Interesting spark gap application

[Kenyan_guy]

I did a tear down on a power analyzer and I noticed the following design. The right of the circuit is conected to a flyback switching converter for AC-DC conversion.

Interestingly, I noticed the attached image(sparkgap.jpg) which seems to be a sparkgap in parallel with an RL parallel circuit of values 220R and 4.7uH. The reversed schematic is in design.png.

The question: Is this a common design pattern to place the sparkgap along the line and what could be the purpose of doing this?

[T3sl4co1l]

Yes, this is also often seen on common mode chokes (these are normal mode chokes, for filtering line noise it would seem).  It's probably done to improve reliability during ESD testing -- if the inductor's insulation breaks down when zapped, it will tend to fail shorted.

Tim

[PartialDischarge]

What power analyzer is that? and could you post a quick teardown with pictures?

[vk6zgo]

That's not an interesting spark-gap-design.
This is an interesting spark-gap-design!
http://vintagehamstation.com/images/n4ggspark7.jpg

[coppercone2]

I was always interested in big multi conductor common mode spark gaps, as far as interesting goes. Imagine matching a big spark gap to a entire bus.

But the application is extremely interesting, because I ran into a chicken and egg protection problem (do you want the filter taking the edge off for the MOVs and stuff? is the x2 cap stronger? how does that inductor fare? etc).

[ebastler]

That's not an interesting spark-gap-design.
This is an interesting spark-gap-design!
http://vintagehamstation.com/images/n4ggspark7.jpg

That is a beautiful piece of equipment! Built with a lot of love, vintage components, and hand-made reconstructed parts, according to the website. 

Can you still operate such a transmitter today, without making a lot of enemies in the ham radio community? My gut feeling would be that it splatters across half the amateur band when transmitting, and at 500W output power it would reach pretty far with a decent antenna?

[T3sl4co1l]

Technically no, because spark transmitters are singled out by the FCC at least, and probably by others too.

With good enough filtering, you might get the trash clean enough to transmit just in-band, but you aren't going to clean up the sidebands without truly heroic efforts (mechanical resonator ran at transmitter power levels?).  The typical in-band waveform is a decaying envelope, which makes the signal easy to hear on AM (no BFO needed) but still not compatible with CW channels as usually defined (500Hz channels, or even less).

Tim

[ebastler]

Technically no, because spark transmitters are singled out by the FCC at least, and probably by others too.

With good enough filtering, you might get the trash clean enough to transmit just in-band, but you aren't going to clean up the sidebands without truly heroic efforts (mechanical resonator ran at transmitter power levels?).  The typical in-band waveform is a decaying envelope, which makes the signal easy to hear on AM (no BFO needed) but still not compatible with CW channels as usually defined (500Hz channels, or even less).

Thanks Tim; that's about what I thought. For Germany, I can't find a specific regulation against spark transmitters. But radio amateurs are obliged to "limit transmissions to the bandwidth required by the respective mode, in line with the state of the art". And for most shortwave bands, there is also an upper limit of 2.7 kHz transmission bandwidth, independent of the mode. So I won't be building a spark gap transmitter any time soon, it seems...

[dmills]

It is in the ITU "Radio Regulations", damped wave transmitters are explicitly banned, which rules out spark transmitters.

However a quick reading does NOT seem to rule out the "Paulsen Arc" as this is a negative resistance oscillator.

Regards, Dan.

[TimFox]

The Poulsen arc, sometimes called the “Federal arc” after the Federal Radio Co. who commercialized it, is a continuous arc discharge, stabilized by a magnetic field.  Since the arc takes a relatively long time to start, Morse code transmission was done with frequency-shift keying, switching a capacitor into the tank circuit with a keying relay.  The negative resistance of the arc circuit connected to the resonant tank circuit results in a continuous undamped oscillation.
The Federal Radio Co. later became ITT, and a magnetic armature from one of their arcs was rescued from a scrapyard to serve in Lawrence’s early cyclotron.

[Kenyan_guy]

Yes, this is also often seen on common mode chokes (these are normal mode chokes, for filtering line noise it would seem).  It's probably done to improve reliability during ESD testing -- if the inductor's insulation breaks down when zapped, it will tend to fail shorted.

Tim

Thanks for the information, yes that reasoning does make sense doing it along the line. I will check the parallel RL characteristics to check the filter cutoff

[Kenyan_guy]

What power analyzer is that? and could you post a quick teardown with pictures?

The egauge pro https://www.egauge.net/

I will send more pictures by Thursday.

[Kenyan_guy]

I was always interested in big multi conductor common mode spark gaps, as far as interesting goes. Imagine matching a big spark gap to a entire bus.

But the application is extremely interesting, because I ran into a chicken and egg protection problem (do you want the filter taking the edge off for the MOVs and stuff? is the x2 cap stronger? how does that inductor fare? etc).

Its all a catch 22! I believe the reasoning given above of the sparkgap preventing the inductor failing short due toinsulator breakdown above is true.

As to the RL filter purpose, all MOVs in the design are ZNR14511U from panasonic whose application tolerances are attached. The purpose of the RL filter may be to act as a high impedance to frequencies above the sensitivity of the downstream components. Example, the transformer, there is also an opamp that performs signal conditioning before passing it off to a 200kHz SAR ADC to sample the state of the voltage on the line for analysis under control of an MCU.

I believe this makes sense, unless anybody has a separate view.

[coppercone2]

i am not sure, but if you have non identical protection components you want a impedance between them (for instance for a GDT to fire before/during the MOV clamp so the MOV does not clamp and prevent the GDT  from firing).


the RL might also prevent inrush into C but IDK why its before the MOV. it must be protecting it some how? maybe they found high frequencies can break it some how (kind of what comes to mind is induction heating is frequency dependent and it heats the surface of objects but idk if this is relevant its just something that comes to mind when I think about it, maybe because I view the MOV as a grainy bunch of balls that are sintered together). maybe its not frequency specific but they just found that Spark gap, resistor, inductor combination restricts energy and is more robust then the MOV and it ends up lasting longer if you set it up that way.  I figure it must be hard to break a inductor and a resistor. I think you would need to test both ways to make sure, it might be very hard to determine what is more protected and against what. it may be that against specific interference functions putting the MOV first would make more sense then after, but I really don't have a clue.

also maybe maybe the MOV acts as a frequency mixer for RF and ends up making it even higher in frequency/worse to deal with, so they try to restrict it.

normally you want the MOV to blow a fuse though

[dmills]

The resistor is a classic way to lower the Q of the inductor and reduce the things tendency to turn a fast transient into a ringy mess, nothing more.

You see something broadly similar in low voltage DC filters for critical things where the objective to to avoid the resonant peaking (There is some discussion in 'advanced black magic').

Regards, Dan.   

[coppercone2]

but is it protecting the MOV?