Tuesday, July 27, 2010

Getting Closer

This weekend's test with the Zilla attached proved to be as uninteresting as I had anticipated. The Zilla supposedly only draws 30 milliamps. What I found supports that. After three days with the Link-10, heater, charger, and finally the Zilla attached, the meter reported that 2.15 kWh's had been drained from the batteries. No different than any of the previous tests.

That leaves only the DC to DC converters left to test. As I've reported before, the ambient temperature here in Phoenix is high enough that the cooling fans on the converters run non-stop. I don't see any point in testing how much that draws off the pack. I'm far more interested in what is drawing power off the pack when the car appears to be doing absolutely nothing. But from the looks of things, that won't be the case until September some time.

For now I know that at least 700 of the Watt hours per day that the meter reports having left the pack is fictitious. I can live with that for now. I still have some other issues to deal with, which are minor in nature, but some of you may find interesting. Right now, I'm in lovely Sedona, so those will have to wait for next week.

Friday, July 23, 2010

Testing: Stranger Still

I've been trying to determine which component is responsible for leeching power from the high voltage system when the car is idle. So far I've found that the Link-10 meter will unfailingly show a 700 Watt/hour per day loss, which I've determined to be largely inaccurate. I've run the same test with the ceramic heater hooked to the system (mind you it was off, just like the rest of the car) and it added nothing to the reported draw off the pack.

I finished a three day test yesterday in which I'd added the Manzanita Micro charger into the mix. At the end of the three day test I looked at meter, and what I found didn't surprise me at all. The meter read that there had been 2.15 kWhs drawn from the pack. That lines up exactly with what I would have expected, and so I can infer that the charger didn't draw any additional energy from the batteries. Again, not to surprising, but I have to test everything if I'm going to be thorough.

What was surprising is what happened when I charged the pack. In the first two tests, I saw that 2.15 kWhs had been drawn off the pack (according to the Link-10). But when I charged it, the charger read the pack as fully charged within minutes and when it finally shut off, it had only added 200 Watt/hours to the pack. So really, there was only 65 Watt/hours per day that were actually consumed. The first two tests had nearly identical numbers. Well, when I charged the batteries this time, the charger said the battery pack reached it's target voltage while I was turning the dial up to increase the current. When it finally kicked off, it had replaced a mere 60 Watt/hours. So, in this test, with only the Link-10, ceramic heater, and charger hooked to the high voltage system, their combined draw was 20 Watt/hours per day.

How could there be such a discrepancy? And how could the three components combined draw less energy than two of them? I think the answer is they didn't really. In spite of the fact that I'm trying to run these tests in the most controlled manner possible, I'm limited by the quality of the equipment. After all, it's not high dollar lab equipment, they're EV components. There is some inaccuracies inherent in each, i.e. the consistent 700 Watt/hour per day error in the Link-10. In this situation, I think we've exposed another inaccuracy in a different component, the charger.

The Manzanita Micro charger is fantastic as a bulk battery charger, and I think it does a terrific job. It is not a high dollar piece of bench lab testing equipment. It uses a potentiometer adjustment, made by the user, to determine the cutoff voltage for charging. I've set it such that the charger starts it's ramp down when the battery pack gets to ~165 VDC. It then ramps up to ~168 VDC before the timer runs out and it turns off completely. Now, the reason I use a "~ " is because it is it's all relatively approximate. Sometimes the timer comes on at 165 VDC, sometimes it comes on at 165.5 VDC. In addition, you may remember I'd said the algorithm it uses for how to ramp the current down is a complete and total mystery. So sometimes it will dump an additional 150 Watt/hours into the pack after the timer starts and sometimes (apparently) as little as 20 Watt/hours.

I guess the short of it is that using the charger as an instrument for doing the fine measurements I'd like to employ on this test is silly. At best I think I can only hope to get an idea of what's drawing current off the pack when the car is off. I've already discovered that the meter is responsible for most of what I've seen simply due to the error inherent within the meter. Am I going to find where that remaining 300 Watt/hours per day are going? I'm not sure, but I'm going to keep trying, and my feeling is that I'll find the DC to DC converters are drawing some of that current.

