Ignition Coil Drive Circuit

[Tom Cloud]

the MSD, Crane, Mallory and other aftermarket are CDI (cap discharge
ignition) systems.  The oem systems I'm familiar with are glorified
point-type using BJT's or FET's instead of points.

the problem with raising the joltage on the primary of a stock
coil is you're gonna let the smoke out -- and it's smoke, not
electrons that conduct current (if you don't beleaf me, try
letting it out and see if the component works innimore)  ;-)

Some of the current (pun) oem coils are wound for much faster
charging rates and the transistor circuit that drives them
is basically a constant current circuit.  this prevents
the coil from overheating and driving all the little smokies
out plus allowing less dwell time ("dwell" being a relic of
the point age, actually we're talking the time required to
build the magnetic field to a desired level -- maybe even
near saturation).

Re: using zener to clip rather than cap -- the cap is reactive.
I.e. it doesn't dissipate or use any energy rather returns it
to the circuit -- that's normally good.  The zener just gets
hot -- that's generally bad.  Secondly, the cap/coil combo
is a tuned circuit which lengthens the coil discharge time.
So, I'd say, in general, you use a cap when you're tuning
the coil (and that's precisely what it does in points-type
systems -- though an artifact of that is that it slows the
primary joltage rise time and suppresses arcing over the
points), and you use a zener when you don't particularly
care (like when you're getting your spark from applying the
joltage and not from removing it -- as in point-type systems).

 **[part of the above is me blowing smoke -- i.e. conjecture.
I know the electronic principles that apply, but I won't
try to say that I know about every little trick of the trade
-- or why someone would misuse a component  ;-)   ]

To summarize what I know ..... older points-type and transistor
ignition systems used 8 to 10 volts (actually, the current)
at typically 4A through the coil.  This allowed it to work
during starting and required a ballast (or dropping) resistor
for running at 12V.  The coil had, of necessity, magnetic
field build-up times that were too long for hi-rpm operation.
One way to overcome that was to use one coil for each cylinder
(or for every two).

Then along came CDI.  It uses faster charging coils (though
earlier units just pulsed oem coils -- but not with great
success) and hits them with a high joltage pulse -- typically
300 volts.  These have more parts and cost more to build --
wonder why we don't see them on oem products ???

Then came the oem answer to "standard" xstor ignition -- the
current-controlled charging circuit.  One way to think of this
might be to say .... imagine taking a coil designed for 3 volts
at 12A and just build a 12A constant current circuit (that's
a BJT/FET with a resistor in the emitter/source ... pretty
simple/cheap).  This would charge the coil faster, as it
would put the full 12 volts (or whatever) on the 3V coil
at the beginning and then decrease the joltage to keep
the current constant.  Still not as efficient as the CDI
(lotsa heat lost in the xstors), but cheeper!

>>>MSD ignitions put out over 400 volts to the primary of the coil (really!).
>>>
>>>Nice hot spark! Make arm move very fast!
>>>
>>
>>I was thinking of something similar, but decided that I would try to focus
>>on finishing off the EFI system first. A good place to start for 12V to
>>What ever supplies are in some of the mid priced Auto audio amps that has a
>>nice PWM supply, they put out a bunch of power, and have a nice case. I
>>have seen some pretty small ones in a magazine, that did something like 58
>>volts at 8 amps. It was not very large and looked prety simple.
>>
>
>BTW, has anybody experimented with building a say 12V->24V 10A 
>dc convertor to power the coil with? Think of all that energy - arc 
>weld with your spark plugs :-) 
>
>The coil may need heatsinking though...  10A * 24V * .5 duty cycle = 
>120W - ouch!

Tom Cloud <cloud at peaches.ph.utexas.edu>