Fuel probe for use with Dynon driver

[lolachampcar]

Can someone advise on the correct capacitive fuel probe with BNC connector for use with the Dynon interface board? My application requires a 1/8" NPT mounting system so an 18" sender with 1/8" NPT mounting would be a bonus.
Thanks,
Biil

 

[dynonsupport]

They're really made for use with the Vans capacitive plates that are used in some RVs, but can work with others that actually need the capacitance between two points measured. If you're looking for a full capacitive probe, with an output that can work with our systems, the ones from Princeton Electronics work well. They're sold at Spruce. They actually output an actual voltage that doesn't require the cap-to-volt converters.

 

[lolachampcar]

Dynon Cap to Voltage Converter Drive Capibility

The Cypress Semi data sheet indicates the Dynon Capacitive to Voltage converter should be able to source and sink current on its outputs. Can the Cap to V converter drive one of the EMS' 10K pull to 5V loads? I ask because I need the "no pull up" inputs for other tasks and still need three fuel tank inputs.

BTW, nice job on the Cap to V converter!

Thanks,
Bill

 

[lolachampcar]

I did a first test with a simple aluminum welding rod suspended in a 3/16 ID aluminum tube. There was not a lot of surface area on the welding rod so the 18" long assembly produced a clean repeatable 170mv full scale swing. What do you normally expect (and what does the Skyview system like) to see for a good empty to full swing in voltage?

 

[lolachampcar]

Changing the series resistor used to charge the probe would increase sensitivity to lower capacitance probes while increasing noise. Apart from noise and stray capacitance affects, are there any other concerns to increasing the size of the series resistor?

 

[lolachampcar]

With respect to driving a 10K ohm pull up load, I have answered my own question. Output filtering is not done in the micro but in an external RC circuit thus there is no drive capibility for the board assembly.

 

[dynonsupport]

Changing the series resistor used to charge the probe would increase sensitivity to lower capacitance probes while increasing noise.  Apart from noise and stray capacitance affects, are there any other concerns to increasing the size of the series resistor?

lolachampcar:

Reinterating - The Dynon Capacitance-to-Voltage Converter is designed for, and intended for use with the fuel level capacitance plates in Van's RV planes. If a customer chooses to buy it and use it on other capacitance fuel level types, Dynon cannot offer any support.

 

[dynonsupport]

Regarding the cal process - strictly speaking, it doesn't stop you from using successive pours that have even the same voltage. But it warns you at some small amount of change (.1V) that you probably don't have a real sensor change. But as long a a voltage is repeatable, it's usable as a cal point. We can resolve down to just over a single millivolt, so if you had 100mV over 16 gallons, say, there'd still be 5ish steps per gallon. 5mV / gallon is about as low as you'd want to go if you want single gallon accuracy (with noise and offsets, that may or may not be usable, and certainly isn't optimal).

 

[lolachampcar]

Where to begin....
The part used on the Dynon capacitive probe driver board is a PSOC part by Cypress.  The interesting thing about PSOC is that it contains both digital and analog resources.  I will not get into details here but will instead concentrate on why that is a good thing.

There are two issues for me.  First, I'm not doing an RV so I need a different range of capacitance for my probes.  Basically, I have a 1/8th inch NPT hole provided at a 45 degree angle to both my wing tanks and the center tank.  The wing probe is 18" while the center is 24".  This means I have two different base line capacitance values, both of which differ from the RV value.  My probes are simply a 3/16” aluminum tube inside a 5/16” aluminum tube.  The dialectic of air is one and Avgass is two so the capacitance of the sensor will be one value when the tank is empty and roughly double when the tank is full.  I say roughly because the “empty” air value of the probe is very small so other natural parasitic capacitance of any sensor board will be a significant part of the total.  This is not a problem but must be considered.

The second issue is that I do not want to be bound by the Skyview limit of only three high impedance inputs (ones with the ability to turn the pull up resistor off).

The solution was to get the PSOC development system and see if I could make use of the current Dynon processor’s native resources.  The standard Dynon part connects one probe plate to 5V through a ½ meg or so resistor.  The other plate is hooked to ground.  In my case, the inner tube is hooked to 5V via the resistor and the outer tube is hooked to ground.  The inner tube is also connected to two processor pins.  One of the pins is configured as an open collector output while the other pin is a comparator input.  The processor simply holds the output low most of the time.  When it is ready to measure, it lets the output go which allows my center tube to charge through the ½ meg resistor.

 

[lolachampcar]

The comparator input is used to capture the amount of time it takes to “charge” the inner rod to a given voltage.

The standard Dynon board’s processor code then outputs a square wave proportional to the charge time. This output goes through a resistor to a couple of 10uF capacitors to integrate the result and produce a voltage for the Skyview system. The source impedance for this output is very high so it can only go into the three pins on the Skyview system that have a no pull up resistor option.

