Hi EVerybody;
Waiting for a Frederikson controller, sounds good. Hope Dan can put it
together, he makes it sound easy? He makes a few good points as I have scene
My Curtis apart at Lee Hart's and the many tiny looking diodes and stuff
look SO small to run a car. Lee pointed out that there ARE BIG diodes,
IGBITS or whatEVer it takes to handle the hi power, over the counter stuff.
The devil is the DETAILS! Protective circuits, we went over on the loooong
drive with Lee to Fla and back. Had him trapped in the car. Ha Ha! He
couldn't escape my "How to build a Controller" quaries. He sure made me
appreciate what goes into them, you just don't throw stuff in a purple/green
box and go! Taint that easy. Do-able, but you will be spending alota time
designing protective circuits as EV controller stuff isn't mainstrean
electronics like TV sets, washing machines. Ya looked at the NEW washers?
All the electronic shit on THEM! Whatthehell is wrong with a cam timer that
they have used since I was a kid, thousands of years ago?!FEH! Good folks
like Otmar have DONE all the safety and control stuff, just for you.
Yeah! I know the shit's expensive, but price out building, say a V-8
from scratch;cast the block, get the crank forged, pour yur own babbbit
bearings,cast aluminum pistons. IF ya had to converyt say EV-1's to gas, and
no cheating; hafta BUILD your own gas engine? Even a one lunger, it would be
a long time and money thing.One lung 6" piston for more torque?No Briggs and
Stratten here<g>!
Anyhow that's my input on controllers, for now. Start savin'
Bob
----- Original Message -----
From: "Dan Frederiksen" <
[email protected]>
To: "Electric Vehicle Discussion List" <
[email protected]>
Sent: Monday, August 20, 2007 2:57 AM
Subject: Re: [EVDL] EV Controllers
> cool, join the club.
> there are a lot of naysayers here. like a crowd of high salary lawyers
> hiding in the closet that comes running out to discourage anyone daring
> to think it can be done relatively easy. the french project you were
> linked is in my opinion a mess simply because the documentation lacks
> concision. certainly not of the kiss school of thought.
>
> although not a full car controller I'm much more enthusiastic about this
> 'core' design: http://zeva.com.au/speedy/ by Ian Hooper (it works
> running a little old trolley car

. he and I seem to agree more on
> design philosophy. it's based on a very easy to work with
> microcontroller and although the design doesn't have the functions
> needed for a full car controller (like current sensing so you dont fry
> stuff and motor rpm so you don't over rev it) it does have a basic
> design that should be possible to extend to a fully working simple and
> cheap design.
> the microcontroller has enough input channels to handle 'everything' so
> the circuit complexity can be kept low. perhaps a few temperature
> sensors too.
>
> I have been able to extract some experience from a sometimes unwilling
> crowd here

and I'm sure I forget some but I can mention a few insights.
> (we should probably have a site with really good controller building
> pointers)
>
> It seems that all commercial controllers use many small transistors in
> parallel instead of fewer larger ones, simple because the price is lower
> that way. the disadvantage seems to be that you have to watch for
> manufacturing differences in them so one doesn't try to take all the
> burden, then fry and take the others down with it causing a spectacular
> cascade failure (a common occurance among first timers from what I
> understand

I figure some math and some overkill might be able to
> handle that problem though.
> The curtis controller uses some 30+ of a very common mosfet BUZ30A of
> form factor TO-220
> http://www.ortodoxism.ro/datasheets/siemens/BUZ30A.pdf
> (very small, not something you would expect could provide power for a car)
>
> The zillas use a similar number of TO-247 form factor IGBTs which is
> also a small component. although several people on this list know which
> component it is they wont tell us. apparantly we don't share experience
> here. (they will give you an ocean of excuses why not. all wrong)
>
> both controllers use a similar number of 'small' capacitors in parallel
> and surprisingly they seem to be greatly underspec'ed currentwise. when
> it comes down to it the batteries carry the bulk of the ripple current
> from the PWM paradigm and apparently it works out ok anyway. the curtis
> even have caps of a particularly poor quality from what I gather

with
> maybe a 40A total ripple current rating on a 550A controller.
> Ian Hooper summed the parts on a curtis once and it has roughly 80$
> worth of power electronics in it.
>
> MOSFETs are faster but more expensive when it comes to higher voltage
> and power than IGBTs. I happen to think we have to move towards higher
> voltage because of significant advantages in efficiency, wire thickness
> etc but there is perhaps some advantage of ease in going with the lower
> curtis voltages because of batteries and typical motor designs. still a
> bit unclear on the motor theory.
>
> A picture inside a curtis: http://www.df-cad.dk/web/ev/Curtis.jpg
> and one of the Zilla2k http://www.df-cad.dk/web/ev/Zilla2kdiscovery.jpg
>
> both controller seems to switch in the 15kHz region. I will be looking
> into switching at a much lower frequency because it costs energy each it
> switches and that is heat inside the transistors which you don't need.
> rough estimate that about half the heatloss in an IGBT at 15kHz is from
> the switching and the loss is linear with frequency. that's my
> impression so far at least.
> The current rating on power transistors are usually higher than you can
> use them at. some say a factor 2. some say a factor 3. depends on
> cooling, frequency etc
>
> there is more but that's what I could think of right now
>
> oh from what I gather the current limit problem is most pronounced at 0
> rpm where the motor is largely a short circuit. especially if locked in
> place as the backemf only start to act like resistance when motor is
> turning (doing work, eating voltage and not just amps)
>
> Lenz's law helps you out a little but the inductance in these motors are
> not high enough to save you for many microseconds so the current sensor
> has to be quick to save you from frying everything

>
> cool going with the DeLorean btw

I considered it myself but they are
> not cheap and not terribly well built so I figured it would be
> expensive. but done right it could be a real crowd pleaser

I'm
> thinking chrome the body work

>
> Dan
>
>
>
>
>
>
[email protected] wrote:
>> I've heard great things about the Zilla controllers but I understand they
>> are both very costly and there is a long waiting list for them. I am an
>> electronics engineer and would consider building my own controller. Are
>> there
>> plans any where? Is there a schematic for the Zillas avaiable?
>>
>> Thank you,
>>
>> Dave Delman
>> 1981 Electric DeLorean Project
>>
>>
>>
>> ************************************** Get a sneak peek of the all-new
>> AOL at
>> http://discover.aol.com/memed/aolcom30tour
>> _______________________________________________
>> For subscription options, see
>> http://lists.sjsu.edu/mailman/listinfo/ev
>>
>>
>
> _______________________________________________
> For subscription options, see
> http://lists.sjsu.edu/mailman/listinfo/ev
>
>
> --
> Internal Virus Database is out-of-date.
> Checked by AVG Free Edition.
> Version: 7.5.476 / Virus Database: 269.11.15/949 - Release Date: 8/12/2007
> 11:03 AM
>
_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev