I have been thinking about batteries for my electric Porche for a while. As weight, power and cost are all part of the equation, I think some compromizes needs to be done. I have not set “absolutely tight budget” (yet) but I’m aiming to get my conversion done with approximately half of the Nissan Leaf cost which is about 40 000 eur in Finland. So, I’m thinking that my rough budget will be 15-20 keur. I’m aiming for moderate performance and going with Soliton 1 DC controller which will give 300 kW of power if battery voltage is 300V and current 1000A. To get to that performance level, battery choice is quite critical. I would prefer to test things with smaller set of batteries first and then decide on final choises after having proper test-data. For now, I have ruled normal Winston (former Thundersky) batteries out because their power/weight-ratio is not suitable for this conversion.

One option to use is Turnigy 5S 5Ah cells. Maximum peak amperage will be limited to 1000A, so that should be used while calculating the pack. For Turnigy cells, I estimated that perhaps 15C would be good limit for peak draw. And for nominal voltage, for now I’m using 250V. I did some calculations in Excel (bold values given, other values calculated) and conclusion was that for 250 kW of power, the total weight is 131 kg for bare batteries, not including connections, casing etc.;

 

Battery voltage (nominal)		250	V
C		15	C
A		1000	A
Ah / pack		5	Ah
Cells per pack		5	pcs
Nominal voltage / cell		3,7	V
Charge voltage / cell		4,1	V
Weight / one pack		0,67	kg
Voltage / pack		18,5	V
Energy / pack		92,5	Wh
# of packs needed in parallel for current		14	packs (P)
# of packs needed in series for voltage		14	packs (S)
Cells in series (BMS-node count)		70	cells (S)
Packs total		196	pcs
Battery capacity		18130	kWh
Charge voltage		287	V
Battery weight		131,32	kg

 

So for a while, this calculation seemed to provide me guidance that I need to order about 200+ battery-packs from Hobbyking or JozzTek. Hobbyking will ship only small number of packs at once so that would require multiple shippings. And I would need to order more packs than I actually need because Hobbyking doesn’t quite test their cells for EV-usage prior shipping. But JozzTek can provide pre-tested packs and will happily ship the packs in one shipment.

All fine – except for one thing - as I have discussed with people about these packs, one concern has rised. And that is their low cycle-count. I have heard that they last only anywhere between 100-300 cycles. I’m not happy to replace whole battery in just few years!

However, I have not tested this myself yet – perhaps I will put my Powerlab 8 to continously cycle few packs and see how they last. But these cycle-counts definitely got me thinking about other options as well.

Another option could be A123 pouch cells. They cost a small fortune if bought from authorized sources (but hopefully not as much as Kokam’s…). But few other EV-builders have ordered a grey-market cells before me and have had a good experiences from them.

So I thought that A123 pouch cells could be good moderate-performance cells and made following calculations;

 

Battery voltage (nominal)		250	V
Peak C		15	C
Continous C		5	C
A		1000	A
Ah / cell		17	Ah
Cells in series per pack		1	pcs
Nominal voltage / cell		3,2	V
Charge voltage / cell		3,6	V
Weight / pack		0,5	kg
Voltage / pack		3,2	V
Capacity / pack		54,4	Wh
# of packs in parallel for current		4	packs (P)
# of packs in series for voltage		79	packs (S)
# of cells in series (BMS node count)		79	cells (S)
Cells total		316	pcs
Battery capacity		17190,4	kWh
Charge voltage		284,4	V
Battery weight		158	kg
Continous ampers		340	A
Peak power		250	kW
Continous power		85	kW

 

I estimated (hopefully concervatively) that grey-market cells have 17Ah capacity and used that as my calculation point. So they would provide about 250 kW with about 160 kg of weight. I would need 316 cells for this pack, in 79S4P configuration. It would of course be better to upgrade the voltage to as close as possible to 340V (Soliton 1 battery voltage limit). As my first actual purchase for this conversion is already done – Eltek Valere 3 kW charger capable of 3kW/10A/250-420V, that would be perfect upgrade to that as well.

