User Tools

Site Tools


GCA79 Contactless ac/dc switching for turnout coil drives with internal timer and feed-back

KICAD 3-D The real thing


By Peter Giling


The first reason to develop this board, is that the the usual coils in a standard switch drive are specificly made for ac. And most digital turnout decoders in the market supply dc to the coils, simply because it is much easier and low cost that way.
Also GCA77 and GCA76 work that way.
The initiation for this was the problems we experienced on the large N-scale layout in our railroad hobby club MSCD. ( see )
Some turnout coils did not respond at all with the used GCA77, but directly driven with ac there was no problem.
Even using a relay-board like GCA107 did not reslly help.
With the conclusion that it is best to use a transformer to power the coils we had to do something extra. With the relay unit GCA107, there where two disadvantages:

  • ac switching always results in 'contact dender' which appears at the wrong moment: at the start of the movement.
  • with the way GC107 is made as an universal switch, it is hard to suppress distorsion from the switching coil.

The previous GCA78 unit provides a nice solution to activate coil switch drives very solidly.
But the mix-up of configuration with two ports, and both needing configurated as pulse output, will create some problems for many railroad hobbyists.
This board GCA79, based on the same principle has some nice extra features.
ON board is a processor which selects the right output and has also timing on board,
switching power to the coil only for a timed period.
Timing can be selected with on board dip switch in 4 different timings
And the distorsion that is generated by the coil, right after switching of the power, is supressed by a VDR on each coil.
Each combination of two coils for 1 turnout has 1 command input: 5V for straight and 0V for round.
This command is given by any port from GCA50(0) , GCA185 , CAN-GC2(a) , GCA_PI02, G-ASH02 and/or most of WIO units.
Depending ON or OFF the GCA79 will activate one of the two coils with the selected timing.
After the set time is counted to zero, the GCA79 sends a logical signal back to the commanding unit.
This commanding and feed-back will be possible via 8 ports of the I/O connector in the same way as used for GCA136

The firmware in this chip makes it impossible to activate coils during power up.
This will initiate a very quiet startup, no high demands for the standard transfomer for the turnout coils.
After this initial starttime, the GCA79 considers the actual value of the command inputs as the last used position,
and set all feed-backs to 'ready, conform theis position.
After this quiet power-up, you will still have the option 'start of day' to be sure that all tunouts are in the right position.
The 5V power for the on-board processor comes from the connected decoder, like GCA50 etc.
Consumption of processor is extremely low: less then 5mA, except when driving a coil.
In that (short) time, consumption will still be less then 15 mA, since only switch will be activated simultaniously.
The supply for the coils could be best the usual 16-18V ac from your standard transformer.
The extra advantage of the used Opto-coupler is abvious: the transformer power is isolated from electronics in the system.
The max current through each coil can be up to 2 amp, so also the Märklin C-rail switches will be fine.
To prefend any damage in this board when short circuit appears, also an automatic resetting fuse is on board.

Why is this a better solution, compared to relays?

A switching relay always has the disadvantage of 'contact dender'.
This causes the initial power to the switch motor to be reduced at first switch on,
causing a sometimes weak respond of the switch motor, and that just at the moment that the most power is needed to start the move.
GCA79 does not have that problem, either dc or ac , the internal Fets in the Opto-coulper give maximum power right from the beginning.
Together with the extremely low internal resistance makes this a better solution, compared to the use of a relay.
Practice has proofed, even with dc, that an old fashioned switch motor works more reliable in all situations.

Another important issue with coil switches.

Each coil where a current is going through (coil is activated), produces a high reverse voltage when switched off.
This will cause many times a distorsion pulse in electronic systems.
A simple and most always good way to supress this, is to connect a diode parallel to the coil (kathode to +).
But many turnout coil drives have a built in switch which then also separates the diode from the coil.
That makes the diode totally useless.
For AC switching, a diode is not possible.
The GCA78/79 are equipped with special VDR resistors, so ac or dc, the high reverse voltage is always suppressed.
But the same issue stands, the VDR will be separated from the coil by the internal switch at the moment it is needed.
Another matter is the fact that this built-in switch of a turnout coil drive makes the start of movement of the turnout sincerely more difficult,
and unfortunately that is exact the moment that the supplied power causes the lowest energie due to way it is constructed.
So for both situations as described above, it is most important to remove the switch, or at least short circuit the contacts.
Specially for DCC or Motorola turnout decoders (except for Digital-Bahns 'Weich-Ei'), this is unacceptable.
Sometimes the switch off command is missing, and the coil will burn out.
There might be many more of turnout coil drives, provided with the vdr, of which I am not aware of, or have no knowledge of it.
Most certain is that diodes will not be used, or the specification for that particular drive is that it is to be used for dc only.
It is mostly not a mentioned detail in user manuals so you will have to check that by yourself.

GCA79 has its own timing, so it is best to use the feed-back to Rocrail.
Then Rocrail will wait on this feed-back before the turnout is given free to pass.
To be sure that a rather small power supply ac or dc is needed for the turnout coils, the GCA79 can only activate 1 turnout at the same time.

It is also possible to use manual switches with the GCA79.
Each turnout needs only 1 double task switch , switching any of the four inputs to either 5V or 0V.
To safe one wire: connect a resistor (10 KOhm) between +5V and the input, and let the switch connect input to zero or left open.
Also there, the timing will be very convenient.

New solution fo DCC / Motorola systems

The use would be very interesting for many railroaders with problems of unreliable functioning
turnout coil drives, The solution for that is coming up. For info see GCA80.

Dip switch settings for timing

dip switch 1 2
0.1 sec OFF OFF
0.25 sec ON OFF
0.5 sec OFF ON
2 sec ON ON

I/O declaration

I/O1 pt value turnout # straight turnout # round turnout # feed-back
when turnout is set
level matches input
1 +5V
2 0V
3 1 1 -
3 0 - 1
4 1 2 -
4 0 - 2
5 1 3 -
5 0 - 3
6 1 4 -
6 0 - 4
7 1/0 1
8 1/0 2
9 1/0 3
10 1/0 4

power supply

The necessary 5V for this board comes via I/O1 connector.
Power for the turnout coils: use standard 16V ac transformer.


The GCA79 is developed with KICAD
Schematic drawing
PCB with parts postions
Bill of materials


The very simple program for the PIC processor is created with PicSimulator IDE from Oshonsoft.
the PicSimulator files, including the *.hex file.
gca79-en.txt · Last modified: 2023/03/21 16:07 by phg