Capacive fuel senders

[marchudson]

I've got the standard vans capacitive fuel senders with your capacitive to voltage convters. I've connected the red wire on the converter to pin 15 of the ems to supply 12vdc. I've connected the white wire to pins 22 and 31 respectively.

When the tank is empty, the pin 31 value is 1.049 volts when doing the calibration. When the tank is full, it reads 1.389 volts.

When setting up the ems for pin 31/22, I selected capacitive as the type of sensor. Should I have selected resistive since the converter is attached? Or is a difference of 0.3 volts normal from empty to full.

Really don't want to fill and drain another tank till I get a solid answer.

Thanks for your time.

 

[Brantel]

I know that with mine the value did not change much from empty to full. I sure wish they did change more because that would allow for better accuracy across the span.

It seems that when Dynon designed those converters, they expected more capacitance change from empty to full than what actually exist. I am sure they did this assuming there might be large variations in the capacitance values from one built to another so they had to leave room for that in their design.

 

[mmarien]

The magnitude of the measurement is probably more related to the sensor then Skyview. In your case the difference is (1.389 - 1.049) = 0.340 or 340 descrete calibration points. If the tank held 340 gallons it may be accurate to one gallon. If it holds 34 gallons then the resolution is 1/10 of a gallon.

In my a/c the different was 0.0781 but it is displayed to four decimal places so I have 781 descrete calibration points. For 130 litres the accuracy is better then 1/10 of a litre.

IMHO the fuel sloshing around is going to create more inaccuracies then the resolution of the measurement. With 80 litres on board I have 120 litres in a climb and 40 in a dive.

 

[Brantel]

The resolution is not determined by the number of decimal places they display the data in... It is determined by the bit resolution of the A/D converter on the input of the box.  

Dynon, what is the resolution of your A/D on the Legendary and Skyview products?

 

[marchudson]

Brantel,  I agree, a larger span would be much better especially given the fact that the converter is supplied 15VDC.  Given that I'm only seeing 0.34V change you would think that noise is going to start having a large play in this small range not to mention repeatability.    My capacitive senders are reading 221pf when empty and 285pf when full.  

I'll give Dynon support a call on Monday morning to verify that I have all the settings correct.  If I do, I might have to build my own capacitance to voltage converter that will indeed scale a little larger.

Hopefully I won't have to resort to that.  

Marc

 

[Brantel]

While at first it may seem like this is a big deal, it has not been in my RV7. I have not suffered much from it.

I use my fuel flow computer way more than the fuel indicators. It is accurate to within .1 gallons at every fillup. I trust it after verifying the fuel visually before each flight.

I thought about ripping into the Dynon sensors to see if I could alter the bias and or span to make them change more but I have not done so yet. They work good enough so I have not bothered.

Maybe Dynon will just share the schematic with us and that would help????

 

[mmarien]

Brantel - I believe all of the Dynon inputs are analog. (Page 4-13 of the Installation Manual). The capacitance to voltage converter is just converting from volume to pressure which are both analog measures.

If the sensors are analog wouldn't the number of calibration points be infinite? limited by the precision of the display device? :-?

 

[marchudson]

mmarien

The inputs are analog. However, the analog to digital (A/D) converter resolution will determine how small of a change in voltage the converter can handle. Also, as you start to make the change in voltage to the inputs smaller and smaller, noise will become an issue as will repeatability.

Given that we are measuring fuel quantity, I would feel more comfortable using the full range of the sensor instead of trying accurately read a small fraction of it.

Bottom line, I just want to know if what I have is going to work and if it is going to reliable indicate 1 gallon of fuel remaining when in fact that is what I have in the tank.

 

[Brantel]

The input is analog 0-5vDC. This analog signal must be converted to digital for the processor inside the display to process the variable.

The capacitive to voltage converter converts the capacitance of the tank sender to a voltage range of 0-5vDC. The problem is the capacitance does not change enough to get 0-5vDC out of the converter.

The analog voltage out of the converter then goes to the display analog input then gets processed thru an A/D converter.

The resolution of an A/D converter is dependant on the number of bits the A/D converter uses.

For a 8 bit A/D converter the 0-5vDC signal would be broken down into 256 digital values or a resolution of 0.01953125 volts. This would be the minimum step size on such a converter.

For a 10 bit A/D converter the 0-5vDC signal would be broken down into 1024 digital values or a resolution of 0.0048828125 volts.

For a 12 bit A/D converter the 0-5vDC signal would be broken down into 4096 digital values or a resolution of 0.001220703125 volts.

I have no idea what type of A/D converter Dynon uses but it is critical to know this in order to calculate the resolution of the fuel indication input.

As you can see if Dynon is using a 8 bit A/D, his .34 volt change would only have 17.408 steps of resolution empty to full. If he had a 21 gallon tank that would be a 1.206 gallon accuracy.

If it is a 10 bit A/D, 0.302 gallon accuracy.

If it is a 12 bit A/D, 0.075 gallon accuracy.

Of course all of this is theory only and there are more variables involved but this would get you close....

 

[mmarien]

Ok - starting to see the light. The converter limits the resolution which is not necessarily the precision that is displayed. I had a look at the data for one of my tanks. The differences between calibration points is:

0.1561
0.2056
0.0719
0.0513
0.0120
0.0086
0.0086
0.0073
0.0074

These are differences between the readings not the actual readings.

If I said that the smallest increment is 0.008 and divided that into the rest I get some x.5's results so I'd have to say the smallest increment is 0.004. Divide 5 volts by 0.004 = 1250. By experiment it looks like at least a 10 bit A/D converter. But it also could be 12 bit with increments of 0.001 given the last four differences.

 

[dynonsupport]

The A/D in both D10/D100 and SkyView products is the same 12 bit converter. Theoretically this gives us .0012V resolution.

The converter is fully microcontroller based and there is no way to change the gain without reprogramming the micro.

 

[marchudson]

Dynon support

The problem is not with the a/d converter. It is working very well with good resolution. The problem with multiple users is with the capacitance to voltage converter.

Brantel is seeing a 0.40 voltage change total from empty to full. See this post for a graphical representation of this. http://www.vansairforce.com/community/showthread.php?t=55492&highlight=Capacitance+values

mmarien is seeing a 0.52 voltage change total from empty to full from his post above.

I am getting a 0.34 voltage change total from empty to full.

Is there any way I could get a copy of the schematic of the capacitance to voltage converter?  I would probably be able to change the gain values given a schematic. If not, what would you suggest?

Thanks. Marc
marchudson99@gmail.com