proof read circuit

[Dr_Doggy]

Hi, I have a black box load that responds similar to that of a spark gap, as in sometimes it has high resistance and sometimes low resistance. I am trying to design a circuit that will auto compensate for these resistances and still provide max current.

Previously i was using my variac and a neon sign transformer, and i have been getting interesting results when manually tuning voltages, but I blew out my secondary coil on neon transformer when i tried to add a rectifier circuit to it, im not sure why it did that, but have a few guesses.
RIP neon sign transformer.

A brief description is that im going to take a MOT transformer(2kv, 2kw on output), attach it to a diode triple circuit to step it up to 6kv, then use a marx generator to ramp it up the rest of the way.. The clever part is at each stage(or every few stages) of the marx spark gaps, i am going to add some diodes so that when my load is conducting it will draw current from the lower end voltages causing ideal(more) current at the output. In the event that my load has a high resistance the marx will ramp up to the next stages.

Here is my idea of how to design and how this design would work, I would like a proof reading, and ideas to make this circuit work & safely....
(sorry for the messy tripler circuit) Also pls verify that Im rating values properly!
thnx!

 

[Tony Stewart]

This is not how to connect a load to a Marx gen.

The output is at the end , not the sides.

Discharge rise time and current is often limited to simulate lightning rise time (1us) and duration time (10us)
Energy is the discharge is 1/2 CV^2 in Joules minus load series R C losses. Typically the power resistors are non-inductive 150 W ( like shock absorbers) under 150 Ohms for big electrostatic generators that multiply up to 150KV in 5 stages.

There is a maximal number of stages often much less than 10 stages, depending on the Cap leakage and charging voltage.
All caps must be Polypropylene or better.

There are never any diodes in a Marx Generator. THe spark gap is varied to determine the output voltage in parallel often with fiberglass linked stepper motor servo.

Dust and humidity is your enemy for creapage corona and leakage current, so keep clean with low residue Isopropyl Alcohol. ( not drug store variety for rubbies)

HV insulators for marx generators are very expensive and 200kV impulse rated (BIL200) porcelain bushing insulators are only good for 50kV unless the surfaces is very very clean, and all wire access points are shielded with 150kV rated plastic sheet insulation then maybe 100kV, otherwise it will bypass due to e-field density from sharp access points and go to the nearest ground. Lightning travels the path of highest chain of capacitance not the lowest resistance and highest potential opposite polarity E-Field.

HVDC will cause slow slow discharges called Partial Discharge that can be as slow as 1 pulse in minute or few and then with increasing voltage, increases Rep Rate like a Unijunction relaxation oscillator then discharges according to the smallest series capacitance since that zone may have the highest voltage as a capacitive voltage divider, unless very leaky (Rp) You can hear PD and corona and also hear it on a SW or AM radio between stations as tick sound.

 

[Dr_Doggy]

Im seeing alot of youtube doing it, my biggest fear is can i use a triple circuit to feed the marx the way i did, I could drop the marks generator in oil to help things even though i enjoy the smell of corona around.

Also thinking about it you maybe right that i dont need the diodes, i have included them as i figure the marx generator is there specifically to ramp up the voltage when a high resistance is present, alternately when there is a low resistance i want to bypass the marx generator to provide the full 6kv 300mA provided at the 10Mohm resistor(which will be changed to a 30k load) but i do not know where the threshold is of when my box will transition, so i have added several taps into the marx to make it more dynamic.
am i wrong with this, would it be easier/help to do by just reducing 1 Mohm resistors to something more adequate?

 

[Tony Stewart]

Corona Ozone can causes headaches or worse. THat's why they redesigned photo-copiers 25 yrs ago to not arc .

It wont work that way. and the caps wont charge up as much and the resistors will burn up.

If you drop the arc pads in oil, you will generate Hydrogen, Methane, Methanol, up to acetylene which gets dissolved in the oil and becomes a safety hazard.

Bad idea.

 

[Dr_Doggy]

so just the capacitors themselves in oil? and i will leave spark gaps on surface.

why wont it work though? when load is high there will be no current in lower end diodes, i just put them there cause without them the high volt spark will feed back to Marx generator

 

[Tony Stewart]

Sparks burn molecules of oil that dissolve combustible gas in oil. For example 4% is the lower explosive limit for H2. Acetylene is higher but makes a bigger bang!
The surface should be made of tungsten or plated steel so they don't corrode. Ozone removal vacuum pipe may be necessary on a continuous basis or use outside.

