Identifying and Dealing with EMI, searching and searching Options

[McMark]

I would like to spend some serious time doing a qualitative analysis of the EMI that is probably going on in my engine bay. I've been searching Google for some write ups that might help me get started, but everything seems to be quick marketing blurbs for every electronics company on the planet. Does anyone have some links or books or anything that might help me?

Here is how I would like to approach the problem:
1. Evaluate my current situation.
2. Identify the sources.
3. Implement effective corrective measures.
4. Judge the effectiveness of my efforts.

This is probably over the top, but being the son of an engineer makes me think this way. (IMG:http://www.914world.com/bbs2/html/emoticons/wink.gif)

So give me your best links and books regarding identifying and dealing with EMI, please.

[GWN7]

Yep, over the top, no need to put the EMI in there..... (IMG:http://www.914world.com/bbs2/html/emoticons/wink.gif)

[lapuwali]

Don't try to reduce the EMI, shield against it. The ignition is going to generate a lot of EMI, no matter what you try.

I'm assuming this is your one-off Kit Carlson setup? If so, make sure the electronics are sealed in a metal box, and that the box is grounded. Twist any sensor wires that come in pairs together, as this reduces their sensitivity to noise. If this isn't enough, or you have a single sensor wire picking up noise, cover it with conducting braid with one end grounded.

Run separate ground wires for sensors showing too much noise (TPS tends to be susceptible to this), rather than running the signal ground through the car body. Again, do the twisted pair thing with these.

[bd1308]

somebody i thought said that having multiple ground points on the car increased noise.....

[McMark]

I'm wondering if I can use my oscilloscope for step 1. But I'm not sure if I should use it in series, in parallel, or just ground to the negative battery post and probe wires. (IMG:http://www.914world.com/bbs2/html/emoticons/confused24.gif)

[TonyAKAVW]

Never underestimate the need for a good source of power. Start off by making sure that your power supply is clean and can deliver sufficient current for the kitcarlson. That means big enough wire, and filter capacitors at the input unless they are present on the kitcarlson circuit board.


After the power supply is guaranteed good and noise free....


An oscilloscope could be a good way to start looking at things. In general you want to use an oscilloscope in parallel. Touch the tip of the probe to the line in question and ground the probe ground to as near a ground as possible. I would guess anything more than 4-5 inches away is probably too much. So that pretty much rules out the battery post as your ground for most things unless you have a very tiny engine.

I would think that rather than starting by finding your sources of noise, find those components which are sensitive to it and figure out how to shield them. EMI from igniton can probably be reduced, but the fact is that you have very high voltage being switched, and thats just bound to cause problems.

lapuwali has excellent suggestions. Shielded twisted pairs is the way to go for sensors, and shield only at one end. bd1308's point about too many ground points can be right. Tiny differentials in voltage from one ground point to another can result in ground loops, and cause noise issues.

I'm not sure how the kitcarlson electronics are designed, but its possible to add components such as capacitors and inductors to filter noise. Depending on the sensor however, the values of these might be critical.

So I would start by looking at what sensors are picking up noise and shield them. Trying to reduce generated EMI is probably not the best awy to go.

I have an interesting article on grounding that I can send you if you are interested.


-Tony

[DNHunt]

Mark

Don't forget proximity. Move things away from the source. Finding EMI in my car was easy. If I took my DVM anywhere near the coil and spark plug wires it would go nuts. That's no way to quantify it that way but it is graphic. I finally changed to 4 MSD plug wires out of a 5.0l Mustang set and routed them toward the flywheel end of the engine bay. They cleaned up a lot. I moved the coil pack to the right side of the engine bay. The only sensor that breaks up now is the CHT.

I'm not sure if this would help or not but, a lot of MS guys starting out have noise problems in their tach imput signal. Some end up adding a filter with a capacitor and a resistor. You might want to search over there. My tach input is sheilded all the way from the VR sensor to the ECU. As has been mentioned only ground 1 end of the sheild. If the cable goes thru a connector, continue the sheild. I pick a pin and solder the sheild to the matching male and female pins

I had a lot of noise in my MAP signal which made it impossible to tune. I mounted it in the engine bay and had wires running through the tunnel to the ECU in the console. The reference voltage is 5v and the signal can be pretty weak. there was a filter on the board so I added a pot and tuned the resistance to get a clean signal.

Dave

[KitCarlson]

Mark, All,

I am back and hope I can be of help. Here are a few details. The KitCarlson EMS draws only 100mA from the supply. No huge wire or special filters needed. A good battery is always required for electronics. A bad or weak battery cell and a mechanical voltage regulator can results spikes and sags. The best ground is at the battery terminal. It avoids problems if the transmission strap is broken (an often neglected part), however you should have a good strap, since the head temperature sensor ground returns there.

