Monday, May 14, 2012

Efficiency I Can't Quite Explain

With the exception of the few months the Z3 has been off the road for repairs or upgrades, I've been driving the car for just over 26 months.  During that time, I've kept detailed records of every charge/discharge cycle of the batteries.  Every time I plug in, I note the mileage on the odometer, and the number of amp hours I've drawn out of the pack.  I then take that data and plug it into a spreadsheet that calculates a number of things for me, including how much money I've saved because I wasn't burning gas, how much the electricity I'm using is costing me, and most interestingly, how many Watt-hours per mile the car is using.

I've reported in the past the the car averages about 320 Watt-hours per mile on surface streets.  That's the number I've used to calculate the range of the car: 19,400 Wh / 320 Wh = 60.625.  This is why I've always stated the car has a 60 mile range.  And I've proved that out once, driving 62 miles on a charge once, in preparation for doing a bottom balance on the batteries.

Of course, once I go on the freeway traveling between 65 and 70 mph, that 320 Watt-hours per mile begins to look like a distant dream.  The aerodynamic drag on the car causes energy consumption to quickly rise to around 420 Watt-hours per mile at 65 miles an hour.  That meant that my round trip to work, a 23 mile journey, which includes 7 miles of surface streets and 16 miles of freeway, averaged between 370 and 380 Watt-hours per mile.  I've made this trip a few hundred times, I know the numbers.

It's no secret that I've had a few problems with the motor in the car.  The balancing putty has come off for a some inexplicable reason, twice.  Just recently it developed a short to the case that no amount of air blown through its guts could resolve.  George Hamstra at Netgain has been a champion through all of this and ultimately had a new motor sent to me.  In addition, we swapped out the brushes from the standard H-49 brushes used for high current applications like drag racing, to H-60 brushes which are better suited for street use.  A cool feature on Helwig H-60 brushes is the split, Red Top design, which helps to ensure better contact on the commutator.

At any rate, I got the new brushes seated in the motor properly, put the car back together and launched it back onto the streets about 3 weeks ago.  The car is my daily driver, so once I began driving it to work and other places, and recording the energy consumption, I was a bit surprised to notice it was more efficient.  At first I thought it was maybe just an anomaly, but it's clear something has caused the car to make much better use of the energy in the batteries.  My round trips to work are now averaging about 280 Watt-hours per mile.  Compare that to the older 370!  That's more than a 25% improvement in efficiency!  I made one trip where the average dropped to 266.

Today I took a bit of a longer trip out to Scottsdale.  A total of 38 miles, with 30 of those miles on the freeway, traveling around 70 mph.  The average consumption for the entire trip was 274 Watt-hours per mile.  In the past, I would have estimated this trip to be a 400+ Watt-hour per mile trip.  But that's not all!  During the entire trip, I had the AC system on (which draws about 9 amps) and of course, I've configured the power steering pump to run all the time now adding another 2 amp continuous draw.  BTW, the change to the power steering is interesting, but that's another post.  So there are more parasitic loads, yet, efficiency is up.

The one thing I haven't done yet is to see what average I would get if I traveled at 40 or 45 mph.   There's no reason to think the gain in efficiency that I'm seeing wouldn't appear there as well, but I simply don't have those numbers yet.

It seems like an obvious conclusion to draw that the increase in efficiency can be attributed to the new motor, or brushes, or a combination of them both.  But I really can't say that with certainty.  Perhaps I'd made some error when installing the drive line in the past which caused some binding or friction that I simply wasn't aware of.  I kind of doubt that, but who knows?  There's no question I've gotten better at disassembling the drive line of the car, but there really isn't much room for error here.  I have no reason to doubt the numbers the meter is giving me; after all, it's the same meeter with the same set up I was using before the motor swap.

What ever the cause, the car does seem to be more efficient.  At an average draw of 280 Watt-hours per mile, it's gone from a 60 mile range to nearly a 70 mile range.  I'll take it.

Update 5/20/2012:

I've continued to see the gains in efficiency I detailed above, but I've realized I've let myself fall victim to insidious creature that is over optimism.  I've always maintained that the Z3 had a 60 mile range.  That was based on the fact that it consumed about 320 - 330 Watt-hours per mile when driving at ~45 mph in normal traffic.  What I really hadn't done is take an average over multiple trips to get a more balanced number, a real world number you can take to the bank.  Well since I've seen this improvement in efficiency, I've gone back  to my spreadsheet to see if I could mine some more useful, accurate data from the numbers.  Here's how it works out...

