Altimeter vs. GPS altitude comparison

Discussion in 'Rocketry Electronics and Software' started by Buckeye, Sep 1, 2018.

  1. Sep 1, 2018 #1

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    After a bunch of file format manipulation, I plotted altitude (AGL) vs. time from an altimeter along with GPS. The altimeter is PerfectFlite MAWD, and the GPS is Big Red Bee 900. Both are several years old and not on the cutting edge of technology. I am quite pleased with the agreement.

    Data was collected from the BRB900 in two ways. One was the recording on the onboard flash memory of the transmitter. The second was to connect the LCD receiver to a com port and interface with u-center software for real-time NMEA recording during flight. u-center is nifty freeware for all kinds of plotting and analysis, making for a decent poor-man's telemetry link. It is fun to replay the data stream after the flight. This pic shows the rocket landing in a woodlot. (It was recovered from the top of a 40 ft tree.)

    20180901_135438.jpg

    Here are the altitude comparisons. The BRB900 transmitter lost GPS lock for the first 12 seconds of flight. The base station receiver lost signal for the first minute or so. The apogee recordings agree to about 700 ft. All my OR simulation variations bracket these values, too. Pretty good.

    Interestingly, the recordings differ the most at high altitude and are very close at low altitude. I am wondering if the lack of temperature correction on the MAWD may be understating the altitude. It was a hot, 85 deg F day, far warmer than the standard atmosphere model.

    Capture1.PNG Capture2.PNG
     
    Last edited: Sep 1, 2018
  2. Sep 2, 2018 #2

    SparkyVTFlyer

    SparkyVTFlyer

    SparkyVTFlyer

    Senior Member

    Joined:
    Feb 8, 2009
    Messages:
    123
    Likes Received:
    1
    Gender:
    Male
    Location:
    Yorktown, VA
    I'm not sure if its the temperature or not. I get similar results using a BMP180 pressure sensor and UBLOX NEO-M8N GPS chipset.

    Baro vs GPS.jpg
     
    Buckeye likes this.
  3. Sep 2, 2018 #3

    tHoagland

    tHoagland

    tHoagland

    Rocket Builder TRF Supporter

    Joined:
    Oct 26, 2016
    Messages:
    160
    Likes Received:
    0
    Edit: I just noticed that things do match at lower altitude so you probably already did this....


    Have you accounted for your launch site elevation? Your altimeter reset ground level at power up and measures altitude AGL, while the GPS is measuring from the modeled WGS84 ellipsoid.


    If you want to compare the AGL height, you'd need to determine the elevation of your launch site, preferably from a Digital Elevation Model (DEM) and subtract that from the GPS data.
     
  4. Sep 2, 2018 #4

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    Yes, I think I am doing this the best I can. The BRB900 transmitter reports the MSL elevation (datum is 34.6 m below the WGS84 ellipsoid for my site) from the NMEA strings. While the rocket is sitting on the pad, I average the MSL readings and call that the "ground." I then subtract this ground from all future MSL readings during rocket flight to achieve AGL.

    Not familiar with DEM. I'll have to look that up. This is fascinating stuff.
     
  5. Sep 2, 2018 #5

    Handeman

    Handeman

    Handeman

    Forum Supporter TRF Supporter

    Joined:
    Jan 18, 2009
    Messages:
    7,248
    Likes Received:
    17
    Location:
    Stafford, VA
    I'm certainly no expert, but I would expect the GPS to be more inaccurate until it has been plotting for a while, when it's closer to ground.
    While doing Geocaching, I've found that GPS tends to improve in accuracy the longer it's turned on and tracking satellites. I assume that has something to do with firmware/software and updating the satellite track data.

    By the KISS principle, a simple pulse transmitter in an RDF system will be much more reliable (won't cut out during acceleration for one thing) then the much more complicated GPS trackers which need multiple satellite inputs, data packet transmission and reception, and other technologies to all work. With that said, there is a minimal learning curve and skill level needed to use the GPS vs an RDF system, and like most things, since the GPS systems have been out for some time, they are becoming cheaper and easier to use even if they are more dependant on more complicated technology. That still doesn't mean they are the most accurate or reliable form of measurement, just the easiest.

    Call me old fashioned, but I still carry a compass in my backpack when I hike, even if I haven't used it in years. I know I can depend on it, it will work and won't have a dead battery when I need it.
     
