What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Capacitive fuel senders, and problems with them

airguy

Unrepentant fanboy
Sponsor
Ok, so I've probably got a great big "Told ya so..." coming for this, and I'll take my lumps as required.

During my build of my fuel tanks, capacitive senders were all the new tech, everyone was talking about them like they were just absolutely gonna be the berries, and I drank the Koolaid. I built my tanks with capacitive senders, and in a wave of unsubstantiated enthusiasm, I did not include the standard float senders as a backup, though they would have happily co-existed in the tank.

The capacitive senders operate by means of an AC voltage applied to a couple of plates in the tank that are insulated from the tank itself, and this forms a capacitor that charges and discharges according to known electrical rules. The fuel is a dielectric between the plates that has a different value from air between the plates, and thus the computer can calculate the level of fuel in the tank from the capacitance characteristics of the plates.

This would probably work fine for most people - but I set my airplane up to run 91E10 autofuel, and it has a different dielectric than 100LL. No problem, I thought - I calibrated them for 100LL, and again for 91E10, and I'll just switch back and forth as needed when I change fuels. Well, as it turns out in practice, I nearly always have some mix of 100LL (due to long trips and fueling enroute) and 91E10 (which I have in the hangar and always top off the tanks with when I shutdown). The result is that I very rarely have a nearly-pure concentration of either 100LL or 91E10 in the tanks, and the capacitive senders are almost always inaccurate to some degree. This has caused me to rely on the fuel flow sender and fuel computer almost entirely, which has fortunately turned out to be incredibly accurate, within 0.5% from my observations over the last 135 hours.

Finally enough was enough - everything else works like it's supposed to on the airplane, and this should too. I ordered replacement access plates, nutplates, and float senders, and then proceeded with surgery this afternoon. I'm going to abandon the capacitive plates in place, no point in trying to remove them, I'm just putting the standard floats in place. They will respond the same no matter what fuel, or mix of fuel, I have in the tanks. Lesson learned and shared for others playing with autofuel.

For the record - it's right at 3 hours to pull a single tank.

IMG_0994_zpsmoajkw1h.jpg
 
Last edited:
I too installed the capacitance plates in my tanks and use the Dynon converter.

Since I only run 100LL, they are dead accurate. Every time I top off, I'm surprised at how good they are.

One thing to think about, you can plug the capacitance senders into to tanks 1 & 2 and make the floats tanks 3 & 4. Not that you really need to with the floats in place.
 
As I understand it, due to dihedral, won't the floats max out before the tanks are full? Just something to be aware of.

I have the capacative sensors and, while very accurate, my calibration was off just a wee bit...at first, they read about 1 gallon high from the mid-point down to a few gallons. I kept careful logs, plotted the curves, ran the interpolants and adjusted them manually in the Dynon menu. Now, they read about 1 gallon *low* from the mid-point on down. Arrrrghhh...I don't want to recalibrate them, as it is a PITA. So I'll probably just live with it for now, since it's "safe", and the fuel flow/totalizer function is dead nuts on.
 
Only problem I've got with the float/Dynon is gauges don't bounce! :p I always knew my Pacer tank was almost empty by the amount of bouncing my fuel gauge needles had to them!.......Actually, the only problem I have with the floats senders on my taildragger is they show about 4 more gallons per tank than they actually have when I'm on the ground with the tail down. But, that just reinforces my resolve to never fully trust a fuel gauge!
 
Float senders are by definition - inaccurate!

I think once you've installed the float senders and gone through the calibration process with them you'll figure out they are even less accurate than your capacity sensors. I'm glad your leaving them in place. Also leave the wiring alone and run new wires for your float sensors. That way you can convert back if you want to.

I, like most RV owners, have the float sensor system. It's accurate at empty (I hope) but inaccurate from full to near empty. I use the float sensor readings to estimate my fuel balance in each tank but rely on fuel flow readings (totalizer function) for actual fuel on board. Like the other commentator- it's much more accurate.

Others can comment.
 
