[Prev][Next][Index][Thread]

Fw: my $1.00 quasi-electrometer and more "electrostatics" experiments



    [The following text is in the "ISO-8859-1" character set]
    [Your display is set for the "US-ASCII" character set]
    [Some characters may be displayed incorrectly]




Hi all,

	Due to Richard Hull's initial experimets on the production of
excess mobile charge (in his case positive) during spark gap coil firing,
I thought I'd try some experiments of my own. It's a very facinating
subject, and I'm not sure we've really pinned the thing down completely.


	First problem: how to detect the charges. Since I don't have a
Keithly 600 series electrometer like Richard, I thought I'd build my own
"electrometer". Though it doesn't have the absolute sensitivity of the 
Keithly, I think it beats out the Keithey on the basis of a
cost/performance
basis ;-). It consists of a $0.30 neon bulb from Radio Shack, a couple of
pieces of aluminum foil and a 6"x6" polyethylene "baggie". The two pieces
of
fiol and the baggie are used to make a low leakage 400pF capacitor to 
store the accumulated charge. The value of 400pF is in no way critical.
The circuit is like this:


GND ---------------(neon bulb)--------------------------> detector input
             |                               |
             |                               |
             ---------| |---------------------
                       400 pF cap


Current flows into the input and charges the cap until the neon
bulb breaks down at about 100volts (or so). The bulb flashes at
a rate dependent on the input current. It takes Q=CVb coulumbs
of charge to make the bulb flash where Vb is the 100Volt breakdown
voltage. If we assume a current of I amps input then:

I=fCVb

Where f is the bulb flash rate. So for f=1flash/sec, C=400pF and Vb=100v,
we get a current of 40 nano amps. If one reduced the cap to 40pF and
allowed
a flash rate of 0.1 per sec, then 400pico amps could be detected. the other
advantage, of course, is that the this circuit is very tolerant of abuse.
Using
a rectified line voltage of about 150V and a 250Mohm resistor I was able to
confirm that the flash rate corresponded to the current as predicted above.

The polarity of the charge accumulated could be determined by
which of the bulb's electrodes light up. The negatively biased electrode
always 
lights up for a single polarity current. If the bulb electrode which goes
to
 the detector INPUT glows then the input current is NEGATIVE. Otherwise
the current is POSITIVE.



Since I live in an apartment where I'm likely to be thrown out by neighbors
who just don't appreciate the niceties of a good spark, I was not able to
use a tesla coil to generate my corona discharges. Instead, I used 3 other
types of supply: a 15KV neon transformer, a transistor driven flyback
oscillator, and an automobile ignition coil (none of which are particularly
loud!)

First, I just used the neon trans. with a diode on one termial to check
things
out and get some data on how well positive and negative corona currents 
flow. I initially used a 4" stainless sphere as a detector and a very thin 
sharp
wire as the source of corona (connected to the diode).

detector/wire	diode		flash/	detector 	comments
distance	polarity		sec	polarity
----------------------------------------------------------------------------

-------------
7"		wire tip (-)	2.3	(-)
8		"		1.5	"
9		"		1.3	"
9*		"		1.3	"	move colletor 2" to the right
10		"		1.0	"

10"		wire tip (+)	0	NA	no flashes seen 
..		"		"	"
..		" 		"	"	would flash with a lump of foil
7		"		"	"	added to detector sphere
6		"		1.1	(+)
5		"		1.8	(+)
4		"		3.0	(+)

Conclusions: for a given polarity of the wire, the same polarity of
charge was accumulated on the detector. Also, note how much difference
there is between the currents for (+) and (-) coronas! Negative coronas
seem to carry much more current for a given electrode potential.

By placing a large plastic sheet in front of the collector, the corona
current was stopped. If a small sheet was used it only attenuated the
current. Placing the wire tip in a 8" plastic tube and placing the sphere
on the other end also seemed to choke off the current for both polarities
though there was no frontal obstruction between the wire and sphere.

Now I removed the diode and just put a wire tip directly on the end of
the neon trans. terminal:

detector/wire	diode		flash/	detector 	comments
distance	polarity		sec	polarity
----------------------------------------------------------------------------

-------------
5"		NA		0	NA  	no flashes seen
4"		NA		0.4	(-)	same wire tip as for diode
"		NA		1.7	(-)	"rounder" wire tip	
chicken to
get closer

Though the DC component was weak, it always appeard as negative
corona current into the detector. If, however, the sphere is placed on 
the neon trans. terminal and the detector is given the wire point, then 
the detector accumulated POSITIVE charge instead. If the capacitor
element is removed so that no time integration is done, then both
of the neon lamp electrodes light up indicating that both polarities
of corona current are present, though of course, the data indicates 
that the negative current prevails.

These experiments were repeated qualitively for the flyback oscillator.
In this case too with the wire tip on the flyback output and the sphere
on the detector, the accumulated charge was NEGATIVE. And by 
exchanging the sphere and wire tip, the detected charge was
POSITIVE.

Finally, for the ignition coil, the polarity of the corona current only
depended on the polarity of the coil's input terminals as far as I 
could tell. But this is not too suprising since ignition coils are not
driven in a "balacnced" manner like the other types of transformers.


So far as my experiments were concerned, the sphere and wire tip
form a leaky diode, thus:


-------------|>|-----------------

----------O    <-----------
sphere         sharp wire tip

Perhaps someone can repeat these experiments since they don't
require much effort.

By the way, I ordered some FET input op-amps from Burr Brown for
$25 apiece. They have a typical  input bias cuurent of 0.040 pico-amp!
This is getting pretty damn close to to the Keithly, eh? The number
is OPA128LM if anyone is interested in making their own FET input
electrometer.


Take care,
-Ed Harris