Automotive DC protection (load dumps and such)

[dmg]

I have to protect the power supply of a controller for automotive 24V systems. That means the device has to operate from 10V to 32V DC. I have to protect the DC input against load dumps, reverse polarity and such. As far as I've seen, one common option for load dumps is to use an automotive TVS. You do a little research, estimate the average alternator resistance, derate that, calculate the energy you'll have to eat, select a proper TVS and place it before power regulators.

However, most automotive DC/DC converters/regulators I see are only rated to Vin < 45/50V with only some expensive ones going up to 60V. All TVS I found that can withstand 32V clamp at 55+ V when there's a high energy dump. So any normal DC/DC converter I'd place after would essentially be fried. This is a cost driven design (they want cheap and good, but fortunately not fast), and I have almost zero experience in automotive. Any recommendations as to how to protect it while still being cost-effective? Any other high energy clamping devices that clamp bellow, say 40V while withstanding 32?

We need to drive some motors and thought about using an L99MD01 driver from ST, but that's only specified for 40V tops on supply rail. It would be very nice to safely integrate it in the design, but ST says that this chip isn't specified for 24V systems, it might work with proper protection but we're on our own.

[Paul Price]

You can easily protect you circuit with an inductor, a N-Chan power MOSFET and a Zener and 1 resistor, 1 electrolytic cap(the cap is already in your circuit).

https://www.eevblog.com/forum/projects/do-laptops-typically-have-battery-over-voltage-protection-strange-project-ideas/

[dmg]

You can easily protect you circuit with an inductor, a N-Chan power MOSFET and a Zener and 1 resistor, 1 electrolytic cap(the cap is already in your circuit).

https://www.eevblog.com/forum/projects/do-laptops-typically-have-battery-over-voltage-protection-strange-project-ideas/

Thanks, but I don't get it very clear from your posts. Maybe schematic pictures got deleted?

This is a 24V system that needs to work from 10V to 32V. Current consumption of my device is about 10A tops worst case ever (normally, under 2A). According to customer they have experimented all sorts of crap in power lines. This is not a crytical system, it can shut itself off during a transient as long as it resets itself to the previous state. I need the protection to:

-Clamp below 40V.
-Work at 32V.
-Provide reverse polarity protection.

This will go iside buses and similar vehicles, and is cost-driven. Probably, due to clamping constraints I'll have to go to an active disconnect approach. The problem? Well, I don't know how much crap to expect, and the more I read the worst. Most papers/appnotes I find start talking about 12V and 24V systems and then diver to 12V only. The worst I've read is that for a 24V system I'll have to survive -600V pulses and 200V dumps.

I've also read (and confirmed with the customer) that operating voltage range has to be 10V to 32V as stated in some standard, but several appnotes for 24V protection clamp even below 30V... My head is a mess now...

[Ian.M]

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

[MagicSmoker]

There are ISO standards for this which are quite comprehensive: ISO7637-2 was used for many years but has recently been supplanted by ISO16750-2.

The easiest solution is wire one or more unidirectional TVS diodes that are approved for one of the above standards (approval for the older one will be much more common) in anti-parallel with the 24V input.

For example, STMicro LDP01-35AY meets your breakdown and clamping voltage requirements for either ISO standard and costs <$2US in moderate quantities. For extra robustness and to better define the clamping voltage you can insert a small value of resistance in series with the 24V supply before the TVS diode; refer to the datasheet to determine an appropriate value, but somewhere between 0.33 and 2 ohms is typical.

You will still need other filter components for noise (especially for EMC compliance) but that one TVS diode solves your load dump problem.

[dmg]

There are ISO standards for this which are quite comprehensive: ISO7637-2 was used for many years but has recently been supplanted by ISO16750-2.

The easiest solution is wire one or more unidirectional TVS diodes that are approved for one of the above standards (approval for the older one will be much more common) in anti-parallel with the 24V input.

