Saturday, October 31, 2009

This Week on the Mini Cooper

The concept of a weekly show requires us to actually do some work each week.  We had a very good week this past, the last week in October.  It rained every day, so no goofing around in the electric Speedster.

We had done a video on the EVision installation.  What we didn't show was all the wires we got in wrong.  We spent a day troubleshooting this and finally got everything hooked up correctly.  The display mounted in the air vent is absolutely gorgeous, and now gives me a way to monitor energy into and out of the pack.

This week's show is largely about DC-DC converters.  Our pack voltage is nominally 375 volts.  We charge to 392, but as soon as you remove the charger, it settles to about 375.  The Mini Cooper is an absolutely AMAZING device.  Normally, automotive manufactures add "features" that they can upsell to customers at additional cost.  The BMW Mini Cooper is quite different. It has DOZENS of "hidden" features you will never know are there.

For example, the engine control unit, termed a DME in BMW parlance, has a power management system.  A fuseblock just off the battery terminal allows it to monitor battery voltage, battery CURRENT and battery temperature.  There are several levels of ON in this car.  There is unswitched power to run the courtesy lights, remote control radio, power windows, doorlocks, etc.  Then if you put the remote in it's little dock, the CAS computer  starts waking things up.  The system has a body K CAN bus, a power train PT CAN bus, a MORE bus for the radios, etc.  And it brings up power in different "levels" or "terminals".

If the engine is running, and the battery starts to provide a lot of 12 vdc power, the DME can actually note this, and will very subtly increase engine RPM to increase alternator output.

This is just one example.  The heated seats are another.  They contain temperature monitors and when you turn on the heated seats, they don't just switch 12 vdc to some heaters.  They monitor the switch at three levels, and the temperature monitor, and power the resistive heating elements with separate pulse width modulators.

This theme is repeated throughout the car.  There are dozens of hidden items.  The air conditioning and heating has temperature sensors at two points in the car, a SOLAR sensor to detect sunlight gain, temp sensors on the heat exchanger and air conditioning evaporator, etc. etc. etc.  How much heat or air conditioning is not exactly a function of hi/med/lo on this car.  It has a separate computer just to calculate how much hot/cold/fresh air to mix.  VERY advanced.  And hard for us country bumpkins to deal with in a way.  But WHAT A CAR.

In any event, in this weeks show we look at some options for replacing the battery and alternator for this car with a pair of DC to DC converters.  We talk about the Brusa model 412.  This monster puts out 1725 watts of power (125 amps at 13.8 vdc from 375 vdc input).  Very capable.  But the falling dollar has likely put it out of reach of most at $3200.

We do describe how to combine three Kelly 125 vdc DC-DC converters to put out about 120 amps for $450.  But what we actually USE in the car is a homebrew DC-DC converter I made for less than $200 using some Vicor DC-DC converter bricks purchased on eBay for $20 each.  We have one 400 watt 12.6 vdc converter to replace the battery providing DC power all the time.  It's fanless, and so doesn't eat much of our pack energy when the car is just sitting.  But when you open the door and start doing stuff, it provides the "heartbeat" power to begin bringing up the systems.

If you hit the START/STOP button, it powers up a second fan cooled 1500 watt DC-DC converter made of the Vicor bricks.  These bricks, nominally 300 volt input, can operate quite well over the range of 180 to 380 vdc.  Three of them will put out 120 amps at 12.8 vdc.

Both DC-DC converters fit nicely into the battery compartment - saving about 20 lbs of weight.  On this car, I roughly calculate 50 pounds weight to an additional 1 mile range.  And they can never "run down".  If they can run down the 40 kW traction pack, you've let it sit a LONG time.

Sounds simple, but it's actually one of those pernicious electric car problems that never quite get solved satisfactorily.  I'm pretty happy with this one.

