Interesting DC "spark gap" LR circuit, with high frequency pulses

[Chef]

Hello

I attached a simple circuit here for analyzing. I'm ready to change, measure anything what are you guys suggesting to find out what's happening here. There is no OU here, I'm posting this only to get help in understanding, and find the errors in my measurements. The circuit has an inductor, one power source, and one non linear switching element, a reed tube used as a hot cathode side spark gap .




The power source is an old tube HV DC power supply. 600v maximum voltage, 300 mA maximum current. It works with charged electrolytic caps also, same pulses appear, but don't have enough capacitor bank to run long, before the voltage drops, and the pulses disappear!
The inductance is 4pcs air coil in series, each ~160mH coil with several thousands of turn. (measured with a digital multi-meter) The resistant of the coils is 989 ohm together.
The 1 ohm resistor is carbon type.
The reed tube made from ferromagnetic material, small one, it's rated maybe 400V 1A DC.  (if used as a switch)

Basically I switching the HV DC power toward  the current measurement resistor through an inductance, controlled by the spark gap. Without the reed tube, when the switch is on, the power is dissipated in the inductance impedance (which is with DC is equal to the resistant of the coil)  and the 1 ohm resistor. During the charging period of the inductor, the current rising as it should be till the steady state reached.

With the reed tube in place, I switching the switch ON, allowing current build up in the inductor, reach the steady state, and turn it OFF. The inductance react to the change in the current, increasing the voltage till the spark jumps the reed tube separated metal plates, and let's the current flow again in the same direction. If the Voltage of the source is high enough, and tuned to the right value for the gap, the spark will stay between the gaps, and the current will flow trough the 1ohm resistor till the source can supply it.

In this state, the spark across the tube metal plates is continuous, the spark act like a series resistor with around 15-30k resistant.Around 5-10mA DC current flowing in the circuit, depending where the source voltage was set( 250v to 300V). 

If the voltage increased  while the circuit running in this state, the cathode side of reed tube starts to heat up, and glow, the anode side stays "cold", and the current increasing. I can increase the voltage and current till the two seperated plates of the reed tube burn together,changing the resistant of the tubes to a short circuit state.

Here comes the interesting part:

So, just to make it clear:

- constant DC current flowing in the circuit through a continuous Spark across
the reed tube terminals, while the cathode terminal glowing, the anode is "cold"

If the voltage is slowly increased (0,1 volt increments) in this condition , there will be a point, where unknown origin pulses will appear. The pulses will come in batches, followed by the circuit steady state current condition. If the voltage level is just right, the pulses will form a continuous wave, for several cycle,and the spark disappear  and the current stops.

The strange thing is, the pulses will appear if the spark is continuous, and I don't play with the voltage level, just wait. I guessed it's because of the heat saturation of the cathode side plate, but sometimes the first pulse will appear only after 15-20 min of steady state. (low value max 15mA current flowing, spark resistant is high)

I see, the pulses appear because of the sudden drop in the resistant across the reed tube terminals.Because of the spark there, with suddenly changing electric field, and the sudden resistant drop, I'm not sure my measurements  technics are good enough in this frequency range. I not included in the drawing of my circuit the stray capacitance's and inductance's.

I attached several pictures, some photos, some from my OWON scope PC software, captured from the running logs. I have  hours of running, recorded in different conditions, with different current measuring resistors, inductance, spark gaps.

So, What's happening in this circuit? I'm interested to find out the cause of the "oscillation", possible spice simulation, suggestion to check, etc.

The Red chanel is the Scope channel 1 100V/div in all pictures, the yellow channel is scope chanel 2 10V/div in all pictures.

The switch on and off picture was made with 100ohm current measurement resistor.

[gyula]

Hello Chef,

I think the oscillations are created by the negative resistance 'component' of the spark gap, i.e. the latter has a V-I curve where current decreases while the voltage increases in a certain voltage range. So this can form an RL oscillator when you just set the working point where the negative resistance can occur.

A neon bulb also has a negative resistance range in its V-I characteristic curve, relaxation oscillators can easily be built from them. See chapter 2 of this old General Electric book:
http://www.introni.it/pdf/GE%20Glow%20Lamp%20Manual.pdf  it may help to shed further info on your oscillator too.

Gyula

[Chef]

Thank You. This document is very helpful. 

I see,I need to make more test. What cause the operating point to change, and drop to the negative resistant region. But seeing how sensitive these glow lamps are, it would be hard to find the exact cause. Even my digital scope could cause changes.

[evolvingape]

Hello

Hello.. are you a sleeper ?

[gyula]

Hello.. are you a sleeper ?

Hi evolvingape,

I would like to understand what did you mean exactly with your question?

Thanks, Gyula

[evolvingape]

What do you think I meant ?

[gyula]

What do you think I meant ?

Well, I want to hear you, this is why I asked.

[evolvingape]

Well, I want to hear you, this is why I asked.

Well, I want to hear your opinion before I tell you, which is why I asked.