For example, STMicro LDP01-35AY meets your breakdown and clamping voltage requirements for either ISO standard and costs <$2US in moderate quantities. For extra robustness and to better define the clamping voltage you can insert a small value of resistance in series with the 24V supply before the TVS diode; refer to the datasheet to determine an appropriate value, but somewhere between 0.33 and 2 ohms is typical.

You will still need other filter components for noise (especially for EMC compliance) but that one TVS diode solves your load dump problem.

Yes, I'm aware of those standards and I've read parts of them (basically extracts that I found online), the problem is the implementations out there that seem to not apply the standard yet guarantee automotiveness. I think I'm going for the load disconnecting approach. As this is not a critical system I'll add an energy reservoir for the microcontroller and try to shut it down preserving RAM during disconnection and just reset to previous state after that. I also like the TVS you suggested. Seems like calculating the series resistor so that clamping current is controlled I can get maximum clamping voltage below 40V.


As for EMC, apart from LC filter at input I also chose SMPS regulators that operate above AM radio. Will check standards. As I said my automotive experience is almost zero. I already designed input protection for all I/O to enable them to be shorted to 24V, ground or to almost anywhere and be able to block a load dump or negative spikes themselves but I'm being slow as I get familiarized with my enemy. Can you point me to relevant EMC standards? Do you know any "cheap-ish" way to set-up the required dump tests?

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

Thanks, I already found that and I was actually simulating the circuit. The only drawback of it is that I'd still need a TVS to avoid frying the poor PMOS or else invest in a super expensive PMOS that can withstand 202 Vsd.

[joeqsmith]

You could open it up rather than clamp it.   What country are you in that you are working with 24V automotive?

[dmg]

You could open it up rather than clamp it.   What country are you in that you are working with 24V automotive?

This ain't for cars, this is for buses and similar heavy vehicles. Here all those vehicles work at 24V. Maybe buses aren't considered automotive? Am I missing something?

I will open but I'll still need to clamp to protect the opening switch. Going HV N-fet for the opening device would mean adding charge-pump and cost, so PFET with clamp.

[MagicSmoker]

... I think I'm going for the load disconnecting approach. As this is not a critical system I'll add an energy reservoir for the microcontroller and try to shut it down preserving RAM during disconnection and just reset to previous state after that. I also like the TVS you suggested. Seems like calculating the series resistor so that clamping current is controlled I can get maximum clamping voltage below 40V.

I recommend clamping vs. opening up the supply because the former approach is simpler, more reliable, operates much faster, takes up less board space and you'll need some sort of transient protection anyway (from other inductive loads being interrupted, and not just the massive attack of a load dump).

Can you point me to relevant EMC standards? Do you know any "cheap-ish" way to set-up the required dump tests?

There are engineers that specialize in compliance and you would do well to hire one for pre-compliance evaluation to help guide your design. The most common for the EU is 2004/104/EC. This page is helpful, in my experience: http://www.autoemc.net/Standards/StandardsMain.htm

[dmg]

... I think I'm going for the load disconnecting approach. As this is not a critical system I'll add an energy reservoir for the microcontroller and try to shut it down preserving RAM during disconnection and just reset to previous state after that. I also like the TVS you suggested. Seems like calculating the series resistor so that clamping current is controlled I can get maximum clamping voltage below 40V.

I recommend clamping vs. opening up the supply because the former approach is simpler, more reliable, operates much faster, takes up less board space and you'll need some sort of transient protection anyway (from other inductive loads being interrupted, and not just the massive attack of a load dump).

Can you point me to relevant EMC standards? Do you know any "cheap-ish" way to set-up the required dump tests?