That upgrade would mean about 94 cells in series and about 338 V as a charging end voltage;

 

Battery voltage (nominal)		300	V
Peak	C	15	C
Continous	C	5	C
Current	limit	1000	A
Ah / cell		17	Ah
Cells in series per pack		1	pcs
Nominal voltage / cell		3,2	V
Charge voltage / cell		3,6	V
Weight / pack		0,5	kg
Voltage / pack		3,2	V
Capacity / pack		54,4	Wh
# of packs in parallel for current		4	packs (P)
# of packs in series for voltage		94	packs (S)
# of cells in series (BMS node count)		94	cells (S)
Cells total		376	pcs
Battery capacity		20454,4	kWh
Charge voltage		338,4	V
Battery weight		188	kg
Continous ampers		340	A
Peak power		300	kW
Continous power		102	kW

 

That sounds reasonable for me. Weight still stays under 200 kg for bare batteries.

With these calculations in mind, I ordered 10 pcs set of A123 cells from Shenzen Victpower. Cells arrived fast (under 1 week delivery time) and I immediately measured open voltages of the cells – all 10 were very close each other – 3.282 – 3.283 V.

I have not yet done any capacity-testing with them but hopefully can do so in few months. It would be perfect to test them at least with 5C power-level, eg. 100A per cell. That should be plenty enough power for normal driving anyway. As the calculation above shows, it would be roughly just above 100 kW of power with 5C.

I’m guessing that it will take anywhere between 15-20 kWh of energy per 100 km. So in theory I could limit my power to 20 kW level and hopefully that will give me about 100 km/h of speed to go for 100 km (100%  DOD, not so good for batteries anyway). But this 15-20 kWh / 100 km should be treated with caution – it’s just a rough guess for now.

Anyhow, these batteries came in soft carton-box and sure enough, few topmost cells had bent a little . But as they all show good voltage, I’m quite satisfied for now. Time and proper cell-level testing will tell how much capacity they have and if they have high self discharge habits.

Few pictures:

Few more images from: http://www.flickr.com/photos/yty/sets/72157629495306153/

That’s all for now. In next post I’ll propably tell you about OpenBMS, an interesting Open Source BMS made with LTC6802(3)-2 IC’s.

Until then – take care and enjoy – it’s almost spring in Finland and sun is finally again shining  !

 

My next big project will be Electric Porsche 944!

I had a little conversion with my wife;

Me; “Hmm, Nissan Leaf seems to be over 40k eur when it’s available…

That’s a lot of money…

I bet I could do an conversion with half that money and have some fun while converting it…”

Wife: “Mhmh…”

Me: “What would be the car you would prefer driving; a Lada, a Beetle or a sports car, say for example Porsche ?”

Wife: “A sports car … A Porsche would definitely be nice …”

Me: “Ok, a Porsche it is then. Model 911 or 944, those are the models I prefer visually.”

Wife: “Ok…”

And so it was decided – we will do an EV-conversion for our own use!

As even old 911′s seem to be way over our budget we decided on 944. Partly because similar conversions already exist, and partly because I have always liked the way Porsche 911 and 944 looks. Perhaps most famous 944 conversion was  done by Tesla Motors current CTO JB Straubel. That is one nice conversion I admired already many years ago and it has been featured on many magazines and blogs. It also featured a nice pusher as a range extender. But I don’t even remotely think that I could get that kind of pusher registered with Finnish/EU laws – but it was a very nice idea anyway.

One active blog I’m currently following is electricporsche.rwaudio.com. The writer there seems to be doing a nice performance sports conversion based on 944. The blog has many good ideas and I hope to do something similar myself.

For now, I’m looking for modest performance with DC-motor and parts-list will be something along these components;

• Soliton 1 or Jr DC – controller (300 kW or 150 kW @ 300V)
• Kostov 11″, most likely 192V version.
• Clutch and transmission will most likely stay as they are.
• Battery option 1) About ~25 kWh batterypack of LFP150/160 cells. They would require much less work but would also add some unecessary mass.
• Battery option 2) About 10-15 kWh batterypack of A123 20Ah pouch cells. Generally I’m not so fond of pouch cells because they require a lot of work for pack building and are delivered in quite fragile format, needing proper mechanical shielding to be used. But they give power nicely and weight less than LFP’s.
• Battery option 3) About 15-20 kWh pack of Turnigy 5S or 6S R/C cells.
• Elithion Lithiumind BMS (or something similar, capable of reporting cell-level data to datalogger)
• Eltek Valere 3 kW CAN charger
• Possibly my own EvUI for displaying they necessary data. But as it only has CAN-bus and Soliton has Ethernet, better choise possibly would be some proper mobile PC with some small display. Or I just go with simple analog or digital meters – this remains to be determined.

