Opportunity
We had been supplying OEM Fuel Quantity Indication Systems to Cirrus Aircraft G3 and G5 aircraft for a year with great success and we had obtained a Supplemental Type Certificate to retrofit this system to legacy Cirrus G3 Aircraft. As we had the STC for the Cirrus it was only a small change to include the G1 and G2 model aircraft into the mix. Generally Cirrus Owners had indicated that the G1 / G2 aircraft had better fuel reporting than the G3 brethren, however we knew that the technology used by these early fuel sensors would eventually fail and without an OEM business and obligation, the prior vendor was free to raise pricing to support this limited production.
We smelled an opportunity.
So we jumped feet first in retrofitting early Cirrus Aircraft. The first indication we had that something was amiss was from a retrofit owner, he reported that the system looked good, but no comment on the accuracy of fuel level information. This report stood in stark contrast to our experience on the retrofitted G3 and new G5 aircraft where fuel level reporting was deemed to be a notable highlight. We had another customer install - same result - Gee it looks great, but no comment on accuracy in reporting. I had heard about issues with the original G1 / G2 fuel level, but I attributed them to the potentiometer technology used to report fuel level.
We still felt we had the best system out there and we sold one more G1/G2 kit. This owner was not so blasé about the accuracy - he noted and documented where the system was showing inaccuracies over his fuel totalizer system. We had his calibration data from the Cirrus Service Center and his report that he lost 5 gallons every time he switched tanks. We were confused - our system seemed to be able to leap tall buildings and burn through steel - inaccuracy in reporting was not something we had seen in the several years it took to develop our fuel level sensor.
Discovery
So we set up the test tank with his calibration data, and started to run tests - the system performed flawlessly for us - it reported the same volume going up or down as we had initially found with calibration. So what gives - We had notated that aircraft structure prevented the Inboard fuel sensor (there are two in each wing for this aircraft Inbd and Outbd) from reaching it's full travel - and we set this limit up with a plexiglass box and tried the recalibration again. Same result - nearly flawless recording, however we did notice one thing - when we depressed the Inbd sensor slightly by 1/4 inch or so we saw a 5 gallon decrease in fuel volume.
The Smoking Gun
We delved a little further - the sensors are actually located in separate tanks - one inbd of the other - the inbd tank (Collector) being much smaller connect to the outbd tank (Main) with fuel lines.
So with fuel being drawn out of the inbd tank the level in that tank must go down 1/4 inch. Well was this true in the aircraft - we had reports for our system and the prior vendors loosing 5 gallons when that wing tank was selected. We knew that it was consistent - if you depress the outbd sensor - the change is progressively smaller and not uniform. Depressing the inbd sensor invariably gave a 5 gallon decrease each and every time, as long as some amount of measurable fuel was present in the main tank. We asked the question "Could this tank level be effected by fuel pump suction" So we modeled the system confirmed it with a little computational fluid dynamics, Sure enough 5 gallons
So why 5 gallons - remember these sensors are in separate tanks see right - but in combination they measure the total fuel volume. When the inbd sensor goes down due to a level change locally in it's smaller tank volume - this amount is subtracted from the larger main tanks - a quarter inch in a 3 gallon tank is peanuts - a 1/4 inch in a 40 gallon long tank is well - 5 gallons. The logical reason the manufacturer used the collector tank for the inbd location of the sensor, is that this location is the best place to measure zero fuel level or the level of fuel just above the unusable amount.
How Do We Fix It.
As we started our work on replacing aircraft fuel sensors in the field - we noticed a lot of subtle bends in the float arms used to optimize the travel on the 60 degree arc potentiometer cards These potentiometer cards are used in legacy fuel quantity systems. We started looking carefully at what this was doing for the sensor, and it was all about geometry or in other words How do you get the most bang for your buck with 60 degrees of potentiometer travel and 60 discrete data points - one per degree. As an example Beechcraft Bonanza goes from least accurate at a full tank to most accurate as the tank drains, this is actually true of a few Cessna Aircraft as well,
It's what tends to happen in a small town.
So why is this true - well - Float sensors have a few limitations -
First and foremost the float has to float
Second as the float swings through an arc it goes from most sensitive at about 45 degrees above horizontal -- to least sensitive
at horizontal -- back again to more sensitive again at 45 degrees below horizontal -
This is a typical Sine function for those who remember trignometry - and this is especially true for float sensors, as increasing fuel is the measured as the sine of the angle the float makes with the measurement system. So these subtle bends in legacy fuel senders were making use of this fact to optimize their reporting capability.
We what if we bent the arms in a similar manner - could we can get the same result, a fuel level output biased to be more and less sensitive as required. We had not crossed that idea, as our patented sensor system had a 180 degree travel limitation and 1000's of arc data points. The
CiES fuel quantity system due to the ability to render small changes into a discrete output, accuracy due to geometry was not needed. Given that - could we utilize geometry for a similar but opposite benefit - could we make the inbd unit less sensitive at the top of the collector tank and mask the small change brought about by fuel being sucked into the engine. Time to run a few numbers. It worked on the spreadsheet - worked being a relative term - it is less sensitive to change at the top of the tank, so out of a 5 gallon change we were now at 3 gallons - not quite good enough. OK this logic seems to offer some potential - what if we made the main outbd tank sensor more sensitive, trying to capture a subtlety in level. Well in this combination, less sensitive collector and more sensitive main tank got down to just under the magic 2 gallon change.
