Common Light My Fire

[SuzyC]

I bought one of those sparky guns to light my gas stove in the kitchen. Came all the way from China. Worked for a while, but started to get cranky and eventually stopped sparking at all.

Now, don't tell me I could just go to the store and buy another.

I took apart the plastic body and the circuit board revealed a 1.5V AA battery to xxxx-V convertor. It is a capacitive-discharge ignition system.

First, a single transistor oscillator creates a 150-250 V P-P pulse that is rectified to charge the .1-uF capacitor. A SCR is then triggered by the simulated 4-layer diode relaxation oscillator three times a second to dump the .1uF cap into the primary of the loopstick HV transformer.

I have been to the household good's store and I could buy a replacement for 6 bucks, but my hobby is electronics and I want to make use of all my studying and get my sparky to light my fire again.

Turns out the potted in epoxy HV transformer had arced over while I was testing it and I need to make a replacement.

What I have on hand to fabricate the new transformer is a core from a AM radio loopstick. I unwound the litz wire from the loopstick and then rewound a tapped primary of 35 turns, the tap at 20-turns using the first 3/4-in of the rod. I made e tap is to choose which number of turns works best.

It is very important to use a low-profile rod core to fit the existing plastic fire-stick enclosure. I have approx 1-in height max to fit the transformer. So, a solenoid wound core is what fits.

The problem now is the secondary. I have taken apart a small relay and now have a spool of AWG 36 enameled wire and a 3.5-in not-yet-used
remainder of the loopstick rod core to wind the secondary on. I have insulated the rod with Teflon plumbers tape to prepare the rod to wind the secondary upon.

See the schematic. The end of the gun has a spark gap to light the gas that is about .25-in wide.

The question is, how many turns to wind on the loopstick core to get the sparky to spark?

[amyk]

Hint: Dielectric strength of air is 3kV/mm.

Note that your replacement transformer will likely need to be potted too, or it will also arc over and self-destruct.

[SuzyC]

I wonder how the hydrogen-butane-methane-oxygen gas mix at the burner top effects the HV needed for sparking?

The protection for the transformer, is the spark gap itself, at .25-in, but I would pot the secondary with a hotwax stick gun once it's working. I don't want to have to re-melt the wax to rewind the core. I need to get it working right before potting.

I figure the Teflon tape(graduated thickness from 1mm to 3mm thick at HV end) will prevent the transformer from arcing over. I plan to wind a single layer of enameled wire, if possible.

In any case, with a single-layer wound solenoid, the arc over would have to be to the loopstick core and an arc-over wouldn't occur to damage the transformer if I apply enough Teflon plumber's tape.

The original transformer failed because one or more turns shorted-out to each other and these shorted turns absorbed the entire pulse energy.

Do I have enough space (3.5-in) to fit these many turns in a single layer?

The original transformer had multi-turn vertically stacked winding sections(6 total) separated by <.5-mm thin plastic, so these wound wired sections were much more likely to short out because the wires were wound in coils where the wires were pressed tightly against each other in each section. So it only took a few arc-overs to damage the enamel insulation. The AWG 38 or AWG 40? or smaller wires had less enamel on them to insulate. At least this is my best guess.

According to the 3KV/mm then I am a trying to arc over .25-in = approx 6 to 7 mm =18KV?? min? That's a 18Kv/250V = 72/1 turns ratio which translates into approx 2200 turns with the 30-turns primary.

However, the tip of the gun makes a spark gap using the pointed end of a wire end sticking out 3mm within a metal cylinder ground.

This pointed wire end creates a strong electrostatic field and the HV pulse generated has a very fast rise time. This must affect the spark potential.

 Doesn't this mean I actually need much less voltage to create a corona effect leading to an arc-over?

Here's my theory: A much lower voltage(3KV or less?)  with a fast time creates(over a period of up to a few seconds) a cloud of ions at the pointed wire tip of the spark gap. This cloud diffuses over the gap width to create a corona leading to a plasma activated area and lightning strikes again!

[IanB]

One lesson that might be learned from this is that it is dangerous to run a high voltage generator without the spark gap present on the output. Basically a spark is going to happen somewhere and the spark doesn't happen at the intended spark gap then it is likely to happen inside the transformer instead and destroy it. So never run such a spark generator with the output open circuit. Always make sure the spark gap is present to absorb the generated energy.

[SuzyC]

IanB, you are now preaching to the saved! I've learnt my lesson!

The original trigger circuit used a 170V approx 4-layer diode to trigger the SCR. The 4-layer diode eventually began to fail and leakage would limit the voltage at the discharge capacitor to around 150-V. in a state of static  equilibrium.

