Circuit Protection on an automotive electronics application 2

[djroseman]

Hi,

I have a requirement to make a protection unit to prevent damage to our telematics system from large voltage and current spikes, both from the vehicle and attempted sabotage by operators.

The normal operating voltage is 24V, but it is possible to have up to 80V or more spikes. My current thinking is to use a clamp or crowbar circuit with inline fuses, which should pop instead of burning out the tracks. Any ideas on this? Also, what sort of current handling can I expect on 100thou PCB tracks in 1oz Cu?

Thanks

Dave Roseman

 

[BrianG]

You don't say what sort of current your equipment normally needs, but you can use an L-C filter combined with transorb clamps to remove spikes effectively. The design does depend on your "acceptable" tolerance on the 24V line under spike conditions.

For longterm overvoltage problems consider a power FET with voltage sensing circuit - some automotive high-side switches already incorporate this kind of technology. This is probably better than an SCR crowbar with fuse.

When you ask about handling current for 100 thou track are you asking about fusing current or normal carrying capacity?

 

[djroseman]

>You don't say what sort of current your equipment normally >needs,

Normal operating current is average 500mA. The problem we have is people deliberately over-volting the unit by connecting supplies to it. I don't know what these are, since it could be anything to hand (if this was 24V batteries in the case of lorry drivers, this could be any multiple of 24V depending on the number of batteries used).

I was thinking of providing protection up to 100V, with maybe 5-10A of current.

>with transorb clamps

Is that like a standard diode clamp cct?

>For longterm overvoltage problems consider a power FET with >voltage sensing circuit

When you say longterm overvoltage protection, do you mean for long periods of excess voltage, or for protection without having to revisit the vehicle to replace fuses etc? Would you know where I could find an example of this circuit?

The 100thou track would be for fusing current, we have suffered in the past with spikes causing the tracks to burn out, hence rendering the board useless. We currently have transient surpressor diodes in place, and we have had occasions where the tracks have burnt out first before blowing the surpressor.

Thanks
Dave

 

[Comcokid]

Questions like this get asked from time to time. In vehicle systems, the transients are generally called "load dump" (google it), but can arise from several different situations. As a specific term, "load dump" refers to the transient arising from a heavly loaded alternator which encounters a sudden interruption in it's load.

Automotive rated voltage regulator ICs can handle transients of 60 to 80 volts at the input. A pass transistor/zener arrangement can also be used. Transient supression such as MOVs, transzorbs, zeners combined with resistors, PTC devices, capacitors, and fuses can also be used ahead of the regulator. It all depends upon how much protection is needed and current used by the device being protected.

thread240-100040
thread67-98147
thread240-98170
thread67-94753

 

[IRstuff]

You should look at Vicor's line of Input Attenuation Modules (IAMs) as a comparable product:

http://www.vicr.com/products/military/filter_modules/mi-iam/

TTFN

 

[itsmoked]

What you need is multiple combined solutions. Essentially you set up your protection to just take high voltages without blowing (briefly) then add a further system which uses the time achieved,(above), to essentially open circuit you system from the supply even if multiple destructive supplies are applied. Then the saboteur cannot even temporarily disable the device.

You say normally the load is 24V @ 500mA?

I presume that since the drivers are trying to destroy these electronics they are somehow monitoring their behavior, (or lack there of?).

Does the device have a display or something so the saboteur can always tell if it is still working?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Warpspeed]

Agree with Keith, there is no really simple solution to this. Once you start to discover how extremely hostile a vehicle electrical system can be, it can be a truly challenging problem.

The short term high voltage spikes and inductive ringing are pretty easy, transorbs and capacitors should cope because the voltage spikes (of either) polarity have very little total energy to dissipate.

Reverse polarity is another problem. When Fred the less than bright auto mechanic installs the battery around the wrong way. Rare, but it can happen. Or jump starting from a second vehicle with the leads reversed !! A fuse and inverse diode works well for this.

Sustained overvoltage, usually from a fast battery charger can be another problem. Although the average dc voltage may be only 15v or so, the repetitive PEAK voltage from the unfiltered rectifier may be up over 20v. This can quickly fry your transorb, zener, or similar shunt voltage regulator. A crowbar circuit and fuse works, but it will then put your unit out of action by blowing the fuse.

Another problem is what is load dump from the alternator. If the alternator is supplying a heavy load, and that load is suddenly switched off, the voltage can rise fairly high until the magnetic field in the alternator adjusts to the new much reduced load. This can be very serious if the engine is being run with the battery disconnected. Again a fairly rare circumstance, but it can happen.

 

[djroseman]

Thanks for all the useful info chaps - I think that I will make an LC filter followed by a crowbar circuit for 30V and 3A current using SCR and zener diode with a transil diode (SMAJ30) in parallel with a resettable 3A fuse. A couple of questions still:

Would a 20Hz LP filter be sufficient on the input to filter out most of the noise?

Is it more advisable to place the fuse before or after the transil diode?

Dave

 

[Warpspeed]

There is no normally accepted standard technique, or "best practice" for doing any of this. Any design of protective circuitry will have its own unique set of advantages, shortcomings and possible failure modes.

 

[felixc]

If you have a 20Hz LPF, then the transil diode will clamp low frequency spikes. It may be better to have the fuse before the diode so the diode may have some (I say some because this type of fuse has a very slow response time) protection too from the fuse. The instant wattage of the diode is high, but too many peaks can melt it.

What is the voltage at which your circuitry works? If the voltage is say under 10 volts, there are switcher circuits targeted at the automotive market that can handle very high voltages and thus reduce the need for complex protection circuits. With such regulators, reverse protection, a resettable fuse and spike filtering are about all you need.

On a design that I've done, the 12 volt was recreated from the 5-volt supply that more easily had all the required filtering and protection.

 

[peebee]

I'd consider designing this thing to handle 120VAC input if you're really worried about sabotage. That would be the most convenient way imaginable to overvoltage this thing (and may be the standard MO already).

 

[itsmoked]

You mean 240.. Since I think we're talking England.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[peebee]

Hmmm... how do you infer England? ... by the 24VDC?

Regardless, energizing this thing from the wall outlet would be the first & easiest way I could think of to fry it.

 

[itsmoked]

"chaps"
"transil diode"

But you're correct that is pretty slim pickins.



Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[danw2]

Naw, I made a similar assumption from "lorry drivers".

Dan

 

[itsmoked]

What's pathetic is I just went thru the whole thread looking for why I assumed it. I didn't see that, danw2. But that's actually probably the original reason why I assumed England.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[peebee]

Jolly good. Carry on. Stiff upper lip, wot.

 

[Warpspeed]

Are they still using imperial volts in England, or are they the the new metric volts ?

 

[zeitghost]

The Cabbage Heads in Brussels want to introduce us to the metric Volts, but we're sticking to the good old Imperial Volts.

Ignoring Johnny Foreigner comes naturally, what.

Mr Gladstone would have sent a gunboat.

 

[rgraceati]

Thanks for all the good thoughts on designing a power supply for a 24V vehicle.

Does anyone know what the official High Volatge sec is? My current design is using a SMD ressatable fuse with a maximum 33V stanfdoff voltage and a FET clamping circuit set for 33V.

The $64,0000 question is: will the DC voltage always be less than 33V? (not incuding transients)