Sunday, November 15, 2015

SparkEV, you are NOT a compliance car

I didn't want to discuss this, because it was so obvious, but it seems some (many?) have mistaken notion that SparkEV is a compliance car. This post is to show some evidence that isn't the case.

Compliance 101

For those not aware, few states in US mandate few percent of total car sales must be zero emission vehicles (ZEV). They also create a market for it such that car companies that sell more ZEV than required can sell the excess sales figure as "ZEV credit". Some gasbags claim that Tesla is making out like bandit for selling ZEV credit than actually selling cars; if you can't see the flawed logic in this argument, please, read it over again. (hint: you must sell ZEV to get ZEV credit!)

There used to be one ZEV mandated state: CA. Then OR joined. Now supposedly there are several ZEV states operating as co-op where credit sold in those states can be applied to over all pool.

Definition of compliance car

As such, the definition of compliance car is simple. It's cars only sold in compliance states (CA and OR for now) only for the purpose of getting ZEV credit for them to sell more gas cars. The reason they don't sell outside of compliance states is due to cost. Even though the car maker can buy ZEV credit from their competitors (eg. Tesla), it does nothing for them while boosting the competitor's bottom line. Even if they're losing some money (or lots of money) by selling ZEV in compliance states, it'll at least bring the people into their dealers and keep the money out of their competitors' bottom line.

One famous compliance car is Fiat500e. Sergio (Fiat CEO) famously said that he hopes people don't buy the 500e, because the company loses $14,000 for every car sold. Whether this is reverse psychology is debatable. But the little Fiat is a down to a tee (cee?) example of a compliance car. Supposedly, the CEO had such disdain for BEV that much of the drive train of 500e were not made by Fiat, but contracted out to Bosche. As such, there's no hope of making money by selling another company's car, especially if that car must use biggest (and most expensive) available battery among mid level EV at the time.

You dirty compliance car!

Some consider compliance cars as a dirty word. They see them as taking away from "true" EV and helping the evil gas car maker that'll make even more CO2. It's like they have some religious aura about disliking compliance car outside of logic and reason.

The simple fact is, compliance cars are just as much EV that doesn't use imported oil. If the compliance car performs better and cost less than "true" EV, by all means, I whole heartedly recommend compliance cars. If they didn't offer this better and low cost EV, would consumers buy more expensive and crappier "true" EV or would they simply buy gas cars? Most would probably buy gas cars, making the situation (imported oil and CO2) far worse.

Compliance cars would be losing money for the car maker. If they made money, they'd have wider sales (see below why not always the case). If the evil car maker is only selling compliance BEV to be able to sell more gas cars, and their BEV is good / cheap, it's all the more reason to buy them. Buying their compliance EV and hurting them financially is a good thing if you happen to dislike the car maker.

Here's an example. Fiat 500e is pretty much the definition of compliance car. But if it came with DC fast charging and cost comparable to SparkEV ($26K, $7K less than current 500e price), it would be a great tiny car: 0-60 in bit over 8 seconds, almost 2 seconds quicker than Leaf, and it comes with 6.6kW L2 instead of SparkEV's 3.3kW. Personally, I like the taller head room of SparkEV, but for smaller people who like "cute" cars, it would've been the best EV. It'd even be a great choice for people who live in areas without much DCFC now (will be coming), but many L2. Compliance or not, such car would be a great buy.

SparkEV compliance test

SparkEV is sold in limited number of state, CA and OR in the beginning, also in MD as of third quarter of 2015. Hmm. Smells like compliance car.

Lesser known is that SparkEV was also sold in Canada to fleet customers from the beginning. Starting in 2016, SparkEV will be sold retail as well in Canada.

SparkEV also has been selling in South Korea since the beginning.

SparkEV is also sold in Mexico.

Just from the fact that SparkEV is sold in places outside of compliance states of US means SparkEV is NOT a compliance car.

SparkEV pricing

Mexico is an interesting link. The price in the web site in Mexican Peso converted to US dollar result in $24K while it's sold as $26K in US. If it's only for compliance to meet ZEV credit in certain US states, why would Chevy sell SparkEV in Mexico for even lower price than US and lose even more money? Are they insane? Or more insane than usual? hint: Iron Duke Camaro!

