Automotive DC protection (load dumps and such)

[brumbarchris]

You are wrong, it is doable. By mid price range TVS. But really we need to know figures and numbers. What pulses does the unit have to withstand? I can understand that the customer does not know about the existence of some common automotive standards. I did deal in the past with some Chinese automotive customers whose whole specification for an ECU was a 4-page word document (as opposed to Daimler for instance who come up with 900+ pages in multiple documents and calling for tens of additional standards).

For the purpose of knowing what to design against, I suggest ISO 16750. Check its chapter 2, it is referring to electrical loads, pulses, voltage ranges etc and includes a specific test for load dump. You will see it is not that bad. Out of the 15 or so design I have been dealing during the past 6 years I have only seen one to completely cut off supply during load dump. This method is not preferred because some customers require the electronics in vehicles to withstand the load dump but also to be functional during it.

If you go through the standard you will know more about load dump. As mentioned before, the key parameter here is the internal resistance of the voltage source delivering the load dump, as that will essentially limit the current through your TVS to a reasonable value which will allow you to dissipate the power on a relatively inexpensive package.

Regards,
Cristian

[uncle_bob]

Hi

It's probably worth taking a look at why / how you get a load dump in the first place. In addition to resistance limiting the current, there inevitably is some inductance creating the spike in the first place. The impedance involved is most certainly non-zero. As far as people not understanding all of this, that's pretty common.

Once upon a time in a land known as Illinois, there was a major company making alternators. We tested (yes indeed we) all of them right at the end of the line. The final test was load dump. By that I mean *final*. There was no check to see if the rectifier bridge had survived the dump .... hmmmm .... You would think that manufacturing the things at a many thousands a week rate would make you familiar with the failure modes.

Bob

[brumbarchris]

The effect of the inductance is only present in real life conditions. It is a parameter that actually helps the electronics dealing with the load dump (as it limits somehow the variation speed of the current owed to the load dump voltage pulse), but it is not a factor determining itself the load dump pulse. This pulse is determined by the inability of the alternator of dealing properly with large load variations (actually some automotive OEMs even give descriptions in their internal standards with regards to the actual conditions in which a load dump can appear: when a highly depleted battery under charge is suddenly disconnected from the alternator; in this case all the electronics remaining supplied by the alternator sees the load dump, as not having to charge the battery with a high current means the alternator is experiencing an abrupt load change).

Anyway, as far as the inductance is concerned, it is only a factor in real life, but when performing testing according to ISO standards (or internal OEM standards) the inductance is not a specified parameter, and the used connecting wires do not have a significant inductance, in general. Which makes sense, I guess, as you would want to see a worst case. In the end, when designing a product you often do not know if it is going to be connected to the alternator by a long or a short wire (in the real vehicle).

Regards,
Cristian

[uncle_bob]

Hi

Indeed, the test is one thing and the real world is something else. That's pretty much always the case. This is by no means unique in that respect. Just about every mil spec ever written is done the same way.

The point is that from a long term reliability standpoint, the test is not what you should go by. The real world has limiters in it that the test does not. Load dump is relatively rare, (but not *that* rare). When it does happen multiple times a day / week / month the pulse is limited. The concern here seems to be more involved with "works in the real world" than with "meets every detail of this or that spec".

Bob

[brumbarchris]

Yes, that seems to be the case, however, I am advising the OP to steer the discussion with his customer or Project Manager into a "meets every detail of this or that spec" type of discussion. Because that can be proven by testing. "Works in the real world" cannot be proved by testing (against what limits?). If there are no hard facts, values and limits against which to test to prove compliance, then you end up with a customer claiming that his load dump is 200V and with "why did you not design it to withstand that?" kind of questions.

In principle, this is one of the reasons (not the only one, of course) for which such standards were developed.

Regards,
Cristian

[eneuro]

To answer your 'Why not use switched batteries? ' question. Limited cyclic life and excessive complexity.

I've 12 year old salvaged automotive car starter  battery and... still no problems to run... controllers when charged for some time
Even more, I've removed oryginal flamable battery from cell phone and... use simple DC-DC to 4.15V and .. make calls directly from this car battery and... in standby mode no need to charge crappy phone for months

So, it is all about power needed, but controllers usually do not draw too much so... there is nothing complex..it is simple thing and more... those batteries can be charged using small solar PV panel on roof, but automotive industry likes... regulations to complicate simple thing$, but still design things to last 5 years or so to fit into waranty period and .. press customers to buy another one 

[uncle_bob]

Yes, that seems to be the case, however, I am advising the OP to steer the discussion with his customer or Project Manager into a "meets every detail of this or that spec" type of discussion. Because that can be proven by testing. "Works in the real world" cannot be proved by testing (against what limits?). If there are no hard facts, values and limits against which to test to prove compliance, then you end up with a customer claiming that his load dump is 200V and with "why did you not design it to withstand that?" kind of questions.

In principle, this is one of the reasons (not the only one, of course) for which such standards were developed.

Regards,
Cristian

Hi

The vibe that I'm getting is that a detailed spec back from the customer is not going to happen in this case. I get involved in enough of those sorts of politics that I have no need to see any more. When I do get involved, I get paid for my time

I've also seen situations where a simple question turns into a "meet every spec ever written " CYA reply  -> 5,000 pages of documents. Since you have a firm  fixed price contract, you just put your company out of business (or at least that business).


Bob