The multiple grounds on the EMS are essential, each one takes the driven current to ground on a separate path. This means the ignition coil and injectors each have their own grounds. This is a good thing. It avoids ground loops. Supply for coil and injectors come from the same places as in the orignial D-jet, sorry, no magic, no free energy for CA. The improved injector system differs in the drive system (sinking instead of sourcing), but that is a long story. The good part is it avoids the grounds on the engine that get dirty and are hard to find buried under the intake.

hmmm...
I think your problem is with spark plug wires. The system is direct fire so you can pull the plug wires on #1 and #3 , leave connected to the coil. Measure continuity end to end it should be 15K to 20K ohms. Repeat with #2 and #4. If you measure "open" you have a break at plug terminal connection and are making a huge spark in a bad place! This will lead to missing and running like c#ap. Recently found this on another test unit on a Turbo engine.

Do not use solid wires! However, I have used them in development without problem. It works because of the resistance in the coil and plug ends. The solid wire radiates much more than carbon wires. No use making things more difficult.

Also the layout of the harness wires and plug wires is important. Avoid close proximity of the plug wires and the harness wires. There is a current loop from ground to plug #1 to the coil out the coil to plug #3 then to ground. Same for the #2 and #4 loop. Placing harness wires under the spark plug wires puts them inside the loop and couples energy. A transformer....

An example of bad is placing the injector harness on the engine tin with the plug wires looping over them. They should be outside the plug wire loop, a very simple change with a huge diffence. Don't worry if you do not understand this.

The real solution is for me to write a good install manual with pictures showing routing prior to going to market. I am just finding not everyone thinks like me, thinking about current loops and mutual coupling for every wire connection.

Please snap some pictures with the aircleaner off showing wiring I might see some fixes.

Dave
aka KitCarlson

[john rogers]

I had some EMF or whatever you want to call it when I first installed my MSD system, coil, etc. I had bundled the wires nice and tight together and the MSD tech said that was a no-no, to leave them loose. I did that and the static in my race radio went away!

[Britain Smith]

QUOTE (KitCarlson @ Aug 19 2005, 06:56 PM)

The multiple grounds on the EMS are essential, each one takes the driven current to ground on a separate path.  This means the ignition coil and injectors each have their own grounds.  This is a good thing.  It avoids ground loops. Supply for coil and injectors come from the same places as in the orignial D-jet, sorry, no magic, no free energy for CA.  The improved injector system differs in the drive system (sinking instead of sourcing), but that is a long story.  The good part is it avoids the grounds on the engine that get dirty and are hard to find buried under the intake.

Are you saying that you should use different grounding points for the injectors and sensors? I had the impression that all the grounds should go to the same point.

-Britain

[McMark]

I think he's saying that they all have individual ground wires that go all the way to the common chassis ground. Multiple ground wires, not multiple ground points.

[914GT]

McMark, one of the most widely used reference books used by beginning EMC engineers is Noise Reduction Techniques in Electronic Systems by Henry Ott. Scopes have limited usefulness for identifying EMI. You can use a scope probe or small loop of wire on the end of a 50 ohm coax to 'sniff' sources of noise, but it may be hard to trigger on anything. A high speed digital scope helps to trigger on transients. I doubt if you have one of those. As mentioned earlier it's very hard to directly probe due to noise pickup on the probe ground lead. It all depends on the frequency of the interference, the higher the frequency the worse it gets. A very useful tool is a wideband current probe and a spectrum analyzer, which I doubt that you have. The suggestions to twist leads and keep noisy circuits away from sensitive circuits is good practice. Clamp-on ferrite cores are sometimes quick easy ways to suppress noise on cables. They help reduce 'common-mode' noise on cables, whether they are the 'source' or 'victim' circuit. These won't work on RFI-suppression spark plug wires however.

[KitCarlson]

Mark,

What type of EMI problem are you having? Should of asked this first. I have not been keeping up the last several months. Could be connection related.

You might want to use the scope to check the voltage of the charging system. A bad alternator diode and weak battery can make a very noisy supply.

Another easy test is to check voltage drops of connections with a meter or scope. An example is place one probe on the ground at battery, the other at an engine metal part. If you see more than a few tenths of a volt AC or DC you have a poor engine ground. The same can be done for supply or other connections. The problem in using this to look for EMI is that the loop of wires you use for the connection pick up radiated EMI and lead to false reading. Test this theory by connecting the scope ground to scope probe tip making a small loop of ~2". You should see noise on the scope in close proximity to an engine. Move it it around to see where the greatest noise is. You do not need to contact any point. Stay at least an inch from the plug wires.

Dave

[Joe Ricard]

I work with a 3 guys that make a living dealing with EMI. Serious pocket protector wearing type Genious guys. I'll ask them what they think but be prepared for more info that you ever wanted to know about Electromagnetic Inteferance.