Electronic Motorized Rotary Switch?

Thread Starter

MRW1962

Joined Jun 29, 2012
18
Hello, our Museum got ahold of a retired railroad signal and would like to set it up to cycle through its 3 indications (green yellow, red) the green cycle would be the longest length of time, then it would repeat the cycle. The bulbs in the signal are 12volt incandescent bulbs probably similar to a turn signal bulb in a car, maybe 15 watts? The signal is position light signal, so it displays a row of 3 bulbs vertically, diagonally or horizontally depending on what level of permission the signal is giving. This means that the center bulb is on no matter which indication is given.

My thoughts are to wire the center bulb directly to a power source and then take the output wire from the center bulb and feed that to the 3 circuits that have 2 more bulbs in series. Then from these I would route the return power through 3 ground paths. This is why I might need a circuit that simulates a 4 position motorized rotary switch. This would give the green circuit ground for 2 switch positions, then yellow for one, and red for one. Then the cycle would repeat.

I'm not sure exactly which path to go, I'm toying with the idea of a DC control timer like a 555 controlling a relay, but am at a loss how to make a simple controller that can control 3 wires through relays. Maybe some kind of sequential IC like a flip-flop?
 

MrChips

Joined Oct 2, 2009
30,712
A picture is worth a thousand words.
Draw pictures showing the light patterns indicating the sequence of lights and the length of time to stay at each stage.
Leave the electronics to us. There are dozens of experts here who would be delighted to present their solutions.
 

John P

Joined Oct 14, 2008
2,025
Yes, Pennsylvania RR position lights?

I would not fool with any fancy wiring. Just operate all the lights off a 12V supply, not switched for the center light and switched with one power FET for each signal indication; you'd wire the two bulbs for each indication in parallel to the FET, so you'd need 3 of them. Make sure they are rated for several times the normal current though, as incandescent lights have a large inrush current when the filament is cold.

You want the 555 timer to drive a counter (could be just a pair of flipflops--no that is not footwear) where 2 of the output states would drive the signal to "Clear" then one state each for "Stop" and "Approach". Or if you want to get fancy, you could have more than 4 states and have a longer "Clear" time versus the other two.
 

Thread Starter

MRW1962

Joined Jun 29, 2012
18
Yes, the signal is the same design as in Tracecom's photo above, but only the upper signal head in the example I am working on.

John P, so I could simply run the center bulb from the 12V and have 3 FETs powering the 3 remaining pairs of bulbs. I will have to depend on your input for such a design, at least as far as the flip-flop circuit goes. If the FET is designed for 12V and double the current would the FET be dissipating a large amount of power? In the future I will replace the bulb with a 3 watt LED. The time for each cycle should be 30 seconds on the green wire (Clear), 15 seconds on the yellow wire (Approach) and 15 seconds on the red wire (Restricting). The cycle would then repeat.

While I wait for your input I'll check out some articles here for 555 timer circuits and flip-flops, maybe I'll understand the suggestions better.

Thanks everyone.
 

tracecom

Joined Apr 16, 2010
3,944
Yes, the signal is the same design as in Tracecom's photo above, but only the upper signal head in the example I am working on.

John P, so I could simply run the center bulb from the 12V and have 3 FETs powering the 3 remaining pairs of bulbs. I will have to depend on your input for such a design, at least as far as the flip-flop circuit goes. If the FET is designed for 12V and double the current would the FET be dissipating a large amount of power? In the future I will replace the bulb with a 3 watt LED. The time for each cycle should be 30 seconds on the green wire (Clear), 15 seconds on the yellow wire (Approach) and 15 seconds on the red wire (Restricting). The cycle would then repeat.

While I wait for your input I'll check out some articles here for 555 timer circuits and flip-flops, maybe I'll understand the suggestions better.

Thanks everyone.
So, this is what you want?
 

Attachments

wayneh

Joined Sep 9, 2010
17,496
If the FET is designed for 12V and double the current would the FET be dissipating a large amount of power?
Nope, once driven "on" they have an internal resistance in the mΩ range. So with even, say, 10A of current (120W at 12V), the power (I^2*R) might be 10^2*0.04 = 4W. That would call for a heat sink, but nothing extraordinary.

Look at the 4017 counter. With each tick of the clock, the output advances to a different pin, which you could route to your MOSFET. So you could use a 15sec clock and just skip a count (from 3 to 5, say) when you want a 30 second on time.
 

tracecom

Joined Apr 16, 2010
3,944
Or you could use three NE555CN monostable circuits in series with the first timed for 30 seconds, the second timed for 15 seconds, and the third timed for 15 seconds. Each output would be used to gate an N channel MOSFET and to trigger the subsequent 555 in a ring configuration.

But the lowest component count would come from using an 8-pin PIC controlling the three MOSFETs. But using a PIC would require a 5V supply. Still could be the cheapest. Lots of options!


To MRW1962. What are you planning to use for your 12V power source?
 
Last edited:

Thread Starter

MRW1962

Joined Jun 29, 2012
18
Tracecom, yes, vertical for 30 sec, diagonal for 15 sec and horizontal for 15 sec. I wonder if making it do even more might be counter productive since most of the people who see it will not appreciate anything more than this basic flow. I will probably just use a regulated power supply with enough current capability to power both the circuit and the bulbs.

