Turnout toggle/capacitor control circuit from the old Atlas forum?

[mmagliaro]

But in this case it latches because of the permanent magnets in the Kato turnouts, right?  It's not related to energizing a coil per se.

It latches because the Kato turnout is just like a latching relay, which also works by magnets.  So you only have to energize the coil for an instant.  It moves the contacts, then the magnet holds them there and you can shut off the power.  And if you turn things off and on, it stays where it was because it is held by the magnet (unlike a conventional relay).
What I'm getting at is, if I use a latching relay for direction control, I'd also like to be able to use a toggle switch, because then the toggle handle itself indicates the direction.   But when the toggle is flipped, I don't want it to unnecessarily keep the relay coil energized, which is just a waste of power.  It's only 50 mA, but still....

[reinhardtjh]

Stillwell!  BCD!  THAT'S IT!
No, I am not especially interested in the LED indication.  What I remembered was that it could be a way to kick a
LATCHING relay one way or the other with a toggle switch, but not have to keep the latching coil energized permanently
after flipping it.  i.e. You flip it, the current pulse fires the latching relay, the capacitor charges up, the current shuts off,
and you are done.  Flip the toggle the other way, the same thing happens.

Okay.. who has the Stillwell circuit info??? That's what I'm looking for.

@mmagliaro

Let's see if this works...

[mmagliaro]

Thank you, John.  Yes, that's the just information I needed.

[mmagliaro]

Oh, ok.  Glad you found it.  The subject line of this thread seemed to imply that you were interested in LEDs as inidicators.

Yeah, I agree.  I changed the thread topic title.

[mmagliaro]

Well, the Stilwell BCD circuits (in all their various incarnations), don't quite work for my application.
They all work on the principle of charging the cap when the relay (turnout coil) is thrown one way, and then DISCHARGING the cap, flowing current in the opposite direction, when the relay/turnout is thrown the other way.

In my case, I have a dual coil latching relay, and throwing it each way requires the same polarity.
(i.e. a center common lead to both coils is at +12v, and then you ground the opposite terminal of
one coil or the other to latch the relay one way or the other).

However, thinking about cap discharge did yield me something simple that works...



How does it work?  When the control switch (S1) is thrown the way shown, initially C2 is not charged, so it presents a dead short, and the relay coil RRelay1 is grounded, so it closes the relay in that direction.  This dead short only lasts about 5 mSec (see the graph provided), because it charges up and then current stops flowing, except for a small 5 mA current through R2.

When you flip s1 the other way, the same thing happens on the opposite coil and the relay closes the other way.    R1 and R2 are there so that the capacitor not currently in use can discharge, so that the next time you throw the relay that way, it will again present a dead short pulse to operate its coil.

R1,R2 discharge the cap quite rapidly.  I rapid-fired the toggle back and forth and could never get the relay to fail to operate.  R1,R2 are large enough that the current that continues to flow through them to the relay coil after it operates is only 5mA.  That's a big improvement over 50 mA, which is what happens if you just use a toggle switch with no circuitry, and let the relay coil stay on all the time.

According to the data sheet for my relay, I should need a 10 mSec pulse to guarantee that the relay operates.  And it is guaranteed to operate at 8.5v (for a 12v coil).   As you can see from the pulse graph, I'm only getting about a 4 mSec pulse with a 47 uF cap.  that is, the time from where the voltage is 0 up to where it is about 3.5v (so the voltage across the coil is 12-3.5 = 8.5), is only about 4 mSec.  So technically, this should not be working with only a 47 uF cap.  220uF would be a safer choice, since that would get me the mandated 10 mSec pulse.