Results matching “conversion”

We've made some progress on a more robust Roadster J1772 conversion. As part of the conversion, we want a circuit that monitors the J1772 proximity pin and cuts the pilot signal when the latch on the connector is released. With such a circuit, a Roadster will behave as a proper J1772-compatible EV and stop the current flow when the J connector's latch is opened, thus preventing any damage to the connector pins which can occur when pulling out the plug while charging.

Cathy and I worked up the basic idea together and got a bunch of help from the EV community. Cathy put in a ton of work selecting components, soliciting feedback, iterating the design, and designing the circuit board. Our solution works without drawing any power from the car, it just uses a tiny bit of power from the incoming line voltage during charging.



We just got the first set of boards back, put one together, and tested it. It works beautifully, performing even better than I had hoped. The response time from when the switch on the connector is pressed until the pilot signal is cut is about 2.2 milliseconds. When hooked up during a charge, there's no perceptible delay between when the J1772 latch is pressed and when the Roadster stops charging.

Even more geeky information is available on Cathy's page of cool details.

In other news, the cable vendor that said they could produce the replacement inlet assembly cable for us took six weeks of excuses and delays to finally say they don't want to do it. So, we're back to the drawing board on that.

Nissan has done a poor job of communicating state of charge to LEAF owners.

The first problem with this display is that you can't tell where you are with a simple glance. Quick: how many bars are there? Imagine if only some are lit up, how long does it take to count them? Once you have counted the bars, you have to divide by 12, or multiply by 8.3%. Like I want to do that while I'm driving! There's a nice number there, 93 miles, but the problem is that number varies wildly based on how you've been driving. Your state of charge might be 40% but the range estimate could be 12 miles if you just reached the top of 4,000-foot pass, or it might be 80 miles if you have been descending from that same pass. Likewise for just getting off of a stretch of 75 mph freeway versus getting onto the freeway after a stretch of 45 mph urban thoroughfare.

Drivers need to know what's in the battery unfiltered by a rating on their recent driving.

This isn't just my opinion, or the opinion of a few old school EV fanatics. I keep hearing from new LEAF owners who after a few weeks of driving realize that the estimated remaining miles on the LEAF dash is not useful. It's not that Nissan did it badly, or that it can be fixed by improving their software, it's not what EV drivers need.

Ford is coming out with the Ford Focus Electric this year and is apparently asking for opinions on what drivers want to see on the dashboard.

First off, Ford should be asking what gas car drivers want to see and putting that in their ads, but they should be asking what experienced EV drivers want to see and put that on the dash. Ford should start with dropping a line to the folks at Plug In America.

When I'm driving, I don't want to see animations or flashy graphics in my main field of view. I'm not watching a movie, I don't need special effects, and I definitely don't need running commentary on my driving. The LAST thing I want to see on the dash is any mention of gasoline. Did the Model T need a gauge showing how many bales of hay had been saved?

Please don't let some gas-driving marketing intern design the dash for an electric vehicle based on talking to other people who haven't owned an electric vehicle.

My wife and I have been driving electric for three years and have logged over 38,000 electric miles. We've done lots of local driving and enough road trips beyond our single charge range that we know what we need.

What I do want to see, in order of importance, is:

1. Speed, preferably numerical, very easy to read at a glance, the biggest number on the screen.
2. After speed, the single most important information an EV driver needs is the state of charge, SOC. This should be conveyed as remaining charge energy, in numerical resolution comparable to a mile's worth of driving, and not mangled by some unknown function of my recent driving and road conditions.
3. Instantaneous energy use. This should be graphical and clearly show whether I'm using or generating energy and how much, even when it's a small amount. Having a number would be nice, but not necessary.
   
4. Trip meter, preferably selectable from several. Having a trip meter that automatically resets after each full charge would be cool, but we still want user-controlled trip meters.
5. Estimated miles remaining based on recent driving is rarely useful, but it would probably be weird to not have it available. Most people think that will be useful until they get used to driving electric. Not having it would be a distracting omission for new owners. It can be on the dash, even on by default, but there should be a way to get rid of it, perhaps making it an alternate to an absolute remaining energy number.

The purpose of showing the state of charge isn't really about figuring out how far you can drive with the current charge. The answer to that question depends on too many factors to ever be a meaningful single number on the dash. Instead, the EV driver needs to answer two simple questions:

1) Do I have enough energy to make it to my destination?
2) If the answer to #1 is "maybe", how do I need to moderate my driving to make it?

Most of the time the answer to #1 is an unconditional "yes". An answer of "no" means it's time to find charging, a condition that should be rare if the car is being used for local driving as intended. If the answer to #1 is "maybe", then I need the best information possible to answer #2.