There are engineers that specialize in compliance and you would do well to hire one for pre-compliance evaluation to help guide your design. The most common for the EU is 2004/104/EC. This page is helpful, in my experience: http://www.autoemc.net/Standards/StandardsMain.htm

Unfortunately even your suggested LDP01-35AY TVS isn't fully compliant to the requirements, it seems to go above 40V in some low energy but high voltage events, and that would fry devices. The 40V absolute maximum clamping voltage comes from another higher-level system requirement that makes very cost-effective to use L99MD01 motor drivers for our application. That part can tolerate 24V levels  and could work in a 24V environment except for the fact that motor supply can't go above 40V for 400ms absolute maximum rating (normal rating is 38V). That makes me have to open and also clamp, but with a higher clamping voltage device thus making it eat less energy. Thanks for the information anyway, I have things much more clear now.

[jc101]

How about the LTC4365 from Linear Tech?

[Halcyon5000]

Hello Boys and Girls,

I've spent considerable time on this subject and have a couple solutions for automotive and industrial equipment.

For a low current <3 amps@24VDC check this out: http://cds.linear.com/docs/en/demo-board-manual/dc1850aabf.pdf
I have tested it using a 200VDC square wave for 400mS and it passed without a scratch. It can be customized to work using various clamping voltages.

For higher current protection circuits >3 amps this process is a bit trickier but do-able. I can share a circuit if anyone is interested.

If you just want to buy something, check this out: http://www.tri-m.com/index.php/products/tri-m/power-ups/hdts1275-db-heavy-duty-transient-protection#downloads

I have also attached a PDF containing a vary well written paper on the load dump transient and TVS diodes. It would also behoove anyone looking into this to examine ISO 7637-2 (google it) for more info on how the load dump is tested.

Cheers!

[dmg]

How about the LTC4365 from Linear Tech?

This IC and similar ones would still require a TVS, in this case a bidirectional one clamping at >-40V so we're back to step 1. Several other surge stoppers from LT say they don't normally need TVS but ISO 7637 pulses say otherwise. I'm not sure on how they plan on dealing with the most demanding -600V pulses... maybe input filtering should take care of them?

It's not just load dumps, it's all the other low energy crap there is that must also be stopped.

To summarize:

-My device has a nominal 24V input, maximum current in very worst case is 10A (probably it could be reduced to less than 5A).
-This input line needs to be reverse-polarity protection (-28V for an indefinite amount of time)
-Working voltage range is 10V to 32 V DC
-There are devices connected to this line, which are the ones who get the most current, that have an absolute maximum supply voltage of 40V
-So any input protection must clamp to +40VDC or less in any circumstance
-Must handle pulses defined in ISO 7637-2 and load dump conditions in ISO16750-2. That means -600V pulses and such.
-Must be cheap-ish
-Can just disconnect itself during transient

My current options:
-Reverse polarity with a power diode with >600V reverse voltage rating + load-disconnect circuit from http://www.ti.com/lit/an/snva717/snva717.pdf adapted to 24V
     -Pros: cheap-ish, simulations say it's realiable
     -Cons: many parts, needs HV PMOS

-Surge-stopper IC's from linear or maxim
    -Pros: integrated, uses NMOS
    -Cons: i think they'd still need TVS even if they say they don't. They say you can absorb those low energy pulses with capacitors and such  but you do the math and simulate according to ISO test procedures and nope.

[joeqsmith]

I assume it is for work and it's a paid job.  Doesn't your company have other engineers that can help you?

[MagicSmoker]

Unfortunately even your suggested LDP01-35AY TVS isn't fully compliant to the requirements, it seems to go above 40V in some low energy but high voltage events, and that would fry devices.

That doesn't make any sense... lower energy in a transient results in a TVS clamping at a lower voltage; it does not matter how high the transient voltage is. The dV/dt of the transient does matter to some extent, but that can easily be slowed with LC filters that will be needed for EMC compliance, anyway.

[dmg]

Unfortunately even your suggested LDP01-35AY TVS isn't fully compliant to the requirements, it seems to go above 40V in some low energy but high voltage events, and that would fry devices.

That doesn't make any sense... lower energy in a transient results in a TVS clamping at a lower voltage; it does not matter how high the transient voltage is. The dV/dt of the transient does matter to some extent, but that can easily be slowed with LC filters that will be needed for EMC compliance, anyway.

Isn't power and not energy what really matters for clamping voltage? I mean, in a low energy but HV transient if HV reaches the TVS and it breaks it will absorb less power and the transient will be less stressful but it'll clamp at a higher voltage for a short period of time. If not, I'm missing something very basic here that I would like to learn. Thank you very much anyway, I'll make simulations with LC filters included.

I assume it is for work and it's a paid job.  Doesn't your company have other engineers that can help you?

We have engineers but the ones I have accessible have zero experience in automotive too. We mostly do radio stuff. How on earth were we assigned this job is a long story, it begun as something for radio that then someone wanted to put in a vehicle and someone said yes to that.

[Andreas]

Hello,

usually the protection is done with specialized (custom specific) ASICs for the protection.
The so called "main relay" cirquit switches off (or limits voltage to around 34V) in case of load dump.

Otherwise you will have to use components which are suitable for about 55-60V transients.

With best regards

Andreas

[max_torque]

Let me add one word of warning:

"You want to be the module on the electrical bus with the HIGHEST voltage possible BEFORE clamping" !!


What i mean is that the sooner you clamp, the more energy your module gets hit with.   It's much much much better to let some other poor sods module get hit with a massive power burst and have your module just sit there and ride out the voltage!

To that end, you want a TVS that clamps as high as you can stand.  A lot of automotive power regulators have significant short term voltage capabilities, often up to 55 or 60v and sometimes even more.  If your system has a low current powersupply path, protect that behind a separate lower rated clamp with a resistive or inductive element upstream to take the lions share of the energy.

If you are required to be MIL rated, then your 24v nominal bus has to be able to live with 200v (when not clamped) spikes

[uncle_bob]

Hi

One thing that has been mentioned in passing, but that is well worth repeating:

You *will* have inductance in the input side of your gizmo. There is pretty much no way to meet EMI without it. In general, EMI will be a much bigger headache than anything involving transients. The dv/dt of the transient works *for* you in the case of an inductor. It simply reflects the bump back so some other poor device can absorb it. That inductor is as much a part of your solution as any of the various diodes and clamping devices.

Bob

[eneuro]

I have to protect the power supply of a controller for automotive 24V systems. That means the device has to operate from 10V to 32V DC.

What about power requirements of this automotive controller?

why not to use two, or better three12V or 6V batteries charged from this 24V system in parallel using cheapy DC/DC, but disconnected from controller powr supply, than when controller need input power it is connected to one of three batteries, while another one is charged from 24V using DC/DC.
We have always 3rd already charged battery, so when voltage in connected to controller goes below some level, we swap connections between charged and discharged.
In next step we swap dicharged battery with charged, so by using lets say 500V AC mosfets switches to reconfigure connections between three batteries and controler input, our controller will never see automotive 24V system and no need to mess with TVs, and fancy protection.
We might need controller powered also when 24V system is switched off, o such floating three battery pack can keep alive our controler for many hours depending on power requirments which were NOT specified there 

[Ian.M]

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

[dmg]

I don't have control about how the power distribution in the vehicle is laid out, I only know it's a 24V system.

Taking your advices and given that I've been told that the vehicles this will go in may or may not have central surge suppressors (cost reasons) I've finally taken a mixed approach.

First, LC filter to take care of EMC and create a high impedance for short transients. Then, a TVS clamping at <100 V to only take the peak of the dump wave. Finally, a load-disconnect circuit with pMOS that can withstand 100V and cuts at 36V (circuit taken and slightly modified from TI's appnote). If switch is open my microcontroller gets a signal via fast interrupt line and immediately stops itself and gets into low power mode. It has an energy reservoir that can power it for almost 1 second in worst case, so I hope the transient will be gone by then.

Probably overnegineering and it drives , but this way my device won't be the one eating the energy of a transient.


