Jump to content
 

Deev

Members
  • Posts

    199
  • Joined

  • Last visited

Everything posted by Deev

  1. Hi Gordon, Yes, a good point, well made. You are quite right, there is nothing specific for traversers in the DCC protocol. What I had anticipated is using turnout ‘thrown’ addresses as the position ‘call’ for the Arduino with turnout ‘closed’ being the corresponding deactivating function (called simultaneously when the next ‘thrown’ command comes). For example point ID 101 thrown is lane 1, ID 107 thrown is lane 7 etc. Do you reckon that this is a viable solution? I had hoped to cover over this messy workaround by using JMRI or RocRail on a PC / tablet. What I had planned is that on startup, the Arduino drives itself to a known position (ie. position 1) which is hard-stopped with a microswitch and then remembers that position. When a ‘thrown’ command is received, say for lane 7, then the code will determine that it needs to move +6 spaces, then log that the traverser is in position 7. If then a command is received for lane 4, then the code will determine that it needs to move -3 places and log it is in position 4. This way gives the added benefit that the traverse can accelerate and decelerate differently if it is moving 1 or 8 spaces. Position confirmation feedback (for the Arduino) is achieved via the servos driving microswitches connected to two of the Analog inputs to the Arduino. As there are spare analog inputs on the Arduino, I had hoped to use a ‘position set’ rotary switch and ‘move’ push button as a manual traverser operating device. I’m interested in your traverser mechanism, I’m surprised that the mechanism from a scanner is powerful enough to move much at all! Can you post a link if you've got one? Dave
  2. I've spent some time trying to decipher the Alex Shepherd Arduino sample code, and can't do it. Instead, I've found this which looks far easier to understand and more attune to what I'm trying to do: http://www.mynabay.c...ino-dcc-monitor The DCC Decoder Library (halfway down the page) shows the data files that are needed. I've uploaded the sample code to my UNO and it does actually compare. Result! My Opto-isolators showed up today, so hopefully I'll get these soldered in to a board and get going with this in the next few days. Dave
  3. The steppers are so accurate that I don't reckon you really need any sort of positional feedback, but it all depends on how accurate your drive mechanism is I suppose. It can all be fine tuned via code anyway. The 12v stepper motors that I'm experimenting with are these: http://proto-pic.co.uk/stepper-motor-with-cable/ with the twin stepper motor shield from Adafruit (I got mine built up for a fiver!). I plan to have 2 x drive mechanisms slaved together, one at each end of the traverser.
  4. Will, Looks like you've done the hard bit, i.e. getting the traverser built and running smoothly. Fitting a stepper motor to your traverser, in theory, should be straightforward. Maybe have a look at http://www.motionco.co.uk at their range of timing pulleys and open belts. I reckon you'll need a hefty stepper to move your traverser. Either than or lots of gearing! Wiring and programming an Arduino to automate it with push-buttons and programming is simple enough. Even with simple scripting those steppers are deadly accurate, I left mine running all night and it was still moving back and forth to the same locations. Clearly though, the key is to minimise backlash in the drive mechanism. I reckon you'd be best off with an Arduino MEGA (http://arduino.cc/en...noBoardMega2560) rather than an UNO like I have, as these have 54 (oh yes) digital I/O ports rather than the 20 that the UNO has (most of which are taken by the motor shield anyway. You can even get LCD screens to plug it as well, if you want to be really flash. Getting it going via DCC is a challenge at the minute, but if you have a spare accessory decoder and a few Opto-Isolators this shouldn't prove to be a problem. I'm amazed that the model railway community hasn't cottoned-on to the Arduino before. By making it DCC compatible and with all that I/O potential it is possible to get a 50-servo driver for 20-25 quid, or a mere 20-servo driver with sound, 20 block feedback, etc., etc. Feel free to post any thoughts, ideas or progress. Dave
  5. The Arduino will act like a DCC accessory decoder, it will just 'listen' for the relevant packets then move the steppers accordingly. It feels a more elegant solution than using a separate DCC decoder. Helpfully there seems to already be a library that does this already, courtesy of MRRWA!
  6. After trying to work out how to control the Arduino via an accessory decoder, it occurred to be that a more effective and versatile way would be to control the Arduino itself via the DCC signal via Pin 2. I've done some length web-research on this and the following links I found to be useful (put here for my own reference really): http://fritzing.org/...a-dcc-decoder/ http://groups.yahoo....s/message/28542 http://mrrwa.org/ http://groups.yahoo....s/message/28542 http://www.oscale.net/en/simpledcc And here are the Arduino libraries (just in case I lose them): MRRwA-2012-06-15.zip Has anyone seen anything similar to this anywhere?
  7. Thanks, there's always enough room for a traverser! The servos can also actuate a switch so that when they are in the 'unlocked' position, power is removed to the traverser table and the traverser is feedback to the DCC station as 'occupied' to prevent locos driving off the edge.
  8. I've spent the last few weeks doing some serious 'to-ing and fro-ing' on what sort of control and drive system I need for my traverser. For the support of the traverser carriage, I've got hold of some lengths of 6mm rod and some corresponding linear bearings. Using four of these and supporting these at both ends and at the mid-point should give me an accurate, smooth and low-resistance movement. The drive system has been a bit more involved, but I've decided on stepper motors having seen them used on similar projects. As my traverser is long, I reckon the best way of driving it is to use 2 steppers that are slaved together so that I don't have to fit a long driveshaft (with its inherent twisting and support issues). To control it I've bought an Arduino Uno board and an Adafruit Motor shield. This allows me to drive 2 x stepper motors as well as 2 x servo motors whilst still having enough spare pins for control from an accessory decoder. I plan to use the servos to act as an alignment lock to help the tracks align, but also as a method of stopping the table moving when the steppers are depowered. the whole setup looks like this: As for the movement drive, I've bought some really cheap MOD1 rack and RC car pinions to see how suitable they are. With all the kit now bought I've been keen to experiment with it. I've put together a simple script that cycles the servo, then moves the rack before cycling the servo again, moving the rack, then cycling the servo before returning the rack to the beginning. I've also coded a 'soft-start' to make things a bit smoother. I'm really surprsied as to how easy it was and reckon it's been a success. All I need to do now is put together the traverser itself! http://youtu.be/TsfmOmQIDxI
