Electrostatic problems

[ian]

Hello smart guys......... got a question for you. Good for the board too.
I make small electronic gadgets for local manufacturers and I've been making
this PIC based gadget for a few years now. It's a small PIC micro on a 4" X 4"
board powered by a 24VAC transformer. Everything is housed in a grounded
metal box. I've made hundreds of these but lately...........
As I make them, I put them in a cardboard box with tape around it. If I place
it in the cardboard box, power it on, and slide it around on the carpet the
micro resets. I'm trying to get to the bottom of the reason but haven't been
able to pin it down yet. I assume it's an electrostatic field on the tape causing
the resets, but I have not modified the design and earlier ones don't reset.
I am currently investigating the solder flux on the board, cleaning the boards
does seem to have some impact, but doesn't totally solve the problem all the time.

Here are the questions though..........

Any idea wy the micro is resetting and anyway to protect it?

If I ground the D.C. on the board to the metal box the resetting stops. But, is it
a good idea to do this? My gut feeling is no. One of the outputs of the board
controls a LED lamp through a 0-10V output. If the LED lamp is also
referenced to ground I'll have ground loop problems. or will I?

[Dean Huster]

What the heck are you doing sliding the thing around on carpet?!!? Haven't you ever been zapped by a doorknob? I'm surprised that you haven't zapped chips other than the fact that they're connected together and to power.

Try powering the thing on a sheet of anti-static plastic (pink poly or some such thing). The resistance of the plastic is probably high enough that it won't affect the circuit.

[haklesup]

A typical cardboard box is conductive enough to dissipate static, any charge obtained by tribocharging the tape (by rubbing it on the carpet) would only reside on the outside surface of the tape and would not exert a significant electric field within the box. You should in any case, package the product in ESD safe packaging inside that box (pink bag or bubble wrap)

First I would ask what normally resets the PIC, how is it designed to reset? If its bringing a particular pin high or low, I might investigate ways in which I could accidently do that in handling or if a static jolt to that pin or trace could trigger a reset. Although you may not have changed your assembly routine, has something about your supply of parts changed. Did this coincide with a new batch of any component particularly the PIC and electrolytic caps.

Grounding the ground of a circuit (center tap of the transformer or DC ground) is a good practice and for DC signals, ground loops are not a problem if there is only one board. The box forms a fariday cage and even a srong zap to the outside usually won't get in. I see no problem making the case and circuit ground the same.

You should always clean the flux. Not only can it cause circuit leakage but over time it corrodes the solder joint and can cause opens. RA flux is worst but even water based flux can corrode copper. No clean flux is the only flux safe to leave on. If you have ever observed flux gunked boards in appliances or electronic products, this is just a sign of low quality and cheapo assembly. I always use water clean flux and wash boards in tap water and dish soap then rince with isopropanol blow dry with compressed air and bake in an oven 50C to 80C until dry. Sure there are better soaps and DI water but only for critical cleaning

[ian]

I figured it out.
I thought it was the flux.
But I cleaned the boards so it wasn't.
HEY. who says I can't slide the electronics around a staticy/staticky/sticky carpet???
Can't I have ANY fun?
Doesn't the grounded metal box count for SOMETHING?
I was using the wrong program. This in itself wasn't the problem but I had the external
interrupt enabled.
But this too, wasn't the problem.
I had the external interrupt pin pulled high by a 10K resistor.
So unless I'm missing something the static pushed its way through the box, through
the 10k pullup to cause resets.
Once I disabled the interrupt all was well in Ianland again.
But I'm still very, very, irked by the fact that the PIC could be so easily affected.
Am I doing something wrong? Should I be designing differently if I want to use
the external interrupt?

[haklesup]

Once a thing like static electricity enters an active circuit and causes a disruptive signal on any line, we call it a transisnt. In most cases, transients can be damped, shunted or filtered by judicuius placement of decoupling capacitors. Since your device was likely sensitive to very narrow pulses (in the ns range) it wouldn't take much to disturb the circuit.

If you had sent the full energy of the ESD pulse into one pin, you wold have had a good chance of a destructive failure. The ESD association is about to publish a new spec dealing with ESD applied to external contacts of a powered circuit such as the usb port on a portable device or an externalantenna input. These specs will allow manufacturers to make more robust products and demonstrate that to potential buyers.

The transient may not have pushed the input pin low against that pullup, it just as well may have pushed VSS up or Vdd down. You would need a bit of time with a high speed scope and a high impedance probe to begin to find the root cause. YOu may not be totally out of the woods yet, any of your other lines could be disrupted but since you only latch that data when needed and no line like an interrupt is active which can stop the whole show on a single random pulse, you probably won't see it again.

You probably can't have too much capacitance across Vdd and VSS. Try 1uF and/or 10uF ceramic chip cap close to the device at each VDD pin to the ground plane or ground node under the part (think short paths) and another 47uF tantalum and/or 100uF electrolytic at the edge of the board near where the power comes onboard. That should quiet a wide range of rise time and energy content. Furthermore, a higher current power supply is harder to disrupt.

[Dimbulb]

you could build a sensitive electrometer such as 9V battery jfet led.
this is used like a sniffer to find where the buildup is.
The fuzzy stuff (static electricity) likes to swarm in and around the inputs.
I have had some success by adding small ferrite rings on signal inputs.
I sometimes prefer build away from the pcb some call this air wire technique.
I found out I needed to keep my signal inputs away from ground planes.
I clean and coat the traces and bare wire with appropriate coating.

[haklesup]

Ferrite beads are an effective ESD suppression technique. They form a low value inductor, this causes the impedance of the wire to go way up when a fast rise time signal is applied to it. Ferrite beads are also used to suppress EMI noise injected into and emitted by circuits via cables.

Ground planes usually don't disturb digital inputs but on analog signals, they can cause unwanted extra capacitance. Generally avoid having a ground plane under a summing junction before an op amp input.