|
Post by Deleted on Nov 6, 2018 22:46:00 GMT
loving your work Don...especially your use of the rotary table....great stuff sir..
Pete
|
|
don9f
Statesman
Les Warnett 9F, Martin Evans “Jinty”, a part built “Austin 7” and now a part built Springbok B1.
Posts: 960
|
Post by don9f on Nov 11, 2018 18:26:33 GMT
Hello again, when I bought the spectacle plate castings I enquired about the availability of "Spectacle Glass", as this is listed in the Steam Workshop catalogue containing all the ex Doug Hewson items. Unfortunately it appears that this is not actually available so I had to look further afield. Having read about O & OO gauge signal construction on the rmweb forum I learned of a supplier that offered a pack of the four relevant colours....red, yellow, blue & green (remember its green for those using white lamps). This was advertised as coloured "film", so I knew it would be rather thin and that it wouldn't be stiff enough used just on its own, so as the spectacle apertures were plenty deep enough, I also ordered some 0.5mm clear acrylic sheet that could sandwich the coloured film between two bits. The appropriate oval and circular shapes of the "sandwiches" were laboriously cut out with a scalpel until everything fitted satisfactorily and after painting the spectacle plates, these were fitted into their recesses from the back and held in place by dabs of araldite on little areas of the brass aperture scraped clean of paint....we'll see how long this lasts! The arms were painted in their relevant colours, masking off as required. They were primed with Halfords Upol Acid 8, then I quickly learned what I should have known already....don't spray Halfords Enamel Black on top of Precision Paints Yellow.... Disaster! I was using what I had to hand and this happened whilst applying the Black to the front of the Distant Arm which was difficult enough to mask anyway because of its "Fishtail" but I had to clean it all off and start again. This time I applied the different paints the other way round and it worked ok! (Did I mention I don't like painting?) The bits were then all assembled with 12ba screws & nuts and are shown in this next photo along with the left over coloured film (you don't get much but it was only £2.50!), plus the Backlight Blinders talked about previously. The screws etc. were subsequently touched in to match. Since finishing the arms as shown above, I've been working on the signal's baseplate....arranging the servo mountings and operating linkages and so on. One of the two constructors I've been following on rmweb refers to this stage as "Installing the Worky Bits" and I've really enjoyed this part as its akin to setting up Model Aircraft/Helicopters etc. which I've done a lot of over the last 40+ years. So next time I'll show how I connected things up so the servos can operate the Balance Levers. After that will come (almost) final assembly and connection of the Balance Levers to the Semaphore Arms ....but then there's still that long ladder to do! cheers Don
|
|
JonL
Elder Statesman
WWSME (Wiltshire)
Posts: 2,906
|
Post by JonL on Nov 11, 2018 19:37:54 GMT
Without anything to demonstrate scale I'd have had no idea these were not full size. Excellent work.
|
|
don9f
Statesman
Les Warnett 9F, Martin Evans “Jinty”, a part built “Austin 7” and now a part built Springbok B1.