  6. Sep 2, 2018 #6

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    By the same token, GPS can get wonky the longer it is turned on, too. While the rocket was waiting on the pad, the coordinates started to drift many meters, then re-centered themselves. In flight, there is no time for user averaging. You get only one GPS measurement every second!

    We are talking about altitude measurement, not finding the rocket. Can RDF do that?
     
    Last edited: Sep 2, 2018
  7. Sep 3, 2018 #7

    Handeman

    Handeman

    Handeman

    Forum Supporter TRF Supporter

    Joined:
    Jan 18, 2009
    Messages:
    7,248
    Likes Received:
    17
    Location:
    Stafford, VA
    No, RDF will not do altitude, although if done right, it can confirm an apogee event.

    The point is that I would trust the purpose built altimeter to be more accurate than a GPS system when measuring altitude.
     
  8. Sep 10, 2018 #8

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    I am no expert on this, but I think GPS can be the better choice. Airplane guys all seem know that barometric altimeters measure pressure, and that pressure is affected by local weather conditions. The more the local weather deviates from the Standard Atmosphere model, the less accurate the barometric altitude prediction. This fact is rarely discussed in this model rocketry forum, however. Guys read their Missileworks, Perfectflite, and Eggtimer devices and assume that is the "actual" flight. Not true.

    Even if the GPS elevation error is 10-20 meters (I read that somewhere), that can be far less than a couple hundred meters of barometric error (like on a very hot day).
     
  9. Sep 10, 2018 #9

    jderimig

    jderimig

    jderimig

    Sponsor TRF Sponsor

    Joined:
    Jan 23, 2009
    Messages:
    2,469
    Likes Received:
    40
    The answer is Sam errors can be on the order of 100 - 150 meters (at 4000m). Compare atmospheric sounding data from buffalo NY and key West Florida and see for yourself.

    However that Sam error is only 2.5percent. There are as large errors in the rest of the chain also. Be happy if your altimeter is within 10 percent of true. Expect 5 percent.
     
    Last edited: Sep 10, 2018
  10. Sep 11, 2018 #10

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    Right. That is why "adjusting" your sim to match "reality" by fudging with drag coefficient is a waste of time. The simulation is already very good in its native form. OR, RS, and RAII all adjust the Standard Atmosphere Model with local conditions, making them more accurate than altimeters in this regard.
     
  11. Sep 11, 2018 #11

    jderimig

    jderimig

    jderimig

    Sponsor TRF Sponsor

    Joined:
    Jan 23, 2009
    Messages:
    2,469
    Likes Received:
    40
    Most of the time the SAM error is small, especially if you are flying 5000 ft or less because the altitude is a difference calculation and some of the error is cancelled out. If you get a significant error in your sim to flights less than 5k then it's more likely your sim is not accurate.
     
  12. Sep 11, 2018 #12

    ksaves2

    ksaves2

    ksaves2

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Nov 25, 2009
    Messages:
    5,339
    Likes Received:
    14
    I wonder if a chipset that uses both GPS and GNSS constellations would have better vertical altitude accuracy. It seems that positioning error is less with both systems being used. Kurt Savegnago
     
  13. Sep 12, 2018 #13

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    What is the vertical accuracy of GPS? My casual Google searching says 10-20 meters, and that was several years ago. My BRB900 says this: "It also programs the high performance 50 channel ublox 7 GPS chipset putting it into "air mode" for reliable altitude reporting (which some similar products do not)" Seems to me that GPS is as good or better than baro in many cases.
     
  14. Sep 12, 2018 #14

    jderimig

    jderimig

    jderimig

    Sponsor TRF Sponsor

    Joined:
    Jan 23, 2009
    Messages:
    2,469
    Likes Received:
    40
    Well a way to find out is to put the GPS on a table and log altitude readings for 24 hours and see what range of reported alts are. During that time you will probably sample most of the constellations.
     
    Buckeye likes this.
  15. Sep 13, 2018 #15

    UhClem

    UhClem

    UhClem

    Well-Known Member

    Joined:
    Feb 6, 2009
    Messages:
    1,301
    Likes Received:
    11
    The problem is with the geometry. To improve the altitude reading you need data from a satellite below the horizon. As it is, the best you can hope for is to get data from one directly overhead and one near the horizon. With a few more scattered around of course.

    VDOP tends to run about twice HDOP.

    Air mode trades tracking for accuracy.

