Saturday, September 5, 2009

Network is failing to email, but here's the SMS contents we got between 1:50 and 2:06 AM

T2.04 COORD 41.86228 -87.65212GPS ALT 14265ft 5mph HDG 131 ALT 17557ft TARGET 7000ft TO 14299ft. FANS OFF. 22F 3.77V HSDPA1900 -89 ECIO 19.0 AT T 310-410 used 2 ballasts. Debug info 268K 87adc 887v 7.47PSI

Below updates are older but came out at the same time:

T23.23 Coord 41.93382 -87.88205gps ALT 666ft 5mpg hdg145 Alt 17342ft Target 7000ft to 14796ft. Fans DOWN. 13F 3.67V GSM1800 -82 snr 2 310-410 used 2 ballasts. Debug info 263K 88adc 897v 7.54PSI

T20.05 Coord 42.11766 -87.95353gps ALT 13625ft 4mph hdg144 Alt 16697ft Target 7000ft to 14299ft. Fans DOWN. 22F 3.83V UMTS1900 14 ecio 23.5 used 2 ballasts. Debug info 268K 91adc 927v 7.74PSI

T19.15 COORD 42.16292 -87.99477GPS ALT 14767ft 3mph HDG148 ALT 13222ft TARGET 7000ft TO 13329ft. Fans off. 40F 4.07V EDGE850 -83 SNR 2 310-410 used 2 ballasts. Debug info 278K 108adc 1101v 8.89PSI

T17.59 COORD 42.23459 -88.03546GPS ALT 13815ft 6mph HDG171 ALT 12667ft TARGET 7000ft TO 11490ft. FANS DOWN. 58F 4.13V GSM1800 -0 snr 1 0-0 used 2 ballasts. Debug info 288K 109adc 1111v 8.97PSI

T17.21 COORD 42.27236 -88.02183 GPS ALT 1814 5mph HDG240 ALT 8741ft TARGET 7000ft TO 10602ft. FANS OFF. 67F 4.24V 0-0 used 2 ballasts. Debug info 293K 132adc 1345v 10.53PSI

T16.45 COORD 42.29093 -87.99460GPS ALT 561ft 2mph HDG259 ALT 3087ft TARGET 7000ft TO 10167ft. FANS OFF. NOT OK TO SLEEP. 85F 4.24V UMTS850 -106 ECIO 17.0 310-410 debug info 298K 168adc 1712v 12.99PSI
at

6 comments:

  1. UnknownSeptember 5, 2009 at 12:21 PM

    So it dropped on the way up and is probably too buoyant now so it is going to have to run the fans all day. Battery is low, so we are going fly high. Well, the concept seems to work...need a post-launch timer or something....

    ReplyDelete
  2. Chris FSeptember 5, 2009 at 1:39 PM

    Yep.

    I hope it's parked over Lake Michigan right now. If it's not, it's probably in Lake Michigan. :) There's nowhere around there it could have gone down without sending out emails.

    ReplyDelete
  3. UnknownSeptember 5, 2009 at 1:58 PM

    I'm not sure which altitude to believe, frankly. I think we need a better sampling algorithm so we don't use bad readings. Looking at the spec again, it was dumb not to put a small bypass cap across the sensor. A tiny one is all we need.

    ReplyDelete
  4. Chris FSeptember 5, 2009 at 3:35 PM

    Across the supply voltage or from the sense voltage to ground?

    We can put a scope on our +5 to see how clean it is. I wouldn't be surprised if it's rock solid. Seems like we get "mostly good" readings. I believe the altimeter over the aGPS anyway. The altimeter seems pretty linear in terms of altitude.

    ReplyDelete
  5. CMTangoSeptember 5, 2009 at 11:07 PM

    Looks like you need some code in there not to drop ballasts unless the altitude is on a downward slope and then poll the changes after it drops one to see what the new slope is and if it should drop another. If you're below your target but gaining altitude, then you don't want to drop another one since you might overshoot your target. It might be good to look in to a bleed valve so you can adjust the buoyancy instead of relying on fans to pull you down.

    Mark

    ReplyDelete
  6. Chris FSeptember 6, 2009 at 2:32 PM

    hey Mark, you described exactly the algorithm it has now.. but it seems like there's enough downdrafts on our way up to fake out the algorithm. Looks like we had 2 false "down readings" between 16:45 and 17:21.

    I think (rightfully) Russ is worried about false altitude readings due to noise from our +5v supply. A capacitor plus smarter polling would help eliminate that.


    while(a_alt < target_low)
    {
    suSleep(2 * 60000, NULL);
    a_alt2 = sbcm_get_altitude();
    if(a_alt2 < a_alt)
    {
    // Going down, better release ballast to balance us.
    sbcm_ballast_drop();
    }
    a_alt = a_alt2;
    }

    ReplyDelete

Subscribe to: Post Comments (Atom)