Interested in a signal building group?

[Eddyg]

Dave.

PM received, thanks. The ladders will be packed up the weekend and sent on Monday probably.

Your prototype looks well. Pity about the ladder length though. I think I'll tale the surplus off the top, extending the 1/4" BSP thread will be a pain if the bottom has to be reduced in length.

Regards

Eddie

[steamysteve]

Dave (& Richard),

Second attempt at this posting!

Thanks for the parcel received this afternoon. I have had a go at making some 'U' bolts. They are all 58mm long with different thread lengths. If the picture gets successfully attached you will see that on the 12.7mm post the top 'U' bolt is threaded 10mm, the middle one 12mm and the bottom one 14mm. Obviously the amount of thread depends on the thickness of the saddles but either the 10 or 12mm would probably be OK.

Please let me know which thread length is required and how many 'U' bolts are required for each post diameter.

Thanks,

Steve.

P1040453 (800x600).jpg (148.07 KiB) Viewed 25433 times

[Dave Noble]

Nice job Steve.

It occurs to me that it may be best if I give you the finished overall size of the U bolts, rather than an estimate of the blank length, the balance lever bracket is 5mm thick, so adding the post (15mm), wire thickness (1.6mm), nut (1.5mm), and a bit of protruding thread (D/2 = 0.8mm), the overall length of the finished U bolt will need to be 24mm.

Thread length shouldn't need to be more than 5mm, a little more won't matter. We'll need 24 of these for a start please.

Another job. There is a circular recess in the bottom of the base to take a weight. My idea is to cast lead weights and to glue them in with resin or Araldite or somesuch. A tapered mould cavity machined into a piece of cast iron should release the castings. Any volunteers? I have the lead and cast iron. If anyone going to Brent's this weekend is interested, I can bring the materials with me.

[steamysteve]

Dave,

My calculation is:
Length of rod around half the post (post diameter plus 1/2 thickness of rod) + (half the post diameter + thickness of saddle + nut + a couple of threads) x 2
So for a 12.7mm Post = PI(12.7+(1.6/2))/2 + ((6.35 + 5 + 1.5 + (1.6/2)) x2) = 48.5mm
and for a 15.0mm Post = PI(15.0+(1.6/2))/2 + ((7.5 + 5 + 1.5 + (1.6/2)) x2) = 54.4mm

I will make a couple of samples of each to check this out! I haven't included an accurate 'bend allowance' calculation which might make a small difference.

You say you want 24 but of which size, or is it a mixture?

Steve.

[Dave Noble]

I would simply compare the length of the original blank with the length overall of the complete U bolt, and adjust as necessary to get to the required 24mm overall. 24 are required for the 15mm post at the moment. Thanks.

I think your first 1.6/2 should be just 1.6

[Eddyg]

Hi Dave and signalling group members,

After the discussions about the spectacle glasses there has been no notification of a resolution to the issues. Has it been sorted? In that respect I was in Hobbycraft the other day and saw some transparent glass paint made by Marabu. It is designed for glass and they claim it's dishwasher proof so it would be OK for our purposes. It can be used on acrylic as well as glass.It's not too expensive - 15ml for about £2.70.

Is it time for an update on progress of the project? I'm conscious we have less than three months to complete and commission the signalling and I would hate to find that everything comes together at the last minute when there are lots of other things going on here.

Could you also give us some idea of what we are letting ourselves in for in terms of cost, I see 3D printed patterns and castings, CNC machining, laser cutting and electronic stuff that I have no experience of buying, so it would be useful to know what might need to be set aside from the pension!

Best Wishes for Xmas and a Happy New Year to all.

Eddie

[Ben Lyons]

Hi All

Dave and I have been busy with the bits for casting ( see picture) some bit Dave is going to cast in white metal as the brass bits are a bit on the expensive side but by using the brass we should get really nice castings. The sceptical arm will be brass Dave wants a slight revision, I will have them cast in Holland.

Picture 2 shows some parts that might be machined Fabricated any offers? lamp bracket 20 off saddle 40 plus at a guess might be two sizes

I have machined all 12 PVC bases but Electric's are asking for 3off 15mm connector holes in each one will do this as soon as I can.

