IEC 61000-4-5 Surge immunity test

[lewis]

I agree with Andy C, ISO 7637-2 is the applicable standard for automotive conducted susceptibility in supply lines. Pulse 5 (load dump pulse) from 7637 is now referenced in ISO 16750-2. The most onerous load dump pulse is 174V, 1R, 350ms for a 24V nominal supply. That's a pain to comply with, almost all standard TVSes will die easily. More info here: http://www.radiocad.co.uk/files/1413/6921/7114/LoadDumpPaper-final.pdf

Buy a copy of these standards. EN/IEC 61000-4-5 is probably only applicable to mains voltage transients, or transients on communication or signal ports, and should not be necessary for power inputs in your application. Triple check whether you really need to comply with it.

[Mad ID]

I agree with Andy C, ISO 7637-2 is the applicable standard for automotive conducted susceptibility in supply lines. Pulse 5 (load dump pulse) from 7637 is now referenced in ISO 16750-2. The most onerous load dump pulse is 174V, 1R, 350ms for a 24V nominal supply. That's a pain to comply with, almost all standard TVSes will die easily. More info here: http://www.radiocad.co.uk/files/1413/6921/7114/LoadDumpPaper-final.pdf

Buy a copy of these standards. EN/IEC 61000-4-5 is probably only applicable to mains voltage transients, or transients on communication or signal ports, and should not be necessary for power inputs in your application. Triple check whether you really need to comply with it.

Hi. At the moment we need the CE declaration. I undestand there are other tests but once again, CE mark (Declaration of conformity) is what we are after at the moment.

You have all been very helpful, thank you very much!

[AndyC_772]

Just to be clear: I (and, it would appear, we) believe that ISO 7637-2 is the applicable test which your equipment would need to pass in order to qualify for a CE mark, and not necessarily IEC 61000-4-5.

Sometimes, determining which tests are applicable to a product can be as difficult as engineering the product to pass them

[Mad ID]

Just to be clear: I (and, it would appear, we) believe that ISO 7637-2 is the applicable test which your equipment would need to pass in order to qualify for a CE mark, and not necessarily IEC 61000-4-5.

Sometimes, determining which tests are applicable to a product can be as difficult as engineering the product to pass them

I will look into it.

Thanks.

[qno]

I have passed the test by placing a P Mosfet in front of a 7805 that switches off when input voltage is > 35V.

The MOSFET has to be able to withstand 200V.

You fall into the category that will reset when the puls is aplied but resumes normal operation when normal power is restored.

If you define this behavior in advance you will pass.

Please set your country.

[Circuitous]

I have passed the test by placing a P Mosfet in front of a 7805 that switches off when input voltage is > 35V.

qno that sounds like a neat approach, do you have an example schematic of that?

[qno]

I can not give you the exact circuit but this datasheet gave me the idea.


http://www.maximintegrated.com/en/app-notes/index.mvp/id/4081

Here is where it is based on.

[Circuitous]

Perfect!  That looks like just what I need for an upcoming project.  Thanks!

[AndersJ]

61000-4-5 says that surge transients are injected between ground and signal lines.
"Ground" in this case means the device chassis, protective ground.

"Signal lines" in my case, and for many other devices are signal ground and the signal itself,
such as 5V digital inputs, 0-10V analog inputs, or an industrial 24V dc signal input, or similar outputs.

Often, as in a PC, and many lab instruments,
signal ground is galvanically connected to chassis/PG.
In these cases,
are surges applied between ground and signal ground?
This implies injecting the surge on a short circuit.
Is this how it is done?

[T3sl4co1l]

I'd have to look at the precise wording of the standard, but I don't see that there would be any harm of it.  The surge generator is designed for open, resistive and short loads.  As is true of any other pin, it should simply not do anything. Right?

Tim

[HackedFridgeMagnet]

Your ground wire will probably come into the DUT from a different wire than the signal ground.
Every wire has an inductance so what is a short circuit at dc may not be for a surge.
Surges are generally modelled as a current source. eg. 50kA 10us up and 350 us down.

The voltages that are generated by the surge through the inductances may cause arcing between different parts of the DUT.
This is one of the things you would be testing for.
Also not all signals are unbalanced with one side grounded.

So put the surge across each signal wire to the ground lead of the DUT.

To create the surge use something like Mikes electric stuff's destructor or whatever he calls it. He has a video where he tests the circuit breaker.
A capacitor bank charged up with high voltage. Obviously very dangerous!
Use an isolated (wireless?) remote to trigger a thryristor to send the surge.
You can measure the surge with a current transformer that attaches to a scope.

[T3sl4co1l]

Er, well, not just ANY bare capacitor, or random wiring, and certainly not a bank of defibrillator capacitors at full charge!!

IEC 61000-4-5 defines a maximum Class 4 of 4kV (into 2 ohms, 2kA) surge.  The time constant into 2 ohms is about 28us (it's not really a time constant because an RLC circuit is used), so about 14uF is needed.

To get the open-circuit waveform, some series inductance and parallel resistance is needed.  Resistance might be around 3 ohm, which means the 1.2us rise time (for this condition) needs around 2uH.

To get the short-circuit waveform, more series inductance is needed (though not much more), a total of double I guess.  The extra parallel resistance means more capacitance will be needed, maybe 24uF.

In the two conditions, sqrt(L/C) is much less than the ESR of the loop thus formed, so it should have no undershoot (the standard allows for 30% phase reversal), and no need for diodes and stuff.  A contactor is probably okay for switching the surge (it's over before the contacts can bounce open again), or thyristors of not too extreme ratings (but do turn them on hard!).

Tim

[jeroen74]

A few years ago I did this very test on a mains-fed lamp controller widget supposed to be mounted in a lantern pole, but up to 8KV. Naturally things got quite explosive once that poor MOV decided it was time to give up around the 25th pulse.

The scope indicated about 1 megawatt of dissipation for a microsecond.

[zare.zz]

how can I find a appropriate TVs diode for pass a IEC61000-4-5 CLASS3(2KV)?

[jonpaul]

Hello: Are you sure that IEC 61000-4-5 Surge immunity test is required for your product? Normally that is for mains connected PSU, not low voltage.

Which surge tester was used?

Generally we avoid MOVs as they are very 'soft" and wear out. Sometimes they catch fire after an overload.

Load dump protection for 28V systems were very large die Zeners like Motorola MR2525.
Vishay   TransZorb diode, https://www.vishay.com/diodes/protection-tvs-esd/


Bon chance,

Jon