Wednesday, April 21, 2010

April 16 Friday Show

This week we digress a bit from vehicle issues and take a look at the progression of LiFePo4 cell offerings over the past two years.  We also do some full 1C discharge and charge tests on the cells.

The first cell in this series is a Thundersky TS-LFP160AH cell we purchased about 18 months ago.  This cell is a Lithium Ion Iron Phosphate cathode cell.  The spec sheet for this cell is available here.  Note that the cell measures 182 x 276.4 x 71 mm in volume and weighs 5.6kg.  It would contain 3.4v x 160AH = 544 watt hours of energy.  With shipping, this cell cost almost exactly $244.


Second cell was a Sky Energy SE-180AHA cell. You can download the spec sheet here. It is only slightly taller at 279.5 mm but is also 182x71 mm and 5.6kg, but offers 180 amp hours at the same voltage - essentially 612 watt-hours of energy. It was purchased at $241 with shipping 9 months ago.

The third cell is a brand new Thundersky TS-LFP-200AH cell. Spec sheet here. This cell features the addition of Yttrium to the cathode material and offers longer cell life cycles - 3000 cycles vice 2000 at 80% DOD and 5000 cycles vice 3000 at 70% DOD. This is a pretty significant extension of life cycle claims. It's dimensions are actually 183 x 276 x 101 mm  or 362 x 256 x 55.5 and quite a bit heavier at 7.3kg. It is also about $240 once delivered to Cape Girardeau MO.




The latest Thundersky cells are quite interesting. Most significantly, they manufacture an unadvertised  variant of this cell at the original 183 x 276 x 71 mm size and 5.6kg weight. But they can sell enough of these without advertising it that there is NO MENTION OF IT on the Thundersky web site.


Because of the quantities of these cells they are now selling, EVCOMPONENTS can get these 276x183x71 mm cells. They do test out to 200AH, though not any more than that. That's 680 watt-hours per cell. http://www.evcomponents.com.  You do have to know to ask

This is rather an amazing occurrence.  In the space of 18 months, we've gone from 160AH to 200AH in the same form factor and at the same price.  That's a 25% increase in energy density, and at least a 50% gain in cycle life.







We developed an Arduino based circuit with some LEM-HASS hall effect current sensors to measure current and AH into and out of the cells. We used our GIgavac contactor box, described in last weeks show, to switch the batteries in and out.

We used a pair of ACME 1500 watt constant current loads to load the cells to a 1C rate, 200AH, 180AH, and 160AH respectively for an hour and recorded the voltages and temperatures at each 2.5% increment by counting 5, 4.5 or 4 AH between each notation.

And then we reversed the process and used a Lambda 10v/200A power supply to do the charge curves the same way.

The result is a series of graphs.  Here is the discharge/charge curve for the original TS 160AH cell.  Click on the image for a larger version.


Here is the results of the Sky Energy 180AH cell.

Here is the results of the new Thundersky 200AH cell.


Because we measured these cells at the same percentage of their total AH capacity, we can compare them. Here is the charge curve comparison of the three cells.



We can do the same thing with the discharge curves.


The interesting area of this test is of course the temperature gain during charge and discharge. The discharge process caused the greatest temperature gain.


But there were some additional temperature gain during the recharge process.


Finally, we're making the entire Microsoft XLS spreadsheet file with all the collected data available for download here.

20 comments:

  1. Jack

    Do you have any idea how TS changed the new (unadvertised) 200Ahr cell? Trimming the weight from 7.3Kg to 5.6Kg is a significant change. There is no mention of Yttrium in the datasheet so I expect this cell is made with the "old" chemistry.

    ReplyDelete
  2. Great show Jack. I think you are correct that a lot of people are interested in these "Battery" videos. I'm also interested in your Arduino device. You are far ahead of most of us in interfacing real world in/outputs to a data display device. Any chance you could devote a show to this aspect of your work? Perhaps post some pictures and schematics of your sensors and input/output pinouts?

    ReplyDelete
  3. No the new cell is Yttrium based. I had this exact discussion with David Kois at EVComponents. I received two cells from him and the data sheet on their web site was from 2007, so I assumed I had received an older chemistry cell and I was a little miffed about it. He told me they were definitely the new chemistry cells and in fact the equivalent of the larger 200AH cell. They only make a few and they can sell all they make without any mention so that's why it's not on the THundersky site.