The spark gap is your diode like DIAC. It will have lower ESR than your diodes any MΩ load on each cap will increase the surge currents required to overcome sag or droop voltage from the R's. 10kV across 1MΩ is 100 Watts !! Basic Ohm's Law.

All caps must be very low ESR as well as your Voltage tripler Diodes. THis limits the performance significantly and your tripler may burn out if arc rep rate is high. Professional units use something like a Vandengraph charge current source to create HV driven by a large HP motor at high RPM.

If you find a successful 100kV unit on the web, attention to details is very important.
When I was working with 100kV DC, in a transformer factory, even the paint on the transformer would give an ESD shock and steel mounted on walls 20 meters away would arc every second going tick tick tick from static discharge from the high E-Field generated by a thin wire going from the 100kV generator box using a 0-50kV with a doubler attachment box. It was an electrifying experience.

Surface charges appear where you least expect them and zap. Even on the steel safety fence around the transformer under test, from the paint being charged up.

 

[Dr_Doggy]

ok, so i should start with transformer , but what is the best way to rectify it and limit( protect it from over current draws)?

 

[Tony Stewart]

Do some more research to complete this project. Don't reinvent the wheel.

The left column are spheres that arc like a diode switch to put the the right column of plastic caps in a large cans from parallel charge to a series string . ( Marx Gen. )
The big blue box supplies the power with hefty braid also used between each stage.

I am surprised this guy doesn't have cataracts yet from the radiation.

 

[Dr_Doggy]

thanks! it seems that my numbers are way off, I only need an arc that will be able to trigger a 3inch spark gap at stp, and idk if i need certain amount of current,

but the other thing i want is when there is low resistance across my gap i want it to draw the power supplied from the 2kv,2kw transformer directly,

basically i want to use the mot, as power/current supply , but the marks generator to kick the spark gap to a conductive(ionized) state, much like a flash tube , but with out the tube , or the gas ,

i put all the extra diodes cause both parts(mot dc out & marx) will connect to spark gap at the same point and i worry that the marks will feedback to the transformer and fry it.....

 

[Dr_Doggy]

k, first maybe let me get my head out of my rear, and start over with this...since i got all my numbers totally wrong....

I have a gap about 2 inches, so at stp i need 3kv/cm or 15kv for the total gap length
and according to the Paschen Curves if i reduce the pressure to 50 torr it works out to 600v/cm

I got a mot but cant id it, however using comparisons i assume it to be the usual 1kw @ 2kv step up...
with the 2kv mot @ 50 torr I can space my gap at about 3cm and still get spark, first I need for my mot is surge protection, I was reading that 1000watts of light bulb in series with mot is the way to go
Also i need DC at output, so that means rectifier, if i use 1n4007 diode, RMS of diode is 700v, while peak is 1000v , which value should i go by?
That also means that i need 3 diode in series * 4(per "leg")
I read that to put diode in series i need parallel capacitor to balance the charge, I have lots of 10kv capacitors(either 10nf or 47nf, will check) will that be enough?

I hope to use those 2 items if i can since i have lots, dont want to use resistors due to power loss inefficiency if i can help it...
will this work?
is there anything im forgetting?

thanks!

 

[JimB]

I have a gap about 2 inches, so at stp i need 3kv/cm or 15kv

Are you sure about that?
The usual rule of thumb for an arc in dry air is 1000V/mm or 25kV per inch.

JimB

 

[Tony Stewart]

Are you sure about that?
The usual rule of thumb for an arc in dry air is 1000V/mm or 25kV per inch.

JimB

I agree, BDV for a smooth large surface is 1kV/mm. This can rise to 3kv/mm in very clean dry air and drop to 600V/mm is moist dusty air. ( I recall Transformer factory Techs telling me the HV impulse Marx generator would arc when ramping up often when the big garage door at the other end of the building opened up after a rain) Marx Generators worth $50k for these Impulse TEsts have calibrated gaps controlled by stepper motors as they read the spinning leather enclosed HV generator charge up the caps in a minute. The gaps often fluctuate with air conditions.

