Protecting a little PIC against over voltage with minimum parts

Comments

• TonyWaite Posts: 219

  2010-04-03 21:41 edited 2010-04-03 21:41

  Hi Rich,
  
  A three-terminal 5V regulator plus two 'optional' input/output caps is nonetheless attractive,
  because of the high-level of protection offered; but a pair of silicon diodes in series would be
  easier and would probably do the job.
  
  Would the 7V be an instantaneous transient or permanently applied? If just a transient, then
  a beefy transzorb diode connected reverse-bias between VDD and VSS could be the answer.
  Experience has taught me that a current-limiting resistor is a wise addition too, in case the
  overvoltage is prolonged.
  
  Regards,
  
  T o n y
• Leon Posts: 7,620

  2010-04-03 21:53 edited 2010-04-03 21:53

  PIC inputs have protective diodes that can take 20 mA, IIRC. I'd use 1k5 or 2k.
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  Leon Heller
  Amateur radio callsign: G1HSM
• W9GFO Posts: 4,010

  2010-04-03 22:32 edited 2010-04-03 22:32

  It is for a servo controller. Nominal 4.8v or 6.0v but a fully charged 5 cel pack is more than 6.0v, not to mention that it would be good if it could handle a 2 cell liPo - but that isn't a requirement.
  
  A regulator would triple the size of the board. It is very tiny, all SMD.
  
  If VDD has a couple diodes ~1.0v drop, does that cause a problem for the input which may be 1v higher?
  
  A 200 ohm resistor on VDD and a 1K on the input would limit the current to about 42ma going through the PIC. If that would not protect it - why not?
  
  Rich H
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  The Simple Servo Tester, a kit from Gadget Gangster.
• kwinn Posts: 8,697

  2010-04-03 22:33 edited 2010-04-03 22:33

  W9GFO, No. If you put a 5 to 6V zener diode in parallel with the PIC in addition to the resistor it would.
• Leon Posts: 7,620

  2010-04-04 00:03 edited 2010-04-04 00:03

  As I said previously, I think that 20 mA is the maximum for the protective diodes. I'd use less than that, to be on the safe side.
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  Leon Heller
  Amateur radio callsign: G1HSM
• Peter Jakacki Posts: 10,193

  2010-04-04 02:57 edited 2010-04-04 02:57

  Rich,
  
  If a chip has an ABSOLUTE maximum then don't even go close to it. However with regards to the supply voltage the biggest problem to put it in a simplistic nutshell is that the precarious isolation between "components" on the same silicon substrate usually consists of less than microscopically thin layers of "insulation" and "reversed biased" wells or junctions. So unlike most I/O pins etc even if you limit the current through the supply you can be upsetting the very foundation of the circuit, not just an I/O pin.
  
  Looking at your reasoning though I am dumbfounded when you say adding a regulator would take up a lot more room. I use tiny SOT-23 pack LDO regulators all the time and they hardly take up any room. You can even forgo the input cap and just use a small 1uf on the output for this type of load.
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  *Peter*
• W9GFO Posts: 4,010

  2010-04-04 04:14 edited 2010-04-04 04:14

  Peter Jakacki said...
  Looking at your reasoning though I am dumbfounded when you say adding a regulator would take up a lot more room. I use tiny SOT-23 pack LDO regulators all the time and they hardly take up any room. You can even forgo the input cap and just use a small 1uf on the output for this type of load.

  
  It was a guess but the space currently occupied by components is less than .04 in. sq. (.2" x .2").
  
  I suppose I could use the other side of the board...
  
  Rich H
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  The Simple Servo Tester, a kit from Gadget Gangster.
• Peter Jakacki Posts: 10,193

  2010-04-04 04:26 edited 2010-04-04 04:26

  It is a bit tiny, isn't it. But at least you won't have to worry about over-voltage then. Zeners etc take up just as much room so it's much better to go LDO instead.
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  *Peter*