Most likely, Chevy probably isn't losing money by selling SparkEV, but making money. Otherwise, why offer for sale in places that do not mandate it? Based on Mexico price, they're probably making $2K to $3K per car in US. This could be more than 10% profit, a profit margin that even Carlos Ghosn, the legendary CEO of Nissan, would be proud of.

By the way, Carlos will be a historic figure, and all those who's shaken his hand or have some memorabilia from/about him is well advised to keep them safe; they will become valuable to your grand kids. No, I'm not suggesting that you cut off the hand that shook Carlos' hand, but don't wash it. Ever!

Even without knowing Mexico pricing, let's deduce how much SparkEV would cost. SparkEV is basically Spark gas version converted to EV. Much of the car's body and chassis is the same, adding only minor cost. Electric motor is made by GM, a derivative of  Chevy Volt motor. Then the gas engine swapped with electric motor is only minor cost difference, if any. Biggest addition is the battery. Assuming $300/kWh, 19kWh battery would cost $5700; let's round up to $6000.

Spark gas 1LT costs $15,000. Assuming 10% extra for swapping gas engine for electric motor and adding $6000 for battery,

$15,000 * 1.1 + $6000 = $22,500

Well, well, what do you know? It's $1500 less than MSRP in Mexico, and $3500 less than MSRP in US. Those are some healthy profits for Chevy!

Why not sell SparkEV throughout US and the world?

From above, it should be clear that Chevy is probably  making money by selling SparkEV. Then the question becomes, why not sell it everywhere? Why only in few markets (some outside compliance states)? Indeed, Chevy had announced much wider sales in the beginning, only to cancel them later. They wouldn't have done that if it was compliance only car. This involves guessing, so I'll present a few.

One reason could be the US federal tax credit for EV. Each car maker is allowed 200,000 EV sold before the tax credit is sunset. Chevy has already used up about 100,000 with the Volt and Cadillac ELR. If they use up all the tax credits before their upcoming Chevy Bolt is released, the fear is $7500 higher price than the competition (Tesla Model 3, new Nissan) will doom the Bolt sales.

SparkEV sold about 950 cars in Apr. 2015 in only 2 states, probably much of it in second half the month when they announced price reduction. Subsequently, they have been constantly sold out in much of the dealers, so the sales figure is less and hard to gauge the demand. But if 950 car in 1/2 month in 2 states (about 20% of US population), nationwide roll out would result in 950 * 2 * 5 * 12 mo = 114,000 cars per year. That'll pretty much wipe out the entire tax credit even before first Bolt rolls out of the assembly line. Even considering 1/5 of that figure (about 35,000 cars sold for 1.5 years before Bolt), that'll eat into tax credit in significant numbers.

Another reason could be that Chevy is being cautious with all battery EV by testing the waters with SparkEV. In case it doesn't hold in limited markets, they're not likely to lose out much. Indeed, releasing in hot climate (Mexico) and cold climate (Canada) is just such test, and it doesn't cost Chevy a dime while they gather valuable data on their BEV performance in the real world. In fact, they probably make some good  profit from conducting the test. And SparkEV drivers benefit from being able to drive what is the best BEV for the money available in the market today. Everyone benefits without too much risk for Chevy.

Then there's the gloomy reason: Chevy isn't serious about BEV, and SparkEV is a ploy to keep green naggers off their back. In Mexico with the awful air quality in some cities, selling SparkEV would be a ploy to get people to Chevy showrooms. In Canada, well, there are some crazy Canadians who dare drive EV there, too (hi Andrew!). MD is where SparkEV motor is made, so that'll be a good PR for them. Basically, Chevy's main reason for SparkEV is meeting compliance requirements, and doing very little else to keep naggers at bay. But even so, having the best EV in the world available at some places is better than not having it at all.

It could be combination of reasons, different reasons, no reason at all, who knows? But above try to make some semblance of reasons why Chevy is doing what it's doing with SparkEV limited release. Sometimes, human mind just keeps asking, "why? why? why?"

But one thing is clear. Mr. Maury Povich, take it away.

SparkEV, you are NOT a compliance car.

Edit Nov. 19, 2015

If SparkEV is a compliance car, how many cars would Chevy need to sell to qualify? We don't know exactly, but we can get some clues from other compliance car sales. According to, Oct 2015 sales report, Fiat 500e sold 5539 cars while SparkEV only sold 2311 cars.

