How do you mean 1A is low enough?I dont see how this is related to the load dump spike.
Takes two to cause a problem, a source and a load. If you simply put an LDO in front of the load -- one that can handle the voltage drop and energy dissipation -- you'd be set, nothing even happens, it just completely rides through it.
Easiest way to do this is a small zener shunt regulator, plus depletion MOS follower. At nominal voltages, the zener isn't in breakdown, so the DMOS gate is at Vin, and it's turned on (DMOS is "normally on" and takes negative Vgs bias to turn off). As input rises above Vz, gate voltage gets clamped, and source voltage follows maybe a volt or two higher (depending on load).
Or a PMOS with control circuitry. Or a proper LDO, which does the same thing, but has a proper voltage reference -- not that you need an accurately clamped voltage, that's just a bonus.
If your input peak truly is 200V, and the load voltage is maximum say 18V, and load current maximum 1A, then the limiter would have to dissipate (200 - 18V) * 1A = 182W peak, for a pulse width of some hundred milliseconds. (It's an exponential decay rather than a square pulse --not as bad as it sounds, but it's still a sizable slug of energy for semiconductors to handle.)
You'd probably be looking at one or a couple TO-220s on heatsinks, or a TO-247 or the like.
If the peak voltage, or the load current, is lower, this looks pretty good.
A switch is definitely an attractive option for higher power levels. PMOS, with less control circuitry, would be fine here. 200V 5A something or other, I suppose. And there are controller chips if you don't mind paying the premium.
Note my first reaction is effectively to increase the load impedance. This is because the nominal load impedance is already higher than the load dump impedance (a few ohms), so we will absorb less than full power already, and can stand to absorb even less by increasing the impedance further. (Using a hand-waving meaning of impedance, such as the average: compare e.g. 12V / 1A = 12 ohm nominal equivalent, versus at full ride-through, 200V / 1A = 200 ohm equivalent.)
Going the other direction, sure, you can absorb the dump to some extent or another -- but you will dissipate far more power, unless you shunt it into a very low impedance (< 20mohm?) so that most of the power is dissipated by the alternator itself.
Which...
Is there a significant advantage to a SCR compared to a FET? And what why is a SCR better for popping a fuse than a TVS? I dont see that much difference in shot term max current when i look into the datasheets of similair sized components.
...If you choose a FET with low enough resistance, yes, that is still fine. You will need to sink some hundreds of amperes to clear a fuse (again, potentially less if you add series resistance to limit short circuit current), which is quite demanding for a MOSFET, but fairly pedestrian for even a D(2)PAK sized SCR.
The SCR doesn't need any additional circuitry, it's a latching device all by itself (just use a zener to set the trigger threshold); a MOSFET needs a latch, and also something to store drive voltage.
Note that, if you use a
shunt regulator rather than a latching crowbar, then you need to bear peak current times nominal voltage drop -- you're lowering rather than raising the equivalent load resistance, and you're not lowering it by very much so it better be one hell of a beefy shunt. Dropping in one of those fat ass TVS diodes looks very attractive versus cooking up a shunt regulator.
![Smiley :)](https://www.eevblog.com/forum/Smileys/default/smiley.gif)
In my case its a early 80's VW LT camper van. Not much electronics in there except for wat i already build in. The van already has a early prototype version of my circuit that runs of the AMS1117 regulator on a arduino. I think it has a single series diode and and a 5A fuse for protection, but that's about it. I'm doing a V2 with an actual PCB and a improved circuit. This is mainly a one-off for myself, but i'll be making a few extra to sell/give away to friends. If it works well, and people are interested i might start selling them. If that doesn't happen i will be making it open source. That is why i'm looking for a affordable but robust solution. My van might be very old, but others may want to install in a even much older camper vans.
I dont need it to support 24V systems. 12V is fine. I'm designing everything to work in the 10 tot 16V range. If a fuse pops at 17V that is fine with me.
Then that'll most likely be 60V load dump. Not nearly as scary. A lot of parts are available to handle it, too! Might just change that '1117 and be done.
Is it
really a 1A load? That's a LOT of Arduinos.
If you have, like, a motor, or some lights, or whatever, something dumb on there, and that's what's drawing most of the load current? You may find it helpful to separate it. Motors aren't going to give a shit about load dump, they can use raw 12V, who cares. Lights will glow brightly and probably not die, or can be designed to handle it more gracefully with less effort. Meanwhile, that Arduino using, whatever, 3.3 or 5V at some 10s of mA, a slightly beefier regulator isn't going to mind passing that while momentarily dropping 50V.
Lots of design opportunities, keep your mind open to them. No need to hide everything inside a black box!
![Wink ;)](https://www.eevblog.com/forum/Smileys/default/wink.gif)
Thanks for sharing. I'm sure there are many standard and not all of them require a full battery disconnect lead dump. Lets be honest, this very rarely happens and when it does happen, its likely that many other electronics die too.
However, i am trying to make it as robust as i can without making things to complicated of expensive.
For one-offs, that are neither terribly expensive, nor terribly mission-critical, it's hard to justify anything beyond basic fire protection (i.e., it shouldn't burst into flames under these conditions), really. You may find that "robust" and "complicated/expensive" just happen to intersect at the point of zero additional design effort.
![Grin ;D](https://www.eevblog.com/forum/Smileys/default/grin.gif)
Tim