In any case, the next component to check is the Zilla controller. How much does it draw? Well it's supposed to be just a few milliamps. Over a 24 hour period, I'd be surprised if that registers on the meter. But I intend to test it anyway. That test began 30 minutes ago.

Stay tuned for the increasingly irrelevant conclusion!

Monday, July 19, 2010

Testing Continued

Before I talk about the results of the most recent tests, I thought everyone might enjoy seeing one of the shots I took of the car, which I sent off for the book the EV Z3 is going to be in. Behold:

Considering the day was a bit cloudy and I don't own any lighting equipment other than a camera mounted flash, which I didn't use, I thought this turned out pretty good. Say what you want about the conversion, but it is a pretty car.

Anyway, back to the test results. So, over the weekend, I left the car idle, in the garage with only the following items attached to the high voltage side: the Link-10 meter, and the ceramic heater. As I mentioned in the last post, there is really no way for the heater to draw any current, but I swore I'd do more thorough tests, and that's what I intend to do. In fact, I decided against the idea of cutting that test short because I didn't expect the addition of the ceramic heater to amount to any difference. After all, if you allow your expectations to influence the method of your testing, your expectations influence and then determine the results of your test.

What did I find, you ask? Pretty much what I expected. The addition of the heater didn't change results at all. Over the time frame, the meter showed that 2.15 kWhs had been consumed. That is bang on what it said when I had only the meter attached. But if you remember back to that test, when I charged the batteries, they only accepted 200 Watt/hours. And after all, that's really what I need to know; how much energy was actually drawn out of the pack. This time was no different. About 3 minutes after I turned on the charger, the battery pack reached it's peak charging voltage and the charger began to ramp down.

By the time it had finished, 220 Watt/hours had been pushed back into the pack. I'm not concerned with that extra 20 Watt/hours that showed up in this test versus the first test. If I did the same test 10 times I'd expect 10 slightly different results because the charger introduces some randomness to the equation. The algorithm it uses to determine how many amps to push out while it ramps back down to zero during the cool down phase is anything but predictable. A 10% difference at this level is not a big deal. Considering the 2.15 kWhs the meter originally stated had been drawn off the pack, that 20 Watt/hours is only about 1%.

On the surface it would appear that I didn't learn anything. But really I did learn that the ceramic heater is definitely not drawing any current when it's off. It's true I expected that, but I also learned that the Link-10 showed the same error in metering 2 tests in a row.

In order to charge the battery pack I had to hook up the charger. That means that right now the Link-10, the ceramic heater and the charger are the only things hooked to the high voltage side. That sounds like a perfect recipe for the next test, which commenced 30 minutes ago.

Friday, July 16, 2010

Not Much to Report

Everything has been running smoothly this past week. The few times I took the car out, it hummed along effortlessly, with no hint of a problem. For those of you not following Arizona's weather closely, it's been hot. Hotter than usual. Yesterday was 116 °F with humidity near 50%. There was no way I was taking the car out in that weather. Remember, no AC.

I'm continuing my tests this weekend to see if I can find which component is leeching power from the batteries when the car is off. So far the only configuration I've tested was to have the Link-10 meter hooked up by itself. You may remember that it reported 700 Watt/hours being drawn off in a 24 hour period.

The Link-10 will remain hooked up the entire time, after all, I need to monitor the pack during the test. But based on the previous test, we know the Link-10 brings along a 700 Watt/hour per day draw to the test. So I expect that if a component is drawing current, I should see more than a 700 Watt/hour per day reading on the meter.

Well, it's time to hook up the next item, and I need to choose which one. I still have to test the Zilla controller, the heating element, the charger and the DC to DC converters. Since the DC to DC converters have been fooled by the hot weather into thinking that they are too hot, their fans run constantly. So I think it's best I skip testing them for now. I think I'll move next to the heater.

Now the ceramic heater draws it's power directly off the traction pack, but it's behind two relays and a manual switch. In other words, in order for it to get any power, the car's HVAC fan has to be on, triggering the first relay. That makes the switch on the dash live. That switch must then be activated in order to trigger the final relay that actually sends power to the ceramic heater. So what are the chances that the heater is drawing any current when the car is idle? I'd say 0%, but I set out to test this methodically, and that's what I'm going to do.