SparkEV is selling for about $7000 cheaper, leasing for $1600 cheaper, it's quicker, it has DCFC. One would think that SparkEV has some other major problems, and thousands of them are sitting at dealer lots. But that isn't the case. SparkEV is constantly sold out. In fact, I'm surprised that they managed to sell even 177 cars in Oct. 2015. Where are they selling? It's not in SoCal as it's been sold out within 250 miles of here for since June of 2015.

Fiat 500e is selling more than twice as many as SparkEV, a car that supposedly loses $14,000 for each sold and their CEO famously quoted as saying they won't sell a single car more than necessary. That suggests Chevy should make lots more SparkEV to meet the demand to meet compliance; having it sold out makes no sense.

Then why not sell more SparkEV? Demand is certainly there judging from sold out dealers. There is no way to know for certain unless you're the decision maker at Chevy. But the clues seem to indicate that they are trying to save federal tax credit for the upcoming Bolt. It's the only logical explanation for selling fewer SparkEV than Fiat 500e.

But then, GM hasn't always been logical, such as when they scrapped the EV1 program that included hybrid models after spending billions on research. Meanwhile, Toyota went on to become the world leader in auto sales, probably thanks to the positive perception brought by the Prius, which was probably crappier than EV1 hybrid.

GM has great engineering talent as shown by EV1 (many Tesla engineers previously worked on EV1), and now the SparkEV. Let's hope that GM executives have caught up to their engineering talent and do the right thing this time and not let the almost sure thing slip through their grasp again.

Tuesday, November 10, 2015

SparkEV destiny

I don't know if this is true, but this post was found in Tesla motors club forum by a user named "Red". If true, Spark was destined to be an EV, and EV is merely fulfilling its destiny. I sincerely hope that it can live on as EV for a long, long time. May you live long and prosper, SparkEV.

"GM wrong about itself

Actually GM is wrong when admitting the Spark was not designed as an EV.

It actually was, just not thoroughly, because it was one of the earliest ones. Spark is basically a reworked Daewoo Matiz, which was originally designed by Italdesign for a partnership between FIAT and the EV drive train pioneer Miro Zoric, who created the first inverters for AC motors, yes, even those in the Tesla. First for industrial use and then for automotive use. He also made first drive trains for GM's EV1, for instance... AC motors were a non existing option for cars before that. They were not controllable. In a way, due to that breakthrough, today batteries are the narrow throat, since AC drive trains are usable and efficient now.

Anyway, what was later known as Matiz, was originally intended to be FIAT's first electric car. A small but zippy city commuter. Due to administrative issues, Gianni Agnelli's desire to have each FIAT model also have an electric version by 2000, was put aside and Daewoo bought the Matiz design and GM later bought Daewoo and renamed Matiz into the Spark. So in terms of legacy, the Spark did start out as a would be EV, but not in the way most would think.

And it was originally supposed to have an AC motor and lead acid batteries, since Mr. Zoric only made rechargeable zinc air batteries (first one in the world) slightly later, in 1997. Because they would take up more space than lad acid, the little car would have lead acid batteries initially and zinc airs were used in buses and trucks. Some trivia "

Thanks, Red. Haven't seen recent posts from you in a while in Forum. I hope you're doing well.

Tuesday, November 3, 2015

Can stock Corvette beat Tesla P90DL in 0-60MPH?

With the recent news about Tesla's announcement that P85D/L usable power is not simple sum of each motor power, it got me curious why much higher power gas cars that weigh less would be slower in 0-60 mph than P85D (or P90DL). The usual argument goes like this.

1. EV has torque available from 0 RPM. But my question is why doesn't the gas car have peak torque available at 0 RPM?

2. EV (P85D/L) does not have to shift gear. But my question is why does gas car have to shift gear?

I thought these are obvious question, but after having some "debate" in forums, it seems people aren't versed in basic Physics and manual transmission to be able to understand what I'm saying. Therefore, if you haven't had basic Physics or you don't understand how manual transmission and clutch works, you should understand them before continuing with this post. It also helps if you know how to pull a wheelie on a motorcycle.