So I created my own trigger relax osc. to trigger the SCR instead. Playing around, I got the spark to arc over .5-in, a few times and to my surprise, the transformer arced-over and failed!

I didn't quit while I was ahead!

[SuzyC]

My poor Sparky!

[IanB]

Might you correct the thread title? The misspelling is bugging me...

[SuzyC]

My Sparky seems 2b a most common electric lighter model. All words are correctly spelled.

Anyone have an idea of what minimum fast-risetime pulsed voltage would often enough spark using a single pointed electrode center wire inline within a cylinder with a spacing of .25-in?

[Zero999]

My Sparky seems 2b a most common electric lighter model. All words are correctly spelled.

Anyone have an idea of what minimum fast-risetime pulsed voltage would often enough spark using a single pointed electrode center wire inline within a cylinder with a spacing of .25-in?

I would have thought that a low duty cycle would make it less likely to arc over, as it takes time for the gap to arc over.

Design your circuit so it produces a much higher voltage, than the minimum, even when the battery is nearly dead. As mentioned above, different gasses have different spark over voltages. If you design it so it works with methane (natural gas) it doesn't mean it'll also work with butane.

The generic guide for spark gaps, in air is 1.1kV/mm. 0.25" = 6.35mm, which would require a voltage of 7kV. If you aim for higher, as 10kV, it's got more of a chance of reliably arcing, especially with other gasses.

[SuzyC]

I see how appropriate your advice is, Hero999, meet the minimum and then consider the worst case of conditions in my design.

Still, I have a very puzzling question. What would be the minimal number of turns on the primary of the HV transformer?  If I reduce the number of turns on the primary, I can also reduce the number of turns on the secondary and the idea of a single-layer secondary looks very practical.
I am guessing that lowering the primary inductance will cause the length of the spark discharge to also shorten and a critical minimum time is needed to create an oscillation that will turn off the SCR.

I was very surprised also to learn that not just any blue spark will ignite the gas. It seems that a certain minimum amount of heat must be generated to reliably cause the gas to ignite. At the same time, the amount of energy delivered to the gas is determined by the energy available stored in the the capacitor and is measured in joules. Energy is proportional to the square of the voltage and directly proportional to the size of the capacitor.

I also note that is takes around .5-sec for the invertor circuit to charge the discharge cap to 150V and almost  a second for the
voltage to charge to  250-V. This clearly means there is an obvious duty-cycle limitation in this circuit of sparks/sec.

My question is now:  if 35-turns might work, why wouldn't 20-turns or 15-turns or even 5-turns work for the primary?

[IanB]

All words are correctly spelled.

If all words were correctly spelled I would not have made the comment 

Come on: a verb meaning go ahead, get moving, keep up

Common: an adjective meaning frequent, commonplace, found everywhere

[chris_leyson]

Thanks for posting the scematic SuzyC. A Xenon trigger transformer might be an option. Spec for a typical 6kV trigger transformer, primary 4uH 0.085 Ohm, Secendary 6.5mH 90 Ohm. The original transformer may well have been intended for use as a Xenon trigger transformer.

[SuzyC]

Corrected Schematic.


My question is now:  if 35-turns might work, why wouldn't 20-turns or 15-turns or even 5-turns work for the primary?

[SuzyC]

IanB: What kind of cheese would you use to get a bear to come to you?

[Zero999]

I see how appropriate your advice is, Hero999, meet the minimum and then consider the worst case of conditions in my design.

Still, I have a very puzzling question. What would be the minimal number of turns on the primary of the HV transformer?  If I reduce the number of turns on the primary, I can also reduce the number of turns on the secondary and the idea of a single-layer secondary looks very practical.
I am guessing that lowering the primary inductance will cause the length of the spark discharge to also shorten and a critical minimum time is needed to create an oscillation that will turn off the SCR.

I was very surprised also to learn that not just any blue spark will ignite the gas. It seems that a certain minimum amount of heat must be generated to reliably cause the gas to ignite. At the same time, the amount of energy delivered to the gas is determined by the energy available stored in the the capacitor and is measured in joules. Energy is proportional to the square of the voltage and directly proportional to the size of the capacitor.

I also note that is takes around .5-sec for the invertor circuit to charge the discharge cap to 150V and almost  a second for the
voltage to charge to  250-V. This clearly means there is an obvious duty-cycle limitation in this circuit of sparks/sec.

My question is now:  if 35-turns might work, why wouldn't 20-turns or 15-turns or even 5-turns work for the primary?

You are right that a certain amount of energy is required to ignite the gas.

