Latching circuit (D-FF?) for LED Switch Track Indicators

I'd just add some .1 caps across the input resistors to snub any spikes that come through.

Boy, this is WAY OVER my head. I can barely pronounce most of the words above.

I forgot to mention, I don't know what LED driver you're using, but you should be able to drive those directly from the flop with a series resistor.

Finally, in looking at the circuit, I think I'd probably want some diodes running from the DC supply to each of the inputs to limit the travel to the power supply voltage.  Probably not necessary, but I'm looking for reliability here.

Hey, John -

Good points. I wondered about the caps, and I still have some that I can use. And the diodes never hurt, so they will go in as well.

Q: Should I ground any of the unused inputs through 10K resistors?

I will not use any LED drivers unless I have to; I need to double-check the output ratings of the FF. The 74HC74 may have a higher output rating, and if I decide to use it, I will have to use a 5VDC supply instead of the 12VDC, and probably add resistors between the switches and the FF to limit the R&S signal inputs.

EDITED TO ADD:

I checked the 7474 specs, and it can source 25ma per output pin - enough to drive the LED's, so that is what I will use.

Roger,

What gets me is that I used to be able to do this stuff as easily as I eating chocolate cake , but now it is like a dream from long ago.

Alex

What is going to wake this up in the proper state when you turn on the power?  It won't know anything until the first actuation on each switch.

What switches are you using, and do you have antiderailing?

Excellent point Dale, and I don't know if there's a ready answer to that one.   Since there is no power to the coils yet, and no feedback sensing to tell you where it is, it is somewhat problematic.

Dale, John –

I have been thinking about that (wakeup state) from the start, and here is my plan – two options.

Option 1. Use a battery to power just the Dual-D FF to keep its last state. The chip by itself should not require much current. I have yet to figure out how to isolate the FF from the LED’s just for the battery backup mode.

Option 2. The 74HC74 will have all inputs either High or Low, and all unused outputs not connected. I will have the R and S inputs High(*) through 10K resistors, and also the D input (do not know if the D will need L or H).

If I use this option for ‘waking up’, I will apply a pulse at the Clock input at the start of the session, and this will cause all Q outputs to go High or Low, depending on the signal at D. Hence, all LED’s will go Green/Normal.

Of course, if I use this option, I have to make sure that when I leave the layout all Switch Tracks have to be Normal to match the LED wake-up condition. But this something I already do now; i.e., at the end of every session, I return all the Switch Tracks to neutral, and if I am not sure when I start the next session, I throw any questionable Switch Tracks to neutral.

Dale – switches? Switch Tracks or turnouts, correct? I have 45 Atlas with their standard machines (dual coils) and 12 RCS with the DZ1000’s. The control panel toggle switches are miniature toggles from Jameco.

(*) NOTE: Since I will be using the 74HC74 instead of the CD4013B, the input levels to it have to be the opposite of what I showed in the circuit diagram. This means that the inputs will be held High (not Low) and the D1 and D2 have to be reversed to provide a Low signal (from the toggle switches when actuated) to the S and R inputs.

Please, please, please – do not be bashful with your constructive critiques, and I thank you for what you have already provided!

Here is my control panel: (Click to enlarge) The LED's will be at the actual Switch Tracks - I have LED holders installed there already - show up as small black circles. The toggles on the track turn the track blocks ON/OFF (Up is ON).

THANK YOU!

Alex

Maybe you could set up a startup or shutdown ROUTE that would normal all the switches.

Dale,

Funny you said that. The last HO layout I had (1987-1993) was 100% computer controlled. I used a Compaq with the 8086 and a whopping 10MB hard disc! The program was Quick Basic and Assembly Language. When I started the session, the first thing the program did was to cycle through all the turnouts and set them to neutral. I had not thought of a Shutdown ROUTE as you suggested, but that will work as well! Thank you.

Today's questions -

Initially, I was going to use a CD4013B, but have decided to instead use a 74HC74 (also a dual D flip flop) because its output can drive the LED’s directly (thanks, John).