I'm going to be using Corvette Z06 as an example gas car for this post, but it could be any car with comparable performance. While Dodge Hellcat has more power, I can't stand Sergio (their CEO) for making stupid comments about EV and not improving on Fiat500e with DCFC or making better EV, so I won't be using it in this post. Sorry Hellcat; I like you, but not your CEO.

Making an EV out of gas car

Gas cars produce peak torque at specific engine RPM, and fully open throttle. Outside of these conditions, torque is far less. One only has to look at power profile curve to see this. Below is 2015 Corvette Z06 power profile.

If there's a way to keep the engine running at 3600 RPM with FULLY open throttle at 0 MPH, full torque 650 ft-lb from the engine would be available from the very start. Is there a way? Of course there is; simply FULLY open the throttle with clutch disengaged, and the engine will happily rev far past 3600 RPM. Hopefully, it has rev limiter so that the engine doesn't fall apart!

So if there's a way to make the engine speed to peak torque RPM from 0 MPH, how do we control the engine speed with FULLY open throttle so that it stays at peak torque of 3600 RPM instead of going to rev limiter? Simple: clutch slip. If the clutch is allowed to engage partially and allowed to slip, the torque at either ends of the clutch shaft must be the same to keep the engine RPM constant. Essentially, the torque at 0 MPH now becomes peak torque of the engine, not the torque at idle.

Some suggested that clutch slip results in loss of torque; if the engine RPM is kept constant, there is no loss of torque. Some suggested power is lost on clutch; yes, but power to wheels is related to wheel RPM, and clutch friction loss does not enter into the picture. Even for Tesla, instantaneous power at 0 MPH would be 0 horsepower (or watts or whatever unit of power you want to call it). This is why I asked for basic Physics as prerequisite for reading this post.

Doing this requires careful control of the clutch. Some have suggested that such fine control over clutch slip is very difficult or not possible. Or is it?

Enter the Motorcycle

Let me back up here, and give an example of this type of activity being done on daily basis by rank amateurs: motorcycle wheelie. Most motorcycles don't have enough power at low RPM to be able to pull a wheelie. The problem is worse for high peak power sports motorcycles as they have very little power in low RPM. Knowing that even 90cc motorcycles can pull a wheelie that sometimes require 2G of acceleration, something is being done to allow that from even tiny engines.

Don't try this at home, kids. The way they pull a wheelie is to rev the engine at sufficiently high RPM, and let the clutch slip to accelerate much quicker than their low RPM engine torque would allow. On high power sports bikes on the freeway speeds, they "blip" the clutch (don't do it!). Once the front wheel rises and the center of mass has shifted sufficiently, less acceleration is needed, and the clutch is allowed to fully engage. At that point, it's a matter of balancing, and very little power is needed.

It's easier to wheelie with high power sports bikes since they have higher center of mass. Doing this on 90cc motorcycle requires careful dance on the clutch lever for much longer time to allow sufficient wheel rise to take place. Indeed, I tried this many times long ago when I had one cylinder, 250cc 300 lb motorcycle that got 15HP peak power and 80 MPG (only in private parking lots with owner's permission; please don't arrest me!). Even something as wimpy as 20 lb/HP vehicle can pull 2G of acceleration, surely something with 650HP engine at 5.4 lb/HP should be able to pull 1.1G for 2.5 seconds, right?

Gears, shemears

Another problem often cited by EV folks is that gas car must change gears. That's not true. If the gas car keeps it in one gear that'll allow it to run at 60 MPH, there is no need to change gears. Corvette would run under 6000RPM at 60 MPH even in first gear. But clutch slip would be the mechanism to match engine's single RPM (peak torque RPM) to variable wheel RPM, not gear change. Looking at Corvette Z06 gearing, it seems 60 MPH would result in 4161 RPM in second gear and 3128 RPM in third gear. The gearing data is from Chevy web site.

The RPM at 60 MPH calculation is from

Let's use third gear and clutch slip to allow peak torque to take place. Second gear means more torque to the wheels, but we have to remember to fully engage the clutch and allow the engine to rev beyond 3600 RPM within the run. For driver sanity, using the third gear would let him (me!) not worry about RPM other than 3600.


There is one potential problem with this approach, and that is traction. Given that Corvette is RWD while P90DL is AWD with intelligent traction control, putting down all that power may simply spin the rear wheels. Again with basic Physics, static coefficient of friction is far higher than kinetic coefficient of friction.

