Exact match. Not showing close matches.
'[OT] Need to build a high voltage spiking circuit.'
Since we are on the topic of electronic circuits....
I need to build a circuit that charges an ignition coil that
discharges at 43,000 V every 100mS. The spark gap
is less than 1/8 of an inch. Current discharge can be in the
micro amps range. This will all be run from a 120VAC source.
The coils I'm familiar with need at least 200mS to charge to full
capacity, thus I will probably need 4 of them if I go that route.
I have experience using the MC3334 type charging/discharge
circuitry controlled by a PIC, but I am thinking that I should be
able to do this discretely, thus not have to regulate down to
5V to use a pic.
Is it too much pain to design transistor multivibrators with some
power on syncronization, or is there an integrated solution available?
|On Tue, 27 Oct 1998 11:04:49 -0700 Craig Lee <ATTCANADA.NET> writes: clee
>Since we are on the topic of electronic circuits....
>I need to build a circuit that charges an ignition coil that
>discharges at 43,000 V every 100mS. The spark gap
>is less than 1/8 of an inch. Current discharge can be in the
>micro amps range. This will all be run from a 120VAC source.
If you have 120VAC, it would be easy to operate the coil in a CDI mode.
Rather than charge the coil with current, you just hit it suddenly with a
few hundred volts from a capacitor. The coil acts as a pulse transformer
instead of having to store energy, so very fast spark rates can be
In case of a 120V AC source, the CDI capacitor could be charged from a
voltage doubler. The input capacitor to the doubler should be small to
limit the charging current. When the coil is to fire, an SCR discharges
the capacitor. Here's my proposed circuit:
| | |
- \ / ( Coil
^ --- SCR ( Primary
| /| (
| | |
C1 is the voltage doubler capacitor, maybe 1 uF (200V). C2 is the CDI
capacitor of 5 uF or so (350V). Both capacitors should be film types,
not electrolytic; C1 because it may charge in reverse and C2 due to the
high pulse current. R limits the flow into C1, especially when the SCR
fires (note the voltage doulber is shorted out then, but only for an
instant* ). As for firing the SCR, a neon bulb or Diac to fire the gate
from another RC circuit would be fine if the timing doesn't have to be
precise. If it does have to be precise, you may need to build a
low-voltage precise timer. Since the gate fo the SCR is connected to
circuit common, firing it isn't difficult.
Obviously the circuit is not isolated from the power line, so it musn't
be touched or connected to external circuits while in operation. But the
1" or longer sparks flying out of the coil should remind you of that.
Install large resistors in parallel with the capacitors (not shown to
avoid cluttering the diagram) to discharge them while not in use.
In a car CDI application, the circuit is the same, but instead of a
voltage doubler, a power converter driven from the 12V supply is used to
charge the capacitor to 300-400V. I'm not sure why they don't connect
the capacitor to one end of the coil and the SCR to the other other than
that 300V would always be present on both coil terminals then.
* The trick to making this circuit work well is SCR selection. If you
use an ordinary "phase control" SCR, all the charge in C2 will discharge
with every spark. Very wasteful. A fast turn-off SCR such as is used in
TV circuits works much better. It will turn off as soon as the coil
primary starts trying to return current to the capacitor (which occurs
close to the peak secondary voltage).
>The coils I'm familiar with need at least 200mS to charge to full
>capacity, thus I will probably need 4 of them if I go that route.
They wouldn't work very well in a car if you could only get 5 sparks per
second... Are you needing sparks every 100 uS (10 KHz rate). A car coil
may have too much stray capacitance to work that fast. The proposed
circuit should give nice fat sparks up to 1KHz or so no problem.
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com/getjuno.html
or call Juno at (800) 654-JUNO [654-5866]
More... (looser matching)
- Last day of these posts
- In 1998
, 1999 only
- New search...