Actually, both my wife and I are excellent cooks, accounting for my unfortunate girth, and Thanksgiving is one of my favorite holidays. Family visit and good food, with none of the preparatory stress of Christmas or the insect life of the 4th of July.
So, I've finally gotten around to posting LAST week's video to the blogosphere. Vanity of vanities - someone might care.
Actually, it was a pretty good show. We ran two cars down to pretty much the bottom of the power pile and took some interesting measurements. I alluded to these earlier.
Speedster the original was run to a full stop. In fact, it stopped about six times on the way back to the shed. I would let it sit for about 5 minutes and the cells recovered enough to go another 150 yards or so. Ultimately we pushed it into our new facility at 601 Morgan Oak. All cells were low but the first six cells on the front string were all actually below 0.500 v measured statically 10 minutes later. So they must have been flat zero in operation.
We are going to rebuild this Speedster anyway with a new motor, different controller, and entirely different battery pack, so I was curious. Put a slow charge on the pack and ALL cells came right back up to perfect working levels. I won't say we did these six any good, but the car is driving nicely and they seem to be perfectly happy at the moment.
The Spyder 550 was more interesting in that we did NOT run it entirely out. The 157 km trip used 176.5 AH of the 180 available and we still had cells above 3.00v. So I sent Matt for a couple of sleds of Stag Beer and by the time he got back at 183 AH they were all nicely below 3.00v.
The list of voltages tells the tale. The cells were very nicely balanced. Understand this pack had 5500 km on it and first rolled last June. It had never been top balanced, and never bottom balanced. The cells were simply installed in the car as received from the manufacturer, strapped up, and charged with the BRUSA NLG511 charger.
One thing that jumped out at me was that the first 10 cells were noticeably lower than the rest, but not randomly so. It almost looked like the first 10 were different. A graph of this makes it clearer.
Initially I thought it might have something to do with cooling and air flow since they were all in the front, but there were actually 12 cells in front, not 10.
Then I remembered the Xantrex episode. We had used a TBS Expert Pro on the Mini Cooper and powered it with the first 10 cells. The unit draws less than 60 ma so how much damage can it do to a 100AH pack? Well plenty as it turns out. It's on 24x7 and after a couple of months, they were 10 AH down. Then I recalled that Duane had been following our progress on this and had replicated our mistake. Eventually he also replicated the solution - a voltage divider across the entire pack. We've since moved to DC-DC converters actually. But the cells remained slightly drained from the originals that hadn't had that treatment.
About 15 amps for 5 minutes dressed it all up. Cell 28 was also a spot down so it got 30 seconds.
That was the fix.
We just flat did not do a show during Thanksgiving week. We have a lot going on with the Mini Cooper, but I've also taken THAT battery pack down to parade rest and we are wanting to look at all the cells while we have it down. That will take some effort as they are much harder to get to than our other cars. Thank goodness I let Brain talk me into the $550 rail/drawer system for the box inside or it would really be difficult. Anyway, we're going to measure all of them this week and tighten all the connections while we're there.
The Rinehart controller has been interesting. It has fought us every step of the way, but I think it's going to get there on the 2009 Mini Cooper Clubman. First, they gave us the wrong motor type code by accident and it did NOT run very well. That fixed, we found an even more unnerving issue. When in electronic REVERSE, if you step on the BRAKE it ACCELERATES the car. It's a little counterintuitive to get it to stop too - let off the brake and sieze the hand brake and haul in on it with all your strength. THAT was exciting. But again, it's simply a software problem and Chris Brune has been very communicative. We isolated it to just REVERSE and he was able to forward us new firmware WITHOUT this little feature literally within four or five hours. That's why they call it software.
The controller is clearly in the prototype area. The hardware design is very attractive. And I'm getting accustomed to the software to configure it. It's actually not bad. The controller is limited to 300 AMPS at up to 360V whereas the TIMS600 was 400A at up to 400V. So it's a bit tamer in actual operation. But that's not entirely a bad thing. As I've preached numerous times, and to little effect with all the power worshipping gearheads, sometimes more is just more. I confess I liked the responsiveness of the TIMS600. But the Rinehart drives the car very nicely and with nothing to apologize for. We're doing more in 2nd gear that we used to do in 3rd - we have the luxury of 6 gears to choose from and so that's all good. It's running nicely at the moment. We are chasing a little gremlin where the controller cuts out entirely. But I think we have traced it to a cable beneath our own 12v DC-DC converter that had worn a spot through to the case and was intermittently shorting the 12v bus - cutting out the contactor.
Chasing intermittents is always interesting because they are, well, intermittent. So if you do something, and it appears to fix it, that's all good unless it was just occupying the good part of the "intermittent" curve by coincidence. But I really do think we have this one.
So we went on strike this week on video and we'll cover all that next week, after we are a little smarter. The time off has given me time to play with a back burner project that is kinda cool, and a bit frustrating. I'm doing it on a Mac OS X machine in Objective C using Xcode and the Interface Builder. This is it's own little world of madness. Mac guys have always been on the EXTREME end of weird. Mostly, they know NOTHING about computers or software design yet somehow manage to achieve superstardom as DIVAS of software design. Just reading the books on the topic is a horror - page after page after page of trivial piff and tosh about absolutely NOTHING all to carry a bare structure of a little bit of information about the system. The information density is EXTREMELY low.
Objective C holds to the hide. It is inherited from NextStep - the OS Steve Jobs made Apple buy as part of his return from the wilderness. The NeXT computer was actually quite interesting in its day. But Objective C is a very strange language. All Object oriented software traces its roots to SmallTalk and LIsp - two very early experiments in object oriented design. Ultimately, C++ and Java emerged as more mature views of this concept. Objective - C is a more direct decendant of SmallTalk. Everything is an object, and it all works on "messages" between objects. The syntactic form of Object.Property that works so well in C++ is replaced by inverted looking messages such as [Object propertyGetter: this one] that I think is just ridiculous on the face of it. You can stack all of this inside bracket set after bracket set. You wind up with programs that are quite short, but with 200 character lines.
And the back and forth between the IDE and the programmatics is just hopeless.
But oddly, you can come up with a reasonably good result in a fairly short period of time, and I suppose that of course gains with familiarity. A lot of iPhone development is going on this way.
In any event, I've been trying to develop a display program that in fairly modular fashion can access serial data streams, and use them to calculate EV interesting things, and then display them in some rational form. So far, I have a Roving Networks Sensor that has 8 16-bit A/D converters in it and reports all the results over bluetooth wireless. I can use this to measure currents and voltages without the usual isolation issues. Then, I have a cheap $36 Globalsat BU 335 GPS unit plugged into the USB port, converted to a serial port with a driver. And I can capture GPS streams.
This gives me speed and its integral, distance.
So I'm trying to glom all that together into something responsive enough and attractive enough to serve as an instrumentation system. But wiring up hardware for it, writing software for it, and learning an entirely new to me operating system and programming language is a bit of a stretch. If I could do just that, I think it would all be cool. But to build EV's, videorecord THAT, edit THOSE and work on a way to display them globally where anyone could actually download one, kind of limits the time I can spend with such programming projects.
In any event, here are a couple of screen captures to show some progress. I'm actually having a lot of fun with it....