I will leave that for others to answer, because I no longer have a layout. But when I did the short-detecting-type-auto-reversers were a big pain in the rear. Now I am primarily interested in seeing the hobby make use of the available technology and in attracting new and younger hobbyists. Bob
Erik, take a look at this: http://downloads.cdn.re-in.de/22500...-LDT_BST_KEHRSCHLEIFENMOD_KSM_SG__MM_DCC_.pdf I think the LDT module does what your proposing, which is to detect the presence of a train before causing a short.
I know a day late and dollar short. Just my two cents and I take change. The best place to start any research is with the manufacturers of said products. Most of the information out there is tainted by misunderstanding as to how something should work or how to wire it in. Finding the right information can be like looking for a needle in a hay stack. Sigh! It's not all bleak. There are videos produced by knowledgeable people on the internet and by professionals in the world of electronics. Mike Fifer, David E. aka Sharkman. These guys can help you sort out the bullpucky. DCC Guy, although not a electronics specialist relies on professionals to help him establish comfortable how-to's. Lot's of help available once you know where to look. Glad to see you solved your own problem. Wishing you the best with it.
Hi lyncher, thank you for that link! Very interesting. Together with BarstowRick you give me a couple very usefull hints. The Littfinski company is a reputable company and what they say, I listen to. Their setup uses current detection at each end, which is really commendable. After TR_Coker's suggestion of a few days ago I did decide to change my prototype intention from 3 to 4 sensors. That way I will be able to have a train reverse too once inside the loop. On top of that it will avoid any risks of short-circuits by have the two "outside" sensors DET-a and DET-b too close or too far away from the isolation. Why do I not opt for current sensing, as a reputable company like LDT uses? Well, because I cannot guarantee yet that the current sensing that I would use would be "foolproof". David Bodnar has some very interesting information and tests on this type of sensor and I am still in the process of testing these. Meanwhile I also did some tests with LDR sensors, but then in a comparative set-up: I use two LDR's per sensor, one for ambient light measurement and on for actual under-the-track measurements. One setup with a 555 timer: http://www.members.optusnet.com.au/satr/LDRdetect.htm Alternative setup with 555 timer, but here I replace R1 with an LDR: (look at the chapter "Using the 555 timer with a photocell") http://home.cogeco.ca/~rpaisley4/xComp555photo.html One setup with a voltage comparator: http://www.qtutrains.com/LDR.html I had astonishing results with this design!! The setup let me detect trains both with dimmed lights as with full lighting, without having to calibrate any resistor. So, is short: two detectors on, the one side (one before and one after the isolation, and the same on the other side of the loop. And this setup will allow for parking tracks within the loop, multiple simultaneous trains inside the loop etc.. And if needed, a fifth detector inside the loop serving as an occupancy detector. Using LDR's these detectors cost me less then a dollar a piece; I collect all parts from Ali-Express. For example, the 555 I get for 10 cents a piece, the LDR's go 50 pieces for a dollar, etc.. All shipping included. I will post pictures, electrical diagrams and any software and/or programming used when ready.
Hi lyncher, now after several months I did some rereading of these posts and I must admit not having seen the point of the Littfinski concept: the use of current detectors allows something that cannot be done with optical detectors, namely detect a train fully inside the loop but which reverses direction. With their current detection setup this can be done with 3 current detectors. With optical detectors, 5 are needed: two on each side of each of both gaps; and a 5th in the middle.
Hi Erik, Thanks for your solution on reverse loop! can you also do diagram of all the hardware connections with board figures not the schematic. ones with all the wiring with attiny85 or arduino? right now its not clear which Relay to use as there are two connection getting switched. also can you implement H-Bridge solution instead of relay? like using motor driver L398 or L293d? that way drawing too much current on controller side can be solved.
while the use of a mechanical relay does work, it is about the slowest option there is .. if you can, go with an electronic option to switch the power on the rails [ darlington pairs or whatever], that way you can add user adjustable pots for thing like reaction time, switching current and the like ..
The wiring of the ATtiny85 is written up in my initial post. I will make a proper drawing, but not right away, I have too much on my to-do list now. The relay needs to be a quadruple pole double throw relay, or 2 double pole double throw relays. Since the ATtiny can sink or source max 20mA an amplifier is required to drive a relay: a simple N-channel mosfet will do. Gate connected to controller output, source to ground and drain to relay coil, with the other side of the relay coil to +12V. For wvgca: using mosfet's as switchers is a good idea, except in this case you will need so many that the effort is not worth it in my opinion. And the difference between 0.2usec versus 50ms switching delay is, in model railroading terms, the equivalent between a space rocket and a NASCAR racer. Meaning: relays are more than adequate, cheap, easy to implement and virtually foolproof for the purpose here in my opinion.
