Advice needed on automotive electircal interfacing

I've completed my first proof-of-concept project, a programmable light sequencer for my car. It's all very cool and should work fine, my quandary is where to put everything.

Ideally I had planned on having the controller in the dash near the signal/brake controls, where I see the indicator led's. From there I had planned on running a single cat5 cable to the back (cleaner than normal auto wiring) where it would connect with another small board with the mosfets to switch the car current to each light.

It occurred to me that there may be line resistance issues I hadn't accounted for, so the other possibility is to have the mosfets up with the controller, and run individual 12v wires for each light. I dont know if 10-20 foot cat5 run would limit current for the mosfet.

Personally the idea of one neat, tidy, easily replaced patch cable appeals to me. Any advice as to why one way would be better than the other, or alternatives, would be great.


The CAT5 cable for signal is going to be a better choice than running your power wires all over the place. A hex buffer (either inverting or non-inverting) with push-pull or tri-state outputs to drive the signal line is a good choice -- cars are very noisy electrical environments and you can use all the help you can get to get a clean signal. Then, use a local driver chip of some kind to drive the MOSFETs. Tri-state or push-pull buffers are a nice choice because they can drive the MOSFET gate both at turn-on and turn-off.

If ordinary buffers don't have the muscle to switch your MOSFETs acceptably quickly, then get some some MOSFET gate drivers. I like the International Rectifier line of these, and they sell all their active parts direct from their website. Their data sheets and app notes are a great resource for all things power MOSFET.

One last note -- car electrical power is generally horrible. Robust power input protection and filtering is a must.

You may need to use screened cable to prevent problems with electrical interference.

Would optocouplers be usefull here placed right before the mosfet. I havent used them before, but it seems like it would be good.

I'm guessing screened = shielded ?

Noise shouldn't be too big of a problem. I have the luxury/curse of replacing all of the electrical in this car. There will be two separate 12v circuits, one for the engine, and one for the rest of the car. I know there will still be some noise, but I can probably get rid of at least 50% of what you'd typically see by just not doing things cheaply like the auto companies.

I'm guessing screened = shielded ?

Yes. I've always called it screened, but I'm in England so perhaps the usual name for it in the USA is shielded? Never noticed the different names until now!

In order to make opto-isolation work, you need separate power and ground supplies for each side of the isolation, and in the given case you'll still have gobs of noise in the cable.

There are other tricks other than shielded cable which might work here. One is to make the logic levels current signals rather than voltage signals. The other is to take advantage of the fact that you're using a twisted-pair cable and run differential signals over the each pair rather than single-ended. Both give you excellent noise immunity, should you need it.

And if you do use a shield, it's remarkably easy to do it wrong. Tim Williams' The Circuit Designer's Companion has a good section on how to do it right, as well as being a generally awesome book on how to make real circuits work in the real world.

I don't want this to sound argumentative, I'm guenuinely interested in the answers (still re-learning here), but read the wrong way my response might sound like I'm just being disagreeable

There would be separate power on each side, and separate ground could be accomplished, but why would I need separate ground.

I'm not sure why you think I would have gobs of noise in the cable.

Could you elaborate a little more on using a shielded cable incorrectly

Don't worry about sounding argumentative -- my response was relatively terse on a number of very complex subjects.

The reason you need separate grounds is because noise and spikes can couple between the systems on the ground. For opto-isolation to be worth anything, the two systems really have to be isolated. Grounding turns out to be a complicated subject as soon as you start to unpack it. The best electronics troubleshooting I've ever gotten is to check the grounds first. It's the problem half the time, so you might as well get it out of the way. There is one and only point in your circuit that is actually a 0 volts -- canonically, the negative terminal of your power supply (battery, in this case). All other points are at some other voltage. Forgetting that has a remarkably high probability of tripping you up.

You'll get noise on the cable because it's ten feet of wire and makes a dandy antenna to pick up any and all radiated noise in its vicinity. And cars have lots.

The trick with shielded cable is that you have to get a very good low impedance ground connection to the shield on one end. the other end needs to be free, in order to avoid creating a ground loop, i.e. an even more effective antenna. And impedance in this case means not just low resistance, but low inductance and capacitance as well. The soldered drain wire really doesn't cut it -- you want a good 360 degree solid connection. A clamp that fits around the shield inside a similarly shielded enclosure is really good.