    EVComponents is doing sufficient quantities now that they are able to get a few of these from Thundersky. I signed up for 40 more.

    ReplyDelete
  4. I should probably document the Arduino thing, and I will. We'll post the code and schematics here surely at some point. The problem is it is very much under development, and so it's change level at the moment rather precludes release.

    It started as an AH counter to drive the old gas gage in the Porsche 356. But since a GPS module is available, might as well do mph and distance travelled, which gives you Miles to Empty pretty easily.

    It does serial communications pretty well out four ports. So I might as well capture the stream from the Curtis controller and have that data.

    There's a bluetooth module available, so I might as well add that so it connect with a computer or iPhone wireless.

    Might as well measure voltage.

    Might as well measure temperature.

    Really needs another computer to capture teh data stream and turn it into a pretty gage display on a touchscreen with the stereo.

    And I blew it up completely yesterday. Since there was so much noise from the controller, my EVision in the Mini is useless and Victor's all pouty because I've been talking badly of him. So I put the Arduino in the Mini and the noise blew it up. I didn't have enough 5v Zeners protecting the analog inputs.

    I did a segment on this six months ago - EMI in electric cars is just horrendous. Building computers to do BMS in the garage does no good because when you put it on the road, it dies from EMI. Well, it did.

    I'm waiting for some more Arduino parts, but I fear it will be the same in the Speedster part duh. It's too noisy an environment to do these things from the get go.

    Jack Rickard

    ReplyDelete
  5. Jack,

    Thanks for the info! Great stuff. To my pofane eye the Sky 180Ah cell looks like a better performer than even the latest TS chemistry. What is your general feeling between the comapnies products?

    Thanks,

    Mike

    PS couldn't open the video link. Could you put it on youtube (so I can look at it on my iPhone ;)

    ReplyDelete
  6. The TS puts a larger amount of power in a smaller package and would seem to imply longer cell life.

    From the charge/discharge curves, I agree the Sky 180 seems more stable, lower temperature, and better behaved.

    As to cathodes, the Sky Energy uses an Alees co-olivine structure, small particles made up of several hundred much smaller particles.

    The TS has added Yttrium to the cathode. This alters the LiFePo4 crystalline structure such that it intercalates slightly better giving the 20C pulse output and the 5000 cycle life.

    Too close to call really. The good news is that both are getting incrementally better at the same price and same size. If they continue to improve 25% every two years, in 127 years we will be doing right nicely with these cells. They have demonstrated sales, money, and motivation.

    I'm looking for a good Mandarin language software program.

    Jack Rickard

    ReplyDelete
  7. Jack, thanks for your continuous research. The progress in LiFePO4 is astonishing. However, there is still one big "issue" with the TS/SE format cells, and that's the non-recommended horizontal mounting position for these cells (i.e. contacts facing sideways not upwards), which render lot of conversions impossible or risky. Do you have some more information about this phenomenon or could you relate this question to your suppliers - is it still valid concern?

    ReplyDelete
  8. I'm not certain it was ever much of a concern. I have "heard" that
    some suppliers have mentioned the cells "should be" upright. I don't
    think they know precisely what will happen if they are operated on
    their sides.

    But here's the theory. The plates are "wetted" with this organic
    liquid electrolyte which actually contains the lithium salts that are
    the intermediary in all of this. It's not precisely filled with it
    like filling a bottle. When we open a cell, almost nothing leaks out
    per se. It is partially absorbed, believe it or not, by the think
    polymer insulator.

    If you have the cell upright, you might expect some settling of this
    liquid toward the bottom of the cell. How much, I don't know. But
    let's say a little. The top of each cell would be a little starved
    for electrolyte while the bottom of each plate would be a little
    oversaturated. But this is a top to bottom phenomenon and applies to
    each and every plate in the cell equally.

    If you layed them down but on edge vertically, nothing much has
    changed. Now one side of the cell is a little starved and the other a
    little oversaturated. But this applies to all foils in the cell
    equally.

    Now if you layed them down flat, you could potentially have a problem.
    Now the foils on the bottom have a greater saturation of electrolyte
    but the foils on the top are a little starved. This would be a kind
    of imbalance within the cell growing over time.

