Synth-DIY Yahoo! Groups Archives

Re: ATMega16 I/O port protection diodes

2010-07-07 by David Reaves

Well, I sent this message well over two weeks ago, so I'm not sure why  
it's being posted now, but, any way... :-)

Why not? The idea is to protect the microcontroller from a non-common  
mode surge on its I/O pins.

That is what an optocoupler is designed to do: as well as isolating,  
in practice, the microcontroller 'sees' any voltage spike only on the  
ground (common) connection, which is easily protected by proper  
earthing practices.

So even if the distant signal supply and the microcontroller power  
supply are on the same power outlet, any surge that comes back through  
the cable would not go into the I/O pin of the micro controller.

David Reaves

Show quoted textHide quoted text

On Wed Jul 7, 2010 12:59 am ((PDT)) "Tim Mitchell" wrote:
> -----Original Message-----
> From: AVR-Chat@yahoogroups.com on behalf of David Reaves
> Sent: Sat 19/06/2010 17:07
> To: AVR-Chat@yahoogroups.com
> Cc: David Reaves
> Subject: [AVR-Chat] Re: ATMega16 I/O port protection diodes
>
>> This is exactly the purpose logic opto-couplers ("opto-isolators")
>> were invented for. Complete isolation, of several kV.
>>
>> You can buy them in SMD and DIP packages, in singles, dual and quad
>> configurations. A couple dollars and you're safe.
>> Very cheap insurance.
>
>
>
>
> No, the original poster was wishing to power an LED (for a railway  
> signal) down a long length of cable, without needing power supply at  
> the far end. An optoisolator is no use in this situation.
>
> -- 
> Tim Mitchell

Re: ATMega16 I/O port protection diodes

2010-07-27 by John

Proper grounding practices can be a bit tricky. 

Years ago we used to encapsulate gas discharge tubes for lighting protection on telephone lines at subscriber premises with the aim of protecting the subscriber, premises and telephone equipment. Lots of spectacular stuff in testing if the batch was bad. Generally would work well on line powered (and earthed) equipment although with a direct hit all bets were off.

Talking to the engineers involved with this there was an issue with surges travelling at different speeds through air, ground and copper. Might have power and phone lines traveling along a road initially with feed offs taking quite different paths to a building a mile off the road (we have active, neutral lines with power ground at the premises). 

Net result is surges of different polarities arriving at marginally different times.

--- In AVR-Chat@yahoogroups.com, David Reaves <rrsounds@...> wrote:

Show quoted textHide quoted text

>
> Well, I sent this message well over two weeks ago, so I'm not sure why  
> it's being posted now, but, any way... :-)
> 
> Why not? The idea is to protect the microcontroller from a non-common  
> mode surge on its I/O pins.
> 
> That is what an optocoupler is designed to do: as well as isolating,  
> in practice, the microcontroller 'sees' any voltage spike only on the  
> ground (common) connection, which is easily protected by proper  
> earthing practices.
> 
> So even if the distant signal supply and the microcontroller power  
> supply are on the same power outlet, any surge that comes back through  
> the cable would not go into the I/O pin of the micro controller.
> 
> David Reaves
> 
> 
> On Wed Jul 7, 2010 12:59 am ((PDT)) "Tim Mitchell" wrote:
> > -----Original Message-----
> > From: AVR-Chat@yahoogroups.com on behalf of David Reaves
> > Sent: Sat 19/06/2010 17:07
> > To: AVR-Chat@yahoogroups.com
> > Cc: David Reaves
> > Subject: [AVR-Chat] Re: ATMega16 I/O port protection diodes
> >
> >> This is exactly the purpose logic opto-couplers ("opto-isolators")
> >> were invented for. Complete isolation, of several kV.
> >>
> >> You can buy them in SMD and DIP packages, in singles, dual and quad
> >> configurations. A couple dollars and you're safe.
> >> Very cheap insurance.
> >
> >
> >
> >
> > No, the original poster was wishing to power an LED (for a railway  
> > signal) down a long length of cable, without needing power supply at  
> > the far end. An optoisolator is no use in this situation.
> >
> > -- 
> > Tim Mitchell
>