One thing is for sure – this is not the final component-list – “final” component-list remains to be seen. On the other EV-projects (eCagiva, eCorolla) I have learned that component choises change and there are always comporomises to be done with performance/range/budget. With eCagiva, I started with very low performance and very low cost and ended up replacing every single electric component with better ones. I hope I don’t do that same mistake second time – this time I would prefer to choose tested and proven components to work with.

But first, we need a donor car. One suitable nice donor vechicle is on the market just now, we are trying to get more details about it. It remains to be seen if we get it or not.

Here are few of my experiences with Kokam 70 Ah pouch cells. Along the way I spent countless hours thinking about cell-clamping and other issues. Final pack turned out pretty good and has now worked for almost one year.

First and most important thing to remember:

!!! TAPE SIDES WITH KAPTON-TAPE !!!

and once again:

!!! TAPE SIDES WITH KAPTON-TAPE !!!

Belive me – they WILL leak if you do not tape them and allow them to contact other cells.

It seems that they have conductive foil left after cutting process at the factory. They will conduct electricity from sides when stacked together. And if you are using aluminum as your battery case material, disaster is ready … I ruined few 70 Ah cells and learned this the hard way.
Kapton-tape applied along the sides of the cells:

Second thing to remember -

!! KEEP THEM AT THEIR PACKING-BOXES AS LONG AS YOU CAN !!

Why ? Because they are very fragile to bending. Be extra careful when handling them. I had one cell ruined only by showing it to visitors at the workshop. Basically the cell was taken from it’s box, displayed to some visitors and put back to it’s box. And after few such “take from box, display, put back to box” events I discoverd that it had a hole in one corner. It seems that foil will very easily break if not handled extra carefully. So I recommend taking them from their box, taping them with Kapton-tape and then directly putting them to their final location at battery-box.

These unfortunate events happened while I was building 2 Kokam battery-packs for 2 different motorbikes – a Husqwarna Supermoto with about 5.2 kWh battery and a Cagiva Freccia (eCagiva) with 5.4 kWh battery. Both of them turned out fine and have quite good performance. They have Agni 95R DC-motors and Kelly DC-controllers. eCagiva has slightly more power with  KDHE-series controller capable of 650A/120V.

Final version of cell clamping had Delrin (POM) clamping system with 4 mm stainless steel threaded rods and 3 mm Aluminum connectors. System doesn’t need any holes to be drilled to the cell tabs. Disadvantage is that it takes more space from between the tabs. Here is one picture of almost last version (last version had that welded clamping replaced with 3D printed ABS-plastic part):

Here is last version, with 3D printed ABS-plastic parts replacing those welded parts:

Those ABS-parts keep those stainless steel metal-parts from rising upwards when clamping is tightened with screws. Clamping force is distributed quite evenly and clamping seems to be adequate. This method does not need any holes to cell tabs. Clamping force is generated with 4 mm stainless steel threaded rod and screws.

Here is one CAD-shot from Alibre showing basic mechanics with aluminum tabs for connecting 50-70 mm2 high-power cable;

Aluminum is 3 mm and was water-cutted. Delrin is 6/12 mm and was  cut with CNC-machine.

Each cell was taped with very thin 100 mm wide 3M double-sided tape:

Tape roll is 100 mm wide:

Here is the finished eCagiva battery-box:

Wires are extra-fine thread high-quality silicone measurement cable. One cable connects to between 2 cells connects. BMS is conected to these cables.

Here are few shots showing actual bikes;

eSuperMoto has Agni 95R at the back-fork. We did some FEA-analysis for powers involved and this 10 mm Aluminum welded motor-stand seemed to withstand those forces well. It of cource adds a bit unsprung mass to back-fork but as it’s very near to “hinge-point” of the fork it does not affect so much (hopefully)…

BMS and controller boxes visible. Controller-box is at the front, BMS-box is between 2 battery-boxes:

eSupermoto “almost ready to run”. 1 battery-box at each side. Controller and BMS at the center. CycleAnalyst as current-meter.

eCagiva at initial testings at local motorbike tuning-shop dynamometer:

First discharge tests were done with a 3-4 kW load. We replaced Agnimotor directly with DIY load-resistor and controlled discharge rate directly by twisting throttle Here is the load we used:

That’s it. Feel free to contact me if you want some more information and/or CAD-drawings of those components described above.

Few more pictures can be seen from: http://www.flickr.com/photos/yty/sets/72157628965069775/with/6735499795

Few months ago I bought an Rixe Cross XC 4.0 bicycle from a local dealer. As a test-drive, I just drove couple rounds around the block and was immadiately sold. They gave me options to choose any Rixe bike and they would do the conversion for electric. I opted for XC 4.0 because it has hydraulic disk-brakes and basic suspension in front fork. I imagined that it would be nice to have a decent bike for winter commuting.

