Filters

[FoxSTI]

Good afternoon folks,

I did not want to bring up an old thread but I want to ask if filters to stabilize a parameters have been incorporated.

the example below references a fuel gauge value that jump around as the fuel moved in the tank.
How can we replicate a slow more natural gauge behavior? I am currently using Display creator. Some mentioned a 2D table? The would still cause some jumping I think.

"Working on the assumption that the C125 setup is the same as the C127, about all you could do would be to increase the filter value to average the reading over a longer time."

Looking forward to responses, thank you !

[adrian]

There is a feature request to add filtering to channel values/bars in DC but at the moment with a C12x Display there isn't any way of doing what you want. I believe this feature wont be added to the next release (current beta on the website) but will be added to the release after that.

If you have a C18x Display then you can just use maths to filter the channel.

[FoxSTI]

I only have a C125. I am showing fuel at a % number.
How do race cars handle his?

[adrian]

Fuel level is extremely difficult to get working nicely because the fuel simply moves around too much. There are a number of different methods OEM Manufacturers use to give you an estimate of fuel level. Some will only read the sensor value when you power up the car and then use Fuel Used calculated from the ECU to adjust the gauge as you drive.

Others will do a similar thing but poll the sensor every couple of minutes and use really heavy filters.

You might find some race cars with fuel level sensors but they are basically useless unless the car has been stationary for a while. Most of the time they will measure the amount of fuel they put in the car and then use either a fuel used calculation or a fuel flow sensor to workout how much fuel is still in the car.

[the_bluester]

I can speak to that one having run through exactly the same issue. Particularly with an older car with a lever and float setup rather than many modern ones with a float moving up and down a resistive post in a tube, the error can be greater than the fuel level.

You could if it is coming from an analog input put an RC filter on the input to slow the response to changes but accuracy will still be average in motion. Worse, the error will generate a lower than actual reading at high fuel loads (The error as the float flaps can go down but up will mean hitting the stop) and more critically, a higher than actual reading at low fuel for the same reason, the float can hit the stop at the bottom which limits the low side of the error but not the high.

I am moving towards a fuel used strategy and knowing how much fuel I have put in the car.

[FoxSTI]

can you please elaborate on that strategy and how it works?
is this what is it used on race cars without an OEM gauge?

I want to know how GT3 cars do it.

[adrian]

GT3 cars have a similar setup to what I explained. They have some way of measuring the amount of fuel going into the car. So calculate the flow rate of the fuel rig and fuel for some amount of time or have the fuel rig on scales so they know the weight of the fuel going in. They might have sight glasses etc etc.

Once they know how much fuel is in the car then they either use 'Fuel Used' calculated by the ECU (challenging to get perfect) or use a fuel flow sensor to workout how much fuel the car is using and then its pretty easy to work out whats left in the car.

[the_bluester]

In my case the ECU (Not a MoTeC one on my car) spits out a fuel used value on the CAN bus. Each time that increments, the dash uses that to increment it's own fuel used value. from my ECU it is not a fuel used pulse but a discrete value of fuel used since the last reset, which can be configured to be on each power up, on a specific condition, never, etc..

I will start a weekend with a known fill either a full tank of a specified amount from a dry tank and a reset on the fuel used value in the dash so when the fuel used display gets within about 10L of the amount of fuel the tank contained plus any top ups, time to give it a drink.

I will be adding a fuel level input as well, but per previous comments that is only really useful if it is well calibrated and the car is stopped on flat ground.

Adrian, regards flow meters, on most cars with a full flow fuel rail, how do they manage that, two flow meters one on the fuel delivery line and one on the regulator return to measure the difference? Are there any differential flow meters that achieve that in one package?

[adrian]

Yes you would need two fuel flow sensors. As for differential sensors, I am not sure. I haven't come across one, but I am certainly no expert in this area.