Whereas the CD4013B could run at 5-15V (or so), the 74HC74 does not like anything over 6v, so I will use a 5VDC regulated supply for it. Also, the 74HC74 requires a negative signal/pulse at the SET and RESET(CLR) inputs.

As I showed in my circuit (first post), I was going to connect the common side of the 18VAC supply (for the Switch Tracks) to ground (the negative of the 5VDC supply), and tap the toggle switches momentary contacts (the other side of the 18VAC supply)  through diodes to the SET and RESET inputs of the D Flip Flop. I measured this voltage, and as expected it is around 9VDC (half of the 18). That 9VDC is the RMS value, and I imagine that the Peak value would be around 13V.

Anyway, for the 74HC74 I need a negative pulse, so the diodes (D1 and D2) would be inverted from what I showed in the circuit diagram.

Two questions:

(1) Based on the above, would I connect the common side of the 18VAC supply to the Plus (+) side instead of the (-) side of the 5VDC supply? This appears correct, but I prefer to ask.

(2) The 9VDC/13Vpeak signal from the toggle switches would be too high for the 74HC74 SET and RESET inputs, so I need a resistor of other means to get that signal down to 5v. Any idea of what value resistor may be required?

I'll post an updated circuit schematic later today.

Thank you once again!

Alex

The diodes will be supplying about 24V peak to the inputs.  Your DC meter values are misleading.

For each input:

R1 10K pullup to +5V

D1 cathode to input, anode to ground

D2 anode to input, cathode to 10K "input" resistor

R2 10K from D2 cathode to switch terminal

The input will sit high normally due to the pullup.  When the switch is closed, the negative swing of the voltage will pull the input down until D2 clamps it at -.6V.

Disclaimer:  I haven't built and tested this circuit.

Dale,

• You are right! The 18 volt is the RMS of the peak, and the true peak is 18 x 1.414, or 25.4V.

• In the original circuit, R1 and R2 served the same function; i.e., pull the inputs down. And D1 and D2 also had the same function – provide a DC signal to the inputs from the AC present at the toggle switches.

Here is the new circuit using both D-FF’s contained in one 74HC74 package.

In the circuit above,

TO1, TO2: Turnouts (Switch Tracks)

S, S2: Control panel toggle switches (Mom-Off-Mom)

R1 > R4: 10KΩ (Reduce voltage for input) *

D1 > D4: IN4004 (provide DC signal for inputs)

R5 > R8: 47KΩ (pull up inputs to +5V) **

R9, R10: 470Ω current limiting for LED’s

IC1: 74HC74 Dual D-Flip Flop with Set and Clear/Reset

LED1, LED2: BiColor LED's

5VDC: Supply for IC1 and LED’s

18VAC: Supply for Switch Tracks (and provide pulse for S and R inputs to IC1)

Questions:

* Will R1 > R4 (10K&Omega be sufficient to reduce the 25V to about 5V?

** Will R5 > R8 (47K&Omega be OK to pull up the inputs to 5V and be small enough to allow the signal from the switches through D1>D4 and R1>R4 pull the inputs down close to 0V?

I will breadboard the circuit as soon as I receive the components (don’t have the 74HC74’s yet), but I want to be as close as possible ‘on paper’ .

Thanks!

Alex

Just thought I'd throw in what I've been using on my Atlas switches, a low-force SPDT Microswitch (Cherry E61-60H (9 grams), Mouser 540-E61-60H). A piece of bamboo barbeque skewer stuck to the lever with CA gives me a little extra clearance.

This works great as long as you have room to squeeze it in. The cover is folded construction paper stuck together with a glue stick.

Regards,

Gary F.

Atlas snap-relays work for this too.

I think you are better off with the clamping diodes I suggested, rather than trying to make a voltage divider.

For stuff like this, I always use clamping diodes to insure the voltages don't exceed the chip ratings.  You never know what's coming in from the rectified AC power source.

