3 aspect model RR signalling help plse

angie1199 said:

Seems like a lot of components compared to my 8, 4 diodes, 2 transistors, 1 resistor and 1 optocoupler.

Now I need to work out the connection of the optocoupler instead of the led and then the pcb for a detector only circuit with 2 connections one side (dcc power) and 3 the other (12V, Gnd, Detect)

Click to expand...

Ok angie,

Do yourself a favor here. Leave your PC board as is. This is a prototype and you should leave a little wiggle room for enhancement and changes. Once an engine runs for half an hour and the LED stays on then add the Opto-isolator. (Try leavingthe LED in the circuit. It makes a good visual indicator of the "detect" condition.) One question. Where are you going to locate the opto isolator. On the detector board or the signal controller board?

I added the diodes to the circuit, changed PNP to NPN and inserted signal transistors instead of power ones.

P

P.

angie1199 said:

Seems like a lot of components compared to my 8, 4 diodes, 2 transistors, 1 resistor and 1 optocoupler.

Now I need to work out the connection of the optocoupler instead of the led and then the pcb for a detector only circuit with 2 connections one side (dcc power) and 3 the other (12V, Gnd, Detect)

Click to expand...

Hi again,

Connectioin correction. You will need 3 on the DCC/track side (remember the tie point. You wil need it) and you will need 2 on the output, Ground and detect from the resistor and possibly the LED. The 12 volts will be connected to one end of the optocoupler (which I assume will be on the signal controler. The other side of the optocoupler will be connected to the detect ouput from the resistor and LED on the detctor board.

Er Um, 8 components?? For the final production unit I make it 2 power diodes, 1 capacitor, 2 transistors, 1 resistor and 1 LED. That's 7.

For your prototype unit where you are determining optimum values you may need a couple more resistors. (1K, 2K, 3.9K and a 4.7K) You can put all three in series with the optocoupler and see if it works. The best way is to get a 5K variable, put a 500 Ohm (or so) in series with it and put them series with the optocoupler and LED. Then crank it up wards until the optocoupler doesnt work any more. Unhook the pot and resistor and measure their combined resistance. Get the closest value resistance below that value. (if the resistance of pot and its resisistor is 5K then get a 4.7K)

Play a bit. Once the optimum values are determined you are down to 6 components depending if you need a capacitor or not.

Finally, the issue about four power diodes. At first we were trying a circuit that used a bridge rectifier. A bridge rectifier in a single capsule with two wires AC going in and two wires DC coming out (shorted) uses 4 power diodes. There is nothing that can be done about that. Using 4 discrete diodes instead of a bridge is a bit silly when you really only need two. However if you really want to use 4 diodes go ahead. But while the engine on your test track is running feel the transistors and see if they get overly warm. The little tykes don't like that. They will work for a while but in the future they might just give up. You see anything over about 0.8 volts will push a little extra current through the NP base/emitr junctions of the little tykes. They'll work but they don't like that too much.

Build the circuit with 4 diodes by all means and while an engine is running feel if the transistors get warm. If they do then try the two diode version.

In anycase when you have a working prototype I suggest you run an engine with it on your test track. Let it run for half an hour and see what gets hot. Then when all is OK, perform the following operation. Put the pot and 500 ohm resistor across the track and crank upthe resistance until you detector ceases to detect (the LED goes out). Remember that the signal transistor version is much less sensitive than the power transistor version. Measure the resistance of the pot (with its 500 ohm resistor in series). THAT is the maximum resistance allowed across your track for the detector to work. And that would have to be the resistance of every axle on your cars. Divide that resistance by the number of resistive axles on the train you intend to pull and use that value to calculate the extra current your DCC power supply will have to provide. You might be surprised.

Let me know.

P

Raiway Pete said:

Connectioin correction. You will need 3 on the DCC/track side (remember the tie point. You wil need it) and you will need 2 on the output, Ground and detect from the resistor and possibly the LED. The 12 volts will be connected to one end of the optocoupler (which I assume will be on the signal controler. The other side of the optocoupler will be connected to the detect ouput from the resistor and LED on the detctor board.

Click to expand...

Don't need the tie. Only one dcc wire is being broken, the one to the rail which is block cut. Tried it and it works.

Raiway Pete said:

Er Um, 8 components?? For the final production unit I make it 2 power diodes, 1 capacitor, 2 transistors, 1 resistor and 1 LED. That's 7.

Click to expand...

Bridge rectifiers are 3 times the price of 4 diodes. I tried the two diode option and it didn't work at all. Will try again tomorrow tho'.

Raiway Pete said:

For your prototype unit where you are determining optimum values you may need a couple more resistors. (1K, 2K, 3.9K and a 4.7K) You can put all three in series with the optocoupler and see if it works. The best way is to get a 5K variable, put a 500 Ohm (or so) in series with it and put them series with the optocoupler and LED. Then crank it up wards until the optocoupler doesnt work any more. Unhook the pot and resistor and measure their combined resistance. Get the closest value resistance below that value. (if the resistance of pot and its resisistor is 5K then get a 4.7K)

Click to expand...

