Altitude and Airspeed errors

[Vern Little(Guest)]

I've chasing down an indicated/true airspeed error in my RV-9A, equipped with a Dynon D10A and a backup airspeed indicator. I'm using the Dynon AOA probe.

Both the ASI and the Dynon agree, but indicate about -5 knots at cruise and +5 knots at stall.

I checked the static system with a manometer. There were no leaks. I calibrated the D10A at 1000, 2000, 3000, 4000 and 5000 feet and got a consistent 4% high reading in altitude. I performed a flight test at 3000, 4000 and 5000 feet (all indicated altitudes with barometer set to the local airport). I compared the GPS derived altitudes, which agreed with the manometer computations within +/- 49 feet worst case (+/- 20 feet typical). All of the readings were not speed sensitive.

My conclusion is that the D10A is not reading altitude correctly, and this may also be the root cause of my airspeed error as well.

The altimeter reads correctly at 0 MSL with a 20 foot compensation.

It looks to me that the unit is not calibrated. It was recently returned for service, and I had asked for complete calibration.

Is it possible to determine if my unit was correctly calibrated by checking the manufacturing records?
Is it possible to calibrate it in the field, so that the aircraft is not grounded waiting for factory calibration?
How are airspeed and altitude computed... does the airspeed use data from the altitude pressure transducer? This would explain the airspeed error as well.

Thanks, Vern Little

 

[khorton]

GPS altitudes will not necessarily compare to barometric altitudes.  They will only be the same if the atmosphere is at standard temperature all the way from the ground to the test altitude, which is never the case.

Did you calibrate the Dynon ASI and the standby ASI using a manometer or pitot-static test set?  If not, that would be useful, so you can determine how much of the speed error is instrument error, and how much is due to other things.  You can get Excel and Openoffice spreadsheets to convert from water manometer readings to ASI indications here: http://www.kilohotel.com/rv8/filemgmt/visit.php?lid=3.

Which method did you use to determine the error in the airspeed?

 

[dynonsupport]

Vern,
The EFIS altitude and airspeed sensors are totally seperate. They of course both use the static pressure input, but the static line splits inside and goes to seperate sensors. We don't use your static pressure from one sensor and then use this to compute altitude as well as airspeed.

You say "Both the ASI and the Dynon agree, but indicate about -5 knots at cruise and +5 knots at stall." What do you mean that they agree? They agree with one another and but are +5/-5 off from some other reading, or that they are close but +5/-5 from one another?

We can't check and see if your unit was correctly calibrated, since we will not ship a unit that is not calibrated right. A unit either passes or fails all of our tests, and if it fails, we repair it. All of our units get calibrated, and then pass a second verification check in the chamber, and then the altitude is hand verified before shipping. It is not possible to calibrate a unit in the field- one of the biggest factors in calibration is temperature, so our chamber passes the unit through about 100 degrees of temperature change.

From what you have described, it appears that the altimiter reads correctly at 0MSL, but is 4% off at 1000-5000 feet, no matter what your airspeed is, even when on the ground. Is this correct?

 

[Vern Little(Guest)]

Dynon has always been a very service oriented company, and very open with it's shortcomings.

I guess it's my turn to admit that I screwed up. I messed up the computations (which proves that a degree in Engineering doesn't prove that you can program Excel). I transposed two digits in a constant that led to this error. [using math derived from EAA website, thanks for the link Kevin].

Therefore, I apologize to Dynon and to anyone else reading this list. Once I corrected my math, the altitude computations provided an average error of less than 1%, which is within experimental error of the manometer system I used.

I also checked the D10A against a conventional altimeter that I tee'd into the plumbing, and they agreed within +/- 30 feet from 0 to 4000 feet using the manometer. The error compared with GPS altitude may be explained by an atmospheric inversion... but that's guessing right now.

I calibrated the D10A and Van's airspeed indicator as well using static pressure (vacuum) only, with the pitot open to ambient pressure. Both instruments agreed with +/- 1 knot over 40-180 knots.

The only variables left (in hunting down my indicated airspeed error) is a leak in the pitot system and/or a pitot position error. I know that the pitot can have up to 15 degrees of angle of attack error and still be fairly accurate, but this assumes that is outside the influence of the wing.

