I did yet another VFSmark on the Toshiba card. Still in the same iQue as was used by the previous tests.
Only significant difference this time was that "File Write" went down from 12-13% to 7%. All other figures are within a few steps in the last digit.
This about ESD and electric/electronic machinery:
Being a professional in this particular field, I've fought this kind of problem for at least 15 years now. I'm working with the design of machinery for the printing business. Everyone who has felt what can happen to paper, when you run it through a hot-fusion, continous web printer at 400+ images per minute, can imagine what kind of ESD phenomena we encounter in these machines.
Sometimes it looks like you can save on the lighting in the building, since that's taken care of by the sparks.
But we've learned the trade, sometimes the hard way, and can now frequently just pass our machinery through the tests, like the EN 61000-4-2 mentioned in the referenced article (it's good, I actually have a printout at work, a copy I use to pester the personnel in our production facilities with sometimes), without any further modifications. Often, they even work at the customers sites, without too much problem with statics.
I'll try to keep away from an in-depth analysis of all kinds of electric noise, and its countermeasures. It's the wrong forum, and I could keep on for hours.
Still, I would like to comment on the wordings about "ground" above. I'll make several simplifications, so if there are other electronic engineers out there, please refrain from commenting and making this as complex as it really is. We'll just feed the laymans with what they need to know.
While it's true that all electric equipment has some kind of ground, there are usually more than one.
If it's a unit that's connected to a grounded AC outlet, that ground is one ground in the machine. But not necessarily the logic ground that the electronics work with. Preferably, it's not.
However, it's this ground that has to eventually swallow all noise that comes into the machine. So, one way or another, there must either be a way for this noise to leave the electronic parts, where it's harmful, or better, to conduct it towards ground before it even reaches the electronics.
Totally enclosed circuits are easier to protect. But the iQue has the antenna connector at one end, the Universal Palm Connector at the other end, and various other exits and openings.
Once I know I had a spark going from my finger to the top of the iQue, which then was in the cradle. Since the cradle contains some kind of ground (even if the charger has no galvanic connection to ground), it will look like a ground to the static discharge.
In an ideal machine (that's how we design our units), such a spark would be conducted by the cover of the machine, to the ground point. The fact that my iQue did a reset when that happened, shows that enough of the energy took the way through the electronics.
All signals with high frequencies, regardless of whether it's the fundamental frequency or a harmonic, can't be satisfactorily conducted in thin wires. As a rule of thumb, we say that if a wire is more than five times a long as it's wide, then it's of no use as far as taking care of high-frequency noise. For that takes a surface.
ESD is a high-frequency phenomena, due to the fast risetimes of these currents. I don't remember exactly, but my spine says that the order of magnitude is 1 ns. So the frequencies involved are in the GHz range.
So PCB:s are nowdays multi-layered. A stable PCB sometimes has a ground plane between all the signal layers, or at least ground planes on both sides as well as in the middle.
If the machine cover is conducting, then it's an excellent conductor for this kind of noise, since it's not a wire, but a surface.
So, had the iQue had a metallic cover, which was integrated with the electric system according to the state of the art, it would
a) Be more resistent to static.
b) Be awfully cold to hold in your hand in the winter.
The last problem can be dealt with by having plastic layers cover the metallic, etc.
Then there's also a plethora of various protective components, as can be seen in the article. Basically, a perfectly shielded unit doesn't need any. But for various reasons, a perfect electric system may render the machine inoperable to the user, so there has to be compromises taken.
That's exactly the point where the engineers at Garmin didn't go long enough in this case, obviously.
I'm in the deepest sympathy with these guys, as well as with all the users, that suffer from the mistake.
I did exactly the same thing, about 13 years ago, due to not knowing better then. We also had to re-build machines already delivered and so on. Sometimes they worked, sometimes not. Uuurgh! It was a mess.
Linear Mode
