EEVblog #678 - What is a PCB Spark Gap?

[max666]

Dave,

I'd really like to see a video or series of videos on input protection just as you suggested. It'd be great to see this for both analog and digital (high speed and low speed, busses, etc). Maybe some tidbits on protecting against impulse/ESD vs slow ramping overvoltage vs long periods of overvoltage.

seconded

[eneuro]

If the board is of any significant age, and has started to accumulate a film of pollutants from the air, moisture absorbed into that film will greatly reduce the breakdown across the board surface itself, which is why slotting the board is such a big help. It really stabilises the breakdown point. It also avoid the carbon track build up when breakdowns are happening, although narrow slots aren't always easy to achieve.

Just made such test ESD gap PCB to connect in parallel to transformer primary coil.

Thinking how to finish this copper surface and make some kind of sharp edges between two sides to make this breakdown voltage >400V <1000V at 230VAC 50Hz mains between two sides in normal conditions.
The best idea for the moment is in my opinion just solder small copper wires about 10mm-15mm long 1mm-1.5mm in diameter, connect those two sides in clossest designed places and then... simply cut them after soldering to leave some small gap.
Those copper wires could be easy to replace if after a few sparks events they were damaged somehow...
Any better ideas how to make those ESD PCB gap small, but still be able to refresh spark hitting area without replacing whole PCB?

[max666]

... Any better ideas how to make those ESD PCB gap small, but still be able to refresh spark hitting area without replacing whole PCB?

Nice PCB spark gap.
But if you plan on something easily replaceable or able to take multiple strikes, I don't think a PCB spark gap is the right tool for the job. I would use a gas discharge tube if you plan for more than the exceptional strike.

[eneuro]

I would use a gas discharge tube if you plan for more than the exceptional strike.

Anyway those  datasheet: Gas Discharge Tube (GDT) Products fom there:
http://www.digikey.com/product-detail/en/CG21000L/F2731-ND/950212

looks like has limited to 400 operations life span or missed something? 

Made also slightly different PCB ESD very simple gap like below, just by soldering 2cm long 2.5mm2 copper wire to PCB and cuted in the middle and adjusted at the height to get given gap size:

Cut copper wires edges are very sharp and gap width can be corrected when a few sparks events will destroy small amount of copper, so it looks like it can be easy to inspect from time to time and make adjustments when needed.
Those sparks shouldn't happen too often.
However, probably it will be good idea to put this thing into thermal cover and isolate somehow, else spark could ignite something in the presence of dangerous gases, etc 

In the case of gas tube I have no idea how to verify its working condition after a few sparks while nothing visible and closed inside not glass  tube 

[SeanB]

It is intended to be used a protective device, so a limit of 400 operations is no problem. You would find that they will last a lot longer with lower energy put in, and in most applications they typically will be run at a reduced energy due to the wiring resistance and inductance. 400 cycles minimum at a defined test pulse is not bad, equates to around 5 years typical in most areas with lightning strikes nearby.

If you are in an area with higher storm frequency you will probably have had the phone lines and local power lines buried decades ago by the utility to keep them from being damaged, so the surges will be a lot smaller and easier to clamp. If not then you will have to cascade the devices to provide protection.

In most cases you will find that the surge protection provided in the package with a new ADSL modem or cordless phone is typically not used, or not connected correctly. Only after the first warranty claim that is refused ( and they do do that now as they get frequent complaints from areas so will ask for that as well for the warranty, no SPD installed ( they can see if it has operated) and no warranty, buy a new and here is your bill for time) are they installed and used.

[JuKu]

So the idea that you want to protect certain nets by giving them a way to arc over to board ground, right?

Kind of, but a spark gap alone is not sufficient.

Q#2 - Is there a benefit to putting a spark gap on a net you want to protect that already had a TVS protecting it?  Is there a benefit to having both?

The purpose of a spark gap is to protect the real protection circuitry (and sometimes, allow for cheaper protection scheme). A spark gap takes 10's of kilovolts away, sometimes single digit kilovolts, but even a few ten volts can kill sensitive circuits.

In a product test we had a spark gap, 100p to gnd and low power schottky diodes to rails. That was cheap and sufficient. 50kV lightning sustained to connector - As the product designer, it was rather scary to look, but the input pin didn't see more than rail voltage. Without the spark gap, the schottkys turned to ash. (And the inside of the IC as well I suppose, even though I didn't see it. But the IC was certainly dead.)