I may not give the heater the full weekend to test it. If I see no change from the 700 Watt/hours I expect from the Link-10 in the first 24 hours, I'll add the next component and start the next test.

Sunday, July 11, 2010

Didn't Make it to Laguna Seca

Well, the REFUEL event at Laguna Seca went off today, and sadly I could not get there to participate. Over the last couple months, I reinstalled the motor and redid the front battery racks, getting the car fit and ready to participate. Where it all fell apart was in trying to arrange transport to get the EV Z3 to the track to participate. I don't own the right equipment and the cost of renting it was just not in the budget.

I'm hoping that there will be similar events here in Arizona in the future, but I'll just have to wait and see. I'd really love to get the car out on a track to see how it does.

On a better note, I took the car out for an extended drive early yesterday and found some interesting numbers. While here in the neighborhood, the car averages about 350 Watt/hours per mile (dismal). However, when out of the neighborhood, on surface streets, it averages about 250 Watt/hours per mile (better). The only thing that I can figure is the constant stopping and starting required by the high number of stop signs in the neighborhood take their toll quickly.

I believe the numbers can be better though. The transmission already has Redline transmission fluid in it, which reduces friction in the drive line. After I sent the transmission away to have the adaptor plate made, I filled it with Redline. I haven't yet put Redline in the differential, so I still need to do that. I don't know how much difference it will make, but it can't hurt. I'm keeping the tires at 45 PSI. The tire's maximum pressure is 51 PSI, so I'm well below the threshold, but they are certainly more firm than the normal 32 PSI.

There are two big things I need to do yet to help with efficiency. First I have to get the front end re-sprung. That won't affect the drive line, but it will get the ride height to the correct level. Once that is done, then I can have the front end aligned and take out any toe-in that's there now. I expect that will make the biggest difference of all.

So, why haven't I had the front end re-sprung yet? A couple reasons really. First, I had to take the car apart recently, you may remember. That hampered my ability to drive it anywhere. Second, the shop that can do it is in the middle of Phoenix, about 23 miles away. Now, the distance isn't a problem for the car, it can do 60 miles. However, the heat is a problem. It's been over a 110 °F lately, which makes for miserable driving when you don't have AC. Second, the little radiator and associated fans used to cool the Zilla have a difficult time keeping it cool in such high temperatures. I could drive the car there early in the morning, but by the time the car was done, I'd have to limp back home in the heat, doing my best to protect the Zilla. At this point, I'm inclined to wait until cooler weather comes so that I can get it there without any risk to the controller. After all, the car works great now, it just uses a bit more energy than I think it needs to.

Oh, and did I mention how much I love that new vacuum pump?

Friday, July 9, 2010

Stranger Still

Let me say that the following series of events falls outside my official testing scheme. Circumstances this week have kept me from continuing along the "add one item at a time to the high voltage system" plan that I want to follow. That said, I find the following thing puzzling and not all together irrelevant.

Tuesday, I had to hook all the connections back up to the high voltage side so I could back the car out of the garage. It then sat idle, with the DC to DC converter fans running the entire time until last night. I took it for a ride and used up 1.7 kWhs. This morning, the meter read that the systems had drawn a total of 2.9 kWhs from the pack. I charged the car, and watched carefully to see how much energy the Link-10 reported going back into the pack.

If it behaved anything like last time, it would have counted back from 2.9 kWhs to around 1 kWh and then the pack would have reached it's peak voltage and the charger would start it's 10 minute decline to 0 current. However, this time it got to 0.0 (meaning the meter thought the charger had replaced all the current that had been used) and continued on. It reached +.02 or 20 W/hours before the charger reached it's voltage limit.

When it was all said and done, the charger had dumped approximately 3.4 kWhs into the pack. That means that the Link-10 had missed 600 W/hours in it's metering duties since it was last charged.

To recap, the last controlled event the meter showed that 2.5 kWhs had been drawn off when it was really closer to 200 W/hours. This time it showed that 2.9 kWhs had been drawn off when it was really closer to 3.4 kWhs. Like I said before, this doesn't fit into my official testing plan, but I think it is significant. It's showing me that the Link-10 is really giving me an approximation of the current drawn off the pack at best. More to follow.

