Shop OBEX P1 Docs P2 Docs Learn Events
RF interference/lesson learned — Parallax Forums

RF interference/lesson learned

yarisboyyarisboy Posts: 245
edited 2011-07-21 09:48 in Propeller 1
Some months ago several forum members tried to help me find out why my digital display would sometimes flicker. I found today that the culprit was that my ribbon cable between my prop and my display board was just too long. After I folded it up to about 6" the flicker went away. Makes me think a local AM station were disrupting communications. I've got the clock pulse for my MC14489 driver chip running at 8K. Comments about my software have been saved for reference when the next gremlin pops up. Thanks for your comments.

Comments

  • Jack BuffingtonJack Buffington Posts: 115
    edited 2011-07-20 16:27
    It is also possible that crosstalk could be the culprit. By folding the wire, you may have changed the characteristics of how the wire was 'crosstalking'. I had a situation years ago where I had a remote control panel that was attached with a six foot tether to the processor. It had me perplexed for several days about what was going on. When I would push one of the buttons, it would function as if a different button was pushed. Crosstalk was the culprit there. I put a low pass filter in there which fixed the cross talk issue. Probably a current limiting resistor would have done just as well.

    If you are using ribbon cable, a good way to deal with cross talk is to make every other conductor be ground.
  • yarisboyyarisboy Posts: 245
    edited 2011-07-20 19:13
    Jack Buffington wrote: »
    It is also possible that crosstalk could be the culprit. By folding the wire, you may have changed the characteristics of how the wire was 'crosstalking'. I had a situation years ago where I had a remote control panel that was attached with a six foot tether to the processor. It had me perplexed for several days about what was going on. When I would push one of the buttons, it would function as if a different button was pushed. Crosstalk was the culprit there. I put a low pass filter in there which fixed the cross talk issue. Probably a current limiting resistor would have done just as well.

    If you are using ribbon cable, a good way to deal with cross talk is to make every other conductor be ground.

    Very good point. I'm using a 10 conductor cable. +5, Ground, Clock, Enable and Data in, and Data out (that I'm not using yet). That will work out perfectly. With the multiple ground trick I could also make them all twisted pairs. I've seen that in commercial equipment at work. Some of them even have a foil wrap over each pair. Come to think of it we have Belden cable spools that way. If the wire were the same awg I could just use the 10-core idc connectors as long as every wire was in the right fork before squeezing the handle and setting the retainer clip. Thanks for jogging my memory w.r.t. electromagnetic effects. Before serial hard drives came out we started seeing some of these tricks with the 40-core IDC cables that came with parallel drives.
  • idbruceidbruce Posts: 6,197
    edited 2011-07-20 20:04
    @Jack Buffington
    If you are using ribbon cable, a good way to deal with cross talk is to make every other conductor be ground.

    Never heard that before, but it sounds like a great tip

    Thanks Jack

    Bruce
  • Toby SeckshundToby Seckshund Posts: 2,027
    edited 2011-07-21 01:47
    A long time ago, I worked for BBC Transmitter Group. They had 198KHz up to UHF units at power levels up to 300KW (individually). Back in about 1980 a mate tried to use his Acorn Atom with one of those 5" floppies ( that we were all just so envious about) at Droitwich. There they had LW and MW ancient beasts that could generate the best part of a mega Watt or so, combined. Unless he used the Atom with a layer of aliminium foil (grounded) around the ribbon then the disk was absolutely useless.

    Some times you can have an interference source that is just below nuisence level but becomes more enthusiastic when some combination of cable lenths becomes that magic 1/4 wavelenth. At Daventry TX station they used to drive the new Jaguar cars around the aerial field. They always used to cut out under the 17MHz arrays and required a tractor tow.

    The trick of using every other wire a GND is all that the 80 way ATA ribbons do, instead of the older 40 way ones.
  • prof_brainoprof_braino Posts: 4,313
    edited 2011-07-21 07:34
    idbruce wrote: »
    Never heard that before, but it sounds like a great tip

    I thought that was the deal with 80 conductor IDE cables. They used to be 40 pin/40 conductor, then the fancier drives started shipping with 80 conductor cables. The connectors are the same as far as the IDE port on the mother board and dirve, but the cable ends themselves maek every other condouctor ground. Anybody know if this is the case or not?
  • Heater.Heater. Posts: 21,230
    edited 2011-07-21 08:50
    Yes, that is the case.

    Years ago I resorted to using a ribbon cable where each signal of twenty signals was carried by a twisted pair. That finally sorted my problems with driving a bus just a bit further than I should have been.
  • TorTor Posts: 2,010
    edited 2011-07-21 09:48
    Heater. wrote: »
    Yes, that is the case.

    Years ago I resorted to using a ribbon cable where each signal of twenty signals was carried by a twisted pair. That finally sorted my problems with driving a bus just a bit further than I should have been.
    Yep, differential signals over twisted pair ribbon cable is the way to go for anything longer than short stubs. There's even rolled-up, shielded twisted-pair ribbon cable that looks like a fat round cable until you take a closer look. We used some of those at work for DMA transfers between machines located on different sides of the computer room at work.

    -Tor
Sign In or Register to comment.