Note that in the video, I mentioned 90K trim resistors to get 12.8 vdc output.  These are the wrong values (different DC-DC converter).  This one uses 140K trim resistors.   I don't care if they are 1% or not.  I just use a good ohmeter and manually match these as closely as possible by picking them out of a band of 100 of them.  It is kind of important they be very close to the same value if you want the three bricks to share nicely.

Also not mentioned was any kind of input fuse.  You really should have a small 10 amp fuse on the +375 volt input.  If one of the bricks goes berserk, it will disconnect you from the pack voltage.

What else?  Well, we have started playing with chargers and charging.  Like the Brusa 412, the Brusa NLG-513 is available for Euros, which has caused the price to go up in dollars.  It's now $3,900.  That's a lot for a charger that will put out 8.5 amps at 400 volts.  We were going to use TWO of them to get 17 amps at 400 volts.  That would let us charge the 40 kW pack in about 6 hours.  An acceptable performance.  But $7800 for chargers?

We're not finding an inexpensive solution here frankly.  A 375 volt system charged to 392 volts requires a charger that will put out 400 volts or so.  That's a pretty rarified area for chargers.  And the ones that can do it, are all pretty much limited to 3.3-3.6 kW.  That implies a 12 hour charge time.  That's probably too long to really be comfortable.  If I roll in at Midnight and can't really count on a full charge until 10:00Am or worse noon, that's not optimum.

But I just haven't found a charger that will "finish" off a pack like the Brusa.  It's Constant Voltage (CV) algorithm is pretty accurate.  And the programmability lets me kind of sneak up on the final charge, finishing off very gently at low power levels.

A LOT of the charger manufacturers are very closed about their "programmable" chargers.  We had a fascinating conversation with DeltaQ, who don't do high power or high voltage anyway.  But they program "charge curves" that you can "select from."  We asked them why they just don't let us program the charge curves ourselves.  They are scared to death of being held liable for damage to batteries. Oh well....

So what we're looking at is a combination of things.  We've been wanting to play with faster charging techniques.  We really don't have a NEED to charge very quickly.  But purportedly, these cells can be charged at 3C or 300 amps.  They can certainly be charged at 1C or 100 amps which would let us charge in an hour.  So I've been planning on how to do that.  Nothing available will charge at 100 amps.

So this week we wired the car with a couple of 1 AWG short cables to the pack terminals.  These cables are terminated with Tweco welding cable quick disconnects.  These are great little devices for connecting high power cables.  The male "pin" which is about 3/8 inch in diameter, has a little cam in it.  A matching cam on the female plug allows you to insert this very large terminal pin, and twist it to lock it.  The cam forces the two faces together providing an excellent very low resistance current path.  And the connection is basically "locked".  If you twist it the other way, the thing pops apart very easily.  I love these things and vastly prefer them to the Anderson Connectors traditionally used for batteries.  I've had several "incidents" with Andersons and do NOT like to use them, although some equipment comes wired with them already and what's to do?

So the car is wired with terminals that would allow cable connections that would carry 400 vdc at 300 amps.  All I have to do is come up with that amount of power somewhere.

Long term, the obvious answer is a "mother" battery pack of 400 vdc.  This could charge all the time.  Pull the car in, connect the mother pack to the car, and it will dump a LOT of power into the pack.  That can get you going again.  Or you can then use the single Brusa to "finish charge" the pack.

An intermediate step is a large, high powered charger.  They're not cheap either, but can be kept in the garage and used on multiple vehicles.  We just received serial number 3 of Manzanita's PFC-75 charger.  They call this a 75 amp charger.  At 400 vdc it cannot deliver 75 amps.  It can DRAW 75 amps at 240 vac, purportedly.  At 400 vdc it can deliver about 38 amps dc charge.  But that's double the 17 amps we would get from TWO Brusas.

The Brusa charger is isolated.  The Manzanita is NOT isolated.  You don't want one of these feeding into the other.  But we think we can plug them BOTH in and do something kind of cool.  If we can set the voltage cutoff on the Manzanita so that it bulk charges up to a certain level and then shuts down, the Brusa can then continue to do the finish charge.  At 38 amps from the Manzanita, and 8.5 amps from the Brusa, we should be looking at 46.5 amps and a total charge time of about 2 hours.  This is also the IDEAL charge rate of 1/2 C for these cells.

In next week's show I'm going to revue the pro's and cons  of this $4400 Manzanita.  It's very, very good, and very bad at the same time.

Long term, I see a mother battery bank, kept charged by the Manzanita, and then a charge function using both the Manzanita and the battery bank in less than an hour.  I'd like to package all of this in another vintage gas pump type package with cables and so forth.  Ultimately, we would combine all of this with a higher voltage PWM controller and some meters to let you "dial in" exactly the voltage and current you want to charge at to do multiple different vehicles conveniently.

In the meantime, we don't have a drive train, and we can't drive the car.  That lets us charge once in a row, and that's not a real good test.  We had previously installed the 4kW electric water heater we will use for heat in the Mini Cooper.  This week we wired it up to the traction pack voltage and the 12 vdc control and pump supply.  The Mini just expects constant hot water from the engine and really doesn't have any control for this.  It MONITORS the temperature to help with the air mixing function, but it doesn't do anything to control it.  Worse, it's kind of an integrated air conditioning/heating system and I can't really find any function that is sufficiently analogous to use to turn on the water heater.  However I approach it, there would be times when the water heater is on and drawing 10 Ah per hour (10% of our pack capacity) when I don't need heat, or not turned on at all when I do.

So we had to go to a manual solution.  We installed a three switch control panel in the center console that simply switches 12vdc from our new fuse block in the engine compartment to 3 different systems.  The first system will be the 4 kW water heater.  The other two are spare for the moment.  I rather believe we'll have other areas where we fail to automate.

So to turn on the heat, you have to also turn on the heated water.  Flip a toggle switch.  They light up with little LED's on the end of the toggle.  And of course flip it off when you don't need heat.  I think this will all work pretty well actually.

So that gives us a LOAD.  It's a bit slow, at 10Ah per hour.  But I do now have a way of draining the traction pack now so we can play with the charging process.

Jack Rickard


  1. Mr. Rickard,

    Have you thought any more on a pwm to handle your mother battery pack? I would be very interested in how you decide to attack this problem and ultimately what you do. Any issues with the speedster and lower temps? I find my TS cells do not like the cold. They seem to sag much more. Keep up the good work both of you.

  2. Jack,
    In the Oct. 28 Friday Show, you say that you use an automotive relay to switch on the paralleled Vicors. In your Oct. 31 blog diagram, you show that relay switching the full 375V. I thought automotive relays are rated only at 30V. Are you actually using some other relay?



  3. I do use an automotive relay. It's a little heavy - a 40 amp 14 volt relay, and yes we are using it to switch 375 volts - at about 5 amps. So far, arcing hasn't been a problem. There is a very wee filter cap in the device, so there is no big inrush situation. And arcing hasn't been a problem.

    No issues with low temps. Probably a little less peppy, but we did a lot of test driving in February down around 20F and the car did much better than the driver. Of course, we had the heater on, and the cells warm up naturally from producing currrent. My experience has been the problems with cell temp are all about getting rid of excess heat, not keeping them warm.

    Yes, more thoughts on the "mother lode" but I need basically a car "controller" for the PWM that will do well over 400 volts and 300 amps. That's a little hard to find at this point. Same old problem, more power.

    We would probably have to start with a series of contactors to select output voltages by tapping in different batteries.

    Jack Rickard

  4. That's all i could come up with as well. Using a curtis that could handle the power and then a trim pot for its input. My voltage is much lower however about 157V so there is no issue finding a controller for it. I suppose using 3 and spitting the load 3 ways would work ultimately with some mucking about at high voltage but you are more of an expert at this than I am. But, how much do you want to spend to charge quickly when you really dont have a need to do so. Perhaps just for testing reasons you could hook up 3 inputs to your pack as well and charge 1/3rd the cells at a time in 20 minutes. (Or get 3 controllers with 3 inputs and do the whole pack at one time since you cant find a high power high voltage one.) That way you could prove the method get a close to fully charged pack in an hour (or 20 minutes if you go the 3 controller way) when needed. Odd i sag so much when it gets cold. I am going to have to check all my connections again.

  5. Sorry, I posted this comments to the wrong place earlier:

    Jack, to do bulk DC-to-DC charging, have you considered using a motor controller, for example a Zilla? Have a higher voltage pack connected to the Zilla's input. Connect the Zilla's output to the lower voltage pack through a large inductor. Set the Zilla's current limit and turn it on. You should be able to get lots of amps flowing that way. A little extra circuitry could shut off the Zilla when a certain voltage was reached--or even just a simple timer if you didn't want to get too fancy.


  6. Hi Jack, I know you tested a bunch of DC-DC converters before coming up with your homebrew Vicor solution. Have you come across anything, short of homebrew, that would handle a 205v nominal LiFePO4 pack? Can you run Kelly's in series?

  7. Jack, if you connecct a controller only to the part of batteries that are higher in voltage than the pile that is to be loaded,the
    controller does not have to be a high voltage type.You dont have to regulate the whole voltage.You may start at the point which
    is the lowest to be expected in the pile
    to be loaded.As everything is in a row,
    the current will stay the same in the
    whole pile except the consumtion of the controller itself.
    (i already posted this a few minutes ago at a very different place in your blog, sorry)

    greetings from germany! (Bavaria)

  8. Me once again!
    This requires knowledge about
    the controller that is used.
    As many controllers switch
    to ground because N-Channel MOSFET
    are better for high currents
    it might be necessary to connect
    both piles of Batteries at the +pole
    and to regulate at the side of the
    Its a bit tricky, but nevertheless
    it should work! but be careful!


  9. Hi Jack, i had just watched your video
    about shunt-balancers!

    You did exactly hit the point!!!!!!!

    (i didnt see it that clear before)
    last week i did some development of a
    balancer system based on a cheap
    attiny13 controller.
    (I want to use an attiny 25 later)
    and of course there is a display for
    undervoltage! Thats the main reason
    for starting the project at all.
    I was missing this basic information
    in so many systems.
    The project is far away from being ready
    at the moment, but you can read the
    actual source-code on my blog.

    The last hex-file is also there.

    Its in german language, but
    if necessary i can translate it
    for You.
    I will implement some kind of
    M-BUS (Meter-Bus) for reading out the
    actual voltage of every cell.
    A Master calls the number and the
    cell that is called will answer
    with its data.thats very easy to program.
    M-bus means, all units are in a row
    The master sends by voltage changes
    the slaves anwer by current change,
    so, only two wires are needed
    for all the informations
    and it is much easier to manage than
    a CAN-BUS.(simple RS232-protocol)
    I also implemented a balancing limit
    and an overcharge-limit
    under- and overvoltage detection use the same pin at the moment
    As you dont like the shunt-balance
    function, simply dont use it or use it for a
    pre-warning for low energy or to
    stop charging.........
    Programming these attinys is so simple.
    And they work from 1.8 to 5.5Volts
    and they have a voltage reference on board.
    You surely will find somebody who
    knows how to do it in your neighborhood.

    The item Mother-Battery is so
    interesting for me, because in the
    sawmill of my parents there is
    waterpower of 25KW and a solar-unit
    of 30kW on sunny days!
    And often all this power is not needed!
    I could get my energy there, but this
    is about 1 mile away,
    so i want to transport it in a trailer.
    (i dont have an EV at the moment)

    Thank You for those wonderful videos!