This way I hope this won't eat the transient

[Mad ID]

I have to protect the power supply of a controller for automotive 24V systems. That means the device has to operate from 10V to 32V DC. I have to protect the DC input against load dumps, reverse polarity and such. As far as I've seen, one common option for load dumps is to use an automotive TVS. You do a little research, estimate the average alternator resistance, derate that, calculate the energy you'll have to eat, select a proper TVS and place it before power regulators.

However, most automotive DC/DC converters/regulators I see are only rated to Vin < 45/50V with only some expensive ones going up to 60V. All TVS I found that can withstand 32V clamp at 55+ V when there's a high energy dump. So any normal DC/DC converter I'd place after would essentially be fried. This is a cost driven design (they want cheap and good, but fortunately not fast), and I have almost zero experience in automotive. Any recommendations as to how to protect it while still being cost-effective? Any other high energy clamping devices that clamp bellow, say 40V while withstanding 32?

We need to drive some motors and thought about using an L99MD01 driver from ST, but that's only specified for 40V tops on supply rail. It would be very nice to safely integrate it in the design, but ST says that this chip isn't specified for 24V systems, it might work with proper protection but we're on our own.

Hi.
I have designed a system which is used in buses/trucks. The first version did not have any protection and the DC/DC regulators were rated for up to 32V only. At least 40 is still installed, it's been a year. Do you know how many got fried? None!!

Then for the next revision I had time to improve my design. I use a 42V DC/DC converters. In front I have a disconnect circuit with 100V P-MOSFET (simple one from TI app note, using LMV431). So everything can withstand at least 100V (I tested it). After the MOSFET I have 3x220uF/50V caps. These caps absorb small energy spikes. The circuit it set to disconnect at 32V.

At least 70 units installed, zero problems.

I think this load dump problem is heavily exaggerated. In 40x1year=40years of operation and with no protection zero failures. In which normal case would it happen that the alternator is giving maximum current and the battery is disconnected? That is like 1 in a many..This 1 in many is probably important for auto manufacturers, but for us? I doubt it..

Correct me if I'm wrong..but this is my experience so far. If 1 in 500 dies after 3 years, who cares? Cheaper to replace probably.

Oh..I also have a LDP01-33AY in front as DNP just in case.

Why 32V? I never measured more than 28.8V on a running vehicle

[brumbarchris]

In 40x1year=40years of operation and with no protection zero failures

With all due respect, that is not representative. Volumes in automotive can range to hundreds of thousands per year and even millions per year for some products. And the ppm failure requirements are often under 100 units. 40 units or 70 units with no fail might just me you got lucky or they never really faced a load dump situation (which admittedly, only rarely occurs)


Anyway, we are approaching this completely wrong, I believe. I do this for a living (automotive HW design) and I think the first question here is what load dump voltage do you have to withstand. Normally a serious automotive customer would indicate a standard like 16750 (chapter 2, electrical loads) and would ask your product to be compliant to that. The mentioned document contains explicit load dump tests that need to be performed in order to prove compliance. It also indicates the exact load dump waveform and (most important of all) indicates the series resistance of the voltage source used to generate the pulse. That series resistance, together with the amplitude and duration of the load dump pulse are the key parameters going into the equation of selecting a suitable protection device (like a TVS).

If no explicit customer requirement points to such a standard, you should put forward a suggestion either to your customer directly or through your project manager and suggest this standard. Then get the customer to accept this suggestion and start your design based on that. Once you have this settled (meaning the standard you need to comply with in terms of load dump) I will be happy to continue the discussion on the topic. Otherwise we are just making suppositions.

Regards,
Cristian

[dmg]

For what I know, in the market this product is for (sorry but no more details) there are customers who aren't even aware of standards and want cheap and others that will ask you for worst. I also know that the main problem the product this is going to substitute had is that... Well, it was cheap but lots of failures and funny behaviours.  Most products i've reviewed from competitors state iso standards compliance in their product datasheet, sometimes against one or several pulses, other times against full spectra of transients.

What I do know now is that doing the math there's no freaking way I can eat a standard transient having to clamp at  <40 V without heavy duty and expensive tvs combinations, so disconnect it is.