Wayneh, good to hear, so basically when "ON" the FET acts like a switch with little wasted power to dissipate. OK, so if I figure out a 555 circuit with a 15 sec pulse timing I can use a 4017 to give a 30-15-15, 30-15-15... timing. Just need to properly interface the 555 to the 4017 and FETs.

Which generic FET might be best assuming a pair of 15W pair of 12V bulbs in parallel? I would probably get them on ebay.

Thanks everyone.
 

wayneh

Joined Sep 9, 2010
17,496
...so basically when "ON" the FET acts like a switch with little wasted power to dissipate. OK, so if I figure out a 555 circuit with a 15 sec pulse timing I can use a 4017 to give a 30-15-15, 30-15-15... timing. Just need to properly interface the 555 to the 4017 and FETs.
All correct.

The 555 output can directly clock the 4017, I believe, and I believe the 4017 can directly drive the gates of the MOSFETs, so it should be pretty straightforward to hook it all up.
Which generic FET might be best assuming a pair of 15W pair of 12V bulbs in parallel? I would probably get them on ebay.
I got myself a bunch of IRF540N MOSFETs from ebay a while back, so that's my default generic MOSFET. There are many other common ones, though, and more modern ones. I'd look for the TO-220 package (which is very common) and a rating of at least 5A continuous duty. There's very little cost to go larger, and not much downside.
 

wayneh

Joined Sep 9, 2010
17,496
Nice. This approach is more versatile if you want to change the relative timings of the 3 states. The 4017 approach would be more versatile - I should say less complex - if you wanted more states in the rotation. For this particular application, it's a bit of a toss up. For a noob, they only need to learn the 555, although the 4017 is pretty darn easy. Heck, even I have figured it out. :D
 

tracecom

Joined Apr 16, 2010
3,944
If I were going to build it for real, I would use an 8-pin μC. For a one-off project, I don't think it matters very much. It was fun seeing my 555 design work.
 

John P

Joined Oct 14, 2008
2,025
The 4017 component (the 4022 would be another possibility, with 8 outputs rather than 10) often gets mentioned here, and in fact I do believe that it's better than my idea; I had something like the 4013 dual flip-flop in mind, and it's workable, but there would have to be some additional logic to drive the FET gates, and you'd be stuck with a 1:1:2 ratio of time intervals.

This device does almost the exact thing you want, but with on-board LEDs rather than FETs driving incandescent bulbs:
http://www.eidusa.com/Electronics_Kits_TRAFFICER_01.htm

Note the way it uses diodes to combine the various outputs of the 4017 so that if any one of several lines is high, it causes the LED to come on. That's what you would use to drive the FET gates, but you'd need to add a pull-down resistor to Gnd so that if none of the combined outputs were high, the gate would see a low voltage level.

Tracecom, your sequence is wrong: imagine a train passing the signal, so it goes from green to red, then as the train moves further away, the signal shows yellow, and finally back to green. That would be G-R-Y not G-Y-R.

One way to deal with the inrush current would be to set up a current-limiting circuit to deliver the 12V to the three switched circuits; only one of the three can be on at a time, so you only have to build this once. It would let you use smaller transistors, but the major benefit would be avoiding the effect you'd get when a bulb pair turns on, where you might find that the inrush current would be more than your power supply could deliver, so the voltage would dip momentarily, and that might have bad effects on your timing and counting circuits. With a current limit, you'd know that the transistors were protected and that you wouldn't see current overloads.
 

tracecom

Joined Apr 16, 2010
3,944
Tracecom, your sequence is wrong: imagine a train passing the signal, so it goes from green to red, then as the train moves further away, the signal shows yellow, and finally back to green. That would be G-R-Y not G-Y-R.
I'll take your word for it; I know nothing about train signals except those at the highway crossings. I just followed the sequence that MRW1962 asked for. :)
 

#12

Joined Nov 30, 2010
18,224
You can use a diode and capacitor (maybe 100uf) to feed the power to the counter and timer so their supply will remain steady if the main voltage sags during an inrush event. It will only sag for a time measured in milliseconds. Plenty of time to confuse a counter chip, but you can keep the sag from getting at the timer with a diode and capacitor in the power line to the thinking parts.
 

Thread Starter

MRW1962

Joined Jun 29, 2012
18
Well, went down to the museum last night and was surprised that the bulbs will not light with my 12V power supply which is 12A! It kills the supply and I have to reset it. When I powered the bulb with a trainset AC power pack they light fine, and from what I saw, 6VAC is bright! I assume that the "inrush" current is tripping a protective circuit in my 12A power supply. The bulbs are of 2 kinds, some labelled 10V and some 12V. What DC voltage would correspond with 6VAC?Based upon this info what would you recommend?

PS I like the triple 555 circuit, what would a single 555 with a 4017 circuit look like?
 

wayneh

Joined Sep 9, 2010
17,496
...what would a single 555 with a 4017 circuit look like?
A little like this. I've borrowed this directly from something I built and I haven't customized it at all for your application, but maybe it will give you the idea.

The 1st timer on the 556 controls the timing of the 2nd timer, to give it a bit of randomness. The 4017 is set to flash an LED on "3" and to reset on "4". Looks like I was using outputs at "0" and "2", skipping "1". Those outputs could go directly to MOSFETs for controlling your lights.

 

Attachments

Last edited:
Top