Testing
 |
| Initial CiES Installation |
Time to run a few tank tests - Eureka - the results from the spreadsheet matched the test tank data. Excellent - time to document the revised sensor and cut metal for TSO senders to replace the ones in the field.
The Graph at the right shows the before condition of the aircraft. You can see that the aircraft is in climb as both sensors are below the totalizer. On the G1/G2 aircraft the senders run for/aft - this tends to accentuate the discrepancy as the senders are real excellent angle indicators.
Notice that the right tank is nearly identical - gauge and totalizer prior to the update. However the left tank shows a combination of fuel level error caused by the fuel draw and induced by climb. I suspect that the fuel flapper valve in the left collector tank is sticky or there is a constriction in the line.
Conclusion
Well we are waiting for a cross country flight to get a real assesment - However preliminary flights and taxi tests demonstrated that the results did carry over to the aircraft. But as in all things aviation - you really want to see how this performs in the field.
First Report - Owner has flown the aircraft down to 20 gallons a side -- the system is accurate and it matches the fuel remaining on the owners dip stick. The fuel level change is in the range of 1 gallon when the tank is selected.
It looks like we have taken a problem and turned it into an opportunity to demonstrate our capability - that's really what we are all about.
Talk with us - let us know the problems you might be experiencing in fuel level reporting - whether you have an Aircraft, Boat, RV, Truck -
We solve problems with fuel level -
Feedback on the the CiES system....
2 - Choice of shop doing the work is very important. With the changes
in wiring harnesses among all the iterations of Cirrus aircraft, there
are a several unique and significant distinctions for each application.
You need a knowledgable, detail-oriented, and thorough installer. I was
fortunate to have one at Leading Edge Aviation.
We had been supplying OEM Fuel Quantity Indication Systems to Cirrus Aircraft G3 and G5 aircraft for a year with great success and we had obtained a Supplemental Type Certificate to retrofit this system to legacy Cirrus G3 Aircraft. As we had the STC for the Cirrus it was only a small change to include the G1 and G2 model aircraft into the mix. Generally Cirrus Owners had indicated that the G1 / G2 aircraft had better fuel reporting than the G3 brethren, however we knew that the technology used by these early fuel sensors would eventually fail and without an OEM business and obligation, the prior vendor was free to raise pricing to support this limited production. We smelled an opportunity. 
So we set up the test tank with his calibration data, and started to run tests - the system performed flawlessly for us - it reported the same volume going up or down as we had initially found with calibration. So what gives - We had notated that aircraft structure prevented the Inboard fuel sensor (there are two in each wing for this aircraft Inbd and Outbd) from reaching it's full travel - and we set this limit up with a plexiglass box and tried the recalibration again. Same result - nearly flawless recording, however we did notice one thing - when we depressed the Inbd sensor slightly by 1/4 inch or so we saw a 5 gallon decrease in fuel volume. 
We delved a little further - the sensors are actually located in separate tanks - one inbd of the other - the inbd tank (Collector) being much smaller connect to the outbd tank (Main) with fuel lines.
So why 5 gallons - remember these sensors are in separate tanks see right - but in combination they measure the total fuel volume. When the inbd sensor goes down due to a level change locally in it's smaller tank volume - this amount is subtracted from the larger main tanks - a quarter inch in a 3 gallon tank is peanuts - a 1/4 inch in a 40 gallon long tank is well - 5 gallons. The logical reason the manufacturer used the collector tank for the inbd location of the sensor, is that this location is the best place to measure zero fuel level or the level of fuel just above the unusable amount.
As we started our work on replacing aircraft fuel sensors in the field - we noticed a lot of subtle bends in the float arms used to optimize the travel on the 60 degree arc potentiometer cards These potentiometer cards are used in legacy fuel quantity systems. We started looking carefully at what this was doing for the sensor, and it was all about geometry or in other words How do you get the most bang for your buck with 60 degrees of potentiometer travel and 60 discrete data points - one per degree. As an example Beechcraft Bonanza goes from least accurate at a full tank to most accurate as the tank drains, this is actually true of a few Cessna Aircraft as well, It's what tends to happen in a small town. 
So why is this true - well - Float sensors have a few limitations -
Time to run a few tank tests - Eureka - the results from the spreadsheet matched the test tank data. Excellent - time to document the revised sensor and cut metal for TSO senders to replace the ones in the field. 
It looks like we have taken a problem and turned it into an opportunity to demonstrate our capability - that's really what we are all about.
If you ever have to remove a sensor to replace a gasket for example. The access panel will not have to be opened.