Surely if it works with a finger, and a loco, then that's it working?

Raiway Pete said:

Finally, the issue about four power diodes. At first we were trying a circuit that used a bridge rectifier. A bridge rectifier in a single capsule with two wires AC going in and two wires DC coming out (shorted) uses 4 power diodes. There is nothing that can be done about that. Using 4 discrete diodes instead of a bridge is a bit silly when you really only need two.

Click to expand...

Tried 2 diodes, didn't work at all.

Raiway Pete said:

In anycase when you have a working prototype I suggest you run an engine with it on your test track. Let it run for half an hour and see what gets hot. Then when all is OK, perform the following operation. Put the pot and 500 ohm resistor across the track and crank upthe resistance until you detector ceases to detect (the LED goes out). Remember that the signal transistor version is much less sensitive than the power transistor version. Measure the resistance of the pot (with its 500 ohm resistor in series). THAT is the maximum resistance allowed across your track for the detector to work. And that would have to be the resistance of every axle on your cars. Divide that resistance by the number of resistive axles on the train you intend to pull and use that value to calculate the extra current your DCC power supply will have to provide. You might be surprised.

Click to expand...

I've added the optocoupler, but although the red signal comes on and the yellow signal comes on on the previous block, the green signals don't go off at all, or even dim.

When I press the actual switch for the red signal, the green goes out on both blocks, the red comes on and the yellow on the previous block comes on too.

are you saying that the block signals are not working now?
lets see the green won't go out at all?
is this before or after connecting the detection circuit?
I though your detecting circuit used and opticouplier to trigger the block signal circuit?
let me know what EXACTLY is happening?
test signal circuit then detection circuit separately then together.
am assuming the schematic is same as posted?

gree led not going out

ve a hunch your opti coupliers are not fully on
connect amp meter as showen on a bread board or ?
I think you mentioned that the opti couplier needs 50ma?? sounds high.
double check your specs.
could maybe be the twin t no turning the led on full(in opti)
test current through the opticouplier.

MrDEB said:

are you saying that the block signals are not working now?
lets see the green won't go out at all?
is this before or after connecting the detection circuit?
I though your detecting circuit used and opticouplier to trigger the block signal circuit?
let me know what EXACTLY is happening?
test signal circuit then detection circuit separately then together.
am assuming the schematic is same as posted?

Click to expand...

Ok, I built the 4 diode rectifed 2 transistor detector as below with a 470R resistor and led. Wired to track 15VAC DCC and 12VDC. This worked. LED lit when finger placed across track. Worked when loco placed on track, even when stationary. Exactly as it should.

https://www.electro-tech-online.com/attachments/working-jpg.36249/

On breadboard next to the detector circuit are two signalling circuits. Both linked so when one block is red the other is yellow, orbth are green.

These signalling circuits had normal push to make swtches and bot worked correctly. Each led has a 390R resistor.

I replaced the led on the detector circuit with the optoisolator (pins 1 to 2, anode to cathode) and connected the detect from the signal circuit to pin 4 (collector), then connected pin 3 (emitter) to ground.

Powered up. Green signals on. No loco on track. All good.

Put loco on track, red signal on one block lit, yellow on other block lit. Green leds did not go out. That's the problem. They should have both gone out when loco detected.

Block Detectors

Hi Guys. I'm new here. I have done several block detector designs for DC & DCC systems.I've just done a 4 channel DCC unit using current sense transformers . Completely isolated from the DCC as it's input is via an in-line transformer and has opto isolator outputs. Can I be of any help?

email me :- steve.verwey@vodamail.co.za

MrDEB said:

double check your specs.
could maybe be the twin t no turning the led on full(in opti)
test current through the opticouplier.

Click to expand...

The led side of the optocoupler is 5mA to 50mA operating range, 1.5V.

**broken link removed**

Will test current later after I wake up.

I varied the resistance across the "track Rails" and noticed that
a 10k resistor = 109ma from battery to collectors of transistors
5k = 213ma
1k = 586ma
I wonder if your pulling enough current to light the opticoupler? after you add a resistor to the led.
insert amp meter to check current draw.

MrDeb, Angie,

First thing to look at.
Accordiong to the data sheet the opto-coupler needs a min of 5 Ma to work. The absolute maximum is 50 Ma. I cases like this I like to go half way and aim for around 25 Ma. However at 10 volts (the Opto LED drops 2) that would mean a series resistance of 400 ohms. That resistance would drastically reduce the sensitivity of the Twin-T. So we increase the resistance to 1K and hope the coupler works at 10Ma. (It should, according to the data sheet).

Second case.

Establish the current required to operate the signal controllers. I assume three are involved. The local detector's for the red and the two adjacent ones for the yellows. Using the switch can you measure the current that the controllers require? Then see if the opto coupler can actually provide it. The opto-coupler acts like a resistance that doesn't quite go to zero. I seem to remember that the yellow input on the controlers go to an emitter. That implies that there are two yellows each requiring a hefty current of up to 20 Ma. That does not include the red which also requires current. I don't know if the opto-isolator can handle that. The data sheet doesn't go into it either or I'm missing something.