    It does that by opening up the loop filters so that the GPS system can track better during high vehicle dynamics. The ublox 7 has several such settings but the highest dynamics one is still for accelerations less than 4G.

    Increasing the loop filter bandwidth decreases accuracy so position error will increase.
     
    Buckeye likes this.
  16. Sep 13, 2018 #16

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    Good info. Thanks.

    This article says GPS elevation error is less than 5m @ 95%, and talks a lot about sub-meter precision. The baro errors are one to two orders of magnitude worse than that, depending on altitude. We are not talking extreme altitudes, either, but regular sport flying. I am gonna process some more data, but GPS looks like the winner to me.
     
  17. Sep 13, 2018 #17

    jderimig

    jderimig

    jderimig

    Sponsor TRF Sponsor

    Joined:
    Jan 23, 2009
    Messages:
    2,469
    Likes Received:
    40
    . ,. Xx r
    But that may not be the error with consumer grade electronics. There is a reason that there is a market for $500 GPS chipsets compared to the $8 dollar ublox chips that we use.
    My table test with a high end rocketry GPS system has a range of 120 meters perfectly still.

    If you average over several minutes the accuracy is much better but you do not have that luxury at apogee. You can't control if you will have a good or bad constellation as your flight approaches the top. Baro is much more accurate because you can correct the result with atmospheric sounding data if needed. You cannot correct a gps reading.
     
    Last edited: Sep 13, 2018
  18. Sep 13, 2018 #18

    jderimig

    jderimig

    jderimig

    Sponsor TRF Sponsor

    Joined:
    Jan 23, 2009
    Messages:
    2,469
    Likes Received:
    40
    This is discussion is good. Because of it I will add the ability to retrieve the pressure reading at apogee (from Marsa alimeters), then you can look up the actual altitude based on the nearest sounding data from http://weather.uwyo.edu/upperair/sounding.html
     
    Donnie likes this.
  19. Sep 13, 2018 #19

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    Good idea, but Adrian beat ya to it on page 12, and I just implemented it in my measurements in post #1. :)

    https://www.featherweightaltimeters.com/uploads/1/0/9/5/109510427/raven_users_manual_2014may20.pdf

    I "undid" the Standard Atmosphere model in the altimeter data and replaced it with the nearest sounding measurements. Now I have nicer agreement between the MAWD and the uncorrected BRB900.

    Capture3.PNG
     
  20. Sep 13, 2018 #20

    jderimig

    jderimig

    jderimig

    Sponsor TRF Sponsor

    Joined:
    Jan 23, 2009
    Messages:
    2,469
    Likes Received:
    40
    So Baro is accurate after all.
     
  21. Sep 14, 2018 #21

    Buckeye

    Buckeye

    Buckeye

    Well-Known Member

    Joined:
    Sep 6, 2009
    Messages:
    1,908
    Likes Received:
    33
    Heh. It is accurate after fixing the atmosphere model. No baro altimeter offers that function out of the box (hint, hint). Who is going to do the fix manually, other than a data dork like me?

    For the casual hobbyists (who don't live in hypothetical air columns at mid latitudes), they are getting ripped off in apogee altitude by 5% or more by default from their altimeter! Then they blame Rocksim ... :rolleyes:
     
    jderimig likes this.
  22. Sep 14, 2018 #22

    jderimig

    jderimig

    jderimig

    Sponsor TRF Sponsor

    Joined:
    Jan 23, 2009
    Messages:
    2,469
    Likes Received:
    40
    Yes if you want a record, fly in cold!
     
  23. Sep 21, 2018 at 4:42 AM #23

    Adrian A

    Adrian A

    Adrian A

    Well-Known Member

    Joined:
    Jan 22, 2009
    Messages:
    2,111
    Likes Received:
    8
    Yep, and the effect is pretty significant even if it's not common rocketry knowledge. On a really hot day the standard atmospheric model will underestimate altitude by more than 10%. Any flight using baro-based altitude with sea-level temperature over 55F is getting an underestimate of the altitude, even if the altimeter functions perfectly. When it's cold and the air is denser than the standard atmosphere assumes, then a rocket at the same actual altitude will get above more air mass, and so experience lower pressure because there is less air above it pressing down.

    Although it's true that a rocket flying in hot weather with less-dense air will benefit from having less aerodynamic drag losses for the same altitude, the gravity losses are the same for hot and cold, and so the cold-weather reporting benefit wins out.
     
    timbucktoo likes this.

Share This Page