Circuit to convert from 12V automotive into a microcontroller Voltage?

[jwhitmore]

Certainly a beginner question ;-) I want to monitor an automotive 12V signal, (so that signal voltage can go as high as 14.5V, say 15V under charging) with a micro-controller. That's easy I thought and considered using a voltage divider and a zener diode for protection of the micro-controller.

Then out of the blue I thought that perhaps the simpler solution would be to use an OpAmp as it will have a very high input impedance, (I think). Then if you have a 5V or 3v3 micro-controller you just have the supply voltage of the OpAmp set to either of those voltages and hey presto. I'm using this for simple logic I want to see when a circuit is switched on or off and measure timings of various circuits.

The more I thought about that solution the more confused I got. It sort of sounds good but OpAmp circuits usually have feedback, so you'd be feeding say 3v3 back against another OpAmp Input which is 14.5V. That sounds like a recipe for disaster. Well maybe not disaster but the OpAmp would be doing all in it's power to make the 3v3 Supply voltage look like 14.5V.

So is that a stupid idea and why, I might learn something? And is there an even better circuit I've not thought of? Remember this is the beginner section so go easy! ;-)

[Ian.M]

Unfortunately, it is a stupid idea, because automotive signal voltages are often *very* noisy and 100V noise spikes will kill the OPAMP very quickly - in fact it will probably die at the first engine start.

You can get specialist OPAMPs with inputs rated for 'over the top' operation up to about 40V above their positive supply rail, but the ones I've seen aren't rated for operation with the inputs below their negative supply rail, so they'd still need protection circuits in front of them.

High impedance isn't particularly desirable anyway - it increases problems with noise pickup, and no significant current increases the risk of contact problems at any connectors.  If the circuit you are monitoring can tolerate it without disturbance to its operation, its usually preferable to draw about 10mA through the potential divider.  Exception: permanently powered circuits, where adding a few extra 10mA draws may result in excessive battery drain if the vehicle is unused for several weeks or longer.

Sorry - its back to a potential divider and clamping, with a series resistor between the clamped divider output and the MCU pin to limit the input current if the clamping is overloaded during a spike and a capacitor to 0V at the pin to filter any HF spikes that make it through the clamping circuit.  However beware of the RC time constant of the filter capacitor with the input resistor network - it needs to be low enough to be small compared to the signal timings you need to measure.

[Benta]

Automotive is a very harsh environment.  During what is called a "load dump", transient voltages up to +/- 60 V can occur. If you check the data sheets for automotive voltage regulators, this is usually what is specified. It can also happen during jump starting.

I agree with Ian that the old-fashioned voltage divider with input protection is the way to go.

[towlerg]

I'm also interested in this, I hope you post a circuit when you get it working.

[jeroen79]

How about an optocoupler?

[Ian.M]

Yes, a load dump transient is *NASTY*.   Unclamped, for a 12V vehicle, it can peak at over 100V and last for 0.4 seconds.  That will put enough energy into a potential divider carrying 10mA  to blow out anything except 1W or higher pulse withstanding resistors (with a 10x, 1 second overload rating - 5x is marginal).   If you drop the current to 5mA, you can get it down to what a 1/4W resistor can probably survive, but it will still need to be pulse withstanding.  Therefore it may be worth looking at alternative input protection circuits that limit the input current rather than clamping the input voltage.

There's a page giving LTspice XVII simulation models for automotive transients, including conveniently tabulated peak voltages, durations etc at: http://www.analog.com/en/technical-articles/ltspice-models-of-iso-7637-2-iso-16750-2-transients.html 
If you are still using LTspice IV you are S.O.L. as it doesn't have the built-in models for the transients, so you have to approximate the waveforms with compound exponential pulse sources etc.  See: https://www.eevblog.com/forum/projects/moddeling-load-dump-pulse-5a-in-ltspice/

An optocoupler brings significant benefits of ground isolation - as automotive systems frequently use the vehicle body for ground return, the combination of high currents and slight corrosion of fasteners and bolted joints or seams can case significant ground offsets.  Worst case: if your device looses its ground connection, it could see negative input transients of several hundred volts on top of a 14V to 15V continuous negative level, relative to your local 0V rail while unpowered.  Alternatively the signal wire ground could end up carrying your device's full operating current or even act as the ground return for some other high current system that has a ground fault.   As ground wires are usually not fused, there is a real risk of a wiring loom fire which can lead to total loss of the vehicle.   Optoisolation avoids all that, as signal grounds can be kept independent.   
 