On the bright side, the car is running beautifully. Although I'm hearing an annoying rattle that I have to chase down.

Monday, July 5, 2010

Of Questionable Metering

As you all know, I decided that I needed to approach some of the small problems I'm having with a much better plan. Ok, a plan, seeing as how I really wasn't using a plan before.

First, I unhooked everything from the high voltage system with the exception of the Link-10 meter. I decided I'd leave it like this through the weekend and check in periodically to find out how much electricity was being consumed out of the pack. I was absolutely astounded to see that in the first day, 24 hours, the meter reported that 700 watt/hours had been drawn off the pack. I figured that simply couldn't be true. That would be the equivalent of a 30 watt bulb burning during the entire 24 hours. If it is being burned off, I don't know where. There's nothing moving, glowing or getting hot.

I left the system configured this way all weekend, and what I found was that the meter, rather reliably, reported 700 Watt/hours for every 24 hours the car sat idle. By the time I got around to charging it tonight, the meter read that 2.5 kWhs had been drawn off the pack. I hooked the charger back up to the battery pack and plugged it in. Now the charger puts out 27 amps at 160 volts, or 4320 watts per hour. At that rate, it should have taken about 35 minutes to get to full charge.

I walked into the house to finish a game I was playing with the family, walked out about 12 minutes later to find the charger had shut off. I measured the voltage on the pack and found that it was at 166.8, which meant it was on it's way down from the 168 that I charge the pack to. I took a look at the meter and it said that I'd only replaced 200 Watt/hours of the 2.5 kWhs that had been pulled out. That could only mean one thing, that energy had never been drawn out of the pack. At the most, 200 Watt/hours had been consumed, and I'm reluctant to think it was even that much.

That means that something in the Link-10 is doing something wrong. If you look at the screen that shows amps that are currently being drawn, it shows 10 milliamps are being drawn off at any time. But the Link-10 is converting that into 700 Watt/hours over a 24 hour period. Clearly a math error or something more serious. I may be contacting Xantrex about this one.

You may remember that I had noticed that when all the systems were hooked up to the high voltage system, I was seeing something closer to 1 kWh a day being drawn off. Well, I can now account for 700 Watt/hours of that in just the internal errors in the Link-10. I now need to hook up each component one at a time to find out what's responsible for the other 300 Watt/hours.

Friday, July 2, 2010

I Spoke Too Soon

All is not well in the land of 12 VDC subsystems. I've always made a point of recording both the good and the bad on the car, and on occasion, the stupid things I've done. Get ready, cause this is going to be entertaining.

I need to give you a little back ground first. After I had taken the car apart, I found out that you're supposed to store these LiFePo4 batteries at a 50% charge if you expect to let them sit idle for over a couple weeks. Well when I started the dis-assembly, the batteries were at 90% capacity. Knowing that the repairs would take several weeks I needed to run them down, but how could I do that without the drive line in place. I hit upon the idea of simply running the heater system and burn it off that way. It worked beautifully. I was even able to predict within 5 minutes when the pack would be drawn down to 50%.

Before I started the heater, I had to connect the batteries I'd taken out of the car to the ones that were still mounted, hook the heater up to the batteries and the DC to DC converters up to the batteries as well. I carefully connected everything up, but I found something unexpected. I had hooked the positive lead to the DC to DC converters, and then put the multi-meter on the negative lead. I don't really know what I was expecting to find, but I can tell you it wasn't what I did find. I found that there was a positive charge on the negative lead. The positive side of the batteries was hooked to the converters, and I could read that positive signal on the negative terminal to the converters. The current was passing through the converters!

Maybe some of you reading this will say "Well of course it does." But I had no idea. That meant that when I hooked it up to the negative terminal, there was a nice big arc. No damage and all worked fine, but it still sucked.

Flash forward to this week. In deciding that I wanted to turn the DC to DC converters off when the car is off, I decided that I'd use two big relays; one on each side, positive and negative. I figured that way when they connected and disconnected at the same time, there'd be no time for a charge to build up in the converters and no arcing. Apparently that was wrong because in about 5 or 6 cycles, I managed to weld the contacts for both relays closed.

