Automotive Electronics Protection Circuit - Thoughts

[JayDee]

Hi Guys,
I would appreciate your thoughs on the following.

Design for a gerneral purpose protection circuit followed by 5 volt regulator for use in Automotive microcontroller applications.
The holy grail would have a small foot print on the PCB, pretty bulluit proof and a sensbile cost! 
...and not hyper complex if I can help it!   

The system should provide a level of protection for.

Sustained over current draw
Transient voltage spikes
Sustained over voltage
Reverse Battery

A preposed cirucit built from individual surface mount parts is.
]
Regulator
Micrel LDO MIC2920A Reg. A 5V 400mA regulator, with reverse protection to -20V, Transient protection to 60V, internal current limiting. Which is a great all round package but I still feel I should provide some front line protection as well. Will also need to provide enough heatsinking on the PCB. It will need the usual caps on the input/ouput but i think I can use 1206 SMD ceramics.

The Front line protection would consist of

Fuse
PTC resettable, 16V, 350mA Hold, 750mA Trip. I would preffer a high operating voltage and PTCs always have a significant gap between the hold and tripping current so you need a serious problem before they will trip But what else to use?
I could use a standard SMD fuse but its 'Return to Base' if any thing happens for repair, which is no good really. And some devices will be potted solid 

Capacitor
A 33uF electrolytic - provides some basic filtering.
But what voltage ?? The lower the voltage gives a smaller the package but I dont want it to pop. The supply will rarely exceed 14V so can I get away with a 25V Cap? Or will a transients blow it ?

Zenner Limiter
18V 3W Zenner Limiter - Anything over ~18v will bleed off, if the over voltage is sustained and excessive, the PTC resettable fuse (FS1) should (eventually!) trip.
In a reverse battery condition the Zenner will conduct forwards and the PTC should trip.

TVS (transient voltage suppressor)
Vishay TransZorb - SMAJ24A 400w, Breakdown 26V, Clamp Voltage 39V. This should catch any of the nasty fast transients.


SO... I'm looking for a robust but realistic design that I can move between various small projects. Is the above overkill / not enough / Just wrong!
Thoughts?

I've not found any 'all in one' protection solutions for automotive applications, anyone heard of one?

Cheers for your input, John. 

[Psi]

Looks good to me, that's more protection than i put into my car electronics.
I just had the fuse and 18V TVS.

[Frangible]

Anything connected to an automotive power system has to tolerate 60V spikes and inverted power.  Much of this happens if someone disconnects the battery from the system with the engine running while the alternator is under load and when they hook the battery up backwards.  Filtering is important to remove alternator hash and all the other noises produced by the switching of inductive loads around the system (e.g. injectors, HV ignition).

[Psi]

I cant say i've heard of anyone who connected a car battery around the wrong way.
I'm sure it happens and ya should protect against it, but spike protection is far more important than the polarity protection.

Over 90% of the car's i've seen use different size terminals for positive and negative, so they just don't fit around the wrong way.

[ejeffrey]

Since you have plenty of headroom for you regulator I would just put a diode in rather than counting on the PTC tripping for reverse polarity protection.  This will also prevent reverse current flow when the supply voltage drops during startup.

You may find you need more input filtering, that depends on your application.

Make sure everything is rated for a wide temperature range, and that you derate the power dissipation for the maximum ambient temperature.  In the worst case you may need a temperature sensor to keep everything in standby if the temperature gets too high.

[ToBeFrank]

I cant say i've heard of anyone who connected a car battery around the wrong way.

I think the typical case of this is when a car needs a jump and the person doing it hooks the cables up the wrong way. I've personally seen this happen.

[Chet T16]

Yeah i've seen it done a few times, only once causing damage but typically that was something fancy (Lotus elise)

[Psi]

I cant say i've heard of anyone who connected a car battery around the wrong way.

I think the typical case of this is when a car needs a jump and the person doing it hooks the cables up the wrong way. I've personally seen this happen.

heh true, that would be quite easy to do.

[JayDee]

@ejeffrey, yeah good point even under cold cranking conditions etc.. and should always have a voltage reasonably over that required by the LDO so a diode would give solid reverse bat protection.

I agree that reverse bat is not common but embarrassing if your device goes POP and everything else in the car recovers.
so I'll certainly have something to protect against it.
do you think a 25V electrolytic will survive any spikes? I am I going to have to put a bigger can on the PC and go 50V or higher?

I think I'm getting close to just getting it on the PCB and then doing some field testing. No one has blown the design out the water so cant be too far off. 
J.

[Simon]

how about a reasonable inductor or resistor in series with the cap. If you know how much current you need you could calculate a decent resistance to have a few volts of drop across it, that way any spike (hi frequency component) would not get much pas the RC low pass filter formed by the resistor and reg input cap

[ChrisKiwi]

I had a look at http://www.vishay.com/diodes/protection-tvs-esd/automotive-tvs/ and found that your suggested TVS is at the lower end of the automotive ratings.  I use to work for a company that made LED lighting for automotive markets who used TVS like the SM8A27.  There are specified limits to what the voltage (not spikes) can get up to.  I can't remember what the "12V" range was but the "24V" range went up to 33V so our multi-volt products worked over the range 9 to 33 V.  Also, you need to consider the energy behind the spikes (hence the SM8A27) as well as the actual peak voltage.  For example, we used S1M diodes on the input lines (supply, control lines), so I'd suggest a diode before the fuse.  For input Cap, I'd use a lower value non polarized ceramic with a higher voltage rating.  You could always use more than one.  Two smaller values might be better than 1 cap of twice the value when you take X_c into account. If the Reg requires some minimum Cap on the input I'd put that right at the input of the Reg.  It was also common to see a low value resistor in the ground line (as the last component before connection to vehicle ground) of 10 ohm, but there was always some debate on the worth of a resister of that value.