Advice on selecting automotive grade components

[oktane]

Hello,

If any of you have experience designing things for cars, your help would be
appreciated.

I would like to know what are the best components to use in a particular design. If you have any advice on protecting a uC in a car, that would be helpful too. The finished unit will mount somewhere inside the passenger compartment, so temperatures should not exceed 150F or so.


Parts I need advice on include:

P-Channel enhancement MOSFET - I'm currently using an IRF9Z34. IRF has a ton of fancy N-channel rugged MOSFET's specifically for automotive use, but I can't find any P-channel ones. I am using the MOSFET to drive a solenoid,
so back EMF and negative transients are a concern. My understanding is that most modern MOSFETs have built in Zener clamps. Is this not enough? Does anyone here use Tranzorb diodes?

Switching voltage regulator, 5.0V - I'm using a LM2931T-5.0. <

http://www.national.com/pf/LM/LM2931.html>

Supposedly this chip is specifically for automotive use and protects against load-dumps, reverse polarity, etc.

BJT Transistors: I hoped a 2N2222A would work fine. Any reason it won't? I shouldn't need to drive more than 250mA or so through the collector.

Thanks,
-oktane

 

[melone]

Oktane,

I design automotive engine controllers for a living. If you email me a schematic with a detailed BOM, I can take a look at it and make a few suggestions. I would offer up some advice right now, but there are so many details that must be known, it is just simplier to look at exactly what you are doing. Please send a detailed description of what you are trying to do, and how you are trying to do it. (I have found that sometimes what you are trying to do and what you are actually doing can be 2 different things)

Good luck with this project, and hopefully I can be of some assistance!!!

Mark_Melone@yahoo.com

 

[oktane]

Hello Mark,

Are you getting this? I sent some more detailed information to you as requested, but have not heard from you.

Thanks,
-oktane

 

[melone]

I am checking some of my current schematics versus what you have and am formulating reasons for why they are different. If the reasons are significant, then I will offer up a suggested alternative.

How soon do you need this problem solved?

 

[oktane]

Well since you are donating your time on this, certainly I don't want you to be under time pressure.

But the sooner I can get information the better, as I would like a prototype in my hands in the next couple of weeks.

Thanks,
0ktane

 

[salec]

Instead of p-MOSFETS, you may want to try automotive high-side switch? It comes all-secured, with logic level command input and with error status output, so you can provide failure indication. I have used Infineon devices, but there are others', to.

 

[oktane]

Salec,

That's not a bad idea. It is a nicely integrated solution, but does add a bit more complexity. Some of the &quot;OMNIFETs&quot; seem to offer similar features in a drop in TO-220 package.

Thanks for the info.

Any news yet Melone?

 

[melone]

I have been reviewing your circuit and am a little concerned about a few things. 1) What vehicle is this circuit going to be installed in? 2) Does this vehicle have an engine / transmission controller already? 3) If so, why would you possibly want to bypass it? I can speak from industry experience, most vehicles since the mid 1990's have had some kind of engine / transmission controller integrated into it. Bypassing the factory module can cause some serious problems (comprimising OBDII, performance, fault detection, auxilliary engine strategy functionality, etc.). 4) Plan on between 3 & 6 amps flowing threw the coil / FET. 5) Do you plan on having any shorted load protection? 6) What happens if there is a reverse battery condition? 7) Are the FET's you chose, overvoltage protected? I assume they are, but did not have a chance to pull the data sheets. What voltage are they clamped to? 8) Assuming that the FET's are clamped, is the clamping active or passive? If power must be applied to the FET for the clamping to work, then you must be VERY careful to protect these devices. 9) How are you going to ground this module? If the ground connection is too long, then you can induce a ground shift into your circuit. How will your logic handle possibly up to 1 volt ground shift?
etc...

I would like the chance to look over this some more. Hopefully, this is at least putting you in the right frame of mind when looking at these types of circuits. Automotive circuits connected to inductive loads can be VERY tricky. I am currently using a self-clamping FET in some of my applications and will check to see if it is a OTS component. Otherwise, I will review the omniFet again. BTW, the high-side driver will not work due to some of the internal brains of the device. I think I know what device Salec is talking about, but it has over & under voltage shut down. If you are controlling the low side of the load, you will never generate enough voltage to get it to work reliably.