Does anyone know why Vans stopped selling the capacitive kit?

My wings are not mounted yet and I installed the capacitive senders. Is it a good idea to also install the float senders while it's fairly easy to do? There might be a good reason Vans stopped selling the capacitive sender kit.
 
Only problem I've got with the float/Dynon is gauges don't bounce! :p I always knew my Pacer tank was almost empty by the amount of bouncing my fuel gauge needles had to them!.......Actually, the only problem I have with the floats senders on my taildragger is they show about 4 more gallons per tank than they actually have when I'm on the ground with the tail down. But, that just reinforces my resolve to never fully trust a fuel gauge!

You partially solve the bounce problem with the Dynon - there is a way to select how much time-averaging is done on the quantity sender between a couple seconds to a couple minutes, the default is a couple minutes to prevent the bounce but you can make it very short.

I think once you've installed the float senders and gone through the calibration process with them you'll figure out they are even less accurate than your capacity sensors. I'm glad your leaving them in place. Also leave the wiring alone and run new wires for your float sensors. That way you can convert back if you want to.

I, like most RV owners, have the float sensor system. It's accurate at empty (I hope) but inaccurate from full to near empty. I use the float sensor readings to estimate my fuel balance in each tank but rely on fuel flow readings (totalizer function) for actual fuel on board. Like the other commentator- it's much more accurate.

Others can comment.

The capacitive senders are quite accurate when you have the calibration right, and for a single type of fuel. My fueling habits are what was killing the accuracy - and I'm not likely to change those fueling habits - so I went back to the standard floats. I'm using floats now in the outboard auxiliary tanks and I'm quite happy with their response, I just need something in the inboards that responds better to a mixed-fuel condition. The capacitive senders would frequently be off by 3-4 gallons with a mixture of 100LL and 91E10, that was unacceptable. The floats will respond the same to the liquid level regardless of composition.

My wings are not mounted yet and I installed the capacitive senders. Is it a good idea to also install the float senders while it's fairly easy to do? There might be a good reason Vans stopped selling the capacitive sender kit.

I would, yes. They will happily co-exist and might save you grief down the road. To be fair, the capacitive senders were good as long as I kept a single type of fuel in the tank - it's just difficult for me to do that when fueling at home with 91E10 and getting 100LL enroute.
 
Last edited:
My wings are not mounted yet and I installed the capacitive senders. Is it a good idea to also install the float senders while it's fairly easy to do? There might be a good reason Vans stopped selling the capacitive sender kit.

IIRC, someone at Van's told me they didn't sell that many.

I love mine and would not want to go backwards in time to floats. The issue the OP had was when switching between fuels.

If unleaded avgas becomes available nation wide, I will recalibrate mine. Doing that is not difficult.
 
+1

IIRC, someone at Van's told me they didn't sell that many.

I love mine and would not want to go backwards in time to floats. The issue the OP had was when switching between fuels.

If unleaded avgas becomes available nation wide, I will recalibrate mine. Doing that is not difficult.

I'm not flying, but I used capacitors and also installed the floats, as a cheap, lightweight back-up. I expect the cap senders to be very accurate and went to great lengths to make sure they were installed correctly.
 
Knowing how accurate my fuel flow sender is, I wish I had bought two and installed one on each tank. It seems to me that measuring fuel used, and subtracting it from what was there at takeoff, while not *directly* reading the level, is far more accurate than my float-type senders.
 
I put Princeton single point set probes in the tanks when I built the Cozy MKIV. They were dead accurate with the original Blue Mountain EFIS and also the Dynon Skyview that replaced it. Still working perfect after 10 years. I installed the Princeton 5 point set probes in my RV-10 with Garmin G3X. I NEVER trusted them and re calibrated the probes numerous times, along with pin-out checks. Sometimes they read accurate and other times they would be all over the place. I ended up making a Fuel Hawk type dip fuel level gauge and always dipped the tanks before each flight and kept track of fuel burn using the red cube. Wished I had the tanks built using the floats.
 