The amount of energy in the capacitor, each cycle is the same E = 0.5CV². The question is how much of it is converted to heat and how much is actually transferred to the secondary side of the transformer? Fewer turns would mean less inductance, which would cause the current to increase faster, resulting in more losses in the winding resistance, assuming you've used the same wire thickness. There also comes a point of core saturation, when increasing the current, no longer results in the same increase in magnetic field. At this point, any increase in current, just heats the coil more. This will be counteracted to some degree by the fact that with fewer turns, there will be less magnetic field generated, for the same current. Another problem with fewer turns, is poorer coupling, so less of the energy will make it to the secondary.

[SuzyC]

Thanks Hero999, and I get your idea about achieving a minimal number of turns: Spare the turns and spoil the rod.

But of course this  info alone doesn't clearly guide me to find out how few primary turns would work.

Amateur electronic engineer that I am, how can I be a little more engineeric and a little less emperical to figure this out?

Can you guide me to find  some way to resolve the characteristics of this loopstick rod to determine how much I can spare the turns?

I have a good oscilloscope. If I do not yet wind the secondary but treat the primary as a simple inductor and trigger the SCR circuit as shown to deliver the storage cap's 150V to 250V  pulses to the core, what could I look for with my 4-chan 200 MHz scope to show me how to use my rod?

I could also wind a few turns upon the secondary and load them with a resistor to determine energy transfer efficiency?
If I observe the waveform on the secondary, could I use some formula to convert the readings to compare with the energy stored in the discharge capacitor?

I guess what I am saying here is that I want to learn more about this  current subject, and this is to me quite exciting!

I hope to learn something about my Sparky before I wind things down on this project.

According to: http://hyperphysics.phy-astr.gsu.edu/hbase/electric/capeng2.html

The energy stored in a capacitor is C*V*V/2, this is quite different from the formula you've presented??

[Zero999]

Thanks Hero999, and I get your idea about achieving a minimal number of turns: Spare the turns and spoil the rod.

But of course this  info alone doesn't clearly guide me to find out how few primary turns would work.

Amateur electronic engineer that I am, how can I be a little more engineeric and a little less emperical to figure this out?

Can you guide me to find  some way to resolve the characteristics of this loopstick rod to determine how much I can spare the turns?

I have a good oscilloscope. If I do not yet wind the secondary but treat the primary as a simple inductor and trigger the SCR circuit as shown to deliver the storage cap's 150V to 250V  pulses to the core, what could I look for with my 4-chan 200 MHz scope to show me how to use my rod?

I could also wind a few turns upon the secondary and load them with a resistor to determine energy transfer efficiency?
If I observe the waveform on the secondary, could I use some formula to convert the readings to compare with the energy stored in the discharge capacitor?

I guess what I am saying here is that I want to learn more about this  current subject, and this is to me quite exciting!

I hope to learn something about my Sparky before I wind things down on this project.

According to: http://hyperphysics.phy-astr.gsu.edu/hbase/electric/capeng2.html

The energy stored in a capacitor is C*V*V/2, this is quite different from the formula you've presented??

I made an error with the formula and have corrected it. Good on you for spotting it.

I don't think a loopstick is really the right thing to use, but will probably work.

I'd cover half of it with the primary, then wind the secondary on the other half. I don't know what the minimum number of primary turns is, especially given the properties of the core are unknown.

I've tried to do a similar thing to this before and had problems with the secondary arcing over. I think you'll get better results by winding it like the original transformer, with several windings separated by plastic insulators. I doubt you'll be able to get enough turns on one layer, to generate sufficient voltage. You may also need to pot the secondary in something, as any moisture can cause arcing. A hard resin or silicone rubber is normally used, but you could try wax, because it can easily be melted, if needs be.

[SuzyC]

I've been searching the web to find info on solenoid transformers and found almost nothing, most info is about air-core,, toroid, and commonly used E-cores or square core transformers that are constructed to close the magnetic circuit.

I find it both mysterious and interesting that a conventional spark coil has the secondary on the outside of the outside of the can, and also found that ignition coil cores were laminated steel rather than ferrite.

I also notice that the primary of the original potted coil was made by potting a rod ferrite core inserted into the multi-layer HV secondary center's plastic form. The primary seemed to be wound on a circular core floating in epoxy within the cylindrical walls and separated by at least 3mm from outer frame.

This architecture seemed to present itself as a very good way to construct a transformer with very  tight magnetic coupling, while still having an open-frame solenoid magnetic circuit, and unlike an ignition coil, the idea here was to place the primary in the center.

