Russell McMahon wrote:
>>> The 78xxx family are not "automotive rated" in the normal sense of 
>>> the
>>> term.
>>> While they can be used with an eg automotive 12V supply as input 
>>> they
>>> would require protection against the normal nasties that occur in 
>>> that
>>> environment. Some (only) of which include load dumps, polarity
>>> reversals, high energy spikes to "most interesting" voltage levels,
>>> ... .
>>>       
>
>   
>>> I'd
>>> start with a series diode for reverse polarity protection and TWO
>>> series resistors with a zener diode clamp to ground after each
>>> resistor. Add a few caps and you have something which will survive
>>> most "events".  We did something similar in a -50V Telecom 
>>> environment
>>> long ago. The two stages of zener clamping do wonders to eliminate
>>> transients which are almost unstoppable with a single stage.
>>>       
>
>   
>>> The beautiful LM2936 would be easier, but its far from cheap.
>>>       
>
>   
>> Because of the -50V to +60V input transient survival protection ?
>>     
>
> Because of all the things in the LM293x data sheets that they say they 
> are protected against which they put there because of what they must 
> have learnt the hard way over the years in automotive environments. 
> Looking at the list in the data sheets gives an excellent guide to 
> what you need to do if you are 'rolling your own'.
>
>   
>> That can be solved on 7805 too.
>>     
>
> Yes.
> As I noted.
> By using the methods I suggested, or something equivalent.
> The user has to decide, as I also noted, whether the end result is 
> small or cheap by their standards.
> A single zener and resistor clamp will NOT remove really energetic 
> fast rise time spikes well enough to let an older style voltage 
> regulator such as the 78xx handle them without passing spikes through 
> to the output. Processor reset systems thrive on spikes. While one can 
> try to brute force filter these you are then into the normal situation 
> where spikes have their own rules with an RF mindset. Using the two 
> zener clamps in series that I described is a proven-in-practice 
> solution. In a telecommunications environment with a -50V battery 
> supply and electromechanical switches you get many of the challenges 
> that will appear in an automotive environment. At least people don't 
> USUALLY insert the exchange battery backwards :-).
>
> As a bonus, the two series resistors necessarily drop some voltage in 
> normal operation and thereby reduce the dissipation in the 78xx. A 
> 7805 needs about 2 V of headroom. As long as you don't want your 
> circuit to hold up during engine cranking (and if you do then there 
> are "problems" ) you can drop about 3 volts in the series resistors 
> and still survive most battery conditions. Having a largish filter cap 
> after the series diode that I mentioned originally will give you 
> holdup during the odd major glitches and also during cranking if 
> dimensioned to suit.
>
>
>   
Yes, Russell, and the series R will also reduce somewhat the energy of 
the spikes which are always present
in an automotive environment. From my law enforcement automotive 
designs, the  input power looked
like this: brute force inductor as spike filter, followed by a small R 
(about 1 ohm) followed by a TVS
(usually PK6xx type) across the input, followed by a 1N4005 diode, then 
a large electrolytic, followed by the LM7805. The
electrolytic was large enough to cover starting droop, and the diode 
prevented discharge back into the battery.
 From that point on, things were quiet. No problems this way.

Watch where you place the inductor;  it needs to be away from the  PCB 
(or away from the active circuitry).

--Bob

>
>         Russell
>
>
>
>   

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist