Diode-matrix control of Kato turnouts

[Atomsk]

Has anyone ever run across an example of a circuit to control multiple Kato turnouts via a "diode matrix" system?

 

All of the diode matrix examples I've seen only apply to twin-coil switch machines, but Kato uses single-coil machines in their switches.

 

 

[cteno4]

Atomsk

 

The BCD point control system has a cap set solutions for doing ladders and such. Even a relay to reset on power on/off with indicator LEDs.

 

Usually George Stillwell emails these out but his email is not working. I can get it to you if you pm me.

Cheers

 

Jeff

[kvp]

The general idea is to have binary capacitor discharge circuits (serial caps with one end on ground) on each turnout and use one relay per turnout to select between supply and ground for the BCD. The coils of these relays can driven by diode logic, the same way twin coil turnouts are driven (bistable relays) or by a low active diode logic in case of morse relays.

 

Ladders and such things can be directly wired though to rotary switches without the relays, by using the various outputs to directly wire power and ground to each BCD.

[Atomsk]

Here's something I cam up with, based on several ideas I found online.  It's not as slick a solution as BCD, but it does give me single-button route control.

 

The limitation here is current.  Each machine draws about 500mA, so my 3.3A power supply can only support about 6 machines.  However, each machine only draws power for half the AC cycle (normal and reverse throw use opposite halves) so I think I can support routes of up to 12 machines, if there are no more than 6 thrown in any one direction.

 

The problem with BCD is the need to have the "input" side of the cap connected to either power or ground at all times, or at least you need to always give it power before grounding it to revers the switch.

 

If anyone has ideas on combining this kind of matrix with BCD, I'd like to hear about it.

Edited May 8, 2016 by Atomsk

[kvp]

BCD and pulsed AC matrices don't mix well, but a bistable twin coil relay for each turnout could convert the short AC pulses into static DC signals. (this also works for controlling light signals)

 

On the other hand, after adding the bistable relays, it's possible to drive them with simple DC signals too.

 

The DC only solution has the added benefit of parallel buffer caps, so you can easily buffer enough power to move more turnouts at the same time than your power supply could supply alone.

[Atomsk]

I'm testing the solution described here, and it looks promising.

 

Figure 1 shows an arrangement of four turnouts that represent five routes.

 

Figure 2 shows a method of controlling all four turnouts with five DPST momentary toggle switches.

 

Using the 12VAC power supply that I mentioned in an earlier post, I constructed a +/- 12VDC bipolar power supply. A 24V, center-tapped transformer would have been the preferred source, but I only had a 2-pole 12VAC "wall wart" to work with. The drawing leaves out some of the peripheral components around the 7812/7912 voltage regulators, but on the actual circuit, I did set them up according to the datasheet's recommended design.

 

Both regulators then charge up two large caps, which provide power to the diode matrix, and from there to the switch coils.

 

To control a single turnout, you only need an SPDT "double momentary" toggle switch.  Connect the center terminal of the switch to one end of the turnout coil, and the other two terminals to the + and - 12V sources.

Edited May 20, 2016 by Atomsk

[NXCALE]

Nice work, please post your test results once completed .

 

Cheers,

[Atomsk]

I've tested the matrix, and it works as expected.  It threw three switches at once with no problems.  Still need to see how many it can throw at once.  May need more than 4700uf, to control the whole layout.

 

Haven't tested the SPDT setup yet, but I don't see any problems, after the tests with the matrix.

I'm assembling "phase 2" (the blue tracks) on my table layout.  I call it The Milford, Davenport, and Franklin, reporting mark MDF (since that's what I made the top of the table out of).  I've fastened down the programming track (green), and I'll work my way around "phase 1" (red), before I start on the permanent wiring.

 

The arrangement at the lower right is based on a "Japanese switchback".  They use these on steeply graded lines, for passing sidings and passenger stations.  It appears that the two spur tracks are on level ground, so that the train won't have to stop on the grade.  Any train can pull into one spur, back across to the other, and then continue in its original direction.  Through trains can run straight through the diamond without stopping.  To differentiate between this and the traditional western "Z" type switchback, I call this configuration by the Japanese "Engrish" term for switchback, "suichi bakko".