Gary,

I have been experimenting with low force limit switches as the one you showed, and have also made my own. I am trying to actuate them with the points to make sure that the points themselves, and not the switch machines or linkages, are properly and fully thrown. I have had good success with my home-made version, but they are difficult to install. In the meantime, I will use the scheme we are discussing, which requires local wiring; just within the control panel.

Dale, John –

My parts were supposed to arrive today, via USPS, but they didn’t, so I wired the input signal circuit to test it, using Dale’s suggestion and John recommendation. (THANKS!)

Here is the circuit (correct, Dale?)

Test bench

Test circuit

(V1, V2, and V3 are the test points, as shown in the circuit diagram, above.)

V1: 18VAC (supposedly), but it actually is 21.1 VAC RMS, about 30V Peak, 60V P-P

I imagine that with a load it would be closer to 18V.

V2 with S1 open. Not significant.

V2 with S1 closed. Not significant.

V3 with S1 open. (Actual input to ground = +5V, as desired.)

V3 with S1 closed - input is active. It is negative from the above state, but it still is above ground level. Will it work?

I can hardly wait to get the parts!

THX!

Alex

I think your 18V supply should be referenced to ground, not +5V.

Well, if you're soliciting "any" comments, it seems that your circuit only indicates what is being commanded.  I'd think you'd want something that indicates the actual position of the switch.  Perhaps some kind of optical, magnetic, or mechanical detector directly driving the LEDs?  This would make the initialization problem moot since no "state" memory is required.

Dale,

"I think your 18V supply should be referenced to ground, not +5V."

I had wondered about this earlier. Just to be clear; you mean one side of the 18VAC connected to ground and not the +5, correct? (As opposed to just the voltage measurement on the existing circuit as it is.)

Stan,

You wrote, "I'd think you'd want something that indicates the actual position of the switch."

And I had written on my last post, in reply to Gary's suggestion of the use of limit switches,"I have been experimenting with low force limit switches as the one you showed, and have also made my own. I am trying to actuate them with the points to make sure that the points themselves, and not the switch machines or linkages, are properly and fully thrown."

Thx!

Alex

what if you did a "power-up-one-shot" that cycles each switch at power up. not a bad test as long as there aren't any cars, trains sitting on a  switch. coupled with position feedback, each throw could be verified for "hangups" - we are getting into computer control here but sweet!

Exactly.  What I'd find interesting is how the circuit should be configured depending on the type of position sensor and effort to modify the switch itself. For example, if using a minimal-throw limit button/switch would there be a no-man's land if the switch is not fully thrown.  In such a situation if the circuit output is digital like your FF, then one light will be on, possibly incorrect, and still have the startup issue.  A variation on the FF output might be able to drive your dual-color LED such that both R and G are driven so you see Orange in such a "fault" condition, or neither are driven so it is dark.

One issue is for non-derailing switches, as soon as a train switches them, your panel has the wrong indications again.

As far as the microswitches on the switch, have you considered a tiny magnet and a reed switch under the switch?  This would give you positive position sensing without the mechanical complexity.

Plandis,

Yes, you got that right! That is the final goal: 100% computer control.

(Which begs the question - Then why the control panel at all? Because I prefer to actuate the turnouts with toggle switches rather than scrolling with the DCS remote as I am doing now, and computer control will be about a year away. Besides, the fully functional control panel will be neat.)

Stan,

Yes, that would be correct, except that if/when I use my sensors (two per turnout), they will drive the LED's directly - No Flip Flops. If the points are in no-man's-land, neither of the sensors will be actuated and the LED's will be off, which would be an indication of an improperly thrown turnout. I have tried my sensors on three Atlas turnouts and they work fine other than it is difficult to adjust them.

