Kato double crossover question

[kvp]

First, i didn't actually reverse engineer it, just found the original description on an old BBS archive written by the original designer before he went offline only with his design. (i wonder why...) The idea is that when the capacitor charges, current goes into it, that creates a flow in one direction, but then it stops as the capacitor is charged. When you ground it, it is discharged and a current flows out flowing through the coil in the other direction, stopping when the capacitor discharges. This ensures that although constant DC voltage is sent through the turnout, it won't burn out as there is no current flowing once the capacitor is charged or discharged. There is a small leakage current in the capacitor as they tend to discharge if left alone, so by keeping it powered it is kept fully charged.

 

The other type of circuit you were describing is for 3 wire dual coil turnouts, that use a central capacitor and momentary switches to pulse the coils and it's extra feature is that its charging is cut off while activated. That design is not usable with bipolar turnouts. (a double pole, double throw modification is possible, but the recharging must still be current limited to avoid tripping the power supply, i built a bunch of those for 3 wire, dual coil viessman signals) For the bipolar, single coil, single capacitor circuit, if you don't keep the capacitor charged after you charged it with the first throw, there won't be enough power in it over time to discharge and it won't throw reliably backwards or at all if it's capacitor is fully depleted itself. In the other position, the capacitor is fully discharged, so it's ready to be charged on the next throw.

 

Please don't try to confuse people with wrong information. (a good example is that a double throw or on-on switch is not to be confused with the nc/no states of a spring loaded single coil morse relay which has a default unpowered state, while the on-on switch is bistable as it stays put in both positions)

 

Edit: i think i explained it way too long and detailed, so let's give this a rest

Edited February 21, 2016 by kvp

[cteno4]

KVP's circuit is what George Stillwell has been putting around for the last decade or so as the BCD circuit. It's been reverse engineered and out there from a lot of sources as well. It uses on-on switches to keep the cap charged and thus not slowly drain with time AFAIK. I've not seen him do a momentary switch for the BCD as it's simpler to use the pulse with the reverse current in the cap. I have seen dual cap circuits in the past for atlas 2 coil machines that used two caps and two momentary switches. But again the single reversing polarity coil of Tomix and kato allow the simple BCD alternative that allows the simple directional lighting to be done as well.

 

Cheers

 

Jeff

[emkay_777]

Sorry guys, all I wanted to know is if circuit is correct.  This is my pictorial interpretation of that circuit with an added switch for an added turnout coil.  Again, I am still awaiting the arrival of my capacitors so I can experiment myself, but in the meantime . . . . .

 

http://s442.photobucket.com/user/jets427/media/bcd%20circuit_zpsxlqz4bfc.jpg.html

Edited February 21, 2016 by emkay_777

[cteno4]

Yep looks good to me.

 

Jeff

[inobu]

The momentary switch was my addition based on this design.

 

 

I felt toggle switches did not look good so I used rocker switched also known as "momentary switches". 

 

emkay,

 

breadboard the circuit and you will see how it works.

 

Inobu

[emkay_777]

I am not so patiently waiting for my components to arrive at which time I will certainly be experimenting. If you use momentary rocker switches, which do look quite nice, would you not require a relay to use led indicators with it?

 

Sent from my Nexus 5 using Tapatalk

[inobu]

I am not so patiently waiting for my components to arrive at which time I will certainly be experimenting. If you use momentary rocker switches, which do look quite nice, would you not require a relay to use led indicators with it?

 

Sent from my Nexus 5 using Tapatalk

LOL,

 

You have the basic operation down. The original Kato switch uses a bent wire that acts as a momentary contact/switch. The momentary switch with the BCD emulates the same concept but isn't really needed. What is does is allow you to visualize the concept. Part of instructing/teaching is provoking thought. Because you understand the basic concept you have reasoned out the issue of holding the LED lit. Because you can see the issue we know that you have an understanding.  

 

As you build your skill set you will be able to rationalize the usage of this setup. I stopped my development of this set up because I came to the conclusion that I wanted more. I wanted computer operated and control and signaling. The time and effort behind building the switch controls, indicator and what I wanted in the future altered my path.

 

Give a man a fish and he will eat for a day. teach him to fish and he will eat for the rest of his life. 

 

In any case I think you will be ok.

 

Inobu

[inobu]

Good luck

 

Inobu

Edited February 21, 2016 by inobu

[kvp]

If you use momentary rocker switches, which do look quite nice, would you not require a relay to use led indicators with it?

Yes, you do. And either you have to wire the turnouts through the realy too or have to use the classic charged cap circuit and the relays at the same time which is quite complex compared to the single switch and cap variant. (Jeff described both) The usage of the relays do allow route control logic though as you are building a relay based logic board (a primitive electromechanical computer). Most people get frustrated at this point and buy something off the shelf though.

 

One interesting variant of the basic circuit is the use of rotary switches. You can get for example a rotary switch which can select 4 positions for 3 outputs. If you have a 3 turnout 4 track ladder, you can wire each turnout and cap pair to one output, then add power or ground on each input. (4x3=12 inputs to 3 outputs) As you rotate to a track number, you set the tunouts to that track. No relays or momentray switches required and programming it is just selecting power or ground for each input. (can be done by trial and error or with a 4x3 track/turnout position table)

 

ps: Last year i posted one of my builds with rokuhan bipolar turnouts where the caps were placed under the main baseboard. The tiny control box just have the switches and a single wire for each turnout with a single common ground. I mounted the caps on a piece of raster board with screw terminals for the turnout wires and a ribbon cable for the control box. This results in a neatly wired layout and requires minimal soldering, but you can even build the basic circuit with screw terminal strips and no soldering at all.