Note that an estimated range is always wrong when it matters because it assumes my driving style and road conditions are going to remain constant. It's basically telling me how I have been driving. I don't care about that. I need the information that will make it clear how I need to be driving for the rest of my trip.

For this reason, the choice of energy unit for the SOC display is critical. I want something more convenient than kWh, something that will not require doing math to interpret the number. If a vehicle has a certain stated nominal range, which corresponds to X Wh per mile (battery-to-wheel), then the ideal energy unit is X Wh. Tesla calls this an "ideal range mile." Call it whatever you like, but it's a very convenient unit of energy as it tells me how much is in the battery and gives me a range goal I can generally meet or even exceed if I need to.

If a car has a nominal range of 100 miles, then SOC percent corresponds to one mile of nominal driving. That's cool, but it doesn't generalize very well. When next year's model has a range of 140 miles, I don't want to have to multiply SOC percent by 1.4 to get nominal miles.

Showing SOC as kWh is even worse. Not only do I have to multiply by some goofy factor, it's a different factor for every car depending on weight and aerodynamics. Showing kWh used as part of a trip meter is awesome, and showing SOC in kWh has a certain appealing geek factor, but I don't want that to be my best-resolution SOC unit.

We'll all be better off if the car companies start showing SOC as nominal miles now.


On the Roadster, an "ideal range mile" is the amount of energy needed to drive one mile on the combined EPA driving cycle and corresponds to driving level highway at about 57 mph in moderate weather. Knowing this number and my miles to destination tells me how I need to drive to make it. This number slowly ticks down as I drive (occasionally ticking up on a long downhill drive), it doesn't fluctuate wildly as I go up and down shallow slopes and small hills. Nominal miles yields a much more reliable idea of remaining charge than an estimated-miles number can.

Having this number enables useful discussions about range and energy use among owners. If someone is planning a trip over the pass from Bellevue to Ellensburg, I can say that I've done that several times: traveling the ~100 miles over the 3,000-foot pass at 60 mph in moderate weather used 113 ideal miles and closer to the 70 mph speed limit used 119. It also makes planning for elevation possible. Every 1,000 feet of climbing uses up about 7 miles of nominal range, and going downhill gives about half of that back. Knowing that simple approximation makes it possible for a driver to plan a trip over a mountain pass just by knowing the required distance and elevation change. If other automakers use the appropriate nominal mile energy unit, these conversations will work across different makes and models, allowing drivers to share approximate energy expectations without a lot of goofy conversion math.

That probably sounds complicated. Just remember, electric vehicles are intended for local driving within their single-charge range. Most of the time the answer to the "do I have enough charge" is "yes, of course you do." It's only for the rare long trip that figuring things out is needed. Having good state of charge information available all the time will allow new drivers to develop experience and insight from their easy local driving that will make it possible for them to figure out which longer trips are practical. It's critical to widespread electric vehicle adoption that automakers get it right.

Cathy and I, with help from Dave Denhart and many others in the Tesla and broader EV communities, have converted our 2008 Roadster and Tesla High Power Wall Connector to use the new industry standard J1772 inlet and plug. This will allow us to charge without an adapter at the tens of thousands of Level 2 charging stations that will be installed in the US by the end of 2011.

What we have is functional and completely reversible, but not ideal; we view this as a version 0.9 conversion. As there are very few J1772 charging stations currently installed, and the numbers probably won't take off until late spring or early summer, we have time to develop a better solution before it actually becomes compelling for Tesla owners to convert in significant numbers. I'm sure Tesla Motors could do a much better job of creating an integrated solution and I would prefer that to having the owner community develop a conversion solution.

We've hear rumors that Tesla is developing an adapter, but are still waiting for official word on what, if any, J1772 solution they will provide. While an adapter would give us a way to charge, we have heard from many owners who would prefer to convert their vehicles and charging equipment to the industry standard rather than leave an expensive adapter vulnerable to theft while charging.

Our effort started last summer when Cathy and I began working with Dave to figure out what it would take to build an adapter that would let a Tesla Roadster charge from the Level 2 J1772 charging stations. We discovered that SAE adopted Tesla's extension to the older J1772 communications standard, so a simple pass-through connector that converts Tesla's charge inlet to the J1772 inlet will allow charging to occur, although there is an issue, which is explained below.

