Shop OBEX P1 Docs P2 Docs Learn Events
AC speed control - Phase angle fired TRIAC? — Parallax Forums

AC speed control - Phase angle fired TRIAC?

MJHanaganMJHanagan Posts: 189
edited 2011-03-08 06:26 in General Discussion
Does anyone have any experience with controlling the phase angle triggering of a TRIAC? I am trying to do some speed control of an AC powered model train (a circa 2004 3-rail Lionel train set) using a Propeller chip. From what I gathered so far the normal power supply that comes with the train set is triggering the output of a TRIAC to essentially "chop" the output from a 20 VAC transformer to produce the speed control of the train. If you advance or retard the phase angle firing you essentially induce a slight potitive of negative offset voltage which triggers the train's bell or whistle into action.

It seems to me that the Propeller chip running at 80 MHz should run plenty fast enough to provide some very nice timing control on a pokey 60 Hz supply voltage. Has any one tried this, or any ideas on how to go about triggering the TRIAC (or DIAC?) to get phase angle controlled output?

Comments

  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-03-06 19:30
    Zero-cross detection is the key concept, you're going to sync your triggering with that.
    Once you trigger it, it stays in conduction till the end of the alternation (like an SCR), but a TRIAC
    can be triggered to conduct on both halves of the cycle.
    There are specialised TRIACs which can be triggered in all four quadrants.

    When controlling inductive loads it's imperative that you include a snubber network.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2011-03-06 20:01
    I've done this with 90VDC motors, and it works well. (The advantage of a DC motor is that you can also control the direction by selecting positive- or negative-going cycles.) As PJ states, you need a good zero-cross detector. And, if you're using an optotriac to trigger the real triac, pick one that does not have zero-cross switching.

    -Phil
  • MJHanaganMJHanagan Posts: 189
    edited 2011-03-07 03:46
    I was thinking about using the output of a voltage comparator as an input to the Propeller to signal the zero-crossing point on the 60 Hz. The based on that timing determine how long to wait before triggering the TRIAC to turn on and off (and yes, I want to use an opto-coupler to trigger the TRIAC to keep things isolated). I figure if this simple process code is running in a single cog it would be able to handle the ±1° phase angle changes with a timing sensitivity of ~45 µsec.

    What does the snubber network do and where do I find out more about it?
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-03-07 04:36
    A snubber is composed of a series RC circuit in parallel with the
    TRIAC's MT1 and MT2.
    Its purpose is much the same as a commutation diode in a DC
    circuit when that involves an inductive load.
  • MJHanaganMJHanagan Posts: 189
    edited 2011-03-07 16:43
    I did a bit of seaching and found an application note on opto-coupler drivers for TRIACS. They discussed a bit about the snubber circuitry for inductive loads (ie motors). They said knowing the "power factor" of the motor is what is needed to make a good estimate of the capacitor and resistor required to do the snubbing. I suspect I will have do some oscilloscope measurements to determine this, right?
  • PliersPliers Posts: 280
    edited 2011-03-07 18:13
    I would go the DC controlled route as mentioned by Phil. Controlling a FET is easy and more precise.
    Your train motor is DC powered, correct?

    Model trains are fun, post some pictures.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-03-07 18:29
    ...

    I'll go out on a limb:
    A suitable snubber, in a general application, would be 100Ω in series with 0.1uF.
    Over-rate those components, use a non-inductive resistor [no wire-wounds].

    Ready-made snubber networks are available, common industrial devices,
    they look like a big mylar cap.
  • MJHanaganMJHanagan Posts: 189
    edited 2011-03-07 19:20
    The Lionel set I have is an AC powered train (as is most 3-rail Lionel setups, I think). If it were a DC system then a simple pulse modulated DC power would be quite simple to build and control. The AC power is a bit more challenging!
  • MJHanaganMJHanagan Posts: 189
    edited 2011-03-07 19:22
    I'll try your 100Ω and 0.1 µF combination and let you know how it does.
  • stamptrolstamptrol Posts: 1,731
    edited 2011-03-08 05:07
    You're getting good advice re the triac.

    As for the train motor itself, many of the old Lionel's ran on AC but the motor was built as a "universal" motor with brushes that can run very nicely on DC. The reversing was carried out with a relay onboard that alternated each time power was applied. By converting to DC, you could have a smoother running system.

    Cheers,
  • MJHanaganMJHanagan Posts: 189
    edited 2011-03-08 06:26
    This is a circa 2004 Lionel train set (NYC Flyer) which I am not certain can be run on DC. I read somewhere that the motor runs on AC and the train sounds are triggered by imposing a DC offset on the AC voltage. Using my trusty "TRMS" volt meter I can measure a DC offset (a positive DC trioggers the bell and negative DC triggers the whistle). So it seems like running it with pure DC may not work.

    I am a bit surprised there isn't more information avaialble on how these things really work and how to build your own power supply. I would have thought all of this had already been done long ago using a Stamp or Propeller micro. I see someone has posted some DCC train code onthe Propeller exchange - but DCC is DC powered trains, not AC powered units.

    Looks like I will have to put a scope on the power unit that came with the set and have a good look at the output myself to see what is going on.
Sign In or Register to comment.