[emkay_777]

Thank you gentlemen for all your input (pun intended) I think I'm seeing clearer now.

KVP, that rotary switch idea is intriguing. So each click of the knob, the cap fires? They must recharge almost instantly then?

 

Sent from my Nexus 5 using Tapatalk

[kvp]

Actually you only change the output for each turnout only once in a ladder, so only charge/discharge each cap once. (sss,dss,dds,ddd->000,100,110,111) The caps charge or discharge as long as it takes to throw the turnout. Faster and it will only throw partially, slower and you just warming the coil. The speed depends on the coil and the capacity, this is why a double crossover with 4 coils need the 2200 uF cap while the single turnouts are happy with something around 1000uF.

[emkay_777]

Well it only took 10 weeks to get my capacitors from China.  I anxiously started in last night and it actually works - except . . . the double crossover.  There doesn't seem to be enough oomph in the 2200uF cap to actuate the switch on the layout but works fine with an identical unused switch I have.  Originally I thought I was using too small a wire gauge (small speaker wire) but it is only 3' distance and works fine with the traditional Kato turnout switch at 5' distance. Only one set of points actuates with the capacitor.  The cap itself is fine as it robustly throws the single crossover (2 switches).  So I'm kind of stumped here.  This might be a dumb question, but can I put a 1000uF cap in series with the 2200uf cap for more power?  Thanks

[Takahama Trainwatcher]

Don't take my word for this, but I would think that putting the capacitors in series would not increase your "power"; putting them in parallel, however, would increase the amount of charge that would be released upon discharge.

Check out http://farside.ph.utexas.edu/teaching/302l/lectures/node46.html which shows that capacitance actually decreases when the capacitors are in series (but adds together when in parallel). The following page, http://farside.ph.utexas.edu/teaching/302l/lectures/node47.html , shows how much energy is stored. One of the equations, W = CV²/2, shows the deal: assuming you have the same voltage V in every configuration (your supplied voltage), the energy stored by the capacitors is proportional to the capacitance C. Therefore you would be aiming to increase the capacitance C if you want more oomph. Parallel would be the way to go with the 2 capacitors.

[kvp]

Around 1000 uF is needed for a single coil. A single crossover has 2 and needs a 2200 uF cap. A double crossover has 4 coils and i would suggest either 4 x 1000 uF or 2 x 2200 uF. Parallel connection (just like with batteries) increases the capacity, while a series connection the maximal voltage.

[inobu]

Voltage drop across the wire needs to be taken into account. A larger gauge wire induces resistance contributing to a voltage drop so larger may not be better.

 

Inobu

[cteno4]

Bummer it took so long! In the last few months I've noticed a wide variation in these small packet deliveries from China. Mine are from 1-6 weeks, most still average about 2 weeks though. Some jumpers just blew here in 4 days!

 

Jeff

[kvp]

Actually thinner has more resistance and thicker has more capacity. The us wire gauge system is reversed.

[emkay_777]

Thanks guys - I put another 2200uF cap in parallel and not much different happened - only one or two points activated.  I'd say I had a faulty switch if it wasn't for the fact that it works fine with the Kato controller. 

[kvp]

Then it is possible that your power source is the limiting factor. What do you use?

[emkay_777]

Oh, that's a good point.  I have a 12V 1.0A power source

[inobu]

You need to engineer your circuit. There are dissipation factors to take into consideration 

 

Think of your capacitor as a cup of water and the wire as a tube. If the tube is too big (diameter/gauge) and long the water will remain in the tube (dissipation) nothing will come out the end. If the tube is too small then the water will back up and spill (another from off loss). There are thresholds for everything. The "4" switches (AKA crossover) are exceeding the capability of your circuit.

 

You have 4 switches but a single input. 

 

Try using a shorter and smaller gauge wire.

 

Keep in mind that the kato switch pack is connected to the power pack which is kicking the switches with 12V 1A

 

Inobu

 

 

not my image.

 

Edited April 3, 2016 by inobu

[kvp]

I think you need around 0.8A per coil. The capacitor method needs a bit more power than the plain DC switch that Kato uses but throwing that too quickly also gets the same result.

[inobu]

emkay,

 

I just tested your configuration and it is beyond the wiring. If your configuration was border line you could play with the wire gauge and length to get it to work but it's no where close. You fine tune it be listening to the chattering of the switch.

 

The crossover needs about 12.79 volts to throw without chattering. 4 individual switched require 11.56 volts to throw.

 

Depending on the capacitor you bought increasing the voltage can help. That's only if the cap is rated to handle higher voltages. When I went through this exercise I bought 16V caps. So I cannot exceed 16V. It appears that the crossover windings are different than the #4 and 6 or so it seems.

 

I'm not sure adding caps will be the solution. This issue starts to venture deeper into electronics with displacement current and such.

 

Inobu

Edited April 3, 2016 by inobu

[emkay_777]

Interesting.  I (also) have a 19V 3.42A power supply.  My capacitors are 1000uF and 2200uF  rated for 25V  105°C.

[kvp]

Ok. First the wires shouldn't matter much if they are under 1-2 meter in length and around the same thickness as the kato wires. Second the voltage should be 12V DC and a 2-3A power supply should work. Third the caps must be connected directly to the turnout and not on the switch end. This means turnout+caps should be in the same location which eliminates most of the voltage drop and stray capacitance issues.

 

ps: it's not magic or special electronics...