    For the time frames of car and cell operation that I see real world, I
    doubt it is very pronounced. But I don't have any way of testing it
    that I can think of that would not take years to prove.

    Jack Rickard

    ReplyDelete
  9. Thanks for the explanation, makes sense. OK, lets further this cells-laying arrangement scenario. In your mind the most precautionary
    mounting scheme would be to avoid laying them bottom flat, right? So perhaps edge vertically position, plus little incline, say just few% ?This should at least hedge against any potential spills/saturation in the upper/conectors area, which are now a bit elevated, if you can picture what I'm talking about.

    ReplyDelete
  10. CocoEV:

    I guess I don't think there IS any issue with spills/saturation in the upper connectors area.

    In truth, I don't as a practical matter see a problem with laying them flat. Do you have any information from anyone that has laid them flat and had a specific problem?

    Jack Rickard

    ReplyDelete
  11. Hello Jack! Thank you for your good work!

    >> I'm looking for a good Mandarin language software program.

    Please try this: http://www.mandarintools.com/dimsum.html

    Greating from Europe by GWL/Power Group
    http://gwl-power.tumblr.com/

    Jan Sedlak

    ReplyDelete
  12. Maybe I missed something, but why are you experimenting with a 1C charge?

    Aren't the charge port requirements for a 1C pack charge more than most homes are able to supply?

    ReplyDelete
  13. Yes, you're probably missing something.

    First, the most common voltage for an EV is probably 144. The most common cell size is 100 amps. That's 14,400 watts. Most homes today have 400 amp service at about 96,000 watts or 200 amp service at half that.

    We have a Manzanita 75 charger on the floor that can do about 15 kw.

    Manzanita has just announced a 225 amp charger - it does require 3 phase.

    You can slowly charge a stationary batter bank and charge your car from that at up to 3C.

    Aker Wade and several others are working on the concept of SAE J1772 Level III charge stations that can provide 400 volts and several hundred amperes for quick charge.

    Most people think the batteries are the problem. I can charge my car now in an hour.

    So if 1C charging appears irrelevant, you are indeed missing something.

    Jack Rickard

    ReplyDelete
  14. Jack, do you have any experience regarding new Yttrium TS cells and their temperature behaviour at cold areas? As I know the old TS cells without Yttrium needed to warm up and/or isolated for winter performance.

    Roger

    ReplyDelete
  15. No. We'll probably do something on that in the fall as winter approaches. We'll simply do a test as above, and then freeze the cells in a 0 degree freezer I have and then repeat the test. That should give us the delta.

    Interesting question.

    Jack Rickard

    ReplyDelete
  16. Actually, you might find this link quite interesting. They have done some low temperature tests on TS cells that I find quite interesting.

    http://www.sahkoautot.fi/paja:akkumittaukset

    ReplyDelete
  17. Jack, Thanks for all the info you have posted. Here's some info to add: I purchased 50 TS-LFP-180 Cells from EV Components which were delivered in Jan 2010. I didn't get to ask Dave before the meltdown about the Yttrium. Is there any way to tell by looking at the cells? The size of my 180 Ah are the same size and weight as your 200 Ah's. 183 x 276 x 71, 5.6kg. I wish I had known about the 200's! Anyway, thanks to you I purchased two extra for spares in the future. Good thing! Keep up the good work. My car's on the road now!

    Jim
    1971 Porsche 914

    ReplyDelete
  18. Jack,you could do us all a favor who are ready to take the plunge to test an Hi-Power comparable cell and integrate the results to these graphs. I posted a link to this article showing your results for all to see.

    From your testing data I've tentatively narrowed it down to Calb (apparent lowest IR & maintained output) and Hi-Power. Being accustomed to sag of epic proportions using US Battery lead, I'm so ready to hit the throttle and have it go like a gasser, not a 1400lb sled! OTOH I would like my investment to last 75K miles or better. That's my lingering question.

    Hope you can find time this winter to add the HP cell.

    YOU'RE THE MAN!

    ReplyDelete
  19. We've played around with the High Power cells and haven't found them sufficiently consistent even in very large 200AH sizes to merit further testing.

    Get a Thundersky (Now Winston) or CALB set. We can't spend time and money testing cells that we know aren't going to make the cut anyway. Besides, they are ugly in white.

    Jack Rickard

    ReplyDelete