[Sparweb]

Hi,
I've been lurking a long time, but still far from a chance to start an EV build of my own. Instead, I've been collecting info from the experts here. Then I realized that some of my research on DIY EV's might actually be worth sharing.

I've been comparing groups of EV's, looking for trends and similarities, and I found a few. I put a lot of the data from the EV Garage into a spreadsheet so that I could compare the battery pack size, motor power, range and so on. I also added data from production EV's for comparison. I hope you find these comparisons interesting.

Both graphs are grouped into four distinct types:
1) DIY lead
2) DIY LiPO
3) Production EV's
4) Tesla (a class of its own).

I plotted trend lines because there's a lot of scatter in the data - but actually a lot less scatter than I expected, and the points for each "technology" are definitely grouped. They also tell a bit of a story of technical progress as EV's become more capable.

The DIY lead is obviously the past, and here's a clear illustration of why. To get any range requires an enormous amount of battery mass, but soon mass of the the batteries overwhelm the mass of the car.

Close together are the DIY LiPO and production EV's. The production EV's have a slight advantage, but there is considerable overlap. It is clear that the DIY builders have access to technology that is similar to the technology being used by the major automakers. Note that I haven't included any DIY builds that use production vehicle drivetrains or batteries. That could be a 5th technology group but it might also have been confusing. So no DIY Tesla drivetrains in a Mazda Miata in these graphs.

Then of course, there's Tesla. By these comparisons they really stand out, although I must give a shout for the proud little Chevy Bolt which has slipped into the Tesla pack by these measures.

If anyone wants to see the original spreadsheet, they can download it here: http://www.sparweb.ca/EV/Design/EVs_Garage_Comparison.xls
I tried to upload the whole thing to Google sheets - but I can't make a graph to save my life on that site. And if anyone finds any errors or would like me to consider something else, I'm all ears.

I hope that plots like these will be helpful to people who are considering what kind of range and pack size they should expect and/or aim for in their conversions. Getting good, reliable data is hard. I have to rely on the claims of builders, but you know, we're not all scientists. Take comfort that we shouldn't believe the range values from the OEM's, either.

For instance, I've seen people suggest that a 3000 pounds DIY car might get 300 Watt-hr per mile, but I haven't seen a credible example of that in a DIY car, but a production EV can do it.

I have also seen people limit their DIY build to about 60 Amp-hours of battery pack in a fairly heavy (4 seat) car. That is fine if they didn't really want a lot of range or had a limited budget, but the experience of others shows that a DIY project can reach 20% or more of the car's weight in batteries, with a payoff in driving range that beats many production EV's.

Overall, I'm still very impressed with what DIY EV builders can achieve. Their results can equal and in some cases even beat the major OEM's is just amazing, and I look forward to the day I can join the club!

 

[ishiwgao]

great stats! very informative.

just a recommendation though. the graph regarding electric car range vs weight, I would suggest you change the y-axis title to energy per range, rather than charge per range (i would prefer energy per distance though).

I'm sure everyone knows what you mean when you say charge, but maybe it's better to use a simpler word rather than an incorrect technical term.

 

[Sparweb]

Done!
Thank you. I realized you were right even before I finished reading your post.

 

[kennybobby]

Great study you have done and interesting data, thanks for sharing your research--it will be a very useful reference.

One annoying thing about microsloft excel graphs is the ugly gray background that makes it difficult to read--but that default can be changed to white or no fill, look for 'format plot area'. You wonder who was the idiot that thought it was a good idea, just try plotting one out and waste ink on that background...

 

[Sparweb]

Thanks kennybobby
I agree that the colors in Excel are bad, but I think they are worse now, with those faded pastels...

 

[Karter2]

Interesting data.
Im not sure that "Wh/mile vs vehicle weight" is a valid comparator as there are so many factors to influence Wh/mile beside weight..
EG...
Aero design
Drive type..AC, DC, gearing, etc
Test routes
Speed
Etc

 

[brian_]

Im not sure that "Wh/mile vs vehicle weight" is a valid comparator as there are so many factors to influence Wh/mile beside weight..

I agree that weight is not a strong factor, which is why the increase in consumption with weight is a shallow slope. The slope in the "blue" category is shown as steep, but there is very little correlation; the Tesla correlation is strong, but the slope is shallow.

Mass affects energy required to accelerate, but EVs make some of that back on deceleration and any vehicle makes some back when coasting down.
Weight (a consequence of mass) increases rolling drag, but aerodynamic drag is a more important factor and adding battery to the same body doesn't change the aero drag.

 

[kennybobby]

i think weight is a first-order factor to consider and definitely a good metric.

Look at the yellow line for Pb batteries, typically the oldest diy projects out there. To get more range additional cells would be added; but this increased the weight and subsequently the energy consumption, as seen by the steep slope.