One experiment I did was shim my pitot mount to increase the AOA by about 1 degree. This increased my low speed indicated stall speed by 2 knots. Today, I changed the AOA by -2 degrees, and I'll flight test this configuration to see if I can get some useful data.

If I can get cruise speeds accurate [derived from the 3-course GPS computation], the indicated stall speed is not critical-- I can just change the range markings. This is easy on my D10A, but more difficult on my Van's ASI.

What concerns me is not that I have an indicated airspeed error, but that it is due to a problem in my plumbing or pitot. I just hate not having everything perfect (or explained).

Vern Little (publically humbled).

 

[dynonsupport]

Vern,
Glad that you figured it out. I'll admit that we've been confused by excel spreadsheets with a wrong number here and there as well!

One idea with the pitot probe- is it located in Van's stock location or is ir moved somewhere else? If it is close to the wing rather than about 6" down, that can cause some problems due to the pressure near the wing. It doesn't sound like a leak in the pitot since leaks basically always make you read low. Reading high in the stall and low in cruise sounds like a pressure wave caused by the wing or a leak in your static system.

Pitot / static system problems can be tricky to figure out. Good luck and let us know if we can help you out with any info about our AOA pitot probe.

 

[Vern Little(Guest)]

Thanks for working on a Sunday night!

My pitot uses a Gretz mount on the bay inboard of the outer inspection panel in the RV-9A wing. This is about 6" further outboard than the Van's stock location.

The RV-9A has a different airfoil than the other types of Van's aircraft, so the pressure wave shape off the wing may be different. I think the adjustments to the angle of attack that I'm making will give some data. The affect of decreasing the pitot tube aoa is to drop the nose of the tube downwoard.

If I can't get good results, I may try an extension to move the pitot lower down (or extend it forward), but these changes are very difficult. It is also possible for me to install the stock Van's pitot and make some comparisons.

I'm not convinced that I don't have a pitot leak, but it's hard to detect because the tube itself is designed to leak through the water drain and I know it also leaks from the seams. There is some possibility of a static position error, but I'm using the Van's recommended location, and the static system has no leaks. Very puzzling indeed.

Does Dynon have any other -9A's flying that use the Dynon pitot? If so, please let me know off-line so that I may contact them and compare notes.

Vern

 

[Vern Little(Guest)]

OK, Here's some more info on my airspeed error. I hope you can help.

I have now tested the pitot-static system for leaks, and calibrate both the altimeter and airspeed indicator on the D10A, both of which are accurate.

However, I could not involve the pitot tube in my leak testing because it was too difficult to seal it up.

I did notice, however, that when I sealed the seams and drain holes on the pitot, and applied pressure to the pitot intake, it leaked out the AOA port.

This seems strange... I thought that the pitot and AOA were independent systems, and not cross-connected. Is this normal behaviour?

Please let me know how the pitot is constructed so I can determine if this is a problem.

Thanks, Vern

 

[dynonsupport]

Hi Vern,

Both the pitot and AOA tubes feature pinhole leaks for moisture control. As you've discovered, you can not completely seal up the pitot to test for leaks. We recommend that you perform your leak check upstream of the Dynon AOA/pitot tube.

If you did manage to seal up all of the vents that let those pinholes vent to the outside, pressure applied to either the AOA or pitot ports would effectively allow air to flow from one pinhole vent to the other since both of those holes vent to a common cavity inside of the pitot tube. In other words, the systems are not cross connected, but if you seal up all of the moisture control vents on the external face of teh pitot, you are effectively creating a semi-sealed chamber inside of the pitot that would allow you to push air between the pitot and aoa lines via their pinholes.

Also note that the pinholes are small enough such that they do not perceptibly affect pitot and aoa readings.

 

[Vern Little(Guest)]

Thanks for the pitot/aoa explanation. It's nice to get the information to eliminate the pitot/aoa as a source of airspeed error.

Note: I taped over the AOA port for the following tests:

I went through the pitot system, and did a leak check upstream of the pitot. I also changed the angle of attack of the pitot tube by shimming the mount so that the nose of the pitot is down by 2 degrees from normal installation (or 3 degrees lower than my previous test). This seemed to lower the indicated stall speed by a couple of knots (52 to 50).

