Cycle Life, Chargers, and Instrumentation

This week we resume after our Thanksgiving break. ;We have a lot of little things going on, and some recurring questions I’ve finally a minute to form some kind of cogent explanation for that won’t sound like total nonsense.

[jwplayer file=”news120310 – iPhone.m4v” hd.file=”” image=”” streamer=”rtmp://” provider=”rtmp” html5_file=” – iPhone.m4v” download_file=””]

The first had to do with cycle life. Understanding how it is calculated is the key to knowing what it IS and that knowledge in turn indicates what you should DO about it.  We received some information last June from CALB – actually some test data and I converted that to a Microsoft Excel graph to make it a bit more visual.
We also address some issues with charging and chargers surrounding how the charger determines the voltage point and where to switch from a constant current mode to a constant voltage mode.
We give a first look at the Current Ways programmable charger.
And we have installed and already since filming the video we have gotten WORKING the EVWorks Zeva2 Fuel Gage driver device.  I LOVE this little unit.  Counts AH and drives a fuel gage to give you a more or less ACCURATE indication of your battery SOC based on AH.
It ALSO provides a pulsed output for instantaneous video – which you can use to display current on your existing tachometer.
Finally, it provides an alarm output to drive an LED if you reach a certain point in DOD.  We’ll eventually use this for switching in a reistor across our throttle to provide a “limp mode”.
jack RIckard

17 thoughts on “Cycle Life, Chargers, and Instrumentation”

  1. Jack,
    Just watched your latest video. I had this exact problem with my charger design. A combination of cable resistance , inductance and noise made voltage measurements very difficult. The solution turned out to be quite simple as you surmised. Turn off the pwm for 1/25th of a second to sample the voltage then reinstate. Measured voltage now differs by less than 0.5v to actual pack voltage. I’ve used it successfully for the last 9 months on agm. This week (hopefully) will see it tested on lithium.

    Damien Maguire

  2. Jack & co:

    Just a word of encouragement on this show.

    Your signal-to-noise ratio jumped up dramatically this week. Terrific information. Even the Madiera discussion was interesting.

    If I’m representative of your viewership (OK, I may not be…) this is a better balance without the 45 extra minutes about Peak Oil, the auto industry or the adoption curve. Those are increasingly noisy conversations, since the underlying facts change so painfully little.

    Charger behavior, battery cycles, beta equipment on your workbench and upcoming product releases to talk about- that’s pure signal to me.

    Great job, boys. More of this…


  3. Thanks Tom:

    We’ve known from the beginning we would have to stumble around a bit to find the right mix. I’ve already in the first few hours heard a dozen comments in my e-mail similar to yours.

    We basically didn’t HAVE a show. We’ve been busy this week, but just a lot of little stuff on different cars. I DID get teh Zeva2 working today and it works quite well. I’m showing EMPTY at 150 AH. My limp mode from the Victron goes into effect at 162. And 180 of course would be dead. Still havent’ implemented the alarm relay in the Zeva but working on it.

    So we just covered some stuff that had been recurring in the mail box, the Zeva, talked a little bit about the Current Ways we’ve been working on with the vendor, that sort of thing. I really thought it was a MESS when we shot it. But when editing, I realized it had a little better pace and as you say “signal to noise” ratio than we’ve had before.

    I’ll work on it. It will probably take a couple of hundred more show iterations, but I think we can get there.

    Aren’t you supposed to be one of our contributing editors or something?


  4. Jack,
    Original design was done by a man named Simon Rafferty on diyelectric forum :

    I made several modifications to the hardware and software. I added isolation of the control board from the power section , input and output contactors , serial data etc. The latest software overcomes the voltage monitoring issue and several other noise related issues that saw the processor glitching.

    the heart of the system is this simple little avr development board from sparkfun :

    The charger has upto 5 stages with voltage , current or timer termination and can be reprogrammed in minutes via the little lcd and buttons or via the rs232 port.

    I have a few videos on my website from the original build. I’m currently doing a little redesign to improve the current and voltage monitoring and make it physically smaller for my motorcycle project.

    So far in over 9 months of daily use its not missed a beat.

    Damien Maguire

  5. Hi Jack !

    I have told you a couple of times that the Elcon chargers can be controlled over the CAN bus.
    I have not tested it but it is supposed to be connected to a BMS. But you can actually use an Arduino or something else to communicate to the charger, get info from it and giv it orders as to how many amps and also the max voltage.

    Here is a guy who has done that.
    Dont let the top balancing BMS boards get in the way of what he has done as of controlling the charger.

    It would of course need som microcontroller and a CANBus interface to do it. But what the heck your a handy guy. The interface specsheet can be found in the Elcon thread at DIY. Would make it very programmable dont you think.

    Best Regards
    /Per Eklund

  6. Loved the Manzanita/BMS post mortem, Jack. I was waiting for you to break into how many shooters were on the grassy knoll.

    This is a common problem with chargers and even lab instrumentation power supplies. A proper charger, in my view, should have both a voltage and a current output settings and it should strictly follow those limits. Just because you reduce your current output it doesn’t mean the charger is free to go over the set voltage. To go even further (than really needed), the charger can also have remote sensors so the charge cable impedance is not a factor by measuring at the point of load.