So far I have driven about 750 km’s with the bike and have been very satisfied with it. My normal commute is about 2 x 10 km (from the city of Tampere to the city of Nokia) and in the summertime I can easily do 2 round-trips (eg. 40+ km) with one charge. But during wintertime I have noticed that range drops dramaticaly.

When I last charged from fully empty battery I had 500 km on the odometer and it was quite a bit warmer than now. From empty to full I charged about 10.484 Ah / 418.4 Wh, end voltage being 42,06V. So the capacity seems to be roughly about 10 Ah.

Measurements were done with simple Turnigy Watt-meter from Hobbyking. It seems to be perfect little device for these measurements.

Now when it’s wintertime I measured my charge after one 20 km roundtrip. It seems to eat about 7.511 Ah / 304.9 Wh after only 20 km of driving! So it seems that during wintertime the bike is using significantly more energy. It’s of course partly because higher resistance when driving through snow – even about 5 cm seems to affect a lot (as every wintercyclist already know). Sometimes roads over here get anything between 5-15 cm’s of snow.

Some capacity loss (or better, voltage drop) is generated also because lower temperatures – lithium batteries do not cope well with cold. So I made a little “blanket” that keeps the battery warm

This seems to affect range, now I get little more range and a little bit better performance – at least until the battery cools down. But as I charge during a workday in my warm workshop and keep the bike in warm place also in nights, this little blanket works fine.

Summer has gone and I found eCagiva EV-conversion to be quite fun to drive. Maximum speed seems to be around 110-120 km/h, acceleration fairly good and overall driving experience very nice. I wish I had an bike with belt-drive though – then it would be totally silent …

Rubber burns quite easily now – it takes only under 7 kW power to get it slipping on asphalt.

During the autumn I have written an new software for monitoring Elithion Lithiummate Pro BMS system. It can read, display and log all the cell level data available. The software makes it easy to monitor the system while test-driving. Of course also workshop testing, balancing and charge-monitoring is very easy with this. Cell voltages, resistances and temperatures can be seen easily. Everything is logged to .txt-file and can be later viewed and graphed with the same software.

Here are some screenshots:

First, some graphs are drawn in real-time;

Everything is available also as a table – from there user can easily lookup each individual value easily.

And finally, there is this special display for those moments that you only want to see few key values with large fonts – such as when you are test-driving those few first drives with your new EV-conversion. Then it’s nice to have few most important readings in large friendly letters in your laptop screen – spelling “Don’t panic” for you

Ah. It was only after about 20 kms of riding that old motor overheated. But I’m not complaining – I was testing the limits of the setup and certainly found them… This happened while waiting for the summer holidays to start. I got lucky and managed to order new motor from Jozz Bikes and new controller from Kelly. Both arrived very fast and so the bike has now new motor and new controller installed in place. Both are quite a bit more powerful than previous units were. Motor is Agnimotor 95R and controller is Kelly KDHE / 122V / 600A. Now the bike is quite a bit more fun to drive – maximum power from the batteries is now about 30 kW and now the bike finally accelerates quite happily

I have also found an new display for EV-use. I call it “EV UI”. It has IP67 protection, CAN bus and colors. I’m using it for all my EV-display needs and hopefully manage to deliver it to some other EV-conversions too. Here is a picture:

I’m using an mobile PLC to read temperature values from the controller and from the motor. Those values are sent to the CAN-bus and the display-device then reads and displays them. I hope this new “EV UI”- display will replace old eGUI mini and eGUI – displays totally. Certainly it’s way more robust, easier to install and easier for me to debug & program. And the price is in the same range – about 450 eur per display. Elithion Lithiumind is going to be the first BMS that I will support with this display. Others may follow.

eCagiva got new batteries and wow – the performance is bit different… Now I can quite easily keep up in traffic and bypass cars. It seems to be quite fun bike to drive now

New batteries are Kokam 70 Ah pouch-cells and the bike has 22 pcs of them. In the first few test rides the raw electric peak-power was somewhere around 14 kW when measured from the battery. A bit of dissapointment was that both motor and the controller heat up very quickly. I think there must be some sort of parameter to tweak in the motor controller (Sigmadrive/DMC 450A/80V 865L). Or I must provide additional cooling for them somehow.