Third case.

Linn Wescott no doubt hit the same problem as you guys. That is why he kept the Twin-T's collector resistor value way up there around 4K. Then he used that tiny current to drive another transistor which pulled in a relay or lit a lamp (exinguished it actually). We now have low current LEDs instead of lamps and are pushing at new frontiers. So let's see what the opto-isolator can do.

angie1199 said:

Ok, I built the 4 diode rectifed 2 transistor detector as below with a 470R resistor and led. Wired to track 15VAC DCC and 12VDC. This worked. LED lit when finger placed across track. Worked when loco placed on track, even when stationary. Exactly as it should.

https://www.electro-tech-online.com/attachments/working-jpg.36249/

On breadboard next to the detector circuit are two signalling circuits. Both linked so when one block is red the other is yellow, orbth are green.

These signalling circuits had normal push to make swtches and bot worked correctly. Each led has a 390R resistor.

I replaced the led on the detector circuit with the optoisolator (pins 1 to 2, anode to cathode) and connected the detect from the signal circuit to pin 4 (collector), then connected pin 3 (emitter) to ground.

Powered up. Green signals on. No loco on track. All good.

Put loco on track, red signal on one block lit, yellow on other block lit. Green leds did not go out. That's the problem. They should have both gone out when loco detected.

Click to expand...

Angie,

If by forcing the detect to ground via a switch turns the greens off then the opto-isolator transistor, for some reason cannot go that far. Can you measure the voltage of the opto-isolator's (detect) ouput? If it's around a volt or two that may not be enough to switch the green transistors off.

thats exactly what I said back at post 284-285.
now how to remedy this. I am thinking op amp again or using a reed switch.
I recall years ago I bought some magnetic reed switches from radio shack, wrapped some wire around it and now I have a magnetic reed switch that draws very little.
worked great. Is this not basically what the current sensing transformers that Condor has in his pic?

MrDEB said:

could maybe be the twin t no turning the led on full(in opti)
test current through the opticouplier.

Click to expand...

I thought the same.

Resistance across Optocoupler Collector/Emitter

resistor on led side
1K resistor = 154R
470R resistor = 134R
390R resistor = 129R
270R resistor = 190R

angie1199 said:

I thought the same.

Resistance across Optocoupler Collector/Emitter

resistor on led side
1K resistor = 154R
470R resistor = 134R
390R resistor = 129R
270R resistor = 190R

Click to expand...

Angie,

Can you increase the green Led driver transistor's bias resistor? Perhaps that might work. Or you could get the opto-isolator to drive a cheapy reed relay that would sink the detector condition to absolute ground. OR you could try dumping the opto-isolator all together and replacing it with a reed relay.

Raiway Pete said:

Can you increase the green Led driver transistor's bias resistor?

Or you could get the opto-isolator to drive a cheapy reed relay that would sink the detector condition to absolute ground.

you could try dumping the opto-isolator all together and replacing it with a reed relay.

Click to expand...

LED's bias resistor just dims the green led, all the time, just dims slightly when loco is present.

Trying to keep away from expensive relays.

What I was thinking, using my logical brain, was to reduce the current/voltage (whatever) to the base of the transistor controlling the green led being on. If there's less current there to start with then surely there's less to beat to switch it off? Or is that too logical?

Might have a go at building the LM393 comparator based detector, I got some of those and it uses 2 power diodes as the bridge

I was playing with different ideas on getting the green base to turn off thus shutting the green led off.
decreasing the current is an option.
you could try homemade reed switches. Will look for and get back to you.

take and wrap some wire around and volia you now have a relay
Reed switch - Wikipedia, the free encyclopedia

possiable solution

found that placing two diodes to the GREEN LED base transistor might work
Also drive the "dector switch" transistor off of the emitter of second transistor.
changing the resistor to the collectors changes results. Maybe a pot but 300 ohm seems to work w/ LED in circuit.
NOTE the LED is for indication only. Don't really need it.

MrDEB said:

found that placing two diodes to the GREEN LED base transistor might work
Also drive the "dector switch" transistor off of the emitter of second transistor.
changing the resistor to the collectors changes results. Maybe a pot but 300 ohm seems to work w/ LED in circuit.
NOTE the LED is for indication only. Don't really need it.

Click to expand...

Ok, so the two diodes option didn't change anything noticeably.

I added a 1N4148 transistor as per image, driving it off the opto isolator. It has now reversed the problem.

Loco Present
Greens off on both blocks, red on yellow on

No loco present
Greens on but red and yellow only dim.

you have it connected wrong

put second 1n4148 diode and eliminate the opticoupler. add an NPN transistor, a base resistor (not really needed as current is really low.
change connection to added transistor , connect to emitter of second transistor (T2 in you diagram)
see my attached pics