Do spark gaps work with battery powered devices?

[sebmadgwick]

I hope this is not a silly question...

A battery powered device (e.g. mp3 player) does not have an earth, only the battery negative terminal.  Would a PCB spark gap provide any protection against ESD strikes?  If so, then why would the spark arc to battery ground given that this is just a arbitrary voltage level, presumably just as arbitrary as the IC pin to be protected?

[SeanB]

No it does not. The spark gap provides a low impedance path for ESD from an input to a ground lead that is attached to a connection that is electrically connected ( by some means like a power cord, phone line, grounded TV antenna) to the earth under you, so that the charge is carried to this ground lead in preference to going through the electronics further inside the device, the voltage generated across the gap being low enough ( though still very high compared to normal operating voltage applied to the port) for the electronics to survive without significant damage to them. Repetitive exposures however will likely cause the gap to fail as either open circuit or a a low resistance.

In a battery powered device the gap is useless as the entire device floats up to the ESD voltage, with only the currents and voltages generated by the capacitance of each part to the earth ground surrounding it. thus you can get damage, but a spark gap will not work to prevent it. You need other methods, that typically operate at lower voltages and act in conjunction with shielding to keep the high voltage transients from causing a voltage drop across internal parts.

[sebmadgwick]

...the entire device floats up to the ESD voltage, ...You need other methods, that typically operate at lower voltages and act in conjunction with shielding...

I have been reading quite a bit about ESD recently but I have not yet come across articles / application notes that specifically discuss battery powered devices.  Are you able elaborate or suggest some links?

This thread is also relevant.

[tggzzz]

I hope this is not a silly question...

A battery powered device (e.g. mp3 player) does not have an earth, only the battery negative terminal.  Would a PCB spark gap provide any protection against ESD strikes?  If so, then why would the spark arc to battery ground given that this is just a arbitrary voltage level, presumably just as arbitrary as the IC pin to be protected?

In theory it will provide a little protection against ESD under some circumstances - but in practice that protection would be insufficient and other techniques are used.

If there are exposed conductors, e.g. for power leads or USB connector, then ESD could strike one conductor and exit via another - even both are "isolated" from "ground". If that happens the electronics would be fried with or without a sparkgap. Instead, protection is via TVS and zener diodes. Where very high reliabililty is required or ESD is likely, then there might be multiple levels of protection, e.g. a spark gap takes the brunt of the discharge, leaving the rest to be dealt with by other means (e.g. TVS/zener diodes).

The concept of a battery having "ground" terminal is fallacious unless that terminal is connected to a spike in the earth. But you already suspect that.

However, the concept of "ground" is itself a theoretical construct that has some use when considering domestic and industrial mains power supplies. The concept can be very misleading when considering electronic equipment - if you doubt that then understand the shape of RF components and remember that nowadays digital circuits operate at microwave frequencies (and computation is done at VHF/UHF rates!).

[tszaboo]

It is not a silly question. If you are charged to 3000V let's say, and you take a battery operated device into your hand, by grabbing the earth, the positive of the battery will be 3003.7V let's say. As the capacity of the actual device to the earth is almost zero (not the capacitors to the "ground"in the device), something like fraction of a picofarad, there was only insignificant amount of current, and no damage. "University people" can actually calculate this capacitance.

Now, if you plug this 3000V into a charger, things get interesting.