www.geocities.com/loudgpz/IgnMod.doc
ject is an ignition module designed to replace
the stock unit for a 1981 Kawasaki GPZ550 motorcycle.
Purpose:
The purpose of this project is to produce a low-cost
alternative to the factory replacement unit. The parts for this project
should not cost much more than $50. A replacement from Kawasaki
can cost well over $500. There are aftermarket suppliers that produce
complete ignition kits, however they often require purchasing different
spark coils. This offsets their price advantage. They also tend to use
optical or Hall-effect pickups that seem to have a slight reliability
disadvantage versus inductive pickups. These designs typically
have no means of providing a dwell angle reduction at idle and thus
the spark coils are heated and the charging system is taxed unnecessarily
at low RPM.
Overview:
This module is designed to work with the stock inductive
pickups and the stock spark coils. Aftermarket coils can be used as
long as the primary side's resistance is not less than 2.0 Ohms.
The stock GPZ550 coils are about 2.4 ohms. Any points-actuated
coils will typically be above 4 ohms and should work, however, with
less than optimal performance. No ballast resistor needs to be
used with this system.
This module will also work as a replacement for other
Kawasaki motorcycles that use the same stock unit as the GPZ550.
The GPZ550 inductive pickups can be retrofitted onto
a KZ650 motorcycle. This module would then work with the pickups
to provide an electronic ignition retrofit for the KZ650. The
original mechanical advance mechanism for the points would be retained,
but the point cam would be replaced with the iron rotor.
Because the GPZ550 has an inline, 4-cylinder engine,
it requires two spark coils, two pickups, and thus, two igniter circuits.
This replacement module (as well as the stock unit) actually contains
two separate ignition circuits acting independently. In the following
analysis, for simplicity, only one circuit will be considered.
For a complete module, the circuit need only be doubled. To save
on parts, however, it is completely acceptable to have both circuits
share the bias section.
Circuit
Analysis:
General Description:
The role of the circuit is simply to switch the spark
coil's primary current "on" and "off". The
control signal is generated by a crankshaft pickup. The pickup
gives a signal for Q1 to turn "on". Q1 then turns "on"
Q2. Q2 turns "on" Q3. Q3 allows current to flow
within the spark coil. When the pickup gives a signal for the
spark to occur, Q1 turns "off", then Q2 turns "off",
and then Q3 turns "off". When Q3 turns "off",
the spark occurs.
It is perhaps best to analyze the circuit in reverse,
starting at the output.
Selecting Q3, Z1, Z2:
To generate a spark from a spark coil, first a voltage
is applied to the primary winding of the coil. This produces a
current that increases with time until a maximum current is reached.
This current generates a magnetic field. Then the current is switched
off abruptly which collapses the magnetic field. The collapsing
magnetic field induces a high-voltage (300v to 700v) on the primary
winding and a much higher voltage (tens of thousands of volts) on the
secondary winding (a.k.a. the high tension winding). The secondary
winding produces the spark.
In this module, Q3 is switching the current to the
primary winding. Q3 is a high-voltage NPN Darlington power transistor.
It needs to be able to withstand the high voltage generated in the primary
windings of the spark coil. Z1 and Z2 help protect Q3 by clamping
the voltage to a safe level for the transistor to withstand.
Q3 is selected as an NTE2317 or ECG2317. It
is rated to handle 450v from collector to emitter. This means
it can stop primary current in the spark coil when the induced primary
voltage is less than 450v. Q3 can handle 15 amps from collector
to emitter. It has a power rating of 105 watts. It has a
minimum current gain of 40. Due to the design of the Darlington-pair
transistor, the forward bias base to emitter voltage is 2 volts.
In order to protect Q3 from induced voltages higher
than 450v, Z1 and Z2 are selected as 150v 5w Zener diodes. At
first it would appear that Z1 and Z2 are clamping the voltage at 300v.
Because Zener diodes have a gradual "knee", they are operating
at a higher voltage than their stated rating would indicate. This
is because there is a momentary spike of current that gives a momentary
high-voltage spike. The sum of the voltage ratings for Z1 and
Z2 are 300v, but they are operating at a combined voltage spike of about
450v. Use of higher voltage Zener diodes would result in damage
to the NTE2317/ ECG2317.
Selecting R5:
Since Q3 is operating as a switch, it needs to operate
in saturation. In saturation, there is the usual voltage drop
of about .3v from collector to emitter. With the operating voltage
at close to 15v, and coil resistance at 2.0 ohms, the maximum collector
current in Q3 (which is also the max. current in the spark coil) is
found by:
Q3 collector current = (15v - .3v) / 2.0 ohm = 7.35a
Q3 is rated to have a minimum gain of 40. To
ensure it is in saturation, a gain of 30 will be assumed. Therefore,
the base current in Q3 must be:
Q3 base current = 7.35a / 30 = .245a
To activate this base current, Q2 will be used as
a switch (in saturation). Therefore, it will have the typical
emitter to collector voltage drop of .3v. As described earlier,
Q3 has a base to emitter voltage drop of 2v when conducting current.
R6 will temporarily be omitted since its current is almost negligible.
Assuming the operating voltage is still 15v, R5 is selected as follows:
R5 = (15v - .3v - 2v) / .245a = 51.8 ohm
For availability purposes, R5 will be selected as
50 ohms. The power in R5 will be significant and is found as follows:
Power in R5 = (15v - .3v - 2v) X (.245a) = 3.1w
To provide a wide safety margin, R5 will be selected
as 10w.
Selecting R6:
R6 is the "pull down" resistor for Q3.
When Q2 is "off", R6 ensures the base voltage of Q3 is well
below 2v to keep Q3 in cutoff. This makes the circuit more stable
and helps eliminate unwanted oscillations. 100 ohms works nicely
for R6. Because the base of Q3 limits the voltage on R6, there
is relatively little current or power in R6.
Current in R6 = 2v / 100 ohm = .02a
Power in R6 = 2v X .02a = .04w
R6 will be selected as 1/4 w.
With the inclusion of R6 and "rounding off"
of R5, some values need adjustment.
Current in R5 = (15v - .3v - 2v) / 50 ohm = .254a
Current in R6 = 2v / 100 ohm = .020a
Q3 base current = .254a - .020a = .234a
Min. Required gain in Q3 = 7.35a / .234a = 31.4
This alters the minimum required gain for Q3, but,
since it is well below the rated minimum for an NTE2317, it is not a
problem.
Selecting Q2:
Q2 is needed to provide gain and a buffer between
Q1 and Q3. Q1 cannot directly drive Q3 because the source (s)
connection of Q1 must be referenced to ground in order to operate correctly
in this circuit. If it directly drove the base circuit of Q3,
the reference would be floating somewhere between 0v and 2v.
Q2 is a common PNP power transistor. It is
available at Radio Shack as a TIP42 power-BJT. The part number
is 276-2027. It was chosen because it is cheap and available.
Selecting R4:
When Q2 is "on", its collector current
is the same as the current in R5. It was calculated at .254a.
In order to get the most reliable switching characteristics in Q1, the
current (from drain to source) in Q1 should be minimized. At the
same time, the current must be high enough to drive Q2 into saturation.
The minimum gain in Q2 is rated at 20. To ensure saturation, the
gain will be assumed to be 11 for the purposes of selecting R4.