09/03/07 Quite a few people have been inquiring as to the status but I had to attend to the work at hand. The last couple weeks have been very educational. It really made me appreciate the wisdom of my decision to keep the powertrain in the dp1 prototype totally stock. This allowed me to just focus on the chassis knowing that the motor and the electricals are a given (well, kind of - the plumbing gave me enough grief). Bikini is the exact opposite because the chassis is a given (well, kind of ;) and I have to 'just' plug the motor in. It has been quite a revelation as to what that really entails and as a result the dp1 production design will be that much better.
John had kindly offered to fly out to assist with the engine startup and I gladly accepted. In preparation for his arrival I was hoping to get all the systems ready and make some good progress on the electricals so we can just fire it up, fine-tune it and hopefully take it for a few laps around the track Friday. In retrospect that was quite naiive. Just finishing up the plumbing took untold hours. The scavenge tubes from the sump to the pump inlet are a real pain so John is now designing a setup that will eliminate those in the dp1 and in future installs in Atoms, Elises and the like. Yes I'm planning to do conversions, e-mail me if you're seriously interested - the cost will be about $50K complete with engine, gearbox, exhaust, intake, ECU, wiring, plumbing, brackets and everything else needed for the installation (and no it won't be street legal). The picture below is of the starter and illustrates the typical clearances involved. This is one of the things I will improve for any 'production' versions but it will always be a tight fit. The full list of planned improvements is long since this install carries a steep learning curve, although I must say all in all it has gone pretty well so far and not all problems encountered were of my making.
When John arrived Wednesday we surveyed the state of things and it still seemed doable to get it running the following day. But the more you get done, the clearer it becomes just how much more there is to do. And that's when things go smoothly...
It so happened that for various reasons quite a few things did not go smoothly. Filling the oil system was easy enough. This being a dry sump we turned the pump by hand to get some oil in the engine. The oil level in the tank dropped indicating the oil was going through the pump and soon there was resistance at the pump indicating back pressure. Cool... Well, not so fast. After the starter was wired up we decided to crank the engine for a while without plugs to circulate the oil through the scavenge system. Quite a bit of cranking but the level in the tank is unchanged and gauge shows no pressure. Faulty gauge? Afterall, about a quart was drawn out of the tank cranking by hand. A bit of brainstorming and then John asks if I plumbed the filter backwards - it has a one-way valve in it. Crawl under to check and sure enough. I built the hoses properly but when I installed them I accidentally switched the input and output ports on the filter. Easy enough to fix and we now have oil pressure.
Crank the engine some more and all seems well. Then M, who is acting as assistant/observer/camera crew, notices a puddle of oil forming upfront. Stop everything, see what's going on. Turns out the brand-new, never-used oil cooler has a leak in it. There is no pressure (the other end goes to the tank which is vented to the atmosphere), it just leaks. I guess I'll have to talk to Brammo and see what they can do about that. In the meantime, lots of disassembly to get the thing out and then I devise a bypass. Finally, oil is done. For the moment.
Then trying to remove a hose from the plastic fuel pump outlet so I can get a clamp on it, I snap off the plastic piece instead! Damn. Getting the fuel pump out is nearly impossible without pulling the engine and I don't have a replacement anyway. Time for plan B. A trip to the hardware store results in some brass tubing and couple kinds of adhesive to try and fix the fitting. Of course it had to be on the pressure side and not the return. That means the joint will be under stress. I glue it the best I can and figure out a way to secure it with safety wire as a backup. That'll be tested once I figure out why fuel pump relay won't turn on.
Taking a break to do a trackday in the MINI on Friday, we get back to the task on Saturday. I try to figure out the car electrical harness (there is no wiring diagram and the connector report e-mailed to me by Brammo only refers to circuit numbers which are not at all descriptive so it proves of little use). Giving up on that we decide to just hotwire the few necessary connections for now and deal with harness integration later. I bypass the immobilizer which gives me a working fuel pump. Pressurizing the fuel system results in a puddle of fuel next to the outlet - the adhesive is not holding. Time for plan C. John thinks there's another type of glue that might do the trick and after a run to the hardware store it's rebonded and re-safetywired. A new pressure test shows a tiny leak that soon stops. It'll do for now.
Clutch is filled and bled. I give up waiting for the preferred gear oil to be delivered (it's been two weeks) and run out to buy some Valvoline synthetic to fill the gearbox. By the end of the day it looks almost ready to start....
Sunday is Johns last day here and we're determined to get it running. The ECU is downloaded with an initial program, a final check of the wiring and fluids, then a test of injectors and coils. One bank of injectors is not firing. All four of them. A couple hours chasing wiring and fixing a couple minor issues. Another test. Still nothing. Apparently all four injectors are dead. Plan D - get the dp1 prototype off the lift and borrow the injectors (helps to be using Hayabusa parts in both cars). This takes some doing but when done the injectors are now ready to go.
By late afternoon we're ready for the big moment. Fuel pump on, ECU power on, camera rolling. Crank it and YES, it fires! Here's the video (11M) of the first start.
I make a couple of phone calls to people who've been waiting to hear the engine run and they're on the way. In the meantime we attempt some initial tuning. All is going well until M notices another oil puddle under the car. Stop everything again, trace it down. The apparent culprit is an oil fitting deep in the engine's V and impossible to get to without disassembly. The paint marks are still aligned meaning it hasn't loosened so it might be that the crush washers are bad or something more serious is wrong. Getting to it will require some work but we don't have a choice. Drain the water, remove one bank of throttle bodies and the coolant pipes. In the meantime people start showing up to see it run. No pressure :)
We get the fitting out, replace the copper washers and retighten. Now all we have to do is reassemble the whole mess. John is talking to one of my friends, others are milling about. So M and I dig into it. She's great to work with and goes between assisting me and doing some of the tasks herself. She later said it felt like being a nurse in an operating room. Well, she was in fact assisting with a birth... :) Before anyone knows it we've got the plumbing hooked up, coolant filled and bled, throttle bodies installed and everything is ready. The engine starts and runs.
From here there is a lot of tuning and detail work. A lot needs to be done before the car is drivable. Integrating engine and car wiring harnesses (which will require figuring out what's what on the car side). Fabricating a lot of exhaust shielding. Fine-tuning the ECU map. Building brackets for some additional gauges. Installing the airbox and finishing the air filter and ducting installation. Fabricating mounts for the coolant overflow tank. Getting the oil cooler fixed and reinstalled. And probably a few things I don't know about yet. But it will get done. My fourth automotive child has been born (counting the Westie, dp1, my client's project and now Bikini).
09/16/07 It drives! There is definitely a steep learning curve here. The parallels with the Westie are quite interesting - that car also took 13 days from first engine start to first drive. This drive, like the Westie, was done minus nose bodywork and minus a working speedo. Much work remains but despite the onset of a typical Northwest drizzling rain the first drive took place as planned. With the disclaimer that the camera sound is all messed up (seems it's picking up interference from the ignition), here's the first drive video (22M file). The engine sounds awesome in real life. And yes, it's the same route as the Westie :) The steam you see is just previously-spilled coolant evaporating. Hopefully.
Initial impressions are very good. The engine is very tractable although the mapping is only ballpark-close, at best. Twin-plate clutch is very usable and much better than the stock Hayabusa one in the Westie. Chassis is a pleasure to drive. The steering is light and direct. The overall Atom experience is as intense as I remember it even just putting around. This is THE perfect motor for the car. I can't wait until I get it a bit more buttoned-down and can open it up a little. Should be an absolute blast! (pun intended ;).
Getting to this point took quite a bit of doing. Once again M was kind enough to assist with the final preparations and it certainly made things go quicker (and more fun too). John provided excellent support over e-mail and phone on the intricacies of ECU programming. Brammo supplied a wiring connection list that I was able to use to perform extended surgery on the harness. All the efforts came together and Bikini is now a real car.
The harness is actually a major challenge but I now have it more or less figured out. Still quite a bit is left to do before I'd call it fully functional. To get everything to fit I had to route the wiring and the return coolant pipe through the upper bulkhead opening. Looks like Lain's house now (a very obscure anime reference, I don't expect many to get it :). It is really quite fortunate that the 'bump' between the seats is as big as it is. Almost looks like it was designed for the purpose. The ECU seems to fit better mounted lower in the tub so that is its new home. Whew...
The engine bay is now crowded but relatively tidy. M spent a lot of time putting shielding over the wire bundles and did a great job of it too. And I FINALLY (knock on wood) have fixed the fuel line fitting. I had tried literally a dozen different adhesives and sealants but nothing seemed to stick to the plastic and under 50psi pressure the fuel would invariably find its way out. The current solution consists of a brass tube pressed into both halves of the broken fitting. Both ends of the tube and the insides of the fittings were coated with a thin layer of aviation-grade gasket maker. Special epoxy for sealing plastic gas tanks was then used in the joint between the fitting halves when they were slid together with the brass tube. Then a fiberglass cloth permeated with JB-Weld was used to wrap them. And finally the whole mess was safety-wired. Now if THIS doesn't hold, I guess I'll just have to bite the bullet and get a new fuel pump. Which is far easier said than done (an immense amount of disassembly is necessary to get it out because the fuel tank sits lower in my car than standard Atom) so I really hope it doesn't come to that. We'll see.
Another major challenge was getting the ECU programmed with a usable map. At first the engine refused to idle. Then, after a lot of fiddling, I was able to get it to hold a steady idle at around 1K rpm. But any application of throttle, especially a slow one, would kill it instantly. Sometimes blipping the gas got past the rough spot (with large clouds of smoke from the exhaust) but usually not. We were stumped. John looked at the logs I downloaded and suggested a couple of things. I tried them with varied but ultimately unsatisfying results. I pulled one of the plugs and it looked pretty much as expected - a bit dark due to over-rich conditions but nothing out of the ordinary.
Some further examination of the logs showed that the ECU was thinking the throttle was staying at 0 when I was opening it just a little. John had reminded me to recalibrate the throttle stops which I thought I had done, but not recently. A quick re-cal and BINGO, engine now revs when throttle is pressed. Imagine that. That's the trouble with adjustable things - there are only a few 'right' settings but an almost infinite array of wrong ones. Like I said, the learning curve is steep.
So with that and countless other things out of the way, the first drive finally took place. A huge amount of work remains still. This really gives me a new appreciation for all the subsystems of a vehicle and how difficult it is to get it right (and conversely how easy it is to mess up). So I'm going to take my time with it. Also, the client project is going to require my full time effort for the next couple of months. It's very much worth it though - we are the beginning of a much-needed change in transportation as we know it. Details shall be revealed at CES in January, until then just take my word for it. It's a pragmatic way to address real-life problems with real-life solutions. And most importantly, it's FUN. The couple slogans I came up with is that 'being green doesn't mean you have to be a pea in a pod' and that 'a good cause should be wrapped in a good package' :). And yes, in some very important ways the client project serves to advance the dp1 effort so it's by no means on hold. Quite the contrary. One thing I can say for certain - life is not boring!
10/08/07 I know, updates have been pretty scarce lately. In part this is due to the fact that I'm working full time on the client project. The other aspect is that when I do get to work on my own stuff, progress is slow because this is mostly detail work and testing/tuning. I have now installed a wideband A/F sensor with a gauge and wired it to the ECU.
The switch right next to the gauge turns ECU closed-loop control of A/F ratio on and off. The ECU is very cool and has a bunch of useful features. One of the most important at this stage is the ability to monitor the A/F ratio and then come up with recommended corrections to the map. This is nice because one can just drive around in closed-loop mode, then tell the ECU to apply and reset the corrections. So it's basically self-learning and can be used to fine-tune the mapping without a dyno, by just using the car. Or that's the theory. So far I haven't done much driving because it's been raining, and then there's the minor issue of registration and insurance... Which I'll address eventually. In the meantime there are a few more tweaks to be made like getting the ECU to turn the radiator fan on and off based on water temp (got 90% of that done, just need to solder the wire now). Then there's the issue of the oil cooler, figuring out the lights, etc. No shortage of things to keep me busy on rainy evenings. Oh, and the dp1 design too (yes, it's still moving forward and there will be a fun dp1 announcement shortly as well).