Sorry for the delay!

 

[salec]

True, it does have over & under voltage and thermal shut down.
Never occured to me as a problem, but I suppose your solenoid must hold, no matter what happens to car electric supply rail.
I used it only to switch incadescent light, which is surely not a critical application. In my design it was important to protect the switch inside a taxi meter (smoked too much ;-) before we embedded one of those), not to ensure uninterrupted supply to the load.

 

[oktane]

The vehicle it is going into is an older one, without much of a transmission brain. The engine ECU is totally oblivious to this unit.

As for shorted load protection, there are already fuses that protect the load and the logic. A reverse battery condition will be of no consequence since the regulator I am using has this provision in mind. I do not think the FET will suffer any damage either, as the internal diode will not allow reverse conduction until voltage reaches some insane value. The FETs I chose do have a zener clamp in them. Not sure of the level, probably 55V or so. I say this since the breakdown voltage of Vds is -55V, and it would make sense to clamp to this level.

http://www.irf.com/product-info/datasheets/data/irf9z34n.pdf

I am not sure of the distinction between active and passive clamping. The zener will allow conduction if the source exceeds a certain level. Maybe you can explain this concept to me.

The ground is through a good chassis ground. There is not much drift from the battery ground. The regulator I am using should compensate for the shift without much problem, or so I think.

I can see how the over/under voltage protection could effect operation through transients. There should be a &quot;dumb&quot; version of the same technology available somewhere in a PTH package!

Thanks for the help guys.

 

[melone]

Actually, there is NO fuse that will protect the FET if the load shorts! Also, are you positive that the ECU does not talk to the transmission? Engine strategy is based on many factors; engine RPM, engine sensors (temp, pressure, etc.), what gear the transmission is in, etc.. I would hate to recommend something and find out that you have just bypass a neccessary input for your engine to run.

Reverse battery is an EXTREMELY critical issue. Think of your FET as having a zener diode placed across the Drain-Source. The orientation of that zener allows the clamping of the voltage. However, if a higher potential is placed on the Source side of the FET (during a reverse battery condition), the diode will turn on! This means that the load can be reversed biased. This will turn on the load!!!

The active vs. passive clamping can be thought of as either having a zener (passive), or some sort of gate control via an op-amp (active). If the device needs to be powered to have the clamping work (active), then it will probably not work in this situation. However, a simple Zener diode should ALWAYS clamp if the potential ever exceeds the breakdown voltage.

With regards to the ground. I am less concerned about your micro working, and more concerned about turn on your load fully! There can be up to a couple of volts difference between the ground at the front and back of the car. Therefore, if you are under a low battery condition (cold weather), and you add a 1 volt to the ground, are you sure that the load will actually turn on? I concede that it will PROBABLY turn on, but if I went to my wife and told her that I am going to modify up her car and it will PROBABLY work, she would tell me to go fly a kite!!!

I will keep pluging away.

 

[oktane]

Oh I see what you're saying about the load shorting. Both the load and the FET are on the same circuit. If for some reason the load were to short to ground, the fuse would blow (not sure how many ms) hopefully protecting the FET. Ideally the FET would have built in overcurrent throttling, but I don't think the one I picked does.

As to the reverse battery, the higher potential currently IS on the source side, since this is a P-channel MOSFET. I am driving the gate using a BJT inverter off the main 12V power rail. During an OFF condition, voltage at the gate is +12V and votage of the source is +12V, resulting in a Vgs=0, and below turn on treshold.

If somehow battery polarity were reversed, you would see either 0V or 12V at the gate and 0V at the source. In any case, Vgs will not go negative enough to allow the FET to turn on. Only thing that should be a concern is transients breaking through the oxide layer, and causing a latch condition. A forward-biased silicon diode could fix this right?

The zener is just a simple diode in these FETs. I see no mention of other circuitry. Although it is true that there will be NO clamping effect unless there are at least a few milliamps of current flowing through the zener, since otherwise it will function as an open circuit. I see no way around that...

The load should turn on fine even with a few volts chopped off the lower end. I have never tried, but most solenoids have a decent range of operation.