Posts: 960
|
Post by don9f on Nov 16, 2018 20:39:48 GMT
Hello again, thanks for the comments and this time I'm going to explain a bit about the servo system that will operate the three different arms on this signal. We do have a couple of semaphore signals at our railway and these quite old and are operated by geared motors driving a cam arrangement with limit switches and so on to stop the motor in the required position for the signal to be "On" or "Off". I chose to use servos instead, as I have loads of these going spare, from dismantled / crashed / otherwise unwanted model aircraft. They do of course require a pulse generator circuit to drive each one but I have already built and tested this as mentioned much earlier in the thread. It is quite simple to do, using easily available timer integrated circuits and Veroboard etc. A typical radio controlled model servo will move through its normal operating range in response to a continuous stream of pulses that vary between 1 and 2 milliseconds in length, repeated roughly every 20 milliseconds. There are other systems nowadays but the way I have tried to describe it is how its been done for many years. They usually need only a nominal 5 volts. Full size semaphore signals are usually operated by wires "pulled" from a Signalbox. A wire would normally run horizontally to the base of the signal where it attaches to one end of a 90 degree Crank, with a short vertical wire then connecting up to the Balance Lever. A further wire then runs up from the Balance Lever and either operates the Arm directly, or in the case of Bracket Signals for instance, indirectly via other Cranks or Rocker Shafts etc. Because the servos are mounted below the baseplate, I didn't need or fit any Cranks at the base of the post. I built a mock up of the lowermost part of the post with the Balance Levers mounted at approximately the right height, so that I could try different servo positions and importantly, establish the correct "throws" to operate the Levers through their correct range of movement. The photo shows what I mean but was of an earlier arrangement than the final one:- For the purpose of this exercise, I was using one of my radio control transmitter / receiver systems that enabled me to control each servo by seperate switches and adjust what are called the "endpoints" of each servo and arrange for them to move the Balance Levers without overdriving them and causing jamming of the linkages. The operating wires are 1.2mm Piano Wire which is flexible enough to provide some "give" in the system in case for when things go wrong and overdriving occurs! There has to be a certain "departure" from the way a full size Slotted Distant is really controlled with this model but thats not important:- After much tinkering, I decided to stack all three servos vertically and then made the proper steel baseplate. Below this plate will be a 4 x 4 x 8 inch box that at the track, will sit below ground level in a concrete pocket. The Signals at our Railway are put out and their electrics plugged in on each running day....they are not left out permanently. The Futaba 3001 servos are pretty old but are very reliable and work fine. The big nut secures the lowermost section of the main signal post, which above the baseplate is 1 inch diameter. This represents the approx. 1 foot diameter tube that later LMS/BR type signals were concreted into. More soon hopefully.... Cheers Don
|
|
don9f
Statesman
Les Warnett 9F, Martin Evans “Jinty”, a part built “Austin 7” and now a part built Springbok B1.
Posts: 960
|
Post by don9f on Nov 20, 2018 21:47:57 GMT
Hi, now that the servo system is all sorted out, it’s time to start the final assembly of the working parts of the signal. Photos of full size upper quadrant signals show a sort of “limiter” arrangement that prevents the semaphore arm from exceeding about 55 degrees above horizontal. I have read somewhere about over-enthusiastic “pulls” from the Signal Box causing the arm to over-travel ! I’m not sure about this but I think there is also some sort of “stop”, to prevent the arm drooping below horizontal when at Danger / Caution. One day I might seek out a signal lying on the ground so I can have a proper look....don’t fancy climbing up one these days ! I made up three adjustable “limiters” with springs (don’t know the proper terminology here) and brackets to go with them that picked up on one of the lamp bracket securing bolts. The photos I’ve seen show these things attached to the crank arm on the backblinders. All the working parts were fully assembled, including the rocker shaft for the miniature arm. The lamps were fitted and the wires for the leds soldered on and the joints covered with a little piece of heat shrink tubing:- The extra bracket seen at the top of the post in the photo above is where the ladder will attach. For the connection between the Balance Levers and the actual arms, I used 1mm piano wire, as I think a scale version of the full size multi-strand wire would be a bit on the delicate side! The Doug Hewson main Spectacle Plate castings have no offset to their crank so I devised a means of attaching the 1mm wires that could allow them to run down the outside face of the plate. They have to run this way in order to clear the stage angles on the way down. I made up a little bending tool to create a small loop at the top end of each wire....not really like the real thing but it works:- The lower end of these wires connect to the Balance Levers and to provide a bit of adjustment, they utilise some threaded clevises from RC aircraft days. These were reduced in length, reprofiled and cross drilled for the 1.2mm pin that connects them to their levers. Short pieces of M2 threaded rod with one end flattened in the vice, then drilled for the 1mm wires screw into the clevises. The wires had a loop formed on the lower end that then hook onto the rods:- Here then is where I’m up to so far....the wires from the Balance Levers to the servos have to be finished off but at last the signal comes to life ! At the moment I am using 3.3kohm resistors for each led, may have to experiment here to get the best effect outdoors. Also to do is the wooden platform for the landing. I haven’t got any video yet and I’m not sure how I could post it to the forum....something to think about over the next few days:- The wiring of the relays that will operate the signal will interlock the movement of the Distant signal, simulating the “Slotting” arrangement, to ensure it can only Clear, when the Stop signal above it is also Cleared....and that it returns to Caution, if the Stop signal returns to Danger. Well I suppose I’d better get on with the ladder....an estimated 42 rungs !!! Cheers for now Don
|
|
|
Post by Deleted on Nov 20, 2018 23:02:10 GMT
What a wonderful piece of miniature railway signaling you have created Don... an inspiration sir....thanks for sharing..