I put Princeton single point set probes in the tanks when I built the Cozy MKIV. They were dead accurate with the original Blue Mountain EFIS and also the Dynon Skyview that replaced it. Still working perfect after 10 years. I installed the Princeton 5 point set probes in my RV-10 with Garmin G3X. I NEVER trusted them and re calibrated the probes numerous times, along with pin-out checks. Sometimes they read accurate and other times they would be all over the place. I ended up making a Fuel Hawk type dip fuel level gauge and always dipped the tanks before each flight and kept track of fuel burn using the red cube. Wished I had the tanks built using the floats.

I used the Princeton 2-set-point converters, and they've been absolutely steady for years.

Itried to use the Dynon converters but didn't have very good luck with them (they used a tiny portion of the 0-5V range). Others have had better luck.
 
I don't remember the dielectric constant of ethanol, but it is WAY higher than the value of approximately 2 for gasoline. It doesn't take much ethanol to throw the calibration off. I have read of a small capacitive plate being installed in the bottom inboard end of the tank that is (almost) always immersed in fuel, to make the system self-calibrating. The extra plate would send a separate signal from the tank, and you would need a special converter and/or software to make use of it. I thought about doing that, but decided to just stick with a steady diet of 100LL.

Also, maybe there is a way to implement a software function in the EIS so you can tell it when the tank is full (or tell it how much fuel is in the tank) and let it compensate from there?
 
Does anybody know where to find a kit or plans for the discontinued vans capacitive kit? I'm getting started on my 7 wings and like the installation of the capacitive sender I have seen in other build websites.
 
Does anybody know where to find a kit or plans for the discontinued vans capacitive kit? I'm getting started on my 7 wings and like the installation of the capacitive sender I have seen in other build websites.



I've got the scanned drawings from the install that I got from Vans, PM me an email address and I'll forward them to you.

I don't remember the dielectric constant of ethanol, but it is WAY higher than the value of approximately 2 for gasoline. It doesn't take much ethanol to throw the calibration off. I have read of a small capacitive plate being installed in the bottom inboard end of the tank that is (almost) always immersed in fuel, to make the system self-calibrating. The extra plate would send a separate signal from the tank, and you would need a special converter and/or software to make use of it. I thought about doing that, but decided to just stick with a steady diet of 100LL.

Also, maybe there is a way to implement a software function in the EIS so you can tell it when the tank is full (or tell it how much fuel is in the tank) and let it compensate from there?

This is probably part of my problem - I was trying to dial in the calibration curves and was not meeting with any success - but the ethanol levels are known to vary in automotive fuel between summer and winter blends, and I crossed that threshold twice while trying to dial in the calibration without taking that into account. The third probe for ethanol content calibration would be a great way to determine that, but at this point I'm not interested in chasing the rabbit any further down the hole, the juice ain't worth the squeeze. I'm going forward with the floats.
 
Last edited:
I WISH mine were accurate

I must be the only guy in the Van's world that has INACCURATE capacitive fuel senders using only 100LL. I've calibrated them a couple of times (I have a Dynon D-180) and I get some warning that the difference between calibration points is small. I just went ahead and continued the calibration and what I get are gauges that are accurate when full but show that I have a lot more fuel than I have when empty. I can't remember exactly how much they are off but they are significantly off. My primary fuel gauge is the totalizer.

It's such a PITA that I haven't readdressed this in years but I've been checking so many things off on my squawk sheet that this issue is starting to bubble up to the top of my list. Maybe I'll get around to solving this issue this year.
 
Last edited:
I must be the only guy in the Van's world that has INACCURATE capacitive fuel senders using only 100LL. I've calibrated them a couple of times (I have a Dynon D-180) and I get some warning that the difference between calibration points is small. I just went ahead and continued the calibration and what I get are gauges that are accurate when full but show that I have a lot more fuel than I have when empty. I can't remember exactly how much they are off but they are significantly off. My primary fuel gauge is the totalizer.

