Ah, you were counting on the "ground around", it makes sense now, but then you must make sure there is some, e.g. your plate mustn't be more than two or three times as wide as the gap distance and mustn't cover the table entirely, otherwise the only ground your emitter will see is the plate, or indeed you must use a ring or something.
But anyway the slit plate experiment is better, it doesn't count on the "ground around" and it works in steady state. It's quite easy to measure a small DC current BTW, all you need is a current to voltage amplifier which will let you measure tiny currents to a virtual ground without an intervening resistor (budget deflates ;-), see http://hyperphysics.phy-astr.gsu.edu/hbase/electronic/opampvar2.html It can be made inexpensively using a cheap FET input (near infinite impedance) op amp such as the TL071. Use e.g. a 1 Meg resistor for feedback, so you'll get -1mV per entering nA. You can protect the virtual ground input against arcing by clamping it to actual ground with two antiparallel diodes. Michel ----- Original Message ----- From: "Horace Heffner" <[EMAIL PROTECTED]> To: <[email protected]> Sent: Tuesday, June 26, 2007 1:02 AM The ground is always around. It always gets some piece of the action unless it is screened. ... Like angels on the point of a pin. Sigh. If you can't accept the ground is around I can see why you think this. ... OK, in that case you need to consider the fact the field of the drop couples to the ground, inducing charge there, as well as the plate. The coupling to the plate increases as the drop approaches the plate and correspondingly decreases with the ground. As the drop falls the current increases in the plate. Similarly, the current into the forming drop varies with the rate of expansion of the surface area of the drop. When a drop breaks off that changes dramatically. An AC signal is generated from the tip. That signal capacitively couples to ground as well as to the plate. It will be seen as stronger in R2 than in R1+R3 below (gosh if R1 and R3 are used we need yet another resistor to measure the summed current. Budget creep begins. 8^) ... What prevents the ions from being attracted radially to the ground ring and thus spread? Also, what prevents the ions from flying apart due to mutual repulsion? ... > Ah, this I agree with, the landing area of the flying charges would > necessarily be a few mm wide so when crossing the border we could > have current on both plates at the same time, which could not > happen with a molecular sized filament. OK, so we agree this is a valid test for a filament. This is good. It seems to me an AC signal could be placed on the filament, if such exists, to make measuring the proportion of signal on each plate easier. Measuring DC currents is comparatively tough. The conductivity of a filament is better than for air. I'll quote some old stuff now for reference. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - "I'm using an old negative ion generator as the power supply. It puts out 10uA maximum, and maybe 10KV to 15KV." at http://amasci.com/weird/unusual/airhard.html. "- I connected a microamp meter in series with the plate. It indicated zero. When I let the other HV wire create one furrow in the mist, the meter indicated zero UA. When I brought the cable close, so there were maybe 50 to 70 furrows being drawn along the mist, the meter started flickering, indicating approx. 0.5uA. These ion- streams, if that's what they are, are each delivering an electric current in the range of 10 nanoamperes or less. Jeeze. No wonder nobody ever notices them." at http://amasci.com/weird/unusual/airexp.html. Assuming a gap of about an inch for "up close" it sounds like the resistance of the filaments is about R = V/I = (10^4 V)/(10^-8 A) = 10^12 ohm. If there were 50 of them though, as above, there would be a .5 uA signal, which would be readily detected. Might take a pre- amp to get it cleanly though. I don't think you would get a signal like that though an air gap from a fine point. Having looked at signals through a 10 m tygon tube of flowing electrolyte, and seeing their dependence on flow rate, I would expect some surprises regarding the signals that would be transmitted through such filaments, assuming they exist. Anyway, moving on ... "In pure water, sensitive equipment can detect a very slight electrical conductivity of 0.055 µS/cm at 25°C." See: http://en.wikipedia.org/wiki/Water#Electrical_conductivity This gives us an estimate of the filament cross sectional area: A = 1/((1E12 ohm/inch)*(0.055E-6 S/cm)) = 4.6E-9 m^2 and radius: r = (A/(2 Pi))^(1/2) = 2.7E-5 m and diameter: d = 2*r = 5.4E-5 m = 5.4 E-3 cm about the thickness of a human hair. So, the filament is over 10,000 water molecules wide if this is correct. If flow is moving at the top end speed Bill estimated, about 10 MPH, we get a flow rate per filament of: F = (4.6E-9 m^2)*(10 MPH) = 2E-8 m^3/s = 0.02 cm^3/sec or about 1.2 cm^3 per minute, which sounds a bit high from the description, but maybe very roughly in the ballpark. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - End old material OK, with a radius of 2.7E-5 m, or less if conductivity of the material is higher than for pure water, a really high frequency signal would be good because it will travel mostly by skin effect. Capacitively coupling the signal into the circuit above the power supply sounds like the cheapest and easiest way to go. A small cap made from some thick glass and some foil would probably work. Could use a signal generator for input, maybe one which can amplitude modulate. Run the signal into a step up transformer. It would be neat to be able to get an audio output from the detection side for experimenting. Feed it into a stereo? Maybe this is overkill. *Something* has to be done to improve the current measuring ability though. Regards, Horace Heffner
