Transistor saturation in coil drive circuits

[Ed Lansinger]

Peter Wales wrote:

>Then the coil has full current passing through it and the coil can absorb no more
>energy >>>and the transistor saturates causing the voltage drop across the
>emitter collector to appear. Hence you know the coil is charged.<<<

I think the emphasised portion of this statement is in error, although if it is an 
observed phenomenon I have a possible explanation for it.  The error could cause 
problems for the DIYer building his/her own ignition system.  The remainder of the 
message regarding the sequence of events as a coil charges is well stated.

The statement seems to imply that there is a causal link between the current reaching 
its maximum and the saturation of the transistor, as if saturation of the transistor 
occurs only when the current stops rising.  This is not the case.  A transistor will be 
driven into saturation immediately upon application of a sufficiently high base current 
Ib >= Ic/hfe, regardless of whether the collector current is rising, falling, or 
remaining steady.

My first guess is that most ignition coil drivers will hit the transistor with a lot of 
base current from the start of charging, immediately throwing it into saturation.  Thus, 
the saturation voltage will appear at the start of charging.  My second guess is that, 
for coil drivers that do not have a current-limiting function, they will supply enough 
base current Ib >= Ic/hfe so that the transistor remains in saturation even when the 
coil current is maxed out (current limited only by the coil's resistance).  I think my 
guesses are reasonable because transistor heat dissipation is minimized if you keep it 
saturated, a good design goal.

Since saturation is not a function of the time derivative of the collector current 
(sophisticated transistor models aside) I would expect to see no sudden jump when the 
coil "stops charging".

Saturation voltage is a weak function of collector current.  For most power transistors 
used for this sort of work I would expect to see something like Vcesat about 1.7 volts 
at Ic=1A and Vcesat about 2.1 volts at Ic=10A.  So, as the coil charges Vce should 
slowly increase from something like 1.7 volts to something like 2 volts.  When the coil 
current ceases to rise you would notice that Vce ceases to rise, too.

If you do notice a sharp rise in Vce at some point, I suspect you are observing one of 
the following phenomenon:

1) The drive circuit has current limiting.  This is usually accomplished by reducing Ib, 
>de<saturating the transistor, which causes Vce to rise sharply.
2) The drive circuit provides constant current to the base, and the coil/collector 
current has built up to a point where Ib >= Ic/hfe is no longer true.  The transistor 
>de<saturates, increasing Vce.  This technique can be used for crude current limiting, 
but with the variation in hfe across different transistors of the same type it's rather 
unpredictable what your final coil current will be unless you tune the circuit for each 
individual transistor.

In either case, there is no point in holding the desaturated condition for any length of 
time.  Current will cease building and transistor heat dissipation will reach a maximum.

This jump, if observed, is no sign that the coil is "fully charged".  It is not a sign 
that the coil current has reached the peak dictated by the supply voltage, Vce, and the 
internal resistance of the coil.  It may or may not occur at a coil current sufficient 
for reliable ignition.   If observed, it is a sign that the >drive circuit< has reached 
some sort of limit and will refuse to charge the coils further.  You can probably charge 
the coils further without changing the supply voltage if you can change the drive 
circuit, something I suppose a DIYer might do...

> Increasing the battery voltage on a 12v battery is not easy.

DIYers interested in reducing the charge time on their coils (for, say, extremely high 
RPM operation) should check out the various batteries now available with both a 12V 
terminal and a 16V terminal and a built-in charger for the 16V section.  I also think 
Jacobs makes a 12V->16V DC-DC converter for just this sort of application.

-------------------------------------------------------
Ed Lansinger
General Motors Powertrain
Powertrain Control Center
Premium V Software & Calibration Group
Milford Proving Ground, Milford, MI
elansi01 at mpg.gmpt.gmeds.com  8-341-3049  (810) 684-3049
The opinions stated above are mine, not my employer's.
-------------------------------------------------------