Here is the first driving video with the new batteries:

I have also made an GPS-logger for testing my conversions. It logs current, voltage and GPS position & speed. All data gets written to the microSD-card as an TXT-file. I’m converting that file to the Google Earth for viewing – here’s quick screenshot of rendering the latest log in Google Earth;

EVlogger data in Google Earth

I’ll tell more about the EVLogger in another post.

Recently I have been struggling to get eCagiva-conversion ready for the summer rides. I haven’t had the time to install new batteries So that’s the main issue right now. And course all the supporting systems need to be upgraded to the new cell-count. Also now I have proper DC/DC (300 W IP67 Sevcon) and an new IP67 3.3 kW charger.

One main system needing to be upgraded is BMS. It needs 5 more BMS-boards since cell count rises from 17 to 22. With that, the voltage rises also from 54,4 V (17*3,2 V) to 81,4 V (22*3,7 V). The BMS works from about 2-4.3 volts so this is not a problem. But mechanical part is. Last summer I drove with exposed BMS-circuits, they were only protected with bikes original fairing. Not an optimum solution at all !

This time I decided to make an proper case for them. Bottom from 2 mm aluminum and “glass” from some clear polycarbonate plastic. It needs to be quite a bit thinner than last version since I don’t have much extra space after new batteries. I opted for PCB-mounting rails and custom 3D printed parts for mounting them.

I made a small video explaining my current prototyping process;

It is so easy to print usable items in plastic. Many times I need some small part for my EV-conversions and usually ABS-plastic is just fine as a material. So I sit down to my computer, desing the parts and just print them. The printing time usually is not a problem at all since I can make other things while the printer is printing.

I can warmly recommend getting or building an 3D printer if you are building an EV or just prototyping something. It’s not an cost issue anymore. You can build an working 3D printer for less than just basic EV motor controller – all the parts for 3D printer cost only about 400€ (~ $550).

One thing I’m waiting is similar DIY printing capability for metal, spesifically to aluminum.

I’m also in the process of building an vacuum plastic molder. The aim is to have plastic heated with heater and then use an vacuum pump to pull soft plastic sheet on top of the mold. When I’m learning something new, I tend to read everything I can from the subject in hand and study the basic issues surronding it. And as for the plastic molding and mold-making I have noticed that deep information can be found from quite unusual places. One great book is made actually for builders of Theater Props! The book is called “The Prop Builder’s Molding & Casting Handbook” and it’s written by Thurston James. It’s a great book and many techniques explained can be directly used in building prototypes for EV’s too. There is an short description for vacuum molding machine and I’m using these plans as my general instructions. Here are few pictures when we were building the heater for the vacuum molder. My son is 5 years and daughter is 2 years in the picture. They are very eager to help in the workshop and it’s actually fun to build things with them

If you can, read every bit of information about the subject in hand and buy the books to your workshop bookshelf. You’ll be glad that you bought them when you need just that one last piece of information / inspiration (and the workshop network is once again down)…

Quite a bit has happened since I last posted.

I have moved all my tools and things back to my own workshop from the TEVC. For one thing, I feel I get more done when I’m working on my own workshop and with familiar tools. I’m tired of moving my toolback around – it’s good to be “on the home-workshop” again

We moved also my current project, the eSupermoto, to my workshop. eSupermoto is basically an Husqwarna Supermotard converted to electric. We choosed some semi-racing parts for it and it should be pretty fun to drive. It has taken longer than anyone expeceted to get here but today we powered up the eSupermoto for the very first time! Everything went fine and motor spinned nicely We powered it only with 55 V / 1.5kW PSU (to be on the safe side of things) – but nevertheless it seemed to work fairly well. Motor spinned quite nicely, Cycleanalyst powered up and registered the current and motor controller responded to RS232 configuration very well. So all in all, we had pretty nice day. Little bit is left to be done, mostly cosmetic things and some more 3D prints to protect few components – such as the fuse-holder/case that separates those 2 battery boxes from each other. All in all, this has been an very good learning experience for us. One thing for sure – the Kokam’s are not the easiest cells to mount and handle. If you are using them, be sure to isolate the edges of each cell with some Kapton-tape – we learned this the hard way…

Anyway, after countless minor details, I’m happy to present you the first few spins of eSupermoto motor:

Today I made my first two 3D prints with my own printer – CupCake CNC. They turned out to be not supergood but hey – they are the first prints after all Hopefully I’ll be able to print my own EV parts and cases someday… It’s looking good now though.