OK, so now my solution for shutting off the DC to DC converters isn't working. I'm thinking that some heavy diodes might solve that problem, but I'm not sure. But it turns out that all of the work I did yesterday was the result of an assumption and I dare say, a stupid one at that.

You see, ever since the car first hit the road, I noticed that sitting idle and off, the meter would show that 500 to 600 W/hours had been consumed in a 12 hour period. Now, there are 5 things hooked to the high voltage side on the car: the Zilla, charger, DC to DC converters, the ceramic heating element and last, the Link 10 meter. I had always assumed that the vampire in the car had to be the DC to DC converters. This past week I even disconnected them so that I could test whether the little 12V SLI battery could handle the car's load while the converters were disconnected. I dutifully measured the progress of the SLI battery on it's march to death (rescuing it well in time). But since I disconnected the converters by tripping the emergency disconnect switch, I got no data on the Link 10 to show if the pack had gone down. I just assumed that it hadn't! That was a mistake.

Last night, the car sat for 12 hours with the SLI battery powering the 12V system and the DC to DC converters off (a flick of one of the relays freed it and broke the connection). When I came out this morning, I found that 500 W/hours had vanished. Well, it clearly couldn't have been the converters, they were off! That's when I realized that I'd been guilty of some very poor troubleshooting practices.

OK, so what to do? Well, I've now tested one of the components that is hooked up to the high voltage system, and found (to my surprise) that it's not the energy vampire. Before I move any further, I have to explore all possibilities. As far as I can work out, the possibilities are either one of the other components is burning off that energy, or the Link 10 is reporting that the electricity has been used, when it really hasn't been. I think it's safe to rule out a short in any given battery as I tested them all after I assembled the car and found they were all the same. If it were a battery, that one would be dead.

So, what to do about it? I think the first thing is to test the Link 10. Right now, the high voltage pack sits in the car with the Link 10 being the only thing connected to it. I should know in 4 or 5 hours if it's going to report any draw down on the pack. If it does, I'm left with the task of determining whether the Link 10 is responsible for using up the energy (unlikely) or simply reporting it wrong. If it reports no change, I'll know that I can hook up one more item. I'll keep doing that until I find the culprit.

So, what have we learned today? Don't make assumptions when troubleshooting! Now if you'll excuse me, I'm going to go write that on a black board a few hundred times.

Thursday, July 1, 2010

12 Volt Dilemma Sorted

Yesterday afternoon, I pointed a large box fan right into the trunk, set it on high and let it run for an hour. I wanted to find out if the fans on the DC to DC converters would turn off if they had enough fresh air. They get fresh air now, but this would be a lot more. The question was, are the fans running continuously because they aren't getting enough fresh air, or because the ambient air isn't cool enough to actually cool them. Well, after an hour, I turned off the hurricane directed at them and found the fans were still humming away. So, 100 °F is simply too warm to cool these devices properly.

Last night I let the car sit with the DC to DC converters on and the 12V SLI battery connected. In the morning, after 12 hours of idle time, I found they had consumed 600 W/hours. To recap, without the SLI battery connected, they consumed 5oo W/hours, and with them off and the SLI battery handling the load, the car consumed 44 W/hours from the battery.

In light of those two discoveries, I made the rather easy decision to hook the DC to DC converters up to a contactor that's triggered with the ignition key. I want them off when the car is not running. It took a few hours to route the wires and such, but in the end it worked out ok. The new contactor is sitting atop the SLI battery in the right corner of the trunk. It's out of the way and should be safe from anything rattling around in the trunk.

The one thing that I'm going to have to watch carefully now is the voltage on the SLI battery. I'm asking more of it now since it will be providing all power to the car when it's off. Sure, it's not much, but in 3.5 days, it will draw that small 33 amp/hour battery down to 10.5 Volts. I have a battery tender that I'll be hooking up to it. When ever I plug in the car to charge it, I'll plug in the battery tender for the SLI battery to keep it healthy for any extended stays in the garage. One extra plug and the energy it will draw is so minute as to be nearly negligible.

This evening, I'll be taking the car out to a nearby park to take some glamor shots. The car is going to be featured in a book about conversions, with a section dedicated to particularly interesting ones. The EV Z3 is going to be in that section. Is that cool or what!?