LVP signals for that matter) to any battery terminal "high" in the pack (say, 96V) will send a high-voltage, high-current spike through the opto-isolator (and LED) on the IsoBatMon. Kerplooey!

I'm going to see if I can work with Bryan from Belktronix to fix this. In the meantime, I'm going to short the OVP signals coming OUT of the IsoBatMon together to force the charging system into low-current mode to prevent overheating any of the shunt resistors.

Onward!

Thursday, December 4, 2008

Fixing the Speedometer, Removing ECU

During the test drive, I noticed that the speedometer wasn't working. It had worked fine before I removed the engine. After researching the Helms manual, I needed to get at the speed-sensor on the transmission to figure out what was going wrong. Fortunately I hadn't yet wired up all the BatMon boards, so removing the firewall batteries to get at the sensor only took a few minutes.


Here's the firewall battery rack with two batteries removed to get at the speed sensor on the transmission. I'm pretty happy with how quickly I could remove the powder-coated hold-downs and remove the batteries.


Here's the non-functional speed-sensor. The debug procedure is to remove this and test the voltage at each of the three wires. It turns out I had a bad ground connection (black wire). The original ground came from a bolt on the engine which no longer exists. The black ground wire actually went up the wiring loom all the way to the engine harness connector, so I could fix it there.


Here's the fix for the problem: Find the all-black(no stripe) wire on the engine harness loom just in front of the driver and ground it to the chassis. This ground wire actually supplies the ground not only for the speed sensor but for the ECU. I used an automotive wire splice (pinkish connector) to splice in a piece of 16 gauge wire to a blue ring terminal on the chassis.


Speaking of the ECU. I guess it's pretty useless now, and I can tap into some of its wires to drive lights on the dash or sense other parts of the car. Since I don't want it drawing extra current (probably not much anyway), I decided to remove it. Here is the passenger side carpet pulled away so I could get at the ECU bolts.


Here's the ECU unbolted on the passenger side floor.


Just for yucks, I opened it up to see the inside. I'll probably sell this if there is a local buyer.


After checking a few other electrical items, I noticed that my 12V car battery was actually quite low (11.8 volts). I'm pretty sure it came this way, since the DC-DC converter appears to be working. I'm leaving it on a 13.6 volt float charge overnight (perhaps the weekend) to see if I can get it back to full.

Next up: wiring up all the BatMon boards (spaghetti, here we come!)

Wednesday, December 3, 2008

Buttoning up the Exterior

Even though all the BatMon boards are not wired up, I was itching for a quick test drive. It was time for me to get the wheels on the car and take it for a short spin.


This is the picture of my front hallway closet. I had originally stored the tires from the Civic in the front hall, but my cats licked the brake dust off the inside, so I had to put them in the closet. They didn't fit that well and kept tumbling out every time I opened the door.


As you can see, they just don't fit that well...
The tire on the lower left is the spare tire for the 914. The batteries prevent me from carrying in the car.


Okay, here we are installing the tires back on the car.



I re-installed the splash guard under the front of the car, but found that the new front battery rack angle-iron interfered with it. Here are pictures of the chunks I cut out of the splash guard to get it to fit around the front rack.

With the wheels on, I used the floor jacks to lower the front end of the car onto the ground. What!?? There's fluid leaking out onto the floor! I realized that there was still coolant in the heater core inside the car that drained out when I lowered the front end. With the front lowered, I used the floor jacks to lower the rear end as well.


Here's the Civic with it's freshly applied "ELECTRIC" emblem on the back! It drove out of the garage quite well.


And here we are, ready for a short test drive. The front suspension is a bit lower than I'd like. There's only about four inches of clearance, but I'm sure it'll work.

At this point I took it for an EV-grin spin around the block. Videos are here.

Many thanks for all the people who have answered my questions and offered support during this project. There still a ways to go, but things are looking good.

Best wishes,
Tim

Routing the LVP/OVP Cables

In addition to the thick, high-current battery cables, we need to route the OVP (over-voltage protection) and LVP (low-voltage protection) wires from the rear BatMon boards to the front BatMon boards. The instructions say to route these away from the high-current cables and away from each other to prevent noise.


I decided to route these through the rubber plug that the fuel tank wires go through. Here is the rear seat lifted up and the fuel tank cover removed. I've cut the fuel gauge wire and fuel pump power cable.


Here are the two shielded OVP/LVP cables routed through the rubber plug with the cover screwed back down. These cables run rearward under the seat backrest into the rear trunk where the batteries are.


Here's the same fuel tank lid, but under the car. This also shows a little of how I routed the high-current cable in the radiator hose along the brake lines to the holes in the sides of the rear trunk.


This shows how I routed the OVP/LVP cables along the underside of the car towards the fuel line brackets. I tucked these cables behind the brake lines and tie-wrapped them in to protect them from road grit.


This shows the OVP/LVP cables coming out of the brake line run at the front of the car and going up into the engine compartment.


I wanted to make sure these lines didn't interfere with any moving parts, like the steering column, so I tie-wrapped them to other non-moving cables to keep them out of the way. Later, I'll connect these to the BatMon boards in the front and rear to enable communication.

Wiring Things Up and Testing

Much of my time was spent wiring up the controller, charger and integrator box from the Belktronix kit.


I had to unbolt the charger in order to get access to the fast-on connectors on the side of the main motor controller.


The Belktronix instructions come with a checklist of things to do before powering up the system. I had my handy Fluke meter out to find any problems.


Here's a top-down look at the vehicle integrator module all wired up. I've just turned on the ignition key and all the proper LEDs have lit in the integrator module.


Another sign of success is that the fan on the controller (near top of picture) is now spinning.


I wired the DC-DC converter right into the alternator input connection, again leveraging the existing main fuse in the under-hood fuse box. The voltage is about 13.8, which shows that the DC-DC working. Note that the potbox sitting on the headlight, still not routed into the passenger compartment.

At this point I felt that EV urge to try spinning the motor.

video
Yay! The hub spins and the controller didn't blow up. This means I actually got several things right. Okay, time to clean up this spaghetti wiring.


Here is the cable for the potbox routed through one of the rubber plugs in the firewall on the passenger side.


The cable from the potbox wasn't long enough to extend all the way into the engine compartment, so I found some shielded 4-wire cable and made an extension cord by using red male and female fast-on connectors. This is a picture of the passenger footwell. The potbox is to the left in the driver footwell and the cable goes forward, through the rubber plug in the prior picture to the driver components.


Her is some of the wiring after a bit of cleanup. I drilled several 1/8" holes in the front edge of the plastic mounting board and used these for tie-wrap holes to bundle the cables together.


Here is a closeup of the bundled cables nearer to the integrator module. It's not pretty, but I think it'll work. Note the BatMon board below, connected to the IsoBatMon circuit with its 7.5A blade fuse. In the lower-left of the photo, you can see the blue male/female fast-on connectors I used to connect the charger to the orange cable going to the charging socket on the outside of the car.