Hello all,

What is the best way to add over voltage protection to 19 inputs?

I should be getting 0-6~ volts input but due to the connector being a mains adapter there is the possibility of it being connected to the mains, this is a design feature I cannot change, so built in protection in the circuity is the only path I can see.

Any suggestions or links you can point me to?

Thanks in advance

Sol

 

The only thing I can think of is zeners and resistors, connected as you'd expect. The number of inputs makes this a minor hassle, but if you use SMD parts, it shouldn't take up that much space. I'll be interested to see what others suggest, though.

 

Cast it into a concrete 1 meter cube. Then let's see them try to plug it in.

 

Heh, thanks sgtwookie

Spacewrench, Is there a name for that technique so I can read up more info?

Thanks both

Sol

 

I was only half-kidding.

You haven't described what the inputs are connected to, with the exception that only a range of 0v-6v is expected. Are these high-impedance inputs?

Clamping diodes or Zeners, and high-value resistors would help the situation, but all diodes have leakage current which might spoil measurements.

All I can do is assume that the connector might accidentally be plugged into 220v 50Hz AC.

Can you better describe the inputs?

 

The connector is a http://www.socapex.com/products.htm#SL61 panel mount connector, wired into shift registers *you may remember from another thread*

Yes, incorrectly connecting could send 230v at 60hz down the lines. The correct voltage is a logic high or low around 5v

Is that any help?

 

OK, I completely forgot about that project.

Clamping diodes and VERY small fast-blow fuses is all I can come up with at the moment.

 

Dam, I was hoping for a way that wouldn't require any work done to the board after over voltage being applied, If it helps, I could narrow the scope to power coming from 6 pins and could be routed on over voltage to 6 other of the pins.

I guess some kind of relay that would switch out the rest of the board on over voltage...

I'm going down the over voltage protection route again as a sizable amount of people try to connect to the mains erroneously with similar circuits.

Sol

 

I think we discussed this in the original thread.

Put a warning on the box, and use sockets for the shift registers, resistor arrays etc. Believe it or not, the shift registers will actually be cheaper than fuses.

[eta]
The original thread:
http://forum.allaboutcircuits.com/showthread.php?t=27922
My very last post on page 4, reply #35:

You can't protect against complete foolishness by a user. You wouldn't be able to use diodes to protect it from that kind of error without preventing it from functioning at all - unless perhaps you used diodes with such limited current handling capacity that they would burn out before the IC outputs fried. Actually, it would be more like a small explosion; their lids would be blown clean off.

I suppose you could use resettable fuses like PPTC's, but that would add complexity and lots of cost. Cheaper to use IC sockets and keep spares around.

 

xxx, I was hoping for a way that wouldn't require any work done to the board after over voltage being applied, If it helps, I could narrow the scope to power coming from 6 pins and could be routed on over voltage to 6 other of the pins.

Have those boards made yet?

I guess some kind of relay that would switch out the rest of the board on over voltage...

Relays would not respond nearly fast enough.

I'm going down the over voltage protection route again as a sizable amount of people try to connect to the mains erroneously with similar circuits.

If you want to make your piece of test equipment more or less "bulletproof", you're going to have to re-think the entire project.

 

If you want to make your piece of test equipment more or less "bulletproof", you're going to have to re-think the entire project.

Dam, I guess your right...

 

OK, so which pins of the cable could possibly be connected to the mains?

And of those pins, which might be hot, neutral and ground?

 

There are many ways to protect i/o lines, the most approriate one depends on the type of line you are dealing with and the type of accidental event that you would like to bear without damages.
Assuming you have to deal with signal inputs in the attached pdf there are some suggestions whose applicability depends on the surge voltage you need to withstand and on the distorsion you can accept on your input signal.

1)
A diode pair in ladder configuration after a series resitor is a common practice useful for relatively slow logic inputs. The more the resistor value is high (usual between 1k and 10k) the more the protection is efficient but fast pulse fronts are filtered out by you input capacitance and stray capacitances. To imporve protection add also a transient voltage limiter on the power supply rail the diodes are pointing to.

2)
when I am desperate.... I omit the diodes and let the bulk protection diodes of the TTL or CMOS ic to do the protection work. In this case a higher resistance and a transient voltage limiter on the IC power supply rail is a good practice.

3)
using a TVS diode to clamp the overvoltages can be a solution, but it is generally more costly if you have many inputs to protect simultaneously. Stray capacitances are higher too.

4)
if you have to deal with analog inputs, a diode limiter is a fast choice, and if you use schottky like BAT54 you can even protect RF signal inputs. In this case the input resistance must be very low, better if substituting the resistor with a ceramic capacitor of few picofarad.

 

Cast it into a concrete 1 meter cube. Then let's see them try to plug it in.

Silly man! You should know that no idiot-proof device is immune to the ingenuity of idiots!

Or...more succinctly......Artificial Intelligence is no match for Authentic Stupidity.