Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
02/11/09 15:41
Modified:
  02/11/09 15:56

Read: times


 
#162286 - Shunt ESD currents to the metall enclosure!
Responding to: ???'s previous message
Todays' micros are indeed less robust than buffers in HCMOS, e.g., just because the higher level of miniaturisation on die. Protecting pn-junctions or Schottky-junctions are smaller and less capable of shunting ESD. Nevertheless, there's no reason to add buffers when using todays ESD and filtering components.

There are two ways to handle ESD: 1. To prevent ESD from hitting the circuit at all, by totally isolating the device, which can be very difficult, or 2. To allow ESD hits, but to shunt the corresponding ESD currents "arround" the sensitive parts.

In your situation I would recommend the following:

Don't forget, that ESD is usually something that wants to flow back to earth. By other words, the ESD event entering your device always tries to leave your enclosure again. So, offer it a proper way to the outside world and your ESD is happy.

Do not connect the cables from keyboard and other circuitry contaminated by ESD directly to the micro. Choose a place a few centimeters apart from sensitive parts, best at the edge of your PCB, where you have a connection between your ground plane and the metal chassis, even if this is only via a (SMD!) cap (or better, SMD-caps!). Have ESD absorbers for all the cables mounted here with the grounds' pins connected to this place. The ESD absorbers can be SMD-caps, SMD-varistors, SMD-transzorbs, or the new SMD-CeraDiodes, which I more and more prefer because of their robustness. Avoid any pig-tailing at the ESD-absorbers at this place!!!!!!!

Now, after this rough limiting you should add a fine limiting: Use some series impedance in conjunction with Schottky clamps to shunt the remaining overvoltage to the supply rails. A BAS70-04 can be used, e.g., which already contains two Schottky diodes in series in a tiny SOT-23 package.

But take care, don't forget SMD decoupling capacitors from Vcc to ground right at the Schottky clamp to allow the current spike to flow to the ground plane and finally to the metal enclosure. So, the Schottky clamps shoud also be mounted near the ESD absorber section.

Take care, this section is highly contaminated with ESD current spikes. Even when using a multilayer-board with several ground planes and an armada of 0603-SMD-absorbers and -caps the noise on this piece of ground will be high enough to erode noise margins. So, keep this section entirely free of any sensitive electronics. Do not route any signal line across this section. Just imagine where the ESD currents will flow and you will be able to fabricate a proper layout.

If you connect your device to another device via cable, choose a shielded cable and connect the shield on both sides to the metal enclosure. By this a Faraday cage is built, which keeps the ESD outside your application. Avoid pig-tailing here too!

You can also see the situation from another point-of-view: By shunting the ESD from the internal keyboard cable to the metall chassis you offer the ESD a much better (lower impedant) way to earth (even if it "sees" earth only via stray capacitance) as if it had to move across your circuit. By this you keep ESD outside of your circuit.

Also avoid "secondary ESD" issues. So, don't route the "hot" keyboard cables too close to the micro.

Kai



List of 9 messages in thread
TopicAuthorDate
to buffer, or not to buffer?            01/01/70 00:00      
   Consider the usage environment            01/01/70 00:00      
   ESD on buttons            01/01/70 00:00      
      HC's and a resistor and a cap...            01/01/70 00:00      
         reply            01/01/70 00:00      
   Shunt ESD currents to the metall enclosure!            01/01/70 00:00      
      Reply            01/01/70 00:00      
         Hae you considered 74S1053?            01/01/70 00:00      
            Reply - 74S1053            01/01/70 00:00      

Back to Subject List