LC filter for automotive power entry.

Go To Last Post
9 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have been struggling for the last few hours trying to find a suitable power inductor for a LC filter at the 12V (car 12v) power entry on my board. The reference design used 150uH/1000uF. I have no problem finding the cap, but finding a large inductor that will take +-15A RMS loads is another story. They are either inexistant or very expensive (like 15$ a pop).

Does anyone with experience on automotive power supplies have a suggestion of a LC combination that works well to filter off most of the noise using lower inductance? 12V entry first goes through a schottky barrier, then I have a 20V 600W TVS clamp, and LC filter would come next.

Or maybe I could wind it myself too, I can find the ferrite, but I don't know much about winding inductors. How many turns and what size wire to use?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You might look for "common mode choke". Another kind is called is "hash filter".

Common mode chokes are two windings on a core. They happen to be wound so that the DC fields cancel out in the core and it does not saturate. These are not rated by inductance value but by attenuation at different frequencies.

You can often find hash filters wound on a simple ferrite core that looks a bit like a big resistor. The wire (heavy gauge) is wound on several layers on the outside.

Did a search on DigiKey just using the term hash filter and came up with at least a page full of entries. Here is an example: it does not fit your current rating but it should provide some ideas...

http://search.digikey.com/script...

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

$15 a pop doesn't sound too bad! The core alone is worth a few $. At a guess, if it is wound on a toroid, the toroid diameter is going to be > 1". Since it is passing DC, you need to watch out for core saturation, therefore you either need a gapped core or a large core.

Amidon used to have some good tables, but these seem to have disappeared. Unfortunately, I design inductors so infrequently, I tend to forget all the equations! There should be plenty of stuff on Google to help. I tends to be an iterative process - choose a core, do the inductance calcs, check the max field vs the core saturation value, iterate until you get a valid result. From that you'll get the number of turns, then you can use tables to choose the wire size. Check to see if it will fit on the core, re-iterate the calcs.

p.s You want to choose a core material that is lossy at high frequencies - iron is particulary lossy.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you both for your suggestions.

Now I have a question. Since the 12V source is already biased and clamped by the Schottky and TVS clamp, do I really need a tuned filter there? Wouldn't a ferrite bead and a big cap do the job? Beads are much easier to get by and much smaller...

Maybe something like this:

http://ca.mouser.com/Search/Prod...

?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well......

It all depends. Are you trying to keep noise in or out? And what kind of noise?

Voltage clamps just keep the input from getting too high (spikes) or negative (polarity reversal). They do little or nothing for general noise. Automotive noise tends to be things like fuel injector transients, ignition noise, noise as various things turn on and off (turn signals, air conditioning controls, and such). It tends to be lower frequency, hence the recommendation for the big inductor.

Beads are designed to filter out noise in the MHz range - say from 5MHz to 50MHz, more or less. They really help keeping processor clock noise inside your box but help very little for vehicle noise. because it is at a lower frequency (except, sometimes, for ignition noise).

Much of this, and how much is needed, really depends on how susceptible your hardware is to external noise. Unfortunately, it IS difficult to quantify and that makes component selection doubly hard.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The 12V source will be feeding a 3-phase buck and it's high power transistors. Calculated draw at 80% efficiency would be about 8A, so I need the path to have very little DC resistance. This is the main problem I have with large inductors, the current demand and the low DC resistance requirements makes them very large physically, and very expensive.

The input to the DC converter is rated to 30V max RMS, it really needs just enough filtering to be safe to the converter. Two big caps after the schottky to get the lower frequencies, and the ferrite bead to filter off the mid range band I think should do the trick, with the TVS catching the occasional high voltage transients. While the bead would keep board noise in, it would also keep power noise out in its working band, yes?

For the record I will also add a relay before the whole thing, controlled with a delayed (5 seconds) ignition signal, in the hopes of avoiding the ignition noise altogether.

Last Edited: Fri. Feb 27, 2009 - 12:27 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Are you filtering the power supply for the micro and a few sensors, or for some power hunger loads?

Often the loads can run off a less well protected power supply.

This significantly decreases the current load for your protected power supply.

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The DC converter is the one from the other thread. :)

Average load will be somewhere around 30A @ 2.5V, converter also serves as current limiter and some more protection (undervoltage detection, short condition, etc..).

The AVR and associated logic are powered from a separate linear regulator.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You need the inductor filter to keep your converter trash from radiating back into the car electrics.

Delaying the startup won't get rid of ignition noise - you get ignition noise as long as the engine is running. When you crank you get the brownout due to the starter motor - I think putting in a relay is probably a waste of time.

Think twice about putting electros in a high current, high temperature auto application. They tend to fail much sooner than you'd like. Mitsubishi had some electros in their ECUs some time back - they were a common fault. I rarely see an electro in late model ECUs these days.

With the TVS - make sure the supply is fused as TVS diodes have a habit of going short circuit. I prefer varistors as these tend to be more resilient - if they explode you know you've had a major transient!