For example, if Corvette did the run with third gear (1.21 * 3.42 final drive = 4.14), and 26.7 inch diameter wheel, driving force at the contact patch of the tires would be

650 ft-lb * 4.14 / (26.7 in / 12in/ft / 2)  = 2418 lb (each rear tire would be 1209 lb)

Corvette weighs 3524 lb. Knowing that Corvette can pull more than 1G, the minimum power needed to break traction would be 3524 lb. Third gear is not enough to break traction, but that's also not enough to push it quicker than P90DL mode. 2418 lb to push 3524 lb car would result in acceleration of only 0.69G, about that of P85D sports mode. Using peak torque RPM will not allow the Corvette to out accelerate P90DL. Scotty, we need more power!

May the force be with Vette

What we need is at least 3524 lb of driving force. Using first gear, 60 MPH would be 5919 RPM; let's say it's 6000 RPM. At this RPM, there would be about 600 550 ft-lb of torque read from the graph. With gearing 2.29 * 3.42 final drive = 7.83,

550 ft-lb * 7.83 / (26.7 in / 12in/ft / 2)  = 3871 lb

3871 / 3524 = 1.1G of acceleration! We are there! If we keep the engine running at 6000 RPM with fully open throttle and modulate the clutch to keep it there while accelerating, Corvette would pull 1.1G from stand-still all the way to 60 MPH. It's 10% quicker than falling.

This assumes traction holds. This is not "popping the clutch" where the torque at the wheel could exceed the traction limit via flywheel momentum change. This is controlled modulation of the clutch, and the torque at the wheels is not allowed to go beyond accelerating at 1.1G. Whether the stock tire would allow this is unknown, but with rear wheel contact patch pushed down from body torque due to acceleration, I'm pretty sure it will hold, especially considering that skid pad (maximum lateral acceleration) is rated at 1.2G.

At 60 MPH, it is 88 ft/sec. 1.1G is 35.42 ft/sec2. This would result in 0-60 MPH time of (drum roll...)

2.48 seconds!

Corvette just beat Tesla P90DL. HOLY SHEEEEEEE! Oh well, it's only in theory in some obscure blog about a "compliance" EV.

Help me, Obi-won

If this was actually carried out, the price we pay to win may make the victory hollow.

First is the question of traction. We have to get the RPM just at 6000, and no less (more would be ok). If the RPM dips any less, there will be more torque, and the rear tires could make plenty of smoke. The advantage of using the peak torque gearing (third gear) was that no matter which way you go, it's not likely to break traction due to less torque, but that wouldn't be quick enough. One can argue that using higher RPM where less torque is available while keeping the car at 1.0G (2.73sec 0-60 mph) may make it more manageable. Maybe, maybe not. That will take experiments to find out.

And those experiments will cost you dearly. For each run, you're essentially dumping lots of power into the clutch, with standstill dumping almost 600 HP to poor clutch. Assuming that clutch response with all that power (aka, heat) is manageable, it will surely reduce its life. It's not clear how long it'll last. One run? Two runs? I suppose Chevy engineers would know, but would you burn up the clutch on each 0-60 MPH run with ~$100K car? Each clutch job may cost more than 3.25 years of SparkEV lease! It's not likely that anyone would do such a thing, and such run would not be valid, would it?

Leia, I am your experiment

Each time a manual transmission car starts from stand-still, it slips clutch. Each gear change involves minute amount of clutch slip. Basically, clutch slip is normal operating mode of manual transmission cars. If that's the case, extreme clutch slip is not out of the operating mode, although it's not something you see everyday (other than motorcycle wheelies; no officer, I did not pull a wheelie!). Without quantifying what amount of clutch slip is legitimate, slipping the clutch all the way to 60MPH would be valid.

Question is, is anyone nutty enough to do it? Well, there are plenty of people nutty enough to drive a 4 door sedan with electric drive train that can take them from 0-60MPH in 2.8 seconds. Worse, there are those who dare drive a "compliance EV" and damn proud of it! I don't see why there wouldn't be some who would burn up their Corvette clutch for bragging rights.