Once we understood the protocol, Cathy and I built and tested a pass-through adapter. When I let the Tesla owner community know about our adapter in mid-September, I wasn't surprised to hear that lots of owners were thinking about those thousands of chargers, but I was surprised how nearly all who expressed an opinion agreed with us that the right way to do this was just to convert the Roadster to use the J1772 inlet. From what I'm hearing from new and prospective owners, it seems to me that many potential Roadster customers are put off by the Tesla plug and this is probably already becoming a barrier to sales.

In the absence of any word from Tesla Motors about a J1772 upgrade path, we've been slowly working toward doing a conversion ourselves. A few weeks ago, we finally obtained an ITT Canon 75A UL-approved inlet and plug pair from Clipper Creek. The plug cord is intended as a replacement cable for Clipper Creek's model CS-100, and carries the same power and signal wires as the TS-70 aka Tesla's High Power Wall Connector (HPWC, formerly the HPC). Clipper Creek also sells a holster for the J1772 plug that can be used to replace the holster for the Tesla plug.

With the necessary hardware in hand, we starting tackling the engineering challenges in getting the inlet mounted inside the Roadster's charge port: there's limited space to work with and the Roadster wasn't designed with the shape of the J1772 plug in mind, so getting the plug and cord to clear the body is tricky. It took a bunch of measuring, brainstorming, numerous experiments, a couple of laser-cut bracket prototypes, some Dremel work on the inlet cup, and then an adapter designed in CAD and printed on a 3D printer to get something functional.

This is what the back of the upgraded inlet port looks like. The blue piece is the mounting plate Cathy designed in CAD and we fabricated on a RepRap 3D printer at Metrix Create:Space.

Here's the work in progress just before installing the J1772 inlet and putting it all back together:

Here's the inlet mounted in the Roadster's charge port:

The ITT Canon cord plugged into the Roadster's charge port:


Charging from our HPWC, now converted to J1772.

The top of the inlet tilts back to angle the J1772 cord up. This works pretty well for the ITT Canon cord with enough clearance at the top of the port that it's easy to slide the plug in and engage the lock, easier than plugging in the Tesla connector in fact. The rubber strain relief on the cord barely rests on the body, plus our Roadster has the paint guard protection there, so I'm not worried about that minor contact damaging the paint.

It's not quite as nice with the plug and cord used by the ChargePoint Coulomb chargers, but I think it's OK for use on the occasional road trip.

In addition to the cable clearance issue, there's another concern with our v0.9 conversion strategy that has to do with the largest difference between the Tesla and J1772 communication protocols.

The Tesla plug uses four contacts: two for power, one for ground and one for the pilot signal. The pilot signal is a low-voltage communications protocol that allows the charging equipment to tell the car the maximum amperage supported and allows the car to ask for the power to be turned on and off. The pilot signal is not connected to the car until the plug is connected and the locking switch is engaged. This switch plays a second role: if the driver tries to remove the plug in the middle of a charge, sliding the switch back interrupts the pilot signal which tells the car to stop charging. This happens very quickly so that the driver cannot get the plug untwisted and removed to break the electrical contacts while current is still flowing. It's important to prevent this because doing so can cause arcing, which would damage the contacts.

Instead of interrupting the pilot signal, J1772 uses a fifth wire for this purpose. Like the Tesla plug, the locking mechanism on the J1772 plug makes the proximity connection, so that when the driver wants to remove the plug and slides the lock it interrupts the proximity connection, thus telling a J1772 car to stop charging immediately (within a tenth of a second). Unfortunately, the locking switch on the J1772 plug doesn't interrupt the pilot signal.

With our v0.9 conversion (or a simple pass-through adapter), the driver can unlock the J1772 plug without the car knowing, and then pull the plug while power is flowing. Cathy and I need to make sure we don't do that. To solve this issue, we need to design a circuit that watches the proximity pin and interrupts the pilot signal when the J1772 plug is unlocked. I don't expect this to be difficult, but we haven't done it yet.

We have already made some improvements in the design. This is version 3 of Cathy's inlet mounting plate design, which we plan to print for our next revision:

In addition to the improved mounting plate, our next steps are:

1) Hope that Tesla Motors provides an official conversion solution before it matters to most owners, thus saving us the remaining steps.

2) Design a circuit to monitor the proximity pin and disconnect the pilot signal when the J1772 plug is unlocked.

3) Test with other J1772 plugs and possibly work on a better solution for cable clearance over the body panel.

4) The 2010 and later Roadsters have the inlet cable assembly connecting to the PEM in a different location. There may also be other differences. We haven't looked into it yet and don't know if it will be more or less difficult to convert than the 2008 Roadsters.

