Isolate DC levels

Discussion in 'The Projects Forum' started by Georacer, Mar 24, 2011.

  1. Georacer

    Thread Starter Moderator

    Nov 25, 2009
    5,142
    1,266
    I am (re)building that slot car project I have submitted on the completed projects forum and I am on a standstill:

    I have some issues with current surges from a switching load (the slot car tracks) causing trouble to the control circuit (a 74XX array). Both these elements have common ground, in order to communicate DC levels with each other.

    I would like to isolate the control circuit from the load without losing the ability to read DC levels from the load. One idea is to use relays that will completely separate the circuits and still be able to read 1 and 0 from the load (I am interested in binary information).

    However, they are clumsy and big in size (I need 16). I can't believe there isn't any smarter way.

    Any suggestions?
    Thanks in advance.
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Have you considered optocouplers?
    They're available in DIP packages, single, dual and quad.
    There is an IR LED on the "emitter" side, and the IR emitter controls a transistor on the output.

    A typical optocoupler like a 4N25 might have an IR emitter rated for a Vf of 1.2v @10mA, you might be able to sink 7 or 8 mA on the output side.
     
  3. Georacer

    Thread Starter Moderator

    Nov 25, 2009
    5,142
    1,266
    Just what I needed! Thank a bunch Sergeant!
     
  4. Georacer

    Thread Starter Moderator

    Nov 25, 2009
    5,142
    1,266
    I 'm posting back to talk about an unexpected event that occurred to me after fiddling a bit with DC isolation.

    As I wrote in the first post, I wanted to isolate electrically two DC circuits. The first one was comprised of many (30+) 74XX ICs and the second was a big mod of a slot car track with added traffic lights. (http://forum.allaboutcircuits.com/showthread.php?t=43588). The 74XX board was supposed to drive two big green LEDs (20mA each) and the bases of the transistors that would power the car motors.

    The problem was that each time a motor would start, the 74XX board would flicker, reset or change state randomly, obviously as a result of harmonic transients from the motors.
    I decided to use a small relay (5V secondary) to drive the track from the 74XX board. The two circuits didn't have one common node and still the transients would mess with my ICs.
    After SftWookie's suggestion I switched the relays with optocouplers and the problem was solved.

    The conclusion I came to is that the primary and the secondary windings of the relay were so tightly connected with mutual inductance that they would act pretty much like a transformer! Correct me if I 'm wrong about this, but this was a pretty unusual effect to me.
     
    Last edited: Mar 29, 2011
  5. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Gee, pretty nice job on a somewhat complex project! :)

    If the "glitches" were occurring when the relay coils were energized, then you did not have enough bypass capacitance near the transistors/relays. If the glitches occurred when the relay coils were turned off, then it is most likely that you did not have reverse-EMF protection diodes across the relay coils.

    The coil(s) in a relay are separate from the contact electrical connections. Unless you connected them together in some way, of course.
     
  6. Georacer

    Thread Starter Moderator

    Nov 25, 2009
    5,142
    1,266
    Thanks for the compliment Sgt! It means a lot coming from you. I might upload some more videos when the project upgrade is complete in the next few days.

    I was using two Sanyou DSY2Y-S-205L relays, without intervening to them internally. The relay's main contact is a selector that connect the output either on Ground or on Vcc. If the coil of the switch of the main contact isn't electrically coupled with the track circuit, then how do you explain the transmission of disturbances?

    I remind that the glitches occurred not when the switch opened or closed, but when the load (car motor) would start with the relay contact feeding Vcc on the tracks.
     
  7. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,536
    I wondered where you got your avitar. :D
     
  8. Georacer

    Thread Starter Moderator

    Nov 25, 2009
    5,142
    1,266
    I don't see how this is relevant to my question, but I googled for a decent picture of Corto Malteze.
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Were the Vcc's common to both the logic and the track supply? That would explain it.

    You have a lot of wires running around the track. You might have been getting coupling between the logic and power wires, particularly if they were twisted together.
     
  10. Georacer

    Thread Starter Moderator

    Nov 25, 2009
    5,142
    1,266
    The two circuits had different supply rails. The 74XX board was powered by a benchtop power supply and the tracks by a small 6V liquid battery.

    If there was coupling caused by twisted wires, wouldn't that persist after I switched to optocouplers?

    I 'm not obsessed with finding what was wrong. I was just asking if someone knew what was happening.
     
  11. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    That's pretty strange.

    I can't remember which relays you used offhand, but they generally have only a single coil.

    Now, I HAVE received relays once in a blue moon that had "stuff" rattling around inside them. This is hard to hear unless you have special equipment. Really small scraps of materials rattle in ultrasonic frequencies, but they can wreak havoc inside something like a relay.
     
  12. Georacer

    Thread Starter Moderator

    Nov 25, 2009
    5,142
    1,266
    Thanks for your time Sgt. In the end of the day I have learned to trust optocouplers when it comes to DC isolation. They can even read multiple levels of voltage, instead of just HIGH and LOW with proper circuitry.

    Have a nice evening.
     
Loading...