Dale,

I tried connecting one side of the 18VAC to ground instead of to the +5V, and that causes the input to be at -5v with no input signal, and at 0v (ground level) with the input signal applied. The 74HC74 requires that the inputs be pulled high with no signal, and then the signal would pull them to ground level. However, it may work anyway since I can always select the Q or not-Q outputs as needed. (I think)

<I sure hope my parts arrrive today so I can finally wire the actual circuit!>

That doesn't make sense.  If you apply only the ground reference wire to the circuit with the switch open, it should not have any effect on the circuit.  Is you 5V supply a fully independent supply?

You are right, Dale; of course it doesn't make sense: I had the scope probes switched around! 

Here is the circuit as it now stands, and with the scope ground clip at ground (duh!) and the probe at the 'input': (Note that V2 is measured across a different place from previously.)

With S1 open, V2 = +5v

With S1 closed, V2 = pulses to ground level

THX!

Alex

PS. No, I did not get my parts yet. The USPS tracking said Nov 30, but so far no luck.

Alex, my Lionel and ZStuff switch motors require a contact closure to ground, which makes anti-derailing a lot easier.  What brand are you using?

Dale,

I have Atlas and RCS. I use a separate 18VAC supply for them, and no anti-derailing. Never had any problems.

Alex

Received the 74HC74 Dual-D Flip-Flops.

Wired one circuit and it works perfectly!

Now to wire 23 chips (46 toggle switches and bi-color LED's)

Thx!

Alex

That should keep you busy for a bit of time.

What do you use to build these on?  I think for that large a circuit, I might break out the wire-wrap gun again.

Send me an email, I may be able to help you with the WW sockets.  You're looking for 14 pin sockets, right?

John, I really appreciate your offer, but if I do go ahead, I will have PCBs made; it just is too much wiring to do it 'long-hand'. I am (learning) using ExpressPCB.

STATUS:

After lots of fiddling and trying different things, here is where I am:

After successful test with one half of the dual-D IC using the prototype board, I hard-wired one complete IC and tested it on the bench and then on the actual control panel. Not good results – the circuit is unstable.

On the test bench –

The A input (Clr) affects B Outputs.

• A inputs cause A outputs High and Low as expected [OK].

• A Clr-A input causes Out-B > High and (not)Out-B > Low, not as expected [Not OK]. The B outputs will remain as such regardless of subsequent A input signals.

B inputs are OK.

• B inputs cause B outputs High and Low as expected [OK].

• B inputs have no effect on A outputs [OK].

On the control panel –

With or without the Switch Track machines (solenoids) in the circuit, A and B outputs are unstable; they will not consistently have predictable output states. Inputs at A will affect B outputs, and vice versa.

Any suggestions on how to make the circuit stable?

Test setup on the bench:

Same board on the control panel:

The control panel toggle switches (only two wired for the LED latch circuit)

The bi-color LED's on the control panel.

Circuit for one Dual D chip. I will need 23 of these, and will have 6 on each of 4 PCB.

Thanks!

Alex

I would start by tying four additional diodes on the set/reset pins to 5V, the anodes going to the IC pins.  You want those diodes to limit the positive excursion past 5V on the input pins going past the IC limits, as well as the existing ones protecting the inputs from going below ground.  Also, the spike filtering capacitors would probably be better directly on the IC pins, allowing the resistor to soak up any large spikes.  Remember, the coils in the switches are putting out a healthy spike at times, those may be causing latch-up of the CMOS chips.

It's always annoying when the lab circuit doesn't work in the real world.

Alex,

Looking good.

I used Express PCB by the way, good outcome.  I got turned on to it by one of my engineer friends who uses them for prototype work.

What do you use for schematic capture?  I use TinyCad, but it's somewhat of a pain to get the components suitable for PCB manufacture.

John,

they have a free program that you use to create either schematics  or the actual PCB layout.  It has a library of a number of components.

The program allows you to directly upload and order the PCB board from them with several options.

I'll have to check their program.  It's key that it allows you to build parts, because it's pretty unlikely they have all the parts I might use.

Yes John, you can build parts because I had to do that for a particulate connector I was using.

thats not 18vac common to ground is it??. Is this legacy or tmcc? I didnt think you could ground the track power common. ???