I posted a current limiting circuit for driving an optocoupler from a 24V signal, protected to +/-100V using cheap 'jellybean' parts a while back.
https://www.eevblog.com/forum/beginners/best-approach-24vdc-digital-signal-with-5v-logic/

If you replace D2 with a 10V Zener, D3 with a 4.7V Zener, and R2 with 390R, it will work with 12V signals, with negligible LED current below  6V and the knee of its characteristic (full on) at 10V, and you'll retain the 100V protectin rating.  Add a 1nF capacitor across D2 to improve its behaviour if exposed to fast high voltage transients.

However worst case automotive transients can exceed +/-100V, so for 'bomb-proof' input protection you'll either need some clamping upstream of it (e.g a unidirectional TVS diode, with a clamping voltage of about 75V so it *ONLY* has to handle rare extreme spikes) or replace Q1 and D1 with parts with a higher voltage rating.  Don't forget to fuse any input wiring that connects to a signal that can source a high short-circuit current at the signal source! 

N.B. Q1 is a Darlington so finding a 'jellybean' HV replacement may be difficult.  An ordinary transistor would need a much smaller R1 to get enough base current without excessive voltage drop, and that would then need to be pulse rated with a higher power Zener.   

[Scrts]

Certainly a beginner question ;-) I want to monitor an automotive 12V signal, (so that signal voltage can go as high as 14.5V, say 15V under charging) with a micro-controller. That's easy I thought and considered using a voltage divider and a zener diode for protection of the micro-controller.

Then out of the blue I thought that perhaps the simpler solution would be to use an OpAmp as it will have a very high input impedance, (I think). Then if you have a 5V or 3v3 micro-controller you just have the supply voltage of the OpAmp set to either of those voltages and hey presto. I'm using this for simple logic I want to see when a circuit is switched on or off and measure timings of various circuits.

The more I thought about that solution the more confused I got. It sort of sounds good but OpAmp circuits usually have feedback, so you'd be feeding say 3v3 back against another OpAmp Input which is 14.5V. That sounds like a recipe for disaster. Well maybe not disaster but the OpAmp would be doing all in it's power to make the 3v3 Supply voltage look like 14.5V.

So is that a stupid idea and why, I might learn something? And is there an even better circuit I've not thought of? Remember this is the beginner section so go easy! ;-)

There is a critical requirement missing here: how much of the current you will need?

[David Hess]

With some exceptions, the operational amplifier input requires the same protection that the microcontroller input would require.  Some operational amplifiers have a common mode input voltage range which extends way above their positive supply but I would still include external shunt protection.

The advantage of using an operational amplifier to buffer the input signal is that its very high input impedance makes input protection and filtering easier.

[jwhitmore]

Firstly OMG and then thank you all for your contribution.

I was planning to power the micro-controller with a SMPS based on a ADP2303ARDZ (http://www.analog.com/media/en/technical-documentation/data-sheets/ADP2302_2303.pdf) protected by a TPSMA6L15A (http://www.littelfuse.com/~/media/electronics/datasheets/tvs_diodes/littelfuse_tvs_diode_tpsma6l_datasheet.pdf.pdf) but now I'm starting to think that TVS Diode might not be enough. I have a prototype board, which I've only tested on a bench power supply so I might just spend my weekend trying to kill it in an automotive environment.

I have tried to get my head around LTSpice and I've tried a number of times, but I'm a SW engineer and I've never made it.

Back to the problem in hand and obviously optoisolators are the way to go, but according to that circuit by Ian.M (Thanks again) I'll still need a bit of protection. I'll have to do some more research on this one.

[jwhitmore]

Can I just add to my own thread again with another SW type question. Given the posts in this thread, and learning about how harsh the automotive is I decided that I'd search for a suitable opto-isolator which would make things simple. I came up with the Broadcom ACPL-87AT-OOOE (https://www.broadcom.com/products/optocouplers/automotive/isolation-amplifiers/acpl-c87at-000e)

Now this is where I get confused by electronics speak. The part is described as "Automotive High Precision DC Voltage Isolation Sensor" that sounds promising but then in a paragraph below it states "The ACPL-C87AT/C87BT features high input impedance and operate with full span of analog input voltage up to 2.46V.  The shutdown feature provides power saving and can be controlled from an external source such as a microprocessor." It's the 2.46V that gets me. What's Automotive about that? Did they just put that on so I'd find it in an Internet search?

