The tale of two saturations...

[FMarrone at aol.com]

Inductor saturation...

An ideal inductor has infinite energy storage capacity and 
never saturates.  A inductor wound on an "air" core 
approximates the ideal inductor.  Unfortunately air has a very 
low permeability (related to inductance for a given amount of 
turns) and the current will quickly  rise to dangerously high 
levels even with low applied EMFs.  Additionally, in order to 
store the required amount of energy (were talking spark 
ignition, right) very high current levels are required in a coil 
wound on an air core.

We get around these problems by introducing some sort of 
high permeability core material, forms of iron being very 
popular.  I could explain the physics but suffice to say 
practical high perm materials DO have a limited energy storage 
capacity and now inductor saturation comes into play.

 For a coil wound on a ferrous core, the practical maximum 
energy storage occurs when the material is deep into saturation 
(in most materials the saturation point is ill defined).  Once in 
saturation the coil behaves as if it were wound on an air core 
and the current abruptly rises until it is limited by the 
resistance of the wire in the coil or by the impedance of the 
driving source (or current limit).

To know what current to limit to for reliable ignition you must 
know the energy level required for said reliability and the 
inductance of the coil (that ole E = L*I^2 again).  In the 
practical world you should probably go for the gusto and hit 
that baby with as high a current limit as the components of 
your system will take.  I suppose if you had a lot of time to 
charge the coil and some trick materials you might melt your 
plug electrodes but most of us don't have to worry about this 
problem.

Time comes into play as V = L dI/dT.  Those who are fond of 
mathematical gymnastics could combine this and the previous 
equation and gain some insight to energy storage vs. coil 
inductance and RPM.



Transistor saturation... 

Vce SAT for most bipolar power transistors range from .2 to 1 
volt.  Darlington pairs (which are very popular for coil drivers) 
range from 1 to 2 volts.  Power mosfets do not have a 
saturation voltage but rather a minimum channel resistance 
and the voltage across their drain and source leads is the 
product of the applied current and that resistance.

Saturation in a bipolar transistor was correctly identified by 
another DYIer as being a function of current gain and applied 
base current.  There is no first order effects of transistor 
saturation that has anything to do with rate of change of the 
collector current (at least not with respect to this discussion).  
For the range of coil charging currents and transistor quality 
we are likely to come up against there should be no problem 
keeping the device in saturation.  With some BIG low gain 
bipolars the base currents required do get pretty high and then 
the drivers driver dissipation can become an issue.

If I were to design an coil driver I would try and drive the base 
with enough current to keep a bipolar device in saturation over 
the normal operating range of the primary circuit.  Doing this 
will assure the minimum amount of power dissipation in the 
driver transistor and make more voltage available for charging 
the coil.  If the characteristics of the coil were such that 
saturation was unavoidable under some normal operating 
condition (usually the case) then I would use some sort of 
current limiting on the driver.  I would set this limit point 
slightly above the expected saturation point of the coil 
(assuming that this current level is reasonable to pass through 
the primary circuitry).

With the driver I designed, as the coil charges the Vce of the 
device would remain constant  (yes I know, there is a Ic 
dependence of Vce) until reaching the saturation point.  At 
saturation Vce will rise to whatever voltage is required to keep 
the primary current at the selected level.  Assuming the 
resistance of the coil and wiring is 1 ohm, the applied voltage is 
12 volts and the selected limit current is 6 amps (values 
selected for integer results, your mileage may vary) the 
transistor Vce junction will now have 6 volts across it.  Since at 
6 volts and 6 amps you are dissipating 36 watts in the 
transistor you better not be spending much time in limit AND 
you should have adequate heat sinking for the output device.

This would be a rather simplistic amplifier and much more 
sophisticated and efficient ones could be and have been 
designed.  On the other hand it would work just fine and the 
more sophisticated versions would not store more energy in 
the coil.  

I hope I haven't waffled too much.  

Frank Marrone at fmarrone at optilink.dsccc.com