It's such a PIA that I haven't readdressed this in years but I've been checking so many things off on my squawk sheet that this issue is starting to bubble up to the top of my list. Maybe I'll get around to solving this issue this year.

This was why I switched out the Dynon capacitance to voltage converter and went with Princeton. The Dynon ones don't allow you designate empty as 0V and full as 5V, and I, too, ended up with almost no difference in voltages as I was adding fuel for the calibration. That made it very inaccurate in use.

Once I swapped to Princeton converters and did the calibrations (which is a PITA because you have to empty a tank, set the Princeton device to E, fill the tank, set the device to F, then *empty the tank* and start the whole EFIS 2-gallon-at-a-time calibration, then do the same thing on the other side....arrrrrgggghhhh) it has worked fine for years (albeit off by about 1 gallon from about 12 gallons on down, it's consistently off by that amount, and I don't have the patience to recalibrate things).
 
I been chasing problems with Van's capacitive senders and Dynon D120 for a long time. Before flight the readings are steady but not always very accurate - often out by up to 2 gallons. After about 10 minutes in the air they become erratic, often scrolling up and down, and giving the D100 a panic attack. I've checked the grounds and the rest of the wiring and recalibrated twice but no improvement. I'm out of ideas.
The red cube is deadly accurate but I want functioning fuel gauges.
Should I put new senders?
 
Last edited:
I been chasing problems with Van's capacitive senders and Dynon D120 for a long time. Before flight the readings are steady but not always very accurate - often out by up to 2 gallons. After about 10 minutes in the air they become erratic, often scrolling up and down, and giving the D100 a panic attack. I've checked the grounds and the rest of the wiring and recalibrated twice but no improvement. I'm out of ideas.
The red cube is deadly accurate but I want functioning fuel gauges.
Should I put new senders?

It's not the senders (if by senders you mean the plates in the tank)...it's the converters, I'll wager. What converters are you using?
 
...
I'm out of ideas.
The red cube is deadly accurate but I want functioning fuel gauges.
Should I put new senders?
With two Red Cubes, and a fuel gauge that works by subtraction of fuel flowing through them, you would have accurate gauges.
 
With two Red Cubes, and a fuel gauge that works by subtraction of fuel flowing through them, you would have accurate gauges.

Unless you had a leak in the system (stuck quick-drain, failed connection, etc.) between the tank and Red Cube.
 
Perhaps Consider a different sender technology

Part of my business and career has involved the application of a variety of sensor technologies. The sensing of liquid levels would seem to be pretty straightforward and often is, unfortunately that is not always the case in vehicles like aircraft. The consequences of having an unreliable fuel level indication can be easily offset through other means but sometimes we humans overlook the obvious and get into trouble, in an aircraft that can be pretty serious.

The small aircraft industry has relied on restive float sending units like the ones in our automobiles since the early days and that has mostly worked because no one trusted them and the analog displays were small and difficult to read. Enter modern glass panels and digital displays. Now we expect resolution down to the 0.1 gal and linearity and accuracy to the fraction of a percent. This is probably not possible with a restive or capacitance sensor within the confines of most budgets.

A technology that I have been looking at for another application is a magnetoresistive sensor. Magneto Resistive - This is another one of those technologies that was described by the one of the giants of physics back in the 1800's, Lord Kelvin.

https://en.wikipedia.org/wiki/Magnetoresistance

Up until the advent of advanced semiconductor technology it has mostly remained practical for use in the laboratory and expensive, complicated instrumentation applications. Today, Honeywell and Allegro Semiconductor have reduced the technology to sensing that are available off the shelf these can be integrated into float type sending units that may be useful in general aircraft.

One company that I have found that has products that use this technology and has TSO'd devices available for several aircraft is CIES in Bend, OR. The last time I talked to them they were trying to gauge interest in a sending unit that would be appropriate to the experimental aircraft market.

https://www.ciescorp.net/

https://www.ciescorp.net/vans-rv-fuel-level.html

- Phil
 
Neat concept, but the only real advantage this will have in our planes over the resistive floats is the size of the float, giving 80 degrees of travel over 70 degrees. We can do that now with a different float. Unless the CIES device amplifies the number of degrees of travel somehow for the B field on the sensor, it's not a major improvement. The resistive sensors currently have plenty of resolution, that's not the problem.