Is half of electronics filtering out marketing speak and the other half "The Art Of Electronics" ?

[Benta]

I'll try to help.
There's automotive in the sense that you have to design for a harsh environment, where transients and under- and over-voltage are normal. This is a challenge for you as designer.
Then there's automotive in the way components are specified. This has to do with operating temperature range, reliability, chemical resistance, lifetime etc.
That a component is "automotive" does not mean can be used directly on a 12 V car system.

You're the designer, you need to think about what can happen and how you can protect your circuit.

[james_s]

If you're just trying to monitor the voltage then a resistor divider should be just fine. If you use sensibly chosen resistors, then the maximum output current can be quite low, even for a fairly substantial spike.

[Zero999]

Are you powering the microcontroller from the 12V automotive system? If so, then look for voltage regulators, with built-in load dump protection.
http://www.onsemi.com/pub/Collateral/LM2931-D.PDF

If that doesn't offer enough protection, then add a series resistor and transient suppression diode or xener, before the regulator. Of course both the resistor and suppression diode will need to be able to withstand the high energy pulse for long enough. A 5W zener and 2W ceramic resistor will probably do.


If all  you want to do is establish whether a 12V signal is high or low, then a potential divider will do. As long as the resistors limit the current to a safe level, then the microcontroller's built-in ESD protection diodes will prevent it from being damaged. In fact it may even be possible to get away with a single resistor and rely on the ESD protection, but that's not good practise, since it makes it more vulnerable to noise.

[malagas_on_fire]

Normaly in Automotive the capacitors are rated for 100V and even more when it electric cars comes in to place.

Chech this paper for more details if using the AEC Q200 for passive components:

http://www.knowlescapacitors.com/getattachment/69a6c2c9-aa2d-47b1-a464-2cf1518cf4c1/AEC-Q200.aspx

[Rick Law]

How about a long time friend - MC34063A / MC33063A / NCV33063A.

Granted, it is an old chip but it will do exactly what you need to power your MCU.

These chips do boost or buck and found in many CLA (Cigarette Lighter Adapter) car cell phone charger.  When I first started with my EE hobby, I disassembled a number of CLA chargers from my collection of old phone supplies so as to learn from them.  Most are MC34063A's.  I found it very easy to make a boost or a buck circuit.  So you can buck it to 5V or 3.3V easy.

While technically, the NCV one is rated for automotive and the MC one is not rated for auto, my Motorola RAZR V3 cell phone car-charger direct from Motorola came equipped with an MC34063A and not the NCV one.

Here are a couple of links to manufacturer datasheet.  I am sure you can find more,
From ON Semiconductor - sold at Sparkfun
^{https://www.sparkfun.com/datasheets/IC/MC34063A.pdf}

TI also make them, those with Q100 designation are rated for auto according to their datasheet here:
^{http://www.ti.com/lit/ds/symlink/mc33063a.pdf}

[james_s]

Normaly in Automotive the capacitors are rated for 100V and even more when it electric cars comes in to place.

Chech this paper for more details if using the AEC Q200 for passive components:

http://www.knowlescapacitors.com/getattachment/69a6c2c9-aa2d-47b1-a464-2cf1518cf4c1/AEC-Q200.aspx

I don't recall seeing 100V capacitors, although most of the "newer" cars I end up working on are late 90s. I know I've replaced 25V capacitors in ECUs on a number of occasions.

[Ian.M]

How about a long time friend - MC34063A / MC33063A / NCV33063A.

......

That still leaves you with a regulator with an Abs. Max input voltage rating of 40V.  If you don't have a TVS diode across its input, it wont survive an unclamped 12V load dump.    It will probably survive a load dump in a modern vehicle that has an alternator with built-in load dump protection.   However there are other possible sources of high voltage spikes so you'll probably need a TVS diode anyway.   If you don't want to have to extremely oversize the TVS diode so it can handle the maximum energy pulse the alternator can produce at full output current (i.e. sized to protect the whole vehicle), you need to limit the maximum current that can reach the device.   A big pulse rated wirewound resistor upstream of the regulator is your friend here though if you need much current for your device, the size and cost of such a resistor rapidly becomes impractical, so rather than TVS clamping, you will need to look at active over-voltage protection that works by disconnecting the input until it falls back into an acceptable range.   

