Automotive Protection for 12/24V systems.

[CM800]

Hi All,

I'm currently working on a preliminary project for a Brushed DC actuator controller.

The unit will be installed in an automotive environment, so will need protection circuitary.

I've found the following Texas Instruments Design for an automotive protection circuit.

We want this system to be reliable, ideally fuseless to reduce maintainance needs in the event of a fault condition.

This will be going in a Police / CT vehicle and while this one specifically isn't mission critical, we would like to treat it as if it is.

What would be recommended for foolhardy protection?

Estimated power for the drive circuitary will be 12 or 24V @ up to 10A

Any suggestions or advice would be greatfully recieved.

We are not opposed to off-the-shelf protection solutions.

[David Hess]

A couple of things come to mind:

1. Fuses are for safety when other protection methods fail.
2. I assume you will have something like an h-bridge driving the motor.  If the h-bridge has overload protection and is derated properly, then it can protect itself without blowing the fuse and the fuse is only needed if the h-bridge fails.
3. That leaves protecting the control circuits which is a much easier task because they draw much less current so series protection is feasible.

The last time I designed something like this, I wish I had included high input voltage protection which disables the device until the input voltage returned to normal.

[Cerebus]

Load dumps are the oft forgotten enemy in vehicular electronics. Assume that your battery supply is, at some point, going to go to 120V for as long as a second and plan accordingly. If the vehicle's electrical system designers have done their job you ought to never see anything above 60V for ~400ms but relying on that would be brave.

This will be going in a Police / CT vehicle ...

Ah, it has to resist the tender ministrations of Constable Plod. Is Chobham armour a permissible  option?  Seriously, the only thing more dangerous to equipment than a copper is an actor - I know, I've worked with both. Assume that any mechanical interfaces (switches, knobs, connectors) will be, at some point, operated with a hammer or a size 13 boot.

[ZaneKaminski]

TI has an okay app note on load dump protection if you are spinning your own PCB: http://www.ti.com/lit/an/snva681a/snva681a.pdf

Their idea is series protection... a pMOS pass transistor with an nMOS to pull down the pMOS gate and turn it on. When the battery voltage exceeds some fixed threshold, an SCR pulls the gate of the nMOS low, so then the pMOS gate goes back up and the pass transistor cuts off, protecting your load from the voltage spike.

You can also just use a big TVS... some TVS diodes are supposedly able to withstand something like a single one second load dump pulse up to 100V every 10 seconds for 10 minutes or something along those lines. I believe NXP/Nexperia have some parts in this category. In my experience this technique is more expensive than the series protection pass pMOS though. So I have never implemented it.

[David Hess]

You can also just use a big TVS... some TVS diodes are supposedly able to withstand something like a single one second load dump pulse up to 100V every 10 seconds for 10 minutes or something along those lines.

Even when it works and recovers, low impedance shunt protection runs the risk of blowing the fuse or causing other damage which from a reliability standpoint is a failure.

[CM800]

Thanks for all the feedback so far guys.

In terms of testing this, rather then spending the $$$ on buying a special test rig. I'm thinking that using capacitor discharges through relays would be a good way to do so. (200V capacitor onto the bus via relay)

...and such along those lines.

Are there any off the shelf solutions for automotive power supply protection? This might be preferable, depending on the volume.

[jbb]

In terms of testing this, rather then spending the $$$ on buying a special test rig. I'm thinking that using capacitor discharges through relays would be a good way to do so. (200V capacitor onto the bus via relay)

I second the use of a fuse with a generous current rating to prevent fires when it all goes wrong. This will protect your electronics, the motor cables and the motor.

At risk of attracting the wrath of fellow posters, I suggest you simulate your input protection. For testing, you can reach out to local EMC testing labs and ask for some debug assistance. It’s very informative to take your kit and be there when it’s tested. You can take several different options along too.

Depending on your BOM cost vs development cost, LT (now Analog Devices) make a line of chips called Surge Stoppers which could be very useful. They essentially do the voltage sensing and NMOS driving for you.

If you’re not planning to surge protect your motor driver, you’ll need a big voltage rating on the H bridge FETs and drivers. So you might want to check out both options (high voltage H bridge vs high current surge protection). Also, the motor control should include current limiting (or preferably an inner current control loop).

[cat87]

Depending on where the unit will be positioned on the car, inside te cab or outside,  or in the engine compartment,  you should also consider some conformal coating to prevent moisture ingress ans defects that can arrive because of vibrations. These are actually the most likely ones to kill an ECU in the long run.

This being said, for electronic protection,  a p-mos and tvs option is what most ecus actually have on them,  together with some emc/noise protection added. As for fuses on the pcb itself...well, sometimes, depending on the designer, they're there,  but most times they aren't.
 You already have fuses in the main fuse panel so any additional ones on the ecu itself is superfluous.

Also,  current monitoring is the main defense against things going crazy in the ECU. Most automotive h bridge chips have this.
Also,  I'd focus more on how noise affects the circuit and undervoltage conditions from like cranking the engine, than those pulsed voltage tests.