Shop OBEX P1 Docs P2 Docs Learn Events
A one-bit error ... — Parallax Forums

A one-bit error ...

Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
edited 2010-05-21 15:36 in General Discussion
... probably from a cosmic particle, but a 26-hour debug turnaround time:

news.discovery.com/space/nasa-finds-cause-of-voyager-glitch.html

And I used to think that waiting for windowed EEPROMS to erase before they could be reprogrammed was bad!

-Phil

Comments

  • stamptrolstamptrol Posts: 1,731
    edited 2010-05-19 16:24
    Each bits does count!!

    Many years ago I was accosted by the corporate IT department who were hell-bent on "taking over" the administration of the PLC's and other controllers on the manufacturing equipment.
    This same guy could not understand how "just a single bit" being wrong could have any effect on the machinery involved.

    Thankfully, cooler heads prevailed and the electrical guys stayed in control.

    Cheers,

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Tom Sisk

    http://www.siskconsult.com
    ·
  • Beau SchwabeBeau Schwabe Posts: 6,568
    edited 2010-05-19 18:07
    Nice one Phil!

    Application dependent, I have realized the value (no pun) of finding a 'fractional bit value' or bias of the LSB rather than truncation.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • rjo_rjo_ Posts: 1,825
    edited 2010-05-19 18:19
    So... Beau...

    If I understand you correctly, what you are saying is that sometimes rocket science should be left to rocket scientists?


    Rich
  • Beau SchwabeBeau Schwabe Posts: 6,568
    edited 2010-05-19 18:40
    rjo_,

    I'm sorry at this bias juncture I need more bits to give a valid deterministic result.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • RavenkallenRavenkallen Posts: 1,057
    edited 2010-05-20 02:45
    Whoa, just one wrong bit and your done for.
  • RDL2004RDL2004 Posts: 2,554
    edited 2010-05-20 09:00
    "The spacecraft is so far away it takes nearly 13 hours for a radio signal from Earth, traveling at the speed of light, to reach it, and another 13 hours to receive a response."

    - wow

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    - Rick
  • BradCBradC Posts: 2,601
    edited 2010-05-20 09:13
    RDL2004 said...
    "The spacecraft is so far away it takes nearly 13 hours for a radio signal from Earth, traveling at the speed of light, to reach it, and another 13 hours to receive a response."

    "Hello, IT support? Yeah, my ping has suddenly blown out to 26 hours. It makes my ssh sessions hell"

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "Are you suggesting coconuts migrate?"
  • skylightskylight Posts: 1,915
    edited 2010-05-20 11:09
    Isn't this taking it a bit too far?
  • John R.John R. Posts: 1,376
    edited 2010-05-20 13:45
    stamptrol said...
    Each bits does count!!

    Many years ago I was accosted by the corporate IT department who were hell-bent on "taking over" the administration of the PLC's and other controllers on the manufacturing equipment.
    This same guy could not understand how "just a single bit" being wrong could have any effect on the machinery involved.

    Thankfully, cooler heads prevailed and the electrical guys stayed in control.

    Cheers,

    The correct approach to the above situation depends entirely on your oganization, and the relative skills of the electrical and IT departments.· When the electrical guys start going after a PLC with a screwdrive with a half inch wide blade, it may be time to call IT.· Likewise, if IT starts by asking "what's a relay", it's time to start looking elsewhere.

    It gets even uglier if you start expanding to "compliance".· Who keeps the sofware licenses and media?· The guys who actually "use" it, or the guys responsible for answering the questions from BSA and/or the US Marshals?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    John R.
    Click here to see my Nomad Build Log
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2010-05-20 15:55
    I can't imagine trying to debug with a 26-hour turnaround. The baudrate at that distance would have to be glacial as well just to make it through the noise. My suspicion is that NASA has a simulator they can use to explore what-if scenarios before even communicating with the space probe. What the article leaves up in the air is whether the bit error is something that can be repaired or worked around. If a stray alpha particle zapped a ROM bit, that would be hard to fix. But knowing which bit was compromised might allow them to at least partially decipher the gibberish they're receiving as a consequence.