  9. A tube, yes of course, that's another problem solved and yes, mounting it upside down would be best for a traverser. I wonder how much backlash there would be with a tensioned direct belt drive? I've spent a bit of time looking at CNC machines and a few seem to be set up like the diagram above, (see below). If anyone was going to suffer with backlash it'd be the CNC people. http://www.youtube.com/watch?v=9QgnD-bgd_4 I don't have anything against using a chain, but belt and belt parts seem to be more plentiful off the shelf without serious modification and it's possible to use a much smaller spur gear with a belt than a chain to give slower speeds without necessitating any step down gearing. As for the points, I don't like solenoids and prefer electrofrog, to satisfy the obsessive conpulsive in me. I plan to use USA-sourced Tam Valley Depot servo drivers for the points on the layout with solid state frog switching which I've procured already. Check out the video. It's all very smooth. http://www.youtube.com/watch?v=AYeHyxWWnwI
  10. Hi Beeman, The main reasons for wanting to construct a traverser are thus: 1) I wish to model the station at Filton Abbeywood, which has SW trains crossing the 'midland' mainline. This would necessitate an extremely complex fiddleyard entrance. 2) I'd like to run as many prototypical length trains as possible, the size of the fiddleyard would occupy space that I don't have. 3) The cost of points / motors / DCC accessory decoders for a 9-lane fiddleyard would be huge, I reckon even with a Arduino stepper motor control, the cost of a traverser should be much less. I've come up with a similar solution to the one you suggested, the same in fact, but using timed belts, arranged at both ends of the table, with support in the middle. Instead of using a continuous belt, I notice from the CNC forums that the below is a preferred method: A worry here is whether a 2900mm shaft will twist and knock the alignment out.
  11. Hi Beeman, Thanks for the informative post - Having read your comments regarding backlash, I see this as being less of a problem on a traverser than a TT. If a worm drive is used at both ends, couldn't a light spring me used to keep the inevitable backlash to one side (like gravity does on a VSU?). Your TT positioning system is novel, and well thought through. Fair play to anyone who can get a VSU to work at all, let alone correctly! One question those, I don't understand how that TT alignment 'key' system works, is the 'key' on a solenoid to line everything up perfectly? I hadn't appreciated just how basic the control of your TT was until I reread the post properly, it really is impressive and something that I'd overlooked at first glance. How close can you get the opto-switch, to 0.5mm say, then take up the slack with the keying device? The problem that I have in creating something similar is that I don't have a massive shed full of old VCRs and microwaves like you seem to have, where do you get all those scrapped yet still useful things from?!!!!???!! Also, what sort of force would be needed to move a loaded 3000mm 9-road traverser on ball bearing runners and would a 5v motor (or two) from a VCR be man-enough to move it at a sensible speed? Dave
  12. Thanks for your comments, all good stuff! I spent my lunchtime researching microcontrollers and it seems like a viable solution would be to use an Arduino controller (http://www.arduino.cc/) with a twin stepper motor shield. This way, the inputs to the microcontroller could be from an accessory decoder (ie. a spare Lenz LS150 that I have) and the outputs could be 2 x stepper motors. Doing it this way would ensure proper alignment at both ends of the traverser, no 'shimmying' to worry about with the table and plenty of 'oomph' to move the table with two motors.
  13. I reckon the key is to come up with a DCC-controlled stepper motor or using JMRI / RocRail to control a stepper motor via USB. I've not seen this done, does anyone have any pointers?
  14. A DCC screwdriver would be handy, but a decoder would be expensive. Even a tame 12v motor would pull 2A! How about using a point decoder to drive 2 x relays to switch power?
  15. I've had a look at this too, and after several hours of 'Google'ing, it looks as though the best way would be to use a dedicated microprocessor and a stepper motor. This can be programmed to move the same amount forth and back every time. The simplest type of drive seems to be a length of M8 rod with a bolt that is connected via a coupling to the motor. The bolt is glued welded to the traverser bed and moves when the rod is rotated. This also has the advantage of locking the bed in place such that it can't move and would give good alignment (better than Lego anyway). Something like this: If you're using microswitches to set the limits and only have two positions, then a simple geared DC motor could be used (e.g. http://www.active-robots.com/motors-wheels/dc-motors). Obviously the motor's polarity is reversed to move between positions. Sounds like a nice interesting project - good luck!
  16. Hi guys, Has anyone ever made one of these monsters? I wish I still had my notes on how to programme the PIC microprocessor! http://www.youtube.com/watch?v=dxRJ9rdyX4c&list=LLPUwc5qPqSpXkhigQXqAHoQ&index=4&feature=plpp_video I want one!! Dave
  17. Cor blimey - those baseboards are a work of art!
  18. Hi Andi, This layout is amazing - aspirational. I notice that you've attached the track straight to the baseboard. Did you glue it and if so is it noisy or have you have any probs? Thanks Dave
×
×
  • Create New...