When going from sharp Point to Neutral gnd plane creates a gradient field. Two Point sources creates an even bigger gradient, which makes it arc at lower voltages and currents. This gradient field promotes a DC charge flow thru the dielectric medium, (double elctric chrage layer effect) which causes partial discharge when dust + Humidity or impurities in oil or air bubble voids in epoxy. When this pulse rate rises, the arc lengthens into a stream, then it reaches a lower BreakDown Voltage (BDV). But if you step the voltage rapidly, there is insufficient time to cause this tiny capacitance charging effect and PD occurs just before BDV unless the medium is flawless( nearly impossible, but depends on rise time and charge time) . The result is like a Unijunction Relaxation Oscillator with Negative Resistance , meaning it will conduct as much current as the partial path ionizes and breaksdown into a very low resistance, except PD is is like a Unijunction in series with a leaky insulator. Rep. rates of PD in oil start around 1 pulse per minute, unless you have more patience, then rise quickly to 100 then 1kHz in the presence of ppm quantities of contamination.

Like SCR's with regenerative feedback, gas ionizes slowly and when conduction occurs, it occurs in sub microsecond(lightning) or sub nano second times ( corona) or pico-second times ( finger ESD) depending on path length and like SCR's will not stop conducting with DC until "quenched" or current drops below the holding current for the medium. ...Hence why your rectifier idea blew up your neon Transformer secondary.

 

[Tony Stewart]

ARe you trying to make a Marx Generator or a HVDC gen or both? Do tell!

What V , I and t values?


Look at my previous photo closely. Those are big Plastic HV caps on the right and spherical gaps on the left which must be triggered by a tiny HV pulse using anything like a magneto coil and wire around one leg to generate a controlled discharge, so that all gaps fire simultaneously and generate a massive arc. If you have continuous DC source stepped up, then you will get a continuous arc. Use chokes to control rise time of current to power the caps and isolate switching rectified charge pulses from discharge currents. ESR of the ionized ARC is always inverse to the current flowing in the arc. Plastic caps of this size have a sub-milliohm ESR, which must be less than your arc contacts.

Contact geometry of a spark plug dictactes the current due to ESR of the ARC. For example internal combustion chamber is large compared to spark plug tips , so the plug gap and contact area and cleanliness dictates combustion trigger current, in addition to high pressure eg 12:1 vs 8:1 typ. and 10:1 on old V-8's .. in Mazda new engines raising the BDV of detonation perhaps to 20kV/mm (est.)

UNder a big vacuum, contact area irregulatities will cause arc location to wander, like lightning due to localized charge repel forces, so high end spark plugs use titanium sharp types for consistent spark control current, and arc length.

 

[Dr_Doggy]

well, i was aiming for a current pump, which is why all the diodes, but now i realize that 140kv is way overkill,
same with the marx altogether, maybe overkill, i though i needed alot more volts...

so just a basic HVDC now,
2kv seems a little low, but ill take the risk, unless a 1-2stage marks would be a simple way of rectifying, i can't pump much under 50 torr

also my electrodes are just 18-22 guage wires for now, but i can change one to a plate if it helps

v=2kv
i=500ma<- transformer max

faster pulses will prolly suit me more, but idk, ideally i dont want to be able to see the charging time delay, so shorter pulses and charge times are prolly better

 

[Tony Stewart]

- your load requirements lack a purpose and method of controlling current efficiently.

- keep in mind the arc trigger is low power , high voltage and the high power can be low voltage for 10~15A, if you use tungsten tips, but much less if using copper due to proximity of >5000'K detonation of air. which generates a lot of "Harmful" RF to your eyes due to modulation and standing waves. ... which is why welders wear deep tinted glasses for Infrared.

- so define your load current and application of the arc and ripple voltage under load.

e.g. 1kW at 10A is 100V , this would represent an arc ESR of 10 Ohms, which is a bit high, would strongly suggest tungsten tips.