As unclamped load dumps are fairly rare - not many poorly maintained older vehicles that are still on the road get retrofitted with modern electronics - it may be accceptable to have to cycle the power to recover from one.  In that case a 120V rated polyfuse upstream of the TVS diode may be satisfactory - it will let though short low energy spikes to be clamped by the TVS diode, but in the event of a full load dump will trip quickly enough to prevent the TVS diode shorting or exploding.

If you only need a lowish current to run a MCU and don't want the hassles of a switching regulator look at the LM317HV  - its capable of withstanding 60V on its input, which means you can tolerate a signficantly higher TVS clamping voltage.   

Back to the signal sensing issue - some automotive signals are fairly benign - if its a sensor that feeds the ECU, and is powered by the ECU, the odds are that the ECU will clamp any high energy power line transients before they reach the sensor.  However, if you are sensing a line feeding any motor or relay, fuel injectors or the ignition coil its unlikely to be so benign.   If you are using a potential divider type input circuit, always consider the worst case transients the input could experience, and make sure the input resistor(s) have a high enough voltage rating and can handle the pulse power at the peak of the transient.  Excessive miniaturisation is *NOT* your friend - its difficult to get a high enough voltage rating in smaller SMD packages.

[Rick Law]

How about a long time friend - MC34063A / MC33063A / NCV33063A.

......

That still leaves you with a regulator with an Abs. Max input voltage rating of 40V.  If you don't have a TVS diode across its input, it wont survive an unclamped 12V load dump.    It will probably survive a load dump in a modern vehicle that has an alternator with built-in load dump protection.
...
...

I am missing something.   So, those CLA phone chargers are time bombs waiting to blow?

Most of them I got from the manufacturers (like Motorola) and directly from the phone companies (ATT / Cingular, T-Mobile, Verizon) with manufacturer's brand.  Most of them have very simple circuits just like the sample circuit in the datasheet.  I have been relying on them rather heavily.  In fact, my USB charger in the car is a Verizon phone charger with original jack replaced by a USB connector.

Damn, you can't trust anyone.  I was expecting that the CLA's buck circuit would have "taken care of it for me."

[jwhitmore]

"A big pulse rated wirewound resistor upstream of the regulator is your friend here though if you need much current for your device, the size and cost of such a resistor rapidly becomes impractical, so rather than TVS clamping, you will need to look at active over-voltage protection that works by disconnecting the input until it falls back into an acceptable range.  "

At present my design doesn't have a resistor in series with the power into my circuit, so that's a mistake on my part. Thanks again for the education everybody. So my design provides USB Host mode to potentially interface with an Android Phone App. So that means I have to provide 500mA at 5V of USB 2.0. On top of that as this is a development uC board to which I wanted to add unknown expansion boards. I thought I'd better throw in a couple of Amps, just to be safe. That was why I chose the ADP2303ARDZ ( http://www.analog.com/media/en/technical-documentation/data-sheets/ADP2302_2303.pdf ) which can output 3 Amps. That and the fact that I could solder it by hand.

So if I went through my workings I want the SMPS to supply 3A at 5V which is 15W. So I'll want 15/12 = 1.25 Amps @ 12V. So  a Resistor of 12/1.25 = 9.6 isn't going to give me any sort of surge protection. :-(

So now I'm heavily leaning on the TPSMA6L15A (http://www.littelfuse.com/~/media/electronics/datasheets/tvs_diodes/littelfuse_tvs_diode_tpsma6l_datasheet.pdf.pdf) so perhaps a 9 might be enough to get me across the line. Given all I've learned here I'd love to just plug me prototype into a device which hit it with a suitable surge. I'm sure that device is the realm of EMC testing so not going to happen, but it'd be great to blow it up and know that yes my design was sh1t.

Perhaps I should have been looking at a linear regulator from the start but those things are inefficient so in my head they're not as good as SMPS. Obviously there's a very good reason for using them, but I'm a human and subject to stupidity. Inefficient and Ugly

[Ian.M]

Your maths is bad.  You have a supply voltage of 14V with the engine running and claim to need 1.25A @12V into your regulator.  That's 2V drop so your resistor cant be greater than 2/1.25 = 1.6 Ohms. 

I'm not even going to consider that you want 15W ( 1.25A @12V) as I know there isn't a snowflake's chance in hell of doing that with a reasonable size and cost series resistor so lets scale it back a bit

500mA @5V out is 2.5W.   Assuming 90% converter efficiency (probably a bit optimistic), that's about 2.8W which @14V in is 200mA.   That's getting into the region where a pulse rated resistor isn't going to be cost-effective.  Even if the MCU board only needs another 500mA (@5V), you'll probably run into trouble with the resistor, but lets run with 5.6W in for the sake of argument.

You *MUST* limit the TPSMA6L15A input voltage at no more than 20V (to stay under its 24V Abs. Max. rating and keep within its working range).  Lets say you need 5.6W into the switching regulator. You don't want more than about 1V drop across the resistor in normal operation, to keep the dissipation reasonably low and the efficiency reasonably high.  5.6W @ (14V -1V) is 0.43A, so you can't make the resistor over 2.3 ohms.   

If you run a sim with a worst case ISO 7637-2 Pulse 5B ^#  , with the vehicle's central clamping system voltage parameter Vs* set to 35V, into a 2.2 ohm resistor feeding a 5.6W constant power load, with 1000uF of bulk input capacitance clamped at 20V, you get a peak dissipation of 68W and about 22.5 joules dissipated in the resistor during the pulse.   To be certain your generic wirewound resistor survives that, its going to need to be a 14W one (x5 overload factor).   If you get a brand name pulse rated one  (x10 overload for 1s), you could use a 7W one.   If you dig deep into the manufacturers datasheets for I²t ratings, as the total overload duration is only about 400ms, you can probably find a 5W one that will survive and *MAY* even find a 3W one. 

However if you test against a worst case ISO 7637 pulse 5A,  (that's the one with no clamping), the peak dissipation is 755W  and about 140 joules dissipated in the resistor, so the odds are your resistor's element is now a rapidly expanding cloud of metal plasma  and its body is hot ceramic shrapnel bouncing around inside your device.     

YMMV with the results above as I cant run LTspice XVII to use its ISO 7637-2 pulse models I'm stuck with LTspice IV so have to use my carefully hand crufted ISO 7637-2 Pulse 5A and 5B sim.  Its as close as I can get it to the spec. but hasn't been independently verified.  Sim attached.

Also its *really* difficult to do high energy clamping at such a low voltage and still pass the x1.5 sustained input overvoltage test.  You need a TVS that's totally off under 18V, with a worst case clamping voltage while carrying the maximum possible peak current the supply can produce (which is equal to the maximum output current the alternator can produce unless you can limit it with a series resistor) that's under 24V.   If the clamping voltage is too high, your switching regulator fries.  If its rated holdoff voltage is too low, the TVS diode can fail during a jump start or if the alternator regulator fails and overvoltages the whole electrical system.  If you are lucky it fails shorted and the supply fuse to your device will blow. If you are unlucky, the TVS diode ruptures and no more protection.

TLDR: Above a couple of watts of load, series resistor + shunt clamping is impractical for such a low clamping voltage.  You'll need to use series MOSFET + over-voltage disconnect controller protection.   Stick a small 100V TVS diode shunting the power in in front of the MOSFET to tame fast HV spikes to what the MOSFET can handle.

# Yes, I know  ISO 7637-2 Pulse 5A and 5B have been superseded, but its what I have sims for.

*WHOOPS!*
I just spotted an error I made in the sim while tidying up the currently commented out 24V pulse 5B.  It should be:
Code: [Select]
.param Uss 70
and the also commented out 12V pulse 5A should be
Code: [Select]
.param Uss 1K
They got swapped. 

[Ian.M]

Given all I've learned here I'd love to just plug me prototype into a device which hit it with a suitable surge. I'm sure that device is the realm of EMC testing so not going to happen, but it'd be great to blow it up and know that yes my design was sh1t.

Physical load dump transient pulse generators don't come cheap or easy.  e.g. the ISO 7637-2 Pulse 5A peaks at 101V with a minimum source impedance of 0.5 Ohms, so that's fractionally over 200A into a dead short.   If you use a less capable pulse generator you'll never be sure whether the failure modes of your D.U.T. that you see are realistic.   Swinging 87V from 14V up 101V, peaking at 200A is extremely challenging - you might get there with basically a class D audio amp design with a 110V DC bus and driving 300A MOSFET modules in a H bridge configuration but I certainly wouldn't want to spend a year or so blowing up MOSFET modules developing it.