Now, on to the good stuff. My first job was to talk to the part on the Dyon board. That turned out to be very easy. I loaded the free PSOC designer software (part used is a PSOC1 part) and plugged the MINIPROG3 (USB programming dongle you can get from DigiKey for $29 part number CY3219) directly into the header on the Dynon board. I had to solder on a five pin header first. I then compiled the DAC example program and sent it out the BNC center conductor. It took all of a half hour and I had a nice zero to 5V saw tooth output on the center BNC conductor. I could now change the code running on the Dynon board.

I’m going to do the exact same thing that the original Dynon code did on the input. That is, I’m going to release an output to let the fuel probe “capacitor” charge and count the amount of time it takes to charge. I will then use a look up table to equate that charge to a DAC (Digital to Analog Converter) voltage output. I will then send that output to the Skyview system using an output pin configuration that lets it drive the Dynon 10K pull up directly.

The above approach allows me to (1) scale the code to my (or any) base probe capacitance and (2) hook the board up to any old input for the Skyview system. The charge resistor can be changed if necessary but I am hoping it will not be necessary. The output filter resistor must be shorted (solder blob if your lazy like me) and those two 10uF capacitors removed.

 

[lolachampcar]

Of course, the code must be changed but that takes all of about two minutes.

I’ll get this up and working and post the PSOC design file with commented code in case someone has a different set of probes they want to use.

I’m also prototyping the dual capacitive oscillator circuit I found in one of the online articles for a completely analog solution (no processor). In addition, I’ve got a bud familiar with the AVR series processor generating some code for the ATTiny to do roughly the same thing. The ATTiny is an 8 pin processor that costs under a buck. It can do the comparator counting input part of the task and will output a PWM which must be run through a resistor into a capacitor then through an op-amp to get the drive necessary to go into any Skyview input pin. It is kinda like the Dynon approach but with simpler parts. I started this path in case there was a gotcha with the Cypress stuff.

I’ll post the results.

 

[lolachampcar]

Update
I've got the DAC working to output any desired voltage.  The drive level can be set to high or "pull up" which is an open collector output with a pull up resistor enabled.  This should work against the Dynon EMS pull up resistor just fine.

I've got the comparitor working and providing an interrupt on the rising edge of the input signal.  I'm handling the interrupt.

The next step is to set the pin configuration correctly so that I can release the pin to charge the fuel sender capacitor then handle the resultant comparitor interrupt.

Progress

 

[lolachampcar]

Update

The cap to voltage board consists of roughly a half meg resistor from 5V to one plate of the level sensor (inner of two concentric tubes in my case) with the other plate tied to ground.  A Cypress microcontroller has two pins attached (pins 1 and 3) to the same plate that has the 1/2 meg resistor.  One of the pins is used by the micro to hold the probe plate to ground while the second pin is the input to the microcontroller's comparitor input.

I now have the comparitor generating an interrupt that sets the "Done" flag.  I am pulling the probe plate to ground, releasing it at the same time I start counting.  I stop counting when the Done flag is set (comparitor trips).  I then pull the center plate back to ground and deal with the count.

For my concentric tube probe, the free air capacitance is about 390 pf which equates to about 170 counts with the processor running at 24 MHz.  The full fuel capacitance should be about double if you believe the dialetric of air is 1 and fuel 2.  If this holds true, the full count should be 340.  I will subtract a value slightly less than the empty count so my count accuratly reflects the difference between empty and full.

I am already turning the processor clock down to 97 kHz from 24 MHz and delaying a little over a second between measurements.

What remains is to confirm the full fuel probe capacitance.  I will then add a digital filter to average readings over a period of time and provide that output to the DAC (digital to analog converter in the microcontroller that outputs a voltage) for input to the Skyview system. I have picked a 6 bit DAC which means there will be 64 distinct available output voltages (64 different tank levels) which should work well for most systems.  There are more than enough resources to use a higher resolution DAC if desired.

There is one hardware rub.  The current Cap to V board uses a digital pin for the output.  One of the analog output pins must be patched to the current output traces on the board.  I will disable the digital pin so it can stay connected without affect.  The resistor in the current digital output path must be shorted and the two 10 uF capacitors on the same path removed.  Once done, there will be a voltage output that can be plumbed into any Skyview EMS input.  If the chosen input has a pull up, it should be set to 10 kOhm.

Once completed and working, I'll post a picture of the board modification and provide the Cypress PSOC design file.  It can be programmed as is or the operating range changed to reflect probes of different capacitance. 

 

[peterk48]

Hi. Building a ONEX with a Skyview panel. Sonex offers two 1/8 NPT fuel probes for their fuel tank. One is capacitive and one is resistive. Which one will work better with my Skyview?

Thx. Pete

 

[jakej]

Pete

The capacitive senders are much more reliable & accurate, mostly.

Jake J

 

[Dynon]

Resistive will work for sure. Capacitive will work if they're an active device that outputs a variable DC voltage within 0-5V, or, if it's a device that needs its capacitance measured, they MIGHT work with the capacitance to voltage converters that we have that are intended for use with the plates built into some Vans fuel tanks. If the cap sensor outputs a frequency (pulses), it won't work.