Noise Filtering

[CMac89]

Are we going to see noise filtering available for selectable inputs, i.e. TPS or inputs applicable for additional VR sensors (wheel speeds)?

[jsmcortina]

There is already noise filtering built in for TPS and VSS.

James

[Reverant]

I mentioned a while ago that I would like to see more filtering for the baro sensor AND for the MAP sensor when the engine is not running. Essentially whatever we are doing for the TPS (tps_ring[] array averaging?) would be enough.

[jsmcortina]

What use is MAP when the engine isn't running?

James

[Reverant]

What use is MAP when the engine isn't running?

James

When customers and tuners see the MAP value jumping around and not stable with the engine not running, they do question the quality of the product.

[jsmcortina]

Quantify "jumping around" and what ECU is this with. It is not something I've ever noticed.

James

[Reverant]

Quantify "jumping around" and what ECU is this with. It is not something I've ever noticed.

James

One of my MS3s. Standard 1K/0.1uF RC filter. If I simulate engine RPM on the bench, it immediately becomes stable. When not running, the MAP signal jumps around by about +/- 1kPa, whereas it's at 0.1kPa when "running". No issue with all the other sensors.

[jsmcortina]

When not running the MAP sensor is just sampled at a time interval. When the engine is running, the sampling is either windowed or averaged over the time period specified by the engine's operation.

James

[Reverant]

Obviously I know that much. Would it be difficult to time average the MAP signal when the engine is not running? I use time averaging when running and it works great. Same for baro.

[sedd]

My baro sensor reading jumps around quite a bit. it will read ok 99% of the time but will spike upwards for a fraction of a second. I changed the sensor and that improved it a bit. I feel this erratic behavior is likely due to wiring and noise. So far it is not worth trying to experiment to eliminate the noise with capacitors and such.

I was hoping to get a really long averaging time (10 seconds or more) that was just for the baro and not linked to map as I think it is now. Since baro doesn't change that quickly 10 seconds should suffice to get rid of the spikes. Of course this is a lazy way to get around the issue, not really a root a cause fix.

As a side note my baro sensor is located back at my exhaust pipes. Since baro correction has to do with exhaust back pressure I wanted to eliminate any effect of low pressure being created at the back side of the car. Since speed (mph) might affect that pressure and it also won't change that quickly I think a very long averaging time would still be ok.

I vote for a separate filter for baro for these reasons. I realize this might require a lot of programming and it might not be important to enough people. Maybe this change could be included in the list for folks to vote on for future upgrades to find out.

[sedd]

I just noticed that the baro signal was now included along with mat, clt and battery. I switched to 10 smoothing, from 50 today. I will see what happens to my baro number and report back on that level of smoothing.

I wonder if I should switch the MAP from 50 to a higher number? I have assumed the smoothing affects the fueling equations and not just how it is shown on logs. Do I want a much faster response for MAP?