You could 'Ghetto' rig something with a modded 200A alternator from a scrapyard, belt driven from a 1 HP motor and a *BIG* flywheel to spin it at about 2/3 of its max rated speed.   You'd need to strip it down and test the main diodes in the diode plate to at least 120V at elevated temperature - if the leakage is excessive you'd need to replace them with higher voltage ones with the same current rating and remove any TVS protection.   You'd also probably need to replace the field current regulator with your own externally powered circuit.   Start with 14000uF of 150V rated capacitors across its output to get the risetime in the right ballpark. Load the alternator output up to 200A with the field current set to keep the terminal voltage at 12V (but don't do that for more than a minute or so or it will overheat) then suddenly remove the load to produce the load dump - think big water cooled bank of load resistors with a job lot of MOSFETS + gate drivers to switch them. The rough pulse shape is set by the alternator physical parameters and the capacitors  Add or remove capacitors to get 5ms 10% to 90% risetime.  Tweak the transient response of your field regulator circuit until you get a 101V peak pulse with the right decay curve to match the standard.   Combine it with a high current DC PSU using some beefy 150V diodes - worst case the one in series with the alternator must pass 200A peak, and feed the result to your D.U.T.   You'll need a really beefy TVS diode across  the alternator output if you want a clamped load dump pulse.   

If you aren't testing high current shunt protection circuits, you don't need the full 200A peak current.   Floating an AWG + a bridge tied load audio amp with +/-55V or greater DC rails, modded for DC coupled input could give you the pulse shape.   You need to float the amp + AWG so you can ground one speaker terminal without shorting the output stage, and you cant use a single ended amp because it woudn't have enough output swing for pulse 5A.  It needs to be DC coupled (+ disable the speaker protection circuit) so you can use it as a 14V regulator to power the D.U.T before and after the pulse.

Perhaps I should have been looking at a linear regulator from the start but those things are inefficient so in my head they're not as good as SMPS. Obviously there's a very good reason for using them, but I'm a human and subject to stupidity. Inefficient and Ugly

If its going to be dissipating more than about 1W a linear regulator is no longer cheap and easy.   Under about 100mA its a reasonable choice in an automotive application.   Over 100mA its probably a dumb idea unless you've got a large chunk of aluminium to heatsink it to that you want to keep warm anyway.

[james_s]

Some of you guys seem to be over-thinking this. Unless you disconnect the battery while the engine is running, you won't get a load dump. While the automotive electrical system is notoriously noisy, it doesn't take any heroic measures to clamp the usual spikes. If you were designing something critical like an ECU or airbag controller then you would want to account for absolute worst case fault conditions but this is typically not how equipment like car audio gear, navigation units, phone chargers and whatnot are made. A regulator with a 40V max input rating will work just fine unless there is something seriously wrong with the electrical system in your car.

[Ian.M]

In an ideal world, where everyone replaces their vehicle batteries as soon as they start to weaken, and regularly maintain their battery terminals, you are correct.

Unfortunately on poorly maintained older vehicles, loose battery terminals or ground straps are not exceptional so load dumps can be quite frequent - with the possibility of one whenever the headlights are switched off or an electric cooling fan shuts off etc. if the battery happens to be making poor contact at that instant.   If this is ever going to be a commercial product, it needs to be sufficiently resistant to load dumps that the odds are significantly in its favour that the ECU will die before your product does, otherwise the warranty claims will hurt your bottom line.

Its a gambling game where you are balancing the costs of minimising the odds of failure against the cost of paying out the claims, and the damage to your reputation.   

Also, try getting compensation for damage (including the phone) due to failure of a cheap after-market in-car phone charger!   The odds are high the importer has long gone and the manufacturer will be untraceable.  As long as it didn't actually catch fire or physically injure you, no one will be interested in helping you pursue your claim.

[Yansi]

Why so much drama in this thread?

http://www.ti.com/lit/an/snva717/snva717.pdf
http://www.ti.com/lit/an/snva681a/snva681a.pdf

(especially the first one).

Slap there an AEC-Q  certified LDO or buck after that and done.

[james_s]

I'm not even talking about cheap aftermarket chargers, quite a few of the OEM chargers are not anything fancy. I've been inside quite a few car audio amps and I don't recall seeing any fancy protection in them. Yes there are a lot of poorly maintained cars out there, but if your car is so poorly maintained that you are getting 100V+ load dumps on the electrical system then the solution is to properly maintain your vehicle. Fix the fault, don't over-engineer everything else unless it's a safety critical item.

I was under the impression that the device under discussion was a hobby project rather than a commercial product?