Original poster: "Peter Terren" <pterren@xxxxxxxxxxxx>The circuit diagram mystery of the polarity detectng LED's is very simple. Just two 10 ohm 1 watt resistors and the two red LED's. One of the resistors is connected to the ground electrode on one end and the spark hits the other end of the resistor. So if 2000 A passes then a 10 ohm non inductive resistor should develop 20,000 V across it, which is enough to spark right across it on the outside. I have seen this happen with higher value resistors. The other 10 ohm resistor is in series with the two LED's which are in parallel but of opposite polarity. This then goes across the first 10 ohm resistor. So why should this work when theoretically the LED's should be receiving half of the 2000A. ie 1000A in each LED in each cycle. We know that we can push current limits in an IGBT to perhaps 100 times the continuous current rating, but this is 50,000 times overload (1000A vs 20mA). So what is happenning? I suspect that the actual current is a lot lower. Added to that is the higher inductance of the loop with the LED and secondary 10 ohm resistor compared with the main 10 ohm on alone.
I knew I wouldn't be able to work out real performance with out some trial and error so that's how I came up with this. I did try the resistors separately to see if they would flash over due to inductance issues but they were OK. The resistor values are bizarre though. For example you cant run the LED's off DC without putting about 4A through the 10 ohm resistor (=40W) and burning it up the resistor up instantly.
It really is bright considering the fact that it is only pulsing in several microsecond pulses for brief periods. I think something is different about the red LED's and not just the lower forward voltage compared to the blue/green LED's. What is also interesting is to use spark gap arrestors which also show the relly fast stuff without burning out. Interestingly, focally they are brighter than the arc itself and a bit more of a point source. In the end I used the LED's but certainly worth thnking of the arrestors for > 1MHz fast stuff.
Peter
Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx> Hi Peter, Cool idea!!I will "guess" two LEDs a resistor and a MOV. The strike current might be 2000A for a few nS which will kill most TVSs and such. It really helps to keep the device voltage low so the instant power is lower too. A capacitor might also work if one could guess the right value.Cheers, Terry At 06:54 AM 9/20/2006, you wrote:OK here is my fabulous "improvement". It is so simple it only took me 4 or 5 hours and a few blown high power ultrafast diodes and TVS's. In the end I settled for this. Only 4 components, two of which are red LED's. Simply, it reads the polarity of a spark. The led closest to the spark is negative ie the spark from the toroid on the left is a negative leader.They then alternate polarity far further than one can see as a spark. Guess how I did it. ie what are the two other components.Photos of the strike which is a negative leader and the alternating polarity of the ringdown are at the bottom of this section on my site.http://tesladownunder.com/HighVoltage.htm#High%20speed%20Tesla%20spark%20photography Peter