    -Phil
  • ercoerco Posts: 20,260
    edited 2010-05-20 16:15
    Twenty six hours is a pretty quick debug, compared to terrestrial precedents in the news now.

    Toyota has been quietly "debugging" (below the radar) their one-bit glitch for about 26 months now. Of course, that bit is the accelerator control, full off or full on...

    What a difference a bit makes!

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·"If you build it, they will come."
  • Beau SchwabeBeau Schwabe Posts: 6,568
    edited 2010-05-20 16:28
    The baud rate could actually be quicker than you think, it's just the propagation delay that 8.7 billion miles of space causes. Even with fiber optic communication here on Earth the 'data bits' can actually be in transit over the fiber cable itself. In other words the length of cable acts like a storage device, because the bit that you have sent on one end has not propagated to the other end and it's still somewhere within the fiber cable. Change the optic medium to 8.7 billion miles of space and think of how many 'data bits' you could have in transit from point A to point B.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2010-05-20 18:00
    Beau,

    Yes, that's true. But the S/N is going to be so low in this case that a fast baud rate is just not practical for error-free comms. If it's more than a few bps, I'd be very surprised.

    -Phil
  • Beau SchwabeBeau Schwabe Posts: 6,568
    edited 2010-05-20 23:27
    Phil,

    True, the number of samples required just to detect the presence or non-presence of a signal will kill the throughput (See the Audio Spectrum analyzer I put together <-- A similar technique is used to increase the signal to noise ratio in space, only in a form similar to spread spectrum using multiple frequency bands.)

    http://forums.parallax.com/showthread.php?p=663985


    ... Still though, you have to admit, this 'bit' error in the memory screams that it was human error and they are trying to make an excuse for it.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • localrogerlocalroger Posts: 3,452
    edited 2010-05-20 23:53
    Why guess? It's online:

    voyager.jpl.nasa.gov/spacecraft/index.html

    Uplink rate is 16 bits per second, downlink at 160 bits/sec for most, 1.4 kbit/sec for some data. The radios operate at 8 GHz, Voyager itself has a 3.7 meter dish and the Earthside dish is 34 meters. Voyager's transmitter puts out 23 watts.
  • HollyMinkowskiHollyMinkowski Posts: 1,398
    edited 2010-05-21 00:02
    WoW!
    13hrs away at light speed.

    It would be interesting to see what sort of signal is
    sent back to the Earth from Voyager. I would guess a simple
    system of turning the transmitter on/off and making slow pulses
    to form a code. Must be very very slow. Data rate up to the
    craft might be faster since they could generate a huge signal
    going that direction.

    And I thought wifi connections from a few miles away were exciting.
  • Martin_HMartin_H Posts: 4,051
    edited 2010-05-21 02:29
    Back in the late 70's and early 80's when the Voyagers were visiting Jupiter and Saturn NASA would have press conferences to show off the pictures. I'd watch them and be enthralled. Particularly the moon Io which looked like a pizza meets a volcano. During one of the press conferences a NASA administrator mentioned that some Ham radio operators had successful received some of the images on their own equipment. An impressive trick at the time.
  • ercoerco Posts: 20,260
    edited 2010-05-21 15:36
    I'm imagining that one of the team members still controlling the Voyagers might be part of the original mission control team. If so, he's an older fellow, ripe for retirement, but still quite happy to be involved in such a long-term project. The workday pace is none too demanding. At 26 hours for a communication cycle, he can send a single ping at noon, go home for the day, and return by 2 pm the next day for the confirmation ping. Not a bad way to fill out your golden years!

    May we all end up in low-pressure positions as technically significant and rewarding!

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·"If you build it, they will come."
Sign In or Register to comment.