  7. I’m glad that my illustrious career of serving as an example for others of what not to do lives on.

    Seriously, this was a terrific show. Jack, I’m really glad you took the time to explain this to your audience, and I sincerely hope that it prevents someone else from stumbling into the same mess I did.

    Since that event, I’ve been charging the batteries to 3.45 volts and often a bit lower than that. Like you’ve described only a few of them actually reach that peak voltage, most are a tad lower, but it just doesn’t matter.

    The other day I ran 106.5 amp/hours out of my 120 amp/hour cells. I took readings on the cells after they’d rested a bit and found that they were all within 0.017 Volts of each other at around 3.240 Volts.

    I did have one that was a tad lower at 3.221 V. That’s likely the one that would go first if I drew the pack down too far. More and more I’m thinking I need to undo the fool hardy attempt at top balancing the pack and do a proper bottom balance.

    Just for the record though, I never said or thought you were wrong about living BMS free. I might have said you were eccentric, maybe even a bit mad, but not wrong.

    Take care.


  8. Jack, I don’t have a Manzanita PFC to test this but why can’t you use a voltage reading off of the output from the charger to set your max voltage? Basically do it as close as possible to where the charger senses voltage. I would think it would behave like my Zivan appears to behave. When it reaches its top voltage (69.7v for 20 cells) it won’t let the voltage go any higher. When I watch the pack voltage, however, the charger starts to cut back the current at around 68V as the voltage rises to 69.6-69.7V. It will taper back until it is pumping a whopping 3-400mA and then shut off. The only benefit I see to running voltage sense wires or reading the voltage between pulses is a slightly, 5 min or so, shorter charge time provided the charging is done to a voltage partway up the voltage curve at the end. It only takes about 10 minutes to finish charging my 200Ah pack once the current starts to ramp down and I’m charging to less than 3.5vpc. Since the current tapers down so far the voltage reading won’t be affected much by the charge current.

    FWIW, I have not seen the ending voltage from my Zivan NG1 vary by more than 0.2V and usually it ends between 69.6 and 69.7V. One time when the batteries were at ~8C it ended at 69.5V. Maybe the Zivan hits its target better than the ElCon? Unfortunately it isn’t user programmable and is only single voltage input.

  9. Yay, starting the show with an electric driving up! That’s one of the real stars of the show. I’d humbly suggest changing the opening scene from the zoom in of the garage to the scene of the mini entering the garage.

    Hmmm, long as I’m on a roll here, change the intro music to rock-and-roll, and show the mini and one of the speedsters burning off the tires and then releasing the brakes, and then going like a rich guy leaving an IRS audit!

    More cars and less wine in any case!

    Love the shows, had to think pretty hard about ways to improve them!

  10. I have no idea of wine. Near my hometown they brew the greatest Beer in the world, called “Flensburger Pilsner” 🙂
    If I got a big donor for you at my blog, I’ll send you a beer crate to Missouri 😉

    I missed the Zivan chargers at your list of chargers.
    My Beetle uses a Zivan NG3 and charge the pack unhappily to 140,5V (3.7V) although the manufacturer Set it up to 136,8 Volts (3.6V).

  11. Hi Jack,

    enjoyed you Dec 5th show a lot.

    Why can’t we just use a standard bench power supply as a charger? Nowadays they’re all current and voltage limited. Digital current/voltage readout is standard, as well as RS232 or USB remote operation. The only feature lacking: No automatic shut-down below the current treshhhold at the end of the CV. On second thought, what damage would it do to hold the battery at its top voltage indefintely?

    Rgds from Austria

  12. I use a benchtop power supply to charge individual cells – a lot. It has a max voltage of 10v but a max current of 200A. I use a high power diode on the output to keep a battery from blowing up the power supply.

    No damage from long term voltage holding. But the PS is not terribly accurate at that either. I never use it unattended.

    Jack Rickard

    1. It’s good to see you in person Jack! Having ‘spoken’ on DIYElectricCar. Very useful information on chargers in the video too.
      My home town (Horsham) in the UK, apparently has the highest consumption of Port Wine & Madeira in the country! I probably buy half of it!
      Best Regards,

        1. Dave Davenport

          I was just re-watching this and realized the linear decline presented in the 500 cycles of the SE cycle data would actually work out to 1000 cycles per 10% loss of capacity. Not sure if it stays linear after 500 cycles, but I think you worked on the theory of a 50% cycle increase for every 10%, but it is only a 1000 cycle increase for each 10% if it stays linear. Still a 9000 cycle life at 10% usage, pretty incredible. And it is absolutely amazing that this data and the graphs they give for these cells represent full 100% charge and discharge cycles. I think most people assume they mean 2000+ cycles if only discharged to 80% DOD, but I think it means 2000+ full cycles until the cell is 80% of original capacity. Now of course we are using them at a higher C-rate than .3 on discharge, but wow, I am still amazed at these cells.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Verified by MonsterInsights