Pete
|
|
|
Post by Roger on Nov 20, 2018 23:40:04 GMT
Absolutely stunning Don, what a wonderful project!
|
|
stevep
Elder Statesman
Posts: 1,070
|
Post by stevep on Nov 21, 2018 9:53:21 GMT
I agree - stunning!
|
|
|
Post by Rob on Nov 21, 2018 20:03:06 GMT
I haven’t got any video yet and I’m not sure how I could post it to the forum....something to think about over the next few days:- I echo the previous comments, the signal is fantastic. Regarding the video, the easiest would be to upload to YouTube. I believe the forum has the ability to embed YouTube videos, and even if that is no longer working, a YouTube link would do.
|
|
don9f
Statesman
Les Warnett 9F, Martin Evans “Jinty”, a part built “Austin 7” and now a part built Springbok B1.
Posts: 960
|
Post by don9f on Nov 21, 2018 20:45:27 GMT
Hi everyone, thanks very much for all the “Likes” and the comments and to Rob re the youtube tip etc. I’m quite chuffed with it myself....it’s taken a lot longer than I thought it would, but it’s been very enjoyable and as people have said, a bit different. I haven’t kept a tally of the costs but the cast parts from SW aren’t cheap and at this time, I don’t plan to build another one.....besides there’s a 9F to finish in the workshop somewhere!
It will be Monday now before I can cut more metal but if the ladder goes ok, I should have it finished by the end of the month, ready for testing up at the track.
Our running season is over now for this year, next run will be a member’s day on New Year’s day, weather permitting.
More photos next week....
Cheers Don
|
|
JonL
Elder Statesman
WWSME (Wiltshire)
Posts: 2,906
|
Post by JonL on Nov 21, 2018 21:25:28 GMT
When you do your next photo could we have something next to it for scale please? I've been struggling to visualise it.
Fantastic work.
|
|
don9f
Statesman
Les Warnett 9F, Martin Evans “Jinty”, a part built “Austin 7” and now a part built Springbok B1.
Posts: 960
|
Post by don9f on Nov 21, 2018 22:20:30 GMT
When you do your next photo could we have something next to it for scale please? I've been struggling to visualise it. Fantastic work. Yes, will do.... Don
|
|
|
Post by terrier060 on Nov 23, 2018 20:23:53 GMT
Fantastic work Don. Now you have to spoil it by weathering it!!!
|
|
don9f
Statesman
Les Warnett 9F, Martin Evans “Jinty”, a part built “Austin 7” and now a part built Springbok B1.
