So, now that the motor is correctly hooked up, I took it out for a test drive. Everything went well enough, but my accelerator is only giving me about half throttle or so. Nathan came over and discovered that I'd attached the potbox in the wrong orientation; that and a little tweaking fixed that problem.
So, time for another test drive. That's much better! Wait, what's that bucking? And the crackling noise? And the motor stink is back, too.
We pulled it into the garage, and jacked it onto the wheelstands. I had asked the EVDL about the motor stink after my last test drive; they said it wasn't typical, and I should check for heat and inspect the brushes and commutator visually.
Fat chance. I think the only way to pull the endcap off the motor is to remove it. Otherwise it's just too close to the side of the engine compartment. A look through the screening didn't show me anything interesting. As for heat, the top brushes did seem to be hot, although the case was not. Of course, heat does rise, so that wasn't convincing.
We decided to watch it in action. This is a touchy point for us, since we were planning to do that when I blew up the motor. If one of us had been under there, he would likely have died from motor parts embedded in the brain. We decided to keep it to a lower speed -- 20mph in third, probably around 2K or so -- and gave it a try. No bucking, no crackling noise, no stink.
Of course, this is a very low-amperage test, since all that's spinning is the wheels. No great weight to be pulled along. We decided to simulate the load with the brakes. The motor lifted up the entire transmission and tray, just from the torque. The crackling noise was barely audible. Then there was a horrible bucking and clicking. A quick sniff discovered the smell of burnt insulation.
We decided that the clicking was obviously the contactors, and that they were probably disconnecting due to the violent bucking. The whole thing seemed a lot like a problem we'd had earlier, with a loose motor connection. Rechecking the motor connections, we discovered that two of the *studs* were loose. We retightened the studs and went for another test drive.
Not much better. We tried a few unlikely things, with no better results. At one point, I saw a few tiny sparks fly from the brushes, but no one else ever observed any. We took an air-conditioning break, where our brains cooled off enough for us to realize that we needed to test the whole loop, one piece at a time, until we reached the problem by elimination. (This is much better than our previous testing method, "Drive It Till It Breaks".) We decided to start with the contactors.
Thank Heaven I had a nice chunk of welding cable handy. Bypassing both contactors (so the only electrical switch in the circuit was our emergency breaker) made the problem go away. Bypassing the old contactor worked, too. We still had the same problem after bypassing the new contactor, though. Sounds like the old one is bad.
But then a recheck showed us that the new contactor was shot, too. This time Nathan noticed all the dashboard lights flashing along with the bucking. Hmmm... could our earlier testing have been a fluke? No, a complete contactor bypass still functioned smoothly. Both contactors must be bad.
Or were they? We took another air conditioning break. With a cooler brain, I came up with an alternative explanation.
Previous readings of the EVDL had warned me about the DC/DC converter I'm using. As the battery pack voltage drops, it starts drawing more and more current to supply the accessories with the power they need. The dashboard lights flashing indicated that the 12V battery was as dead as a doornail, and since it'd had plenty of time to recharge, there was definitely something seriously wrong with it; maybe it was shorted.
So, that would mean the converter would be constantly pulling juice to recharge the dead battery. As we drew more current for the motor, the battery pack voltage would drop, causing the converter to draw MORE juice. At some point, there wouldn't be any juice left, causing the converter to shut off, causing the contactors to release. Then the batteries would jump back up to a reasonable voltage, and the vicious cycle would start all over. The current draw could even burn off some insulation in the converter, which is right next to the motor anyway.
To test, I got a new battery from the Discount Auto Parts down the street. I considered an Optima Yellow Top (I didn't even know they came in that size!), but it was three times as expensive as the regular battery.
With a new accessory battery, it was time for another test drive. There was no bucking, and no motor smell. Problem solved! Time for a road test!
Yup, all solved, except for the crackling noise. Unfortunately, we'd used up all the juice in our testing. We'd disconnected the E-Meter, so it was always reading full. I limped home, praying I wasn't killing my pack; when I pulled in and reconnected the E-Meter, we were down to 117V. (That's very bad. When a block is at 10V, it's extremely discharged; this is 9.75V per block. I may have damaged something.)
We agreed to hook stuff up and try another test drive the next day. Nathan wouldn't be able to make it, but I would have an hour or so before heading off to the Labor Day party in Daytona.
The next day I decided to take Silent E to the grocery store. I'd made longer trips in the past, notably to the municipal pool, which is about twice as far. Granted, I hadn't been able to adjust the charger, so I didn't have a topped-off charge, but I was showing 143V. Should get me there and back. (Edit 2007-09-14: Note that this is less than 12V per block, and that "full" is 13.2V per block. This low charge may account largely for the short range I experienced.)
I loaded up the kids and headed off. I noticed that Silent E seemed awfully sluggish, like something was holding me back. It also didn't seem to be able to coast like my hatchback does. At a later light, I discovered that I needed to readjust my throttlepot: the engine was always pulling, just a little bit.
By that time, I was already having trouble. Voltage was dropping below 100V during acceleration, and I couldn't get past 30mph in any gear. I coasted to stop at a nearby Subway and started walking around the plaza in search of electrical outlets.
I let the kids get started with the shopping and went back to the car. There was a guy just exiting from the Subway's back door. I figured that I couldn't lose anything by asking, and he agreed to let me plug the car in. He ran the wire right into the restaurant, and I ran it around the back walls so it wouldn't be a trip hazard.
I showed off the car, then went shopping. We returned after about half an hour or so. We managed to get home, but just barely. It's slightly uphill, too. I put Silent E on the charger and went to Daytona, disheartened.
Maybe when I dropped that battery, I actually killed it, even though it still shows 12V. Maybe I reversed a cell (or more) during the testing. Maybe there's something dragging in the driveline.
We'll know more next Sunday, when Nathan comes over to help debug. Until then, I'm going to be adjusting the charger, so I get a really good charge.