 

[Tony Stewart]

i.e. if you rectify into 120 Hz 10 ohm load and you want to reduce the sag in voltage that may extinguish the arc... to say 10% then you need RC=5T for T=83ms, C=415/10= 42mF. ,, which is why that Marz generator in my photo has such big arcs. You can generate 50-100kV with lots of silicone insulation and thin 8mm rubber wire with a stepup, low power HV 1:1000 small tesla coil to initiate an arc with two needle tungsten tips for about 0.5 kV/mm then the voltage drops to your 1kW power supply which sustains the arc. Thus I would expect 100kV/0.5kV/mm =20cm arc , which would be blinding bright with 1kW of luminous arc and very dangerous for X-rays due to sub-ns arc modulation transient current, if discontinuous.

Philips makes a 1kW plasma bulb in a thumb size quartz glass bulb. The power supply is 100x more expensive than the bulb and must be housed in a Faraday cage.

I have used 0-50kV DC generators for Hipot testing megawatt transformers with a voltage doubler that was the size of kitchen garbage can. It electrified the epoxy paint on the steel transformer tank, so that sliding your finger along it, you could feel the ESD pulses. The safety steel cage outside the test area also had a few kV on it and a steel girder on the wall 30m away which only for cosmetic mounting of product switch gear was continuously pulsing a very loud arc snap every 10 seconds or so due to the high gradient E field and slow build up charge on the floating horizontal wallmounted girder. HVDC is fascinating. The big toroidal donuts you may see on HV feed lines spreads out the E field evenly to raise the BDV in air.

 

[Dr_Doggy]

that is a tough question, there is no load, besides the spark gap, the old neon transformer seemed to work great, but there was prolly alot of losses that wern't noticed, just to re-mention my gap is only 2-3 inch long, 140kv was a huge miscalculation when i added one too many 0's...

I wonder how neon signs work , could I run a neon light(filled with neon) on dc pulses(1)? as i understand the way it works is that the charge time or frequency is fast enough that the (current * volts) is absorbed for light energy, and the volts are high enough to ionize the particles, while the off time allows them to release, thats kinda what i want the gap to do,
also a question is that my foot long(1 foot just a guess to the volume of bulb for old transformer) neon bulb needs 14kv, I have been reading that the ionization potential required for element neon is about 22v, does that mean that i multiply 22 * volume factor, or 14000/22 = 636(volume *pressure) of gas(2)?....
Also what if i wanted to turn most of that neon in said bulb a different color, ie ..or ..have those atoms not just get stripped of the first valance, but instead strip the second one as well, the second ionization voltage required is 41volts, would that mean i just need to double voltage OR* 41 * 636 = 26076 volts(3)???? (also i realize that would mean I need to sustain the same current for same amount of light)....

 

[Tony Stewart]

that is a tough question, there is no load, besides the spark gap, the old neon transformer seemed to work great, but there was prolly alot of losses that wern't noticed, just to re-mention my gap is only 2-3 inch long, 140kv was a huge miscalculation when i added one too many 0's...

I wonder how neon signs work , could I run a neon light(filled with neon) on dc pulses(1)? as i understand the way it works is that the charge time or frequency is fast enough that the (current * volts) is absorbed for light energy, and the volts are high enough to ionize the particles, while the off time allows them to release, thats kinda what i want the gap to do,
also a question is that my foot long(1 foot just a guess to the volume of bulb for old transformer) neon bulb needs 14kv, I have been reading that the ionization potential required for element neon is about 22v, does that mean that i multiply 22 * volume factor, or 14000/22 = 636(volume *pressure) of gas(2)?....
Also what if i wanted to turn most of that neon in said bulb a different color, ie ..or ..have those atoms not just get stripped of the first valance, but instead strip the second one as well, the second ionization voltage required is 41volts, would that mean i just need to double voltage OR* 41 * 636 = 26076 volts(3)???? (also i realize that would mean I need to sustain the same current for same amount of light)....

• colour is determined by gas
• current limit in AC is due to V/(2pi*f*L)+ V/ESR
• current. limit in DC due to series V/R+ V/ESR
• for V across tube after ionization, and ESR of tube during arc

 

[Dr_Doggy]

will 120v light bulbs on the primary side be sufficient for current limiting the secondary in spark discharge conditions?

is there anything i can/should do on the secondary circuit, besides rectifier and smoothing cap?

 

[Tony Stewart]

1) maybe, depending W size thus R ratio
2) PTC's are cheap and handy to prevent thermal overload, ICL's are hand to prevent inrush current on primary or secondary