5) Before recommending an unofficial conversion to other owners, we'll need to find out how this will impact our warranty. Tesla Motors has been cooperative with our efforts: they sold our group an inlet cable assembly so that we could do the conversion reversibly. We hope they will continue to be supportive rather than forcing us to wait until our warranties expire before being able to effortlessly access standard J1772 public charging stations.

We are about to see a mass deployment of public level 2 SAE J1772 charging stations, over 14,000 from The EV Project alone. This compares to fewer than 100 public Tesla charging stations (240V/70A High Power Connectors, aka HPCs). Over the next 12 months, I expect that the availability of level 2 J1772 chargers will totally overwhelm all other charger types.

While most of these 240V chargers will be limited to 30A or 32A, J1772 chargers capable of supplying 240V/70A are available from Clipper Creek with many other vendors also working on charging stations.

Teaming up with a number of other Tesla owners and members of the broader EV community, Cathy and I have been looking into what would be required to bring J1772 charging support to the Roadster community.

The good news is that Tesla and J1772 use the same communications protocol to establish the connection and start/stop charging. This didn't happen by accident. Tesla Motors was involved early on in the development of the J1772 spec. But the Roadster was designed before the new J1772 committee even got going, so the Tesla charging protocol was designed based on the old J1772 specification which used the Avcon connectors and limited charging to 40 amps. Tesla extended this protocol up to 70 amps, and successfully lobbied the J1772 committee to adopt this extension. Cathy and I have confirmed that the SAE J1772 JAN2010 spec exactly matches the amp limit waveforms produced by the Tesla HPC at all amperage limits from 12A to 70A.

So, the Tesla Roadster uses the same communications protocol as J1772. (Except for the button on the HPC that can be used to start charging; I don't know how that works.) The only barrier to charging a Roadster from a J1772 station is the Tesla plug. We confirmed this by building a proof-of-concept adapter and using it to charge our Roadster at a Level 2 J1772 charging station in Olympia, WA, last Friday (Sept. 10, 2010).

We'd like to thank Dave Denhart, Rich Kaethler, Chad Schwitters, Martin Eberhard, and Dave Kois for helping us with this proof-of-concept project. Thanks also to Jim Blaisdell of Charge Northwest for helping us find a level 2 charging station and getting us a ChargePoint Network card overnight. Our crude adapter is not a robust solution. As you can see it's quite bulky (since we didn't want to cut the cable to a working Tesla plug) and isn't watertight enough for general outdoor use.

When the Roadster was entering production, there was no standard J1772 plug, so Tesla had to design their own. That was a necessary step, but now that the final standard uses a different plug, I think we need to find a real solution to this incompatibility. As I see it, there are at least 4 possible solutions:

1. An upgrade to switch both the Roadster and HPC to use J1772 connectors.
2. A compact adapter that converts J1772 to the Tesla connector.
3. A new pigtail for Tesla's universal mobile connector (UMC).
4. A new pigtail that requires purchasing a re-engineered UMC.

A new pig tail for the current UMC (solution 3) isn't very appealing as the UMC is limited to 40A, cutting us off from any 70A J1772 chargers, while also requiring us to stuff a large, heavy, awkward cable into our trunks just to charge at a station that is guaranteed to have a cable that will reach our charge port. It's also not nice for those of us who have already invested in a different mobile connector, like the original MC240 or the RFMC. Solution 4 is even worse than 3 as it shares all of the problems and it would require everyone to purchase a new mobile connector.

A compact adapter (solution 2) is better in that it could support the full 70A charging and also be quite compact, little more than a J1772 receptacle and a Tesla plug. It will still be quite expensive as it requires a Tesla plug. My guess is that it would cost at least $1,200 retail, based on what Tesla charges for the MC240 and UMC. It also has the downside of being an obvious target for malicious theft when the car is left charging unattended. Nissan Leaf owners won't have to leave an expensive, unsecured device dangling from their cars when charging, why should we?

Full conversion to J1772 (solution 1) sounds radical until you see a J1772 receptacle. It's very close to the size and shape of the inlet in the Tesla charge port. Once I saw that, it required zero imagination to picture a Tesla Roadster with a J1772 receptacle in place of the proprietary Tesla charge inlet.

The downside of solution 1 is that it would also require replacing the plug on our home chargers (HPC or mobile connector). This could be done by either replacing the cable, or by using the old Tesla inlet and a J1772 cable to make a Tesla-to-J1772 converter.

The retail cost of an ITT Canon UL-certified J1772 receptacle and cable pair rated for 75A is $825 from Current EV Tech. I don't know of anyone else selling these newly-available connectors, but I do expect it to be a competitive market much larger than just Roadster owners. Even adding in reasonable labor costs, it seems to me that converting a Roadster and HPC should be near or below the cost of a J1772-to-Tesla adapter.