Now where are my girlfriends, one with Chemical engineering degree who own bunch of Corvettes and another with Electrical engineering degree who own bunch of Tesla P90DL? Are you reading this? ;-)

Edit Nov 4, 2015

After sleeping over this (don't we all dream about cars and/or motorcycles in our sleep?), I think sanity check is in order. We know Corvette's 0-60mph is 2.95 sec. If we assume that clutch is fully engaged as soon as possible off the line and first gear is maintained, we can calculate how long it'll take for 0-60. Then we can compare to actual run time to see how close the calculations are.

Unfortunately, the torque data isn't available, only the graph. We can "eye-ball" some values from the graph and make rough calculations. How far off can we be? As far off as the eye ball and my biases allow!

The torque number we are after is average  torque from lowest to 6000 RPM. It's pretty hard to tell, especially only 2 numbers are available: 650 ft-lb at 3600 RPM and 533 ft-lb at 6400 RPM converted from peak horsepower number. I don't know, lets just guess 450 ft-lb. While it could be higher, torque above 600 ft-lb would result in greater than 1.2G skid pad limit in first gear and resulting tire spin, so one would have to try to stay below it. Assuming 550 ft-lb is used as peak torque, using 450 ft-lb as average would not be unreasonable.

450 ft-lb * 7.83 / (26.7 in / 12in/ft / 2)  = 3167 lb
3167 / 3524 = 0.9G
88 ft/sec / (0.9G * 32.2 ft/sec2) = 3.03 seconds

We are in the ballpark to actual experiment data with our calculations, so I feel more confident that clutch slip method could yield quicker 0-60 mph than P90DL.

Edit Nov. 5, 2015

I just can't seem to leave it alone, can I. I've been accused of being ignorant of Physics and not knowing how manual transmission and clutch works. Of course, both are wrong. It seems people who don't understand basic Physics or who haven't driven manual transmission in many years forget things. Here's a very simple way to picture the concept I present: riding the clutch.

When you were a beginner manual transmission driver, the fear you had while stopped on a hill is rolling back and hitting the car behind you. What you did (or what I did) was to partially engage the clutch while revving the engine so that the car doesn't roll back. Obviously, this wears out the clutch very quickly. What is significant is that you are accelerating via fighting gravity through clutch slip and higher power of the engine than idle can provide.

For steeper hills, you needed higher RPM. 45 degree hill would be about the maximum on most cars due to traction. Then the power needed would be? Physics homework for you! If the hill is 90 degrees and there is traction (fly paper tires?), you'd need 1G of acceleration to avoid rolling back.

The proposal I make in this post is simply "riding the clutch" taken to extreme levels. We know that it can be done at lower power levels; beginners do it all the time. The question of whether it can be done at extreme level requires (expensive) experiments. At least the calculations I present show that it's possible. Now I need to go find girlfriends with ChemE and bunch of Corvettes. BRB.

Edit: 2016-10-20

There's no question that Corvette has enough power to out accelerate Tesla P90DL with proper abuse of the clutch. But the question was if the tires can hold traction. There's an excellent video by Engineering Explained that stock tires probably won't hold traction, and limit it to about 0.8g in RWD cars like Corvette. That is probably why Corvette is rated only for 2.95 sec 0-60 MPH, and even that's very optimistic.

Video does an excellent explanation of normal force and shifting center of mass. With shorter wheelbase, peak acceleration would be much higher. That's why lateral acceleration (as in cornering) on Corvette is 1.2g and probably far larger than linear acceleration since width of the car (left to right) is much shorter than length (front to rear).

So the conclusion is that Stock Corvette will not out accelerate Tesla P90DL only due to lack of traction. But given flypaper tires of infinite traction, (if there is such a thing), appropriate clutch abuse (and destruction) would allow Corvette to out accelerate even P100DL that's rated for 0-60 MPH in 2.5 seconds.

Many arguments given by EV folks that gas cars lack "instantaneous torque" or that gear shifting make them slow are not true. Rather, gas cars are slow for most parts due to drivers' inability or unwillingness abuse the car to the extreme. Even with capable driver (aka, abuser), traction limit would not allow stock gas cars' tires on RWD configuration to be quicker than intelligent AWD Tesla.

Which begs the question, is there a tire combination that will allow properly abused Corvette to be quicker than P90DL or even P100DL? I really need to make more money if I'm to satisfy my curiosity. I should setup a "go fund me" page asking for tens of billion dollars so I could take over GM to be able to run these experiments. Anyone care to donate GM to me?