Maker Pro
Maker Pro

Short pulse detection

J

JimW

Jan 1, 1970
0
Hi all,

I'm using a spark gap and length of coax to create short, high voltage
pulses, 2-6KV 40-300nS range (Blumlein line). Currently I ramp up the
voltage on the power supply using a potentiometer, wait until I hear
the spark then turn the power off. This system works ok but I was
hoping to automate it some how.

I had planned to use an RC circuit connected to a length of wire
wrapped around the coax (12 times) - in principal, when the pulse
occurs a voltage is induced which charges the capacitor. After the
pulse, the cap discharges over a longer time(mS) and is able to be
detected by a microprocessor etc.

So far I've been unable to find any R-C combination that gives me a
sufficiently long decay time. I tried 10uF and 4.7Kohm which should
give a time constant of 47mS – but in practice this gave me a no more
than a few uS of detectable voltage. I was wondering if any one has an
better ideas or perhaps where my idea is going wrong?

Thanks for any help,

Jim W
 
J

John Larkin

Jan 1, 1970
0
Hi all,

I'm using a spark gap and length of coax to create short, high voltage
pulses, 2-6KV 40-300nS range (Blumlein line). Currently I ramp up the
voltage on the power supply using a potentiometer, wait until I hear
the spark then turn the power off. This system works ok but I was
hoping to automate it some how.

I had planned to use an RC circuit connected to a length of wire
wrapped around the coax (12 times) - in principal, when the pulse
occurs a voltage is induced which charges the capacitor. After the
pulse, the cap discharges over a longer time(mS) and is able to be
detected by a microprocessor etc.

Is there a diode in the circuit? The pickup coil will produce a fast
ringing with no DC component, and that won't charge the capacitor.

It would be easier to separate the detection from the timing
functions. A pickup of some sort could fire a one-shot which would
time out whatever interval you need.

Or have an oscillator (or a pushbutton) set a flipflop that starts the
charging, and have a picked-off fire pulse reset it.

John
 
A

Active8

Jan 1, 1970
0
Is there a diode in the circuit? The pickup coil will produce a fast
ringing with no DC component, and that won't charge the capacitor.

It would be easier to separate the detection from the timing
functions. A pickup of some sort could fire a one-shot which would
time out whatever interval you need.

Or have an oscillator (or a pushbutton) set a flipflop that starts the
charging, and have a picked-off fire pulse reset it.

John

Does a coil wrapped that way work because of the high voltage
involved? The fields will be at right angles. I'd try something more
along the lines of wrapping some turns (or one) on the end of one
half of a clothespin and clamping that onto the line.

Maybe just a short wire parallel to the line, like a directional
coupler.
 
M

mike

Jan 1, 1970
0
JimW said:
Hi all,

I'm using a spark gap and length of coax to create short, high voltage
pulses, 2-6KV 40-300nS range (Blumlein line). Currently I ramp up the
voltage on the power supply using a potentiometer, wait until I hear
the spark then turn the power off. This system works ok but I was
hoping to automate it some how.

I had planned to use an RC circuit connected to a length of wire
wrapped around the coax (12 times) - in principal, when the pulse
occurs a voltage is induced which charges the capacitor. After the
pulse, the cap discharges over a longer time(mS) and is able to be
detected by a microprocessor etc.

So far I've been unable to find any R-C combination that gives me a
sufficiently long decay time. I tried 10uF and 4.7Kohm which should
give a time constant of 47mS – but in practice this gave me a no more
than a few uS of detectable voltage. I was wondering if any one has an
better ideas or perhaps where my idea is going wrong?

Thanks for any help,

Jim W

It's often easier to control something than to measure it.
IF you charge up a cap to just below the arc threshold, you can trigger
the arc with another smaller arc. That way, you have controlled trigger
time and controlled energy stored in the cap.
If you need a sustained arc, you can always use a timer from the trigger
to terminate it.
If you can accomplish your task without measuring anything, you're
better off.
mike


--
Return address is VALID.
Wanted, PCMCIA SCSI Card for HP m820 CDRW.
FS 500MHz Tek DSOscilloscope TDS540 Make Offer
http://nm7u.tripod.com/homepage/te.html
Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
Bunch of stuff For Sale and Wanted at the link below.
http://www.geocities.com/SiliconValley/Monitor/4710/
 
J

John Larkin

Jan 1, 1970
0
Does a coil wrapped that way work because of the high voltage
involved? The fields will be at right angles. I'd try something more
along the lines of wrapping some turns (or one) on the end of one
half of a clothespin and clamping that onto the line.

Right; the magnetic coupling is theoretically zero. But a zap like
this will get into *anything*.
Maybe just a short wire parallel to the line, like a directional
coupler.

No coupling in theory, if the currents in the center conductor and
the shield balance. In practise, probably lots.

Or maybe tap into the shield two places a few inches apart.

John
 
J

JimW

Jan 1, 1970
0
mike said:
It's often easier to control something than to measure it.
IF you charge up a cap to just below the arc threshold, you can trigger
the arc with another smaller arc. That way, you have controlled trigger
time and controlled energy stored in the cap.
If you need a sustained arc, you can always use a timer from the trigger
to terminate it.
If you can accomplish your task without measuring anything, you're
better off.
mike

Thanks guys for some very helpfull comments.

I was a bit concerned that the wire wrap wouldnt detect anything due
to the coax screen but in practice it does. With the system described
in my original post I got a very messy +/- 120V signal from the 12
turns of wire over a few hundred nS. Using a diode and a zener I
removed the negative component and hoped the zener would clap the
discharge voltage to 20V.

Ill keep trying with what ive got for ther moment, if i have no joy
ill give the Mikes idea a go.

Thanks for the help!

Jim
 
A

Active8

Jan 1, 1970
0
Right; the magnetic coupling is theoretically zero. But a zap like
this will get into *anything*.


No coupling in theory, if the currents in the center conductor and
the shield balance. In practise, probably lots.

Or maybe tap into the shield two places a few inches apart.
Definitely, except if it's Al braid, it'll be impossible to solder.

For unbalanced, I've threaded a piece of insulated 24 awg under the
braid to get coupling. It's a rig-up, but you don't need ferrites or
striplines to get a quick RF sample. That might even work here since
even though the line is balanced, all that RF noise should get some
usable standing waves going, eh?
 
M

Marc H.Popek

Jan 1, 1970
0
If you want to sample the activity of the charging bluemlein and spark gap
discharge system, you need a differential signal "b-dot" to reject the
E-field and only respond to the dI/dT (dB/dT) signal. Our group routinely
extracts 1 nS dB/dT signal from pulsed power systems that YOU DESCRIBE.

MARCO
 
Top