I have been told that Tesla Motors is investigating ways to bring J1772 support to the Roadster which may include either a compact stand alone adapter (option 2) or a J1772 pig tail for the Tesla universal mobile connector (I'm not sure if this is option 3 or 4). They are early in the process and not promising anything at this point. From what I have heard, Tesla Motors is not interested in providing a full J1772 conversion (option 1) and hasn't even committed to supporting J1772 on the Model S.

It's possible the full J1772 conversion could be done even if Tesla Motors doesn't give us an official way to do it. I expect our group will continue exploring ideas in case we have to tackle the problem ourselves.

We are several months away from having a significant number of Level 2 J1772 chargers installed in metro areas targeted by The EV Project, and even further away in other areas of the US. There's plenty of time left for both Tesla Motors and the owner community to explore possible solutions, but I believe this will soon be an important issue for every Roadster owner who wants to be able to take advantage of the soon-to-be pervasive J1772 charging infrastructure to conveniently drive beyond the Roadster's single charge range.

The Tesla Roadster offers a wide variety of charging options, from 120V/12A up to 240V/70A. Charging at higher voltage and current charges faster, but most of the time charging speed isn't an issue. If you drive a typical commute and charge at night, even the lowest power will get the car fully charged overnight. At least with the early Roadster firmware, charging at 120V was pretty inefficient because of the fixed charger overhead, but what about charging at 240V at various amperage limts? My theory was that charging at higher current is more efficient because you spend less time paying the charging overhead, but another owner challenged that assumption with the theory that higher current is less efficient because it generates more heat and thus increases the amount of energy spent keeping the battery pack cool.

Another aspect of charging is that for any given current setting, the Roadster will charge steadily at that current until it gets near the top of the charge, at which point it will start to taper off. This reduces your charge rate near the top of the pack. This aspect of charging isn't documented in the owners manual.

If I don't care about charging time, what's the best amperage for energy efficient charging? If I'm on a road trip and want to squeeze the most range out of time spent at a charging stop, how should I space my stops and how long should I charge at each one? I've collected enough data to shed some light on these questions.

Methodology

I performed a series of charges at various current levels from relatively low battery states up to a full standard mode charge. For each charge, I collected time, voltage and amperage once per minute, and state of charge once every 10 minutes. From that, I can compute energy used during every segment of the charge and the total energy used.

To track our energy use for driving, we have a dedicated electric meter for each of our EVs. To validate my energy calculations, I verified that the total energy calculated per charge matches the value computed from the meter readings.

All charging was done overnight in cool weather with a 2008 Roadster. The 16A charge was done with firmware version "3.5.17 15", all other runs were done with firmware version "3.4.17 15". The 16A charge stopped at a lower state of charge (96%, 188 IM) than I normally see (98%, 193 IM). I don't know if this is due to the lower current limit, the new firmware, or a one-time fluke.

Charge Rate	Wh per Std %	Wh per Ideal Mile
120V - 12A	807	414
120V - 16A	723	371
240V - 16A	589	306
240V - 24A	544	282
240V - 32A	527	274
240V - 40A	512	266
240V - 48A	524	272
240V - 70A	516	268

Charge Rate	Ideal Miles per Hour	Current Tapering Begins At:
		Std %	Std IM	Range %	Range IM
120V - 12A	3.3
120V - 16A	5.1
240V - 16A	13	93	179	82	205
240V - 24A	20	94	180	82	205
240V - 32A	28	93	178	82	207
240V - 40A	36	93	178	81	204
240V - 48A	42	91	174	80	201
240V - 70A	61	84	161	75	188

Let's assume I want to get the most range for time spent charging, and don't need to charge all the way to the top. From the above table we see that if I'm charging at 48A or lower, I can expect to see the charging rate start to taper off at around 80% or a bit over 200 ideal miles (range mode). If I'm lucky enough to be charging at 70A on the road, my charge rate will start dropping around 75% or 188 ideal miles. I'll keep charging above 40A until I hit that 80%/200IM mark, so if my next charging stop is only 40A, I may as well keep charging to that point.

I'm sure there's some variation from car to car, and the pack and ambient temperatures will change charging behavior, so don't plan your trip to depend on these exact values, but this is at least a rough guide.

Charging Profile Graphs

Let's start by looking at how the state of charge varies over time using different current limits at 240V. All charges are standard mode all the way up and normalized so that all the charge sessions are shown from the same starting point, around 36%.