EDIT - I thought perhaps I had an "Ah-HA!" moment looking at the drawings for the Vans-supplied Stewart Warner float senders, because when the sender float arm is bent as shown and installed as shown on the drawing, you get only about 60-65 degrees of movement across the rheostat. I thought it would be simple enough to change the way the float arm is bent and increase that to 90 degrees or better, and sure enough that's fairly simple to do. Then I thought about the rheostat itself... so I called Stewart Warner technical support and asked the question, sure enough the rheostat itself is limited to 70 degrees rotation. Vans design actually does a fairly good job of using all that without encountering the deadband on the ends, so no realistic improvement is available with that sender.
 
Last edited:
RV7A Flyer, I did of course mean the converters. They are the Dynon supplied converters. Doesn't seem like there is much to go wrong with them, but I'm not an electronics guy.
 
RV7A Flyer, I did of course mean the converters. They are the Dynon supplied converters. Doesn't seem like there is much to go wrong with them, but I'm not an electronics guy.

As noted earlier, since you can't set the endpoints, you can (like on an RV) end up with the empty-to-full range using a very, very small range of voltages, rather than the whole 0-5V. On mine, I think almost 3/4 of the calibration points showed NO difference on the display (which shows 2 decimal places). Granted, the actual data (which you can see via other menus) did have some slight differences in the 3rd and 4th digits, it still seems less than optimal to use such a tiny fraction of the available data range. And, indeed, they were inaccurate in the mid-range.

So, I recommend using the Princeton 2-setpoint converters.
 
As noted earlier, since you can't set the endpoints, you can (like on an RV) end up with the empty-to-full range using a very, very small range of voltages, rather than the whole 0-5V. On mine, I think almost 3/4 of the calibration points showed NO difference on the display (which shows 2 decimal places). Granted, the actual data (which you can see via other menus) did have some slight differences in the 3rd and 4th digits, it still seems less than optimal to use such a tiny fraction of the available data range. And, indeed, they were inaccurate in the mid-range.

So, I recommend using the Princeton 2-setpoint converters.

Thanks, I think I will.
 
The holy grail for me would be self calibrating gauges. They say it can't be done, but I don't understand why.
It seems to me that you would just need a simple sensor to determine which tank is selected. It could be very similar to the magnetic door sensors and added to the selector valve.
With tank selection known, fuel added is a known quantity, fuel burned is a known quantity(from red cube), leaving simple math of fuel remaining. The computer should be able to calculate what the reading should be versus what the current reading actually is, and automatically make the requisite voltage change corrections itself. Even fuel tank temperature could be added to the mix to really refine. It could then store averages over myriad flights and constantly compare and refine itself from its stored history, presenting its own performance grade similar to how the magnetometer is calibrated.
 
Last edited:
The holy grail for me would be self calibrating gauges. They say it can't be done, but I don't understand why.
It seems to me that you would just need a simple sensor to determine which tank is selected. It could be very similar to the magnetic door sensors and added to the selector valve.
With tank selection known, fuel added is a known quantity, fuel burned is a known quantity(from red cube), leaving simple math of fuel remaining. The computer should be able to calculate what the reading should be versus what the current reading actually is, and automatically make the requisite voltage change corrections itself. Even fuel tank temperature could be added to the mix to really refine. It could then store averages over myriad flights and constantly compare and refine itself from its stored history, presenting its own performance grade similar to how the magnetometer is calibrated.

And then you could add a Kalman filter to it... :)
 
in the meantime, manually noting Zulu time for each tank switch and processing logged tank readings and fuel flows offline could eliminate the two gallons at a time fandango....
 
And then there's that pesky 14 CFR part 25.1337 to consider...

