Diesel Fuel Level Sensor

Pieter Potgieter33

Other Parts Discussed in Thread: FDC1004

Our company has developed a Capacitive fuel level sensor, using two stainless steel tubes the diamenter of the inner tube is a=5mm and the outer tube is b=15mm.

The length of the tube is L = 2300mm and according to this formula Capacitance is: C = ((2*pi*Eo*Er) / (ln (b/a) ) * L. (The Er for diesel is about 2.)

This then equates to an empty tank C = 100pF and a full tank is double that (Er=2) = 200pF. - (real world we get about 120pF to 240pF)

Currently we are using a RC type oscillator that generates a dutycycle that is converted to a DC signal - we are using 32kHz oscillator for this. and a 30k resistor.

However this circuit is very sensitive to temperature effects and also we need to use very high spec resistors and caps.

The current circuit is only 1% accurate - we are looking at upgrading this circuit to use the FDC1004.

Requirements:

- I need to measure at lease 0.1% accurate - that means on the 2300mm lenght I want to be able to measure 2.3mm or better - 1mm is about 10 liters of diesel.

- Probe's outer tube is grounded - so this means single ended measurement?

- Electronics sits directly on the probe - goes out via micro and rs485

Question:

1. Can I put the sensor probe in series with a 15pF NPO type capacitor - and then subtract the 15pF offset using the FDC1004 - this will give me a range from 0 - 15pF - that is full scale for the chip.

What is a recommended way to get the 100-200pF probe capacitance into the 15pF range of the FDC1004 ? Will my Series cap plan work? (combined capacitance of 200pF(probe max) and 15pF = 13.9535.

Will it be better to rather us a 200pF cap in series so the range is then (200 + 100)=66pF to (100+100)=50pF.

I can then use the offset subtraction to get to 0-15pF ?

We only need to measure relative capacitance - calibration will take care of absolute values.

2. Please explain the temperature graph on page 7 - the gain drift vs temperature:

Does this mean for temperature drift from say 20 degC to 40DegC there is a 200ppm drift - what does that mean? is it 0.02% of 16bit value? 13.1072 clicks on the A2D?

Thanks in advance for the help - if anything is unclear please ask me to clarify

Pieter

over 10 years ago

0 David Wang11 over 10 years ago

TI__Expert 7400 points

Hi Pieter:

Thanks for your interest in the FDC1004. We have done initial tests with the FDC1004 and diesel and it is possible to detect the change in capacitance. your sensitivity/resolution will be based on the size of the sensors. We have not done extensive tests for capacitive tube approach, so I will not be able to comment on the sensitivity based on that.

To answer your questions"

Can I put the sensor probe in series with a 15pF NPO type capacitor - and then subtract the 15pF offset using the FDC1004 - this will give me a range from 0 - 15pF - that is full scale for the chip?

ANSWER: The FDC1004 has the ability to offset a single ended measurement configuration using the CAPDAC. The max offset that is capable is 100pF.

What is a recommended way to get the 100-200pF probe capacitance into the 15pF range of the FDC1004 ? Will my Series cap plan work? (combined capacitance of 200pF(probe max) and 15pF = 13.9535.

ANSWER: Technically, it is possible to put a cap in series with the sensor (which is connected to one of the channels CINx). It will work, but the issue is that depending on the capacitance you are trying to offset, in your case is, 100-200pF, and getting it into a 15pF, you would need possibly a 5pF cap in series with your sensor. Having such a small cap compared to your sensor capacitance and sensitivity will dilute your signal. The change that you expect to see will be less than what you actually see and measure. You will need to verify this with your setup since I do not know how much it would change your sensitivity. All I can expect is that your signal will be diluted.

Will it be better to rather us a 200pF cap in series so the range is then (200 + 100)=66pF to (100+100)=50pF. I can then use the offset subtraction to get to 0-15pF ?

ANSWER: This will probably affect your signal much less than using a 5-15pF cap in series.

Please explain the temperature graph on page 7 - the gain drift vs temperature. Does this mean for temperature drift from say 20 degC to 40DegC there is a 200ppm drift - what does that mean? is it 0.02% of 16bit value? 13.1072 clicks on the A2D?

ANSWER: The way that it is obtained is that for each temperature, 4 capacitances are measured (1pF, 3pF, 8pF, 12pF). The slope is calculated with these 4 points, and then gain error % and ppm value are calculated. So effectively, ppm/C is based on capacitance change not on the 16 bit value.

Hope that helps

Thanks

-David Wang

Capacitive Sensing Applications

0 David Wang11 over 10 years ago in reply to David Wang11

TI__Expert 7400 points

Also please check out our liquid level app note (www.ti.com/.../snoa925.pdf) and Out-of-Phase liquid level solution (www.ti.com/.../TIDA-00317) reference design.

Thanks
-David Wang
Capacitive Sensing Applications

0 Pieter Potgieter33 over 10 years ago in reply to David Wang11

Prodigy 30 points

Hi David
Thanks for your reply - I have ordered a dev kit and will evaluate the Chip with our probe - I do hope it works because it will save a lot of discrete analog electronics.
Just to be clear - a 200ppm value on a test cap of say 8pF will be 0.0002 * 8pF = 1.6fF error due to temperature?

0 David Wang11 over 10 years ago in reply to Pieter Potgieter33

TI__Expert 7400 points

Yes that is correct for the gain error drift.

Thanks
-David Wang
Capacitive Sensing Applications

Sensors

Sensors forum

Diesel Fuel Level Sensor