Now, my indicated stall speeds are virtually identical to the CAFE report on the RV-9A, and about 1 KIAS higher than the markings on Van's airspeed indicator (which does not quite match their specification).

Therefore, I'm satisfied now with my slow-flight calibration.

As for cruise speed indications, I think I'm still indicating a bit slow (about 2%), but it's very difficult to calibrate in crowded airspace with unstable air. At this point, I'll repeat the test with the AOA port open, then I'm just going to make a correction chart and declare success. After all, in cruise GPS ground speed is more useful information.

In conclusion, the angle of attack of the pitot does seem to make a difference at slow speeds. It may not be significant, but I found it helpful to get my indicated airpeed agree with the CAFE CAS.

Vern

 

[vlittle]

I thought this discussion had run it's course, but I now have some new information.

After calibrating my D10A and ASI on the ground using a water manometer, I was satisfied that both were accurate within 1% (experimental error).

Even though in-flight calibration showed the TAS indication of the Dynon to be about 3-4% low (based on several triangular GPS runs), I was satisfied just doing the adjustment in flight.

It turns out that in flight, the Dynon seems to be about 3% lower indicating than my external ASI for IAS.

Yesterday at the Van's homecoming, I spoke with another -9A builder who has the same setup, with the pitot in the same location. He sees almost exactly the same errors that I see... the D10A is indicating low with respect to his ASI, and his GPS calibration shows about 3-4% low.

This does not seem to be a coincidence, and we are wondering if there is something about the hardware or software of the D10A that may affect the airspeed indication. My first guess is that engine vibration may be introducing pressure pulses into the transducer which are not properly filtered. Alternatively, is there an error in instrument temperature compensation?

Dynon knows best, but it would be worth checking the bench calibration vs. in-flight performance.

Vern Little

 

[dynonsupport]

Are you saying that when you use a manometer on the ground, your EFIS and ASI agree, but once in flight, they do not? And also, that nothing is changing in the plumbing between these two situations, except for the end of the pitot tube being hooked up to a manometer?

Additionally, are both sourced off the same static?

 

[vlittle]

Are you saying that when you use a manometer on the ground, your EFIS and ASI agree, but once in flight, they do not? And also, that nothing is changing in the plumbing between these two situations, except for the end of the pitot tube being hooked up to a manometer?

Additionally, are both sourced off the same static?

Actually, when I did the airspeed calibration, I left the pitot open, and applied a vacuum to the static port using the manometer. This should be equivalent.

Both agree on the ground, but not in flight. The dynamics of the ASI and the Dynon are different, so it's necessary to be in stable conditions to notice the error.

It's possible that vibration is affecting the ASI, but I doubt it. That's why my question about the D10A's performance with vibration.

Vern

 

[khorton]

Vern,

When you did your testing on the ground, did you take readings both with airspeed increasing between readings, and airspeed decreasing?  I noted some hysterisis in my mechanical ASI.  I expect that with the engine running, the mechanical ASI would go to about the midpoint in the hysterisis band.  This would result in a slightly different ASI to EFIS difference than was seen on the ground.

 

[dynonsupport]

We don't have any vibration issues that we know of, and we have done vibration testing to look for this sort of thing. You could try gliding with the engine near idle to remove (or at least change the characteristics of) the engine vibration.

 

[vlittle]

Vern,

When you did your testing on the ground, did you take readings both with airspeed increasing between readings, and airspeed decreasing? I noted some hysterisis in my mechanical ASI. I expect that with the engine running, the mechanical ASI would go to about the midpoint in the hysterisis band. This would result in a slightly different ASI to EFIS difference than was seen on the ground.

Kevin, I didn't do this. Instead, I tapped on the ASI until it stabilized, and compared this reading with the Dynon. The tapping had the effect of giving me the same readings going up as down (in a couple of test cases). It is possible that this has an effect in flight, but I'm seeing consistently higher readings with the D10A than the ASI.

I also did some best glide speed tests today with the engine at idle, and the Dynon was consistently a couple of knots higher (at 72 KIAS) than the ASI.

I'd be interested in anyone else who has any data to share. There are a lot of variables at play here.

Vern

 

[vlittle]

Previous post should have said "constantly lower readings from the Dynon than the ASI"
Vern