There is much info about transformers in the most common shapes and much about solenoid inductors, and many scholarly articles about transformers in general to very advanced, complicated and difficult to understand pulse transformers such as were constructed to power klystrons at CERN.

Info that explains the detail of a sophisticated closed magnetic-loop transformer design presents formulas that expresses design using  what is known abut a core's permeability, magnetic flux density, saturation, hysteresis of core materials, eddy currents, energy, work, power, ampere-turns, reflected inductance, primary and secondary coupling coefficients, core losses, types of specialized ferrites, magnetic curie points, reluctance, inductance, and efficiency, isolation of windings, etc. .

I remain in the dark, too stubborn to light a match to make dinner and spend my time cursing the darkness and whining about not knowing how to wind my wires. I do this rather than lighting one match, starved intellectually, kept from understanding the math that reveals the hidden secrets of designing a solenoid transformer using engineering rather than trail and error techniques.

A new Sparky is on it's way from China, sent to me free of freight charges, sent by means of Chinese international economy airmail, and guaranteed to arrive before Mar. 31, 2018.
.
Will I find a way to wind my windings before then?

In the meantime, I am trying to memorize the info in this excellent .pdf published by the IEEE.
FUNDAMENTALS OF
MAGNETICS DESIGN:
INDUCTORS AND TRANSFORMERS
Thursday September 15, 2011
Arthur Williams
Chief Scientist
Telebyte Inc.

[John Heath]

Jim Morison would turn over in his grave. It's " come on BABY light my fire " , gees. I am liking the second stage with the SCR pounding the last transformer with all it has but what's with the weeny first stage with the fancy oscillator and stuff. This does not need to win a  Nobel prize . We only need a little spark. I understand there is male pride here where we can not under any circumstances  go back to the store and buy another 6 dollar lighter. I am thinking kick ass in the first stage the same way as the second stage.  Drive the first stage like a flyback transformer with lots of current then fast off. An AM antenna ferrite has a fair saturation point and who cares about the battery life. This will give some beef for the second stage to work on . In short brute force and ignorance. Don't worry if it does not look cool. The main point is to get Mr Sparky up and running again.

[Zero999]

Re-reading this thread, I've missed a couple of things:

Ignition coils use laminated steel, rather than ferrite, because it has a higher permeability. It also has higher hysteresis and eddy current losses, but those aren't important at low frequencies.

The main capacitor charger circuit, could be simplified to used a blocking oscillator, similar to the type used to charge the main capacitor in a zenon strobe circuit.

Photo flash trigger transformers only output a tiny amount of energy, probably not enough, for this application. If you look at the capacitor discharge on the primary, you'll see it's tiny, compared to the sparky.

[SuzyC]

Hero999. The schematic of my Sparky does use a blocking oscillator to charge the CD cap, probably the same transformer and circuit as xeon flash cameras use. The capacitor at the primary does discharge a fairly large .1uf cap at up to 250V and that is not a tiny amount of power to be used to be released instantly into  the HV transformer primary to make a spark.

Energy=C*V*V/2.

Didn't the Fab Four, as well as Jim Morrison sing about this?

Hey Joule, don't let me down!
Make a spark,
We'll all get to eat,
If I can make dinner..

Remember,
To light the gas fast,
And makes things better!

[cdev]

I have a piezo flame starter that is 100% human powered.

It uses a piezo crystal which creates a high voltage when momentarily deformed.

[Zero999]

Hero999. The schematic of my Sparky does use a blocking oscillator to charge the CD cap, probably the same transformer and circuit as xeon flash cameras use. The capacitor at the primary does discharge a fairly large .1uf cap at up to 250V and that is not a tiny amount of power to be used to be released instantly into  the HV transformer primary to make a spark.

Energy=C*V*V/2.

Hey Joules, don't let me down!
Make a spark,
So I can make dinner..

You're right, it's a blocking oscillator. I didn't look at the schematic properly.

It looks like I was wrong about the xenon trigger transformer too: they use similar size capacitors to what you're using. I've Googled and most are probably cost more than your sparky. This is the cheapest I could find, but the seller doesn't provide much information.
https://www.ebay.co.uk/i/272408094685?chn=ps&adgroupid=13585920426&rlsatarget=pla-142405566786&abcId=&adtype=pla&merchantid=113383089&poi=&googleloc=9046203&device=c&campaignid=207297426&crdt=0

[SuzyC]

Hero999,

The solution you've found,
in a flash,
Belatedly,
costs too much cash.

My Blinky is cheaper,
But on the blinker.

A new one is ordered,
But won't arrive soon,
Coming from China,
I t might get here faster,
Than one or two moons.

But I still want to learn about how design a HV solenoid transformer using a loopstick!