Or not. I believe Part 25 only applies to transport category airplanes. There's also 91.205, but this applies only to aircraft with a standard category airworthiness certificate. As near I can tell there's actually no regulation that requires EAB aircraft to have fuel gauges? Not saying its a good idea, but I'd rather than a good totalizer than a bad fuel gauge.
 
I'm having trouble finding the Princeton capacitance to volts convertors.
Spruce sells the Princeton kits but they don't have the convertors as a separate item. Any ideas?
 
voltage converter

Are the converters you are looking for similar to, or the same as, the Dynon converters? I have a set of Dynon converters that I will not be using.
 
Are the converters you are looking for similar to, or the same as, the Dynon converters? I have a set of Dynon converters that I will not be using.

No. The Dynon converters have no capability of setting the voltages at the endpoints (empty/full), which the Princeton ones do (to ensure using the full 0-5V range, rather than the tiny portion of it that the Dynon ones use in this installation).
 
So I was thinking about the mix of fuels issue and the resultant inaccuracy of the sensor.

If the range is adjustable to allow for the variance in capacitance then as long as two fuels are calibrated separately then operating with either fuel should provide accurate results. When fuels are mixed this approach doesn't work. Depending on the majority of a particular fuel and the basis for the calibration then the errors will be smaller or larger. This error may be unacceptable in many cases.

Now what about if you had the ability to reset the full level either in your sensor converter or your measurement device. At full fuel level you have the maximum capacitance for the given fuel mix that is actually in the tank. This measurement could then be used as a scaling parameter for the actual calibration that was created when the tank was initially calibrated. The pilot could reset the full capacitance at any time to reapply scaling to the calibration.

No one that I know of allows this type of operation but it seems like it would resolve the fuel mix inaccuracy problem. I'd have to go through real calculations to confirm it, but it seems like it would work. Maybe a new software feature.
 
Ray, check with your EFIS manufacture for the answer to your question.

With the Dynon (and others) you have to tell it you filled the tanks, although the gauges read a new level.

If they allowed you to enter the number of gallons in each tank, it would be easy for them to calculate a scaling number.
 
Bill, yeah exactly I think they could take care of this problem with a minor feature addition. I'm not super familiar yet with all the supported operations since my efis is a big open space at the moment.
 
Best kit or plans for capacitor fuel sender system?

Hi VAF,

I want to install cap fuel senders in my RV-7 SB kit however I've discovered that Vans doesn't sell them any more. I've looked at the Princestons from Air Spruce but I'm not sold on them.

Does anyone know of any Vans-like kits supplied elsewhere or have some plans I can get a copy of?

Cheers,
Gav
 
Hi VAF,

I want to install cap fuel senders in my RV-7 SB kit however I've discovered that Vans doesn't sell them any more. I've looked at the Princestons from Air Spruce but I'm not sold on them.

Does anyone know of any Vans-like kits supplied elsewhere or have some plans I can get a copy of?

Cheers,
Gav

Lynn will probably pop in at some point but he was starting to fabricate some and even offered to send me the delrin in exchange for feedback on them. I decided to go with floats since most say they are reliable enough with the glass panels and you can mix your fuels which I know I am going to do.
 
inaccurate readings from capacitive senders

We have the capacitive senders in our RV-7A and are 17 hours into Phase 1 testing. The fuel level readings were fairly close (within a gallon or two) until recently. Now the readings will go up when fuel is added, but don't go down very much as fuel is used, e.g. only a decrease of a gallon is shown (either tank) when several gallons have actually been used from a tank. Has anyone seen this, and what might be done about it?
 
We have the capacitive senders in our RV-7A and are 17 hours into Phase 1 testing. The fuel level readings were fairly close (within a gallon or two) until recently. Now the readings will go up when fuel is added, but don't go down very much as fuel is used, e.g. only a decrease of a gallon is shown (either tank) when several gallons have actually been used from a tank. Has anyone seen this, and what might be done about it?

Need more info...what sort of display/EFIS do you have (including EMS), and what C2V converters are you using?
 
Back
Top