For starters, I considered the Hamamatsu 35w
L2173 short arc xenon lamp.
It requires a short startup pulse of 15KV, after which it will ignite and after a current surge from 50v it needs 3.5A at 11V.
I came up with this little cirquit. It is a 3Vin-50Vout current/voltage regulator with a series injection igniter. It uses an extra coil, T2, to transfer the HV spike to the lamp, this way there is not so much inductance and resistance in the power path.
I chose the LTC1871 for the low Vin and high Vout, it may be better to use a different chip, and an extra boost cirquit for the trigger, but I figured the powerful 1871@50v would save me the space of some large startup capacitance. See
Nighthunter patent for an idea of the current surge of the startup.
Here is how I thought the startup sequence would be:
LTC1871 reaches 50v, set by R2 and R3 to the fb pin. After a short delay provided by C3, R4 and R5, comparator detects that Vfb>INA146 output, and its output is grounded.
Comparator drains the gate of FET Q2 to ground, which pulls current through the ignition coil T1 primary. high voltage charges C5 (nighthunter has a resistor there, not sure what it's for?) until C5 is released through spark gap into coil T2 which sends a 15kv spike through the lamp. (in the nh the whole part between T1 secodary and T2 primary is unconnected to anything else. Is that a better idea)
Comparator also drains C4, which has the effect of lowering the gain on INA146 until C4 is recharged through R7, R8 and R9. This increases the current limit of 1871 during startup.
Comparator also drains C3 through R6. This resets the comparator until C3 is recharged through R4. The comparator hysteresis is set by R11. These components set trigger pulse width and retrig rate. There could be several trigs before C5 is charged high enough for the spark gap to close. Possibly the trig cirquit could be made with no spark gap, but then the coil T1 and Q2 would have to be sized for conducting the full energy content of the ignition spike in one go? The smallest spark gap I could High Energy Devices BX15.0, 15KV, 10.5x10.5mm + connections.
What do you think, may it work? are these worthwhile ideas?
will the components be protected from the 15kv? will the coils be large? can I use the same toroid for both coils?