With the availability of LiFePO, diy conversions could be done with less weight in the pack. For the blue trend line there is a dramatic change in slope compared to the yellow, but the general principle applies that a lighter car uses less energy per distance.

The OEMs can do better than diy, but the slope is still in the same direction: more weight uses more. Even for the "tesla" class, heavy uses more.

so from these charts it's clear to see that weight is a primary factor for consideration of average expected range and energy consumption. Plenty of same-route testing of this has been done over the years by jack over at EVTV.me for various conversions, various pack sizes and cell types. He came up with the 10% rule of thumb on consumption versus weight: a well-done 2500 lb Li-cell conversion could expect to get 250 W-h/mile

 

[jonescg]

Here is an interesting link:
https://www.elektroauto-news.net/2017/gewicht-keine-entscheidende-rolle-reichweite

Basically weight doesn't really impact range - rather the amount of battery. Because batteries can range from lousy (lead) to outstanding (Li-NMC) weight is something of a secondary issue.

Energy density (Wh/l) and specific energy (Wh/kg) are what determines a long range car, and Wh/l is probably the more important. If it doesn't fit, it won't matter how heavy it is.

Some cells offer great power density (W/l) and specific power (W/kg) but always at the expense of energy.

Below is a little charge I've been compiling. Basically you can have power or energy, but the bleeding edge is both.

 

[ishiwgao]

Here is an interesting link:
https://www.elektroauto-news.net/2017/gewicht-keine-entscheidende-rolle-reichweite
Below is a little charge I've been compiling. Basically you can have power or energy, but the bleeding edge is both.

all the batteries seem to lie on a line that lies between 100Wh/kg 3000W/kg and 280Wh/kg 300W/kg, except for FEP9050135. May I know what cell is that?
tried googling for that but came up empty

 

[jonescg]

all the batteries seem to lie on a line that lies between 100Wh/kg 3000W/kg and 280Wh/kg 300W/kg, except for FEP9050135. May I know what cell is that?
tried googling for that but came up empty

It's the same pouch cell I use in Voltron Evo - 30C continuous, 60C burst. Its energy density is not the greatest, but at least the low impedance makes for a fairly efficient cell.
https://endless-sphere.com/forums/viewtopic.php?p=1269272#p1269272

 

[Karter2]

Basically you can have power or energy, but the bleeding edge is both.

Yes, but few need that "bleeding edge" combination.
For example that 18650BE , with a low specific power, is basicly the cell Tesla use ..to give one of the best performing EVs available in terms of power and range.

 

[brian_]

He came up with the 10% rule of thumb on consumption versus weight: a well-done 2500 lb Li-cell conversion could expect to get 250 W-h/mile

As a really rough guide - which is what a "rule of thumb" is for - the weight serves as an overall indication of size. The heavier EV will typically be wider and taller as well, so it has more aero drag and more of everything that consumes energy.

 

[Sparweb]

Thanks for all the feedback.
I'm surprised at the attention you have given to the 18650 cell. Definitely a critical factor in the construction of so many LiPo battery packs, but is it the chemistry, or the packaging of these cells that matters most?

In my study, it didn't seem to matter what system voltage was chosen; AC or DC. With the vast majority in the ~140-200 DCV range there are not enough very high/very low examples to use for comparison.

If anyone knows of some well-documented examples in the forum or the EV Garage of high-voltage and/or AC drives I'd love to see how they look relative to the typical DC drives. I did find a couple, but they are extreme examples with very powerful traction motors (~300kW) so I don't think they are representative.

 

[jonescg]

I work in a field where we use 18650s, so I have done a stack of work on them recently. Data was easy to come by.

The chemistry has made the biggest difference moving from LiFePO4 to LiCoO2, LiNiMnCoO2 and LiNCA. But cylindrical Li-ion cells consistently offer better energy density (by volume) than pouches.

DC motors are less efficient than AC motors, but you could argue with the money saved by installing a DC motor you can afford more battery

My race bike is high voltage AC but it's probably not much use to your comparisons (being a motorcycle, and a racing machine, efficiency was the #3 priority).

 

[Sparweb]

i think weight is a first-order factor to consider and definitely a good metric.
Plenty of same-route testing of this has been done over the years by jack over at EVTV.me for various conversions, various pack sizes and cell types. He came up with the 10% rule of thumb on consumption versus weight: a well-done 2500 lb Li-cell conversion could expect to get 250 W-h/mile

I haven't watched enough of EVTV to have found any episodes with road mileage tests, but since you mention it, I'll go back and have a look.

About the 10% rule of thumb: my collection of numbers doesn't support that. No claim to a scientific basis here, but of what I've seen on EVTV, they can't either. My list comes from builder's own claims, and of the 20 cars I put in, only 4 can claim they hit the 10% rule. The rest are like 12% to 13% or more.