[max666]

random question ::

can a NE2 or similar neon tube be adapted to be used to arrest small spurious surges? they are after all rated to trigger at 125v? 

http://youtu.be/KXbz2KUr8HQ?t=9m2s

But when you say "rated to trigger at 125v" I hope you don't mean for using them to protect mains 120 V? Because the trigger voltage of neon tubes can vary substantially, even depending on environmental conditions. Quoting from wikipedia here "Most small neon (indicator-sized) lamps, such as the common NE-2, break down at between 90 and 110 volts".
Also keep in mind once a neon tube is triggered it won't shut off until the voltage falls below about a third of the striking voltage.

[max666]

Damn it! I did deliberately burn out a NE-2 (or at least very similar) once, but I can't remember the values.
The only thing I can say for sure was that my electrophoresis power supply didn't have enough current to destroy it and it goes up to 400 mA. So I paralleled it up with my lab supply (with a reverse protection diode of course) and I further increased current slowly until the electrodes melted. 
I think the current was more than 1 A and it couldn't have been more than 2 A, because that's the max of the lab supply. What's really interesting is that the electrodes started melting from the top (at the bottom are the leads which allow heat to be better carried away), and forming a little ball at the end of each electrode, which moved down and got bigger as the melting went on until the two balls touched and shorted out the leads. Of course that was a really slow burn out, so surge behaviour might be completely different. That shorting behaviour however would be really great for a protection device, because with a fuse it would separate the circuit if the neon lamp got overloaded, but again I don’t know if it does that for surge overload reliably.

If you want me to replicate the burn out or test something similar let me know, I still have some of the old neon lamps lying around.

[T3sl4co1l]

Oooh, that sounds fun.

Once upon a time -- think I was about 10, I had some neon lights laying around; of course, you only power these with a resistor.  You use a large resistor for dim, or a small resistor for bright intensity.  Well, I figured, how bright can they get?  One neon, two little metal leads, two slots in the wall receptacle.... *POP*

I recall never finding any of the glass, only a single rod electrode in my lap.

Under slightly less strenuous conditions (like discharging larger and larger capacitors into them), the glow discharge turns from a brighter orange glow, to "sparkier" colors like blue -- presumably, all the gas is becoming singly ionized, and higher ionization states are being produced, with corresponding color changes.  I forget how many uF (and V) it takes before damage starts happening; I think the blue spark threshold is around 200V and either 1uF or 10uF?

[eneuro]

Also keep in mind once a neon tube is triggered it won't shut off until the voltage falls below about a third of the striking voltage.

Just wondering if another PCB on top will create enclosure with edges filled with solder, etc, than a few strikes inside this thing can increase temperature and lower air break voltage, I guess? 

It was a little bit art work, but final esd gap like this could have copper traces cut on laser CNC and put inside two PCBs to (like one bottom here) spot welded using 1000A pulses, to glue everything together and catch air at given humidity inside 
This way ESD from sharp edges could hit into 1mm copper plate walls inside between PCB top/bottom walls 

In this prototype, probably will add on top another PCB with 4 holes and screw them together.

Now thinking, how to test this thing-probably will use transformer from microoven teardown and add a few stages of a few kV voltage doublers to make something like preconditioning-hit a few HV strikes to ensure this ESD gap copper rings soldered later will enable breakdown at lower voltage later 
Probably this is not needed if 1mm CNC copper plates will be used instead of this temporary copper rods and solder art work 

[SeanB]

Or you could just use a gas discharge tube that likely will have both higher surge capacity, higher current handling and more importantly a much more predictable breakover voltage that does not change with ambient air pressure, temperature, composition and humidity.

[sxanthony06]

I have a few questions about the material of the first PCB showed in this video. I am studying Electrical Engineering and we need to build high voltage circuits that aren't suitable for breadboards. I have seen this type of board in power supplies we have at the university. So my question is: is this a normal FR4 laminate like they sell on DigiKey and other vendors with the top layer coated with a light yellow soldermask with black silkscreen on it or is this entirely another kind of material? If the latter, where can I buy it?

[EEVblog]

I have a few questions about the material of the first PCB showed in this video. I am studying Electrical Engineering and we need to build high voltage circuits that aren't suitable for breadboards. I have seen this type of board in power supplies we have at the university. So my question is: is this a normal FR4 laminate like they sell on DigiKey and other vendors with the top layer coated with a light yellow soldermask with black silkscreen on it or is this entirely another kind of material? If the latter, where can I buy it?

Just standard FR4 type PCB material.
For really high voltage stuff you use routed out slots, so you get genuine clearance instead of creepage paths across the board.