Posts: 960
|
Post by don9f on Dec 15, 2018 17:31:19 GMT
Electrics.... Hi, as mentioned before, I wanted to use old model aircraft servos to operate this semaphore signal. We have two other semaphores in use on our miniature railway that are quite old and they both use motor/gearbox combinations, along with cams, microswitches etc. and I just thought servos would be easier and I have loads of them. Acknowledging all the good advice about using micro-controllers and so on, I wouldn’t know how to go about it from scratch, so I built this little timer board that uses two 556 timer ic’s each containing two timers, four in total. The standard servos I am using can move through a total of approx 90 degrees and the two end points, or any position in between, is determined by the length of a “train” of pulses that is sent normally by a radio receiver to each servo in an aircraft, but in the case of this signal, by my timer board. Each servo “expects” to see a pulse of between 1milli-second to 2 milliseconds long, repeated approx every 20 milli-seconds. These particular servos are therefore analogue controlled by “time”, many thousands are still in use for model applications but newer technology uses digital control....but the end result is the same. One of the four timers “triggers” the other three every 20 milli-seconds, to send their pulse to their particular servo. These three timers are each set to send one pulse every time they are triggered and if we consider just one of the three timers, the length of its pulse at any particular time, depends on the value of a certain resistance in its circuit....a resistance which can be varied by a relay contact, shorting some of it out. Thus by energising/de-energising it’s control relay, the servo moves from one end point to the other. As the servos are quite old, their speed is not great and I didn’t need to fit any sort of “servo slow” circuitry and the movement of the semaphore arms looks just right! The blue items with screw adjusters are variable resistors that form part of the “resistance” mentioned above and are necessary to trim the actual end points the servo moves between. These obviously have to correspond (via the linkages) to the signal being “On” or “Off” in each case. The connection points are for the servos themselves, power supplies and links to the three relays etc. As servos can be quite “noisy” (in electrical terms) I found it best to feed them with their own independent 5v dc supply, using a standard switching power supply device. This is to help keep the timing side of things nice and stable. The actual timers are fed from a 7805 regulator circuit, which also feeds the illumination LEDs. Again, blue things to adjust the LED brightness. The power for all of this comes from the nominal 12v dc that runs to each signal around the track. Here is the relay section wired up ready to fit, one relay for each of the three semaphore arms. These operate at 12v and interface with the controls and interlocking that is located in the signal box. All these parts fit under the signal baseplate and will eventually all be boxed in and the “box” fits into a concrete pocket where the signal is positioned where it then plugs into the system. Next time I will show how it all fits together and with any luck, show a short video of the signal working..... Cheers Don
|
|
|
Post by Roger on Dec 15, 2018 23:53:35 GMT
Hi Don, That's a very neatly made proper 'old school' solid design. Much as I like microcontrollers, it will be a lot easier for future generations to understand and maintain this sort of solution if it were ever to start playing up.
Out of interest, a microcontroller solution would probably use optical isolation instead of relays, software to create the pulse trains, and perhaps a couple of buttons to enable you to set the end positions so they could be retained in the internal Flash memory. None of this is difficult, but you need to buy a debugger for say £100, and obsolescence is always going to be a problem. That sort of solution is fine for a commercial product with a relatively short life, say ten years or less, but becomes increasingly problematical further down the line. Hopefully someone will still be enjoying your superb signals in 50 years or more.
|
|
|
Post by delaplume on Dec 16, 2018 0:42:58 GMT
WOW !!.......
I'm very cautiously going to make a suggestion because I am in complete AWE of your workmanship.........
The Blue lens were used because the old oil lamps originally used had a flame which was Yellow in colour......In your photos the "Green" looks too much like Blue..............If I've missed that in your write-up then please accept my apologies.....
|
|
don9f
Statesman
Les Warnett 9F, Martin Evans “Jinty”, a part built “Austin 7” and now a part built Springbok B1.
Posts: 960
|
Post by don9f on Dec 16, 2018 17:01:59 GMT
Thanks Roger & Alan for your comments and yes I would have liked to use a micro controller (use several in my flying models) but especially now I don’t need to “slow down” the servo movement, I’m happy with what I’ve done.
I agree the spectacles look a bit too blue....it’s because the “warm white” LEDs I've used aren’t yellow enough but I’m not going to change them now. The optional green film supplied in the pack I bought (for “white” light) was in my opinion, too green so I went with the blue. Maybe it’s the photos but I don’t think they look bad really. They were taken in a room illuminated by LED light, so maybe that’s distorted the colours a bit? I might try one taken in the dark..... In reality it doesn’t matter because we only ever run in daylight....I could probably turn the lamps off and no one at the track would notice!
Cheers Don
|
|