Since the car was put on the road, up until the motor/brush replacement, it has averaged 376 Watt hours per mile.  That is the real world average.  Sure there were many trips that were better, but there were also many that were worse.  I can't get out of my neighborhood without consuming something like 480 Watt-hours per mile.  It's all about stopping and starting.  With no regenerative braking, stop signs and stop lights really affect your range.  The more of them per mile, the worse your range.  Well, there's 6 stop signs on one of the routes it takes to get from my house to the main streets, so you can imagine what that does to energy consumption.

The average after the motor/brush swap has dropped to 321 Watt-hours per mile.  Compare that to the old 376, and you note a 14.7% improvement.  That is huge!  Interestingly, and I've mentioned this above, virtually all of the trips I've made in the car since putting it back on the road have been on the freeway at 65 - 70 mph.  Meaning, that as more trips on surface streets are recorded I expect that 321 number to drop even further.

The bottom line is that I've been misrepresenting what the car's range really is by skewing the data toward the happier, more optimistic lower numbers.  Not intentionally or maliciously mind you.  The old average was 51.6, the average now seems to be 60.4.  A painful thing to admit, but there you have it.  Just as I could have easily squeezed 60 miles out of the car before, sticking to surface streets and avoiding stops, I expect I could squeeze 70 miles out of it now, doing the same. 

I've brought all this to Jack Rickard's attention at EVTV, and just like me, he was skeptical and ultimately amazed.  He's been doing some tests swapping out the original H49 brushes for the split, Red Top H60's and he's finding the same results.  I have a feeling this is going to be a hot topic in the EV community for a while.


Padraic said...



Tim Catellier said...

Thanks Padraic, but I'd sure like to know why. It would be nice to tell the community what's going on.

Padraic said...

Well, it seems your pedantic record keeping has paid off. In this week’s show (18/05/12), Jack demonstrated that your new brush type had shown an increase in efficiency in the few test that he carried out on his warp 9”. He also showed an interesting video from Helwig explaining and demonstrating the advantages of the red top.
Did you ever write a manual for your car with pictures? I would love to do a similar project, but with 100 Amphr batteries for a 40-50 mile range with a less powerful moter.


Jack Rickard said...

It appears the myth is busted and the question in a paradox in an enigma is solved.

In fact, your updated figures more closely match the 14.39% efficiency gain we had on the test bench.

It was a lot of work, but I think we have discovered something really quite useful to EV enthusiasts everywhere. It is rare to come up with a $150 "fix" that can get you 14.39% further down the road on the same charge, or conversely, reduce your energy costs by 15 cents on the dollar. But unless I'm missing something huge, that is what you've done.

We are calling it the Catellier Effect. Your careful attention to detail and data recording is absolutely the cause. Again and again, congratulations on a great build. You would not learn this much driving a Toyota Tim.

Jack Rickard

Tim Catellier said...

Thanks Jack. It stacks up rather neatly, that your test results and the numbers I'm seeing in my data coincide so closely. I still find myself shaking my head in amazement when I put the numbers in and find that I've just done a 30 mile trip averaging less than 300 Watt/hours per mile.

I really do hope others are able to duplicate this result. It would be great if I were able to return something to a community that has helped me so much.

Thanks for taking the time run tests on the crazy lead I threw you.

Ruckus said...

After seeing the results I am putting together a bulk order for the 8" Advanced DC motor brushes. If anyone wants to join, the price goes down rather quickly the more we get. I have a thread in the motor section of the DIY electric car forum.

Andyj said...

Tim have you noticed any change in your consumption on the old brushes between the damp days and the dry?

Seems many/or some brushes depend on humidity. Have a toot at your data. I'd love to know.

It's interesting, a motor that gives 88% efficiency under tested load/speed conditions can be improved almost 15% in the real world by a change of brushes.

Al Gore said "The science is settled" and he's got a hundred years of data but its not settled!

Tim Catellier said...


Though I've kept detailed records on many data points, ambient humidity wasn't one of them. What I can tell you is that most of the year is pretty dry here in AZ with humidity averaging between 30% and 45%. Late July through mid September tends to be more humid averaging between 65 and 75%, and on bad days it's in the 80's (with 110°F heat).

I went back to my data and took a sample during that time of year that we call the monsoons and found the car averaged 385 Watt/hours per mile; no different from any other sample.

Humidity may make a difference, but apparently it there needs to be a wider variance than that to be appreciable.

Andyj said...

Thanks for your good and considered reply Tim.
It's something I wanted to know after reading this webpage below. It tells of the importance of humidity. Whether it concerns wear only.. I forget now:-