Ben

Brass bits 2.JPG (11.78 KiB) Viewed 25273 times

brass bits.JPG (25.32 KiB) Viewed 25273 times

[Ben Lyons]

Hi all

I forgot to say I have sent Dave masters for the 2 and 3 way lever Brackets and will send on Monday new masters for the Smaller saddle and the lamp saddle, ( There are lots of Bits this has turned into a bigger job than we thought !)
Ben

[Ben Lyons]

Hi All

Also sent to Dave are 2 sets resins for the lamp bodies and Lids will add picture later.

Ben

[Dave Noble]

First picture shows the completed spindles, first spindle brackets, and first backblinder arms. The remainder of the arms should be with us mid January.

PC230058.JPG (74.57 KiB) Viewed 25216 times

Ben has almost finished the bases, the sockets have arrived, and I have the Perspex for the bottom covers, I just need to get it to Chris for cutting.

PC230060.JPG (63.37 KiB) Viewed 25216 times

Second picture shows assembled parts. Arms need drilling and slitting, any volunteers?

Steve Birch is on with the U bolts, and I've decided how to do the lights and coloured filters, all parts are to hand, just need to sort out the lamp pattern (shown in picture). I have lots of lamp lenses. Weights still needed. Patterns for all small saddles, crank bases and lamp lid are being made, and should be ready immediately after Christmas. Lamp bracket pattern should be ready soon after. Etchers are being a bit slow, but first sample for arms and diamond sign saddles is promised for first week in January. Spectacle plates may be a problem, I'll try them in whitemetal, if that doesn't work we'll have to go to expensive lost wax brass.

No news on the electronics. Jonny, Richard?

Merry Christmas, one and all.

[Ben Lyons]

Hi Dave

Been thinking on the lamp given the small edges around the lid and thus lip is it worth thinking about making lamp body and lip as one casting.
Make the circular boss at the base a stepped peg that has the bulb in it , the lamp then fits down and onto the peg, makes changing the bulb easy just lift off the lamp and there it is.

Brass spec plates like one shown where some £25 from Holland but there are GOOD

Merry Christmas to all.

[Richard Phillips]

No news on the electronics. Jonny, Richard?

I'll explain the point I am at. Mostly trying to find time to get anything done!

So when it was originally suggested I got involved in the programming of an programmable electronics board, Johnny suggested the Ardunino Leonardo, I haven't been to previous Shilden events but it has been used in the past? Anyway, that board seems to now be superseded, and there is an equivalent board available with a similar number of output lines. but much smaller, the Genuino (alternate non-US branding) micro which I'm currently working with.

It was suggested we use RS485 for communications - a 2-wire bus, and there are a few good reasons for this - Ethernet might be nice, but there's a good "overhead" for that. USB is out because of the distances involved and hubs required etc. RS485 could be used a a standard serial communication, and we could invent our own "protocol", but I'm favouring modbus, there are libraries available that handle all the communication issues we might fall into, ie multiple messages on the bus, etc, it's a defined set of calls that we can adapt for our purposes - and in theory could use a standard modbus controller to do our control - although I can't see how we'd map functions, but that's a discussion for another day!

Problem 1 was to not consume all available Arduino input lines just to give each signal a unique address. I used a in-line DIP switch to set a binary address - we've got a sort-of limit of 31 in the modbus standard (excluding the master ie "the signalbox") which fits within the addressing capability of the 8 way DIP switch - I was thinking I can use the other two "spare" switches to have an auto-manual mode, so these signals can all be driven without needing to be controlled by modbus, so could be used at home with standard external switches. This is run through a Shift register so we use only 3 input lines - this has been tested and works fine.

Problem 2 is testing modbus - and despite some initial issues with the native USB chip of the Genuino Micro versus the more common serial emulation of the Duo, I've successfully sent control messages over modbus from a DUO to a Micro using a variable resistor to drive a servo.

I'm at the stage of marrying up the two bit of code, and trying to find time over Christmas - it's all prototyped on a breadboard, and I have a second Micro to prove it's not a fluke and I need to set up a "signalbox" controller - the DUO I have to send instructions but I'm running out of connectors!

At this stage I'm thinking one program for all signals, because to have 30 different programs might cause havoc. The only real difference between them is the number of arms, so the commands from the signal box are essentially "Signal 7 arm 1 ON". We'll have to rely on the signalbox to have the smarts to control the locking situation and probably have to override the mechanical slotting on the signal posts as the boards will be fairly dumb to the mechanical situation, especially if the programs aren't customised. It will reduce the complexity.

I'm using a fairly widely available prefabricated RS285 board, but we could use a separate MAX 485 chip with some external components. Because there are some external components to make this setup work and the power coming in to drive all this lot the Micro will need to be mounted on a PCB.

Approximate parts list per signal:

3x 9 pin(?) DIN connectors (in - out- through)
1x ardunio/genuno Micro
1x servo (per arm)
1x 3 pin connector (per arm for servo connections)
1x bulb (per arm signal lamp) + resistor
1x Custom Printed PCB
1x 9 way DIL DIP switch
1x CD4021B Shift Register
8x 10K resistors
1x 24V Voltage regulator - may require additional caps/resistors
1x MAX 485 chup/prefab board

[Dave Noble]

Thanks for the detailed explanation Richard. A couple of points.

You say we only need 3 input lines. I hate to disrupt things, but we were assured that we needed only two signal wires, two for power, and three for manual control, so we now have 50 off, seven pin sockets to suit. Do we now need 9 pin sockets?

The lamps will have 6v grain of rice bulbs. Running on 5v (from the board) gives a nice paraffiny light, so no resistor necessary. Chris should be able to cut the "glasses" from the filters used on theatrical lights. Ben has seen them and is happy with the colours I've chosen. The material is very cheap, £6 for 21" x 48". We should be able to incorporate a decent lamp holder in the lamp itself. They only take 0.1W.

You say a typical address would be "Signal 7 arm 1". As you know, I'm an electronics dunce, but isn't that two pieces of information, would it be simpler to just call it "Arm 17" or whatever, and ignore the post?

Which brings me on to Signal Identification Plates, like this
http://www.selectequip.co.uk/controlled ... on-plate-1 It says they are positioned below the signal to which they relate, like this
https://commons.wikimedia.org/wiki/File ... 367620.jpg But, many photos only show one plate per post, irrespective of the number of arms. I imagined we would add plates for each arm to correspond to Jonny's signalling plan, thus simply identifying each signal, and where it goes. Is this OK? Is this correct?

There are some good detail shots of BR signals at http://www.rmweb.co.uk/community/index. ... d-details/

[Eddyg]

Dave,

Since the number on a signal is used to identify its location among other things, only one plate is fixed to a signal post regardless of the number of signal arms carried. The plate shows the number of the primary route signal only to avoid confusion. The signal box will have the primary route number marked against the numbers of any other signals on the post so the signalman can identify the signal(s) that the man on the ground is looking at.

Eddie

[Richard Phillips]

Thanks for the detailed explanation Richard. A couple of points.

You say we only need 3 input lines. I hate to disrupt things, but we were assured that we needed only two signal wires, two for power, and three for manual control, so we now have 50 off, seven pin sockets to suit. Do we now need 9 pin sockets?

I'm using three lines to set an identifier, this is done with a DIP switch and is not an external input that is coming in via the sockets.

So yes, two for power, and two for the bus, the rest can be used for anything, although the mentioned "manual mode" could be used to control the signal with external switches by applying 5volts to the line - they'll need a 5 volt feed though at the if not supplied off board through, so that would need to be borne in mind.

The lamps will have 6v grain of rice bulbs. Running on 5v (from the board) gives a nice paraffiny light, so no resistor necessary. Chris should be able to cut the "glasses" from the filters used on theatrical lights. Ben has seen them and is happy with the colours I've chosen. The material is very cheap, £6 for 21" x 48". We should be able to incorporate a decent lamp holder in the lamp itself. They only take 0.1W.

Okay, but i've not seen the spec I jest slightly here as we perhaps ought to keep an accessible list of decisions online or something, as I do feel like I'm making arbitrary decisions sometimes as I've not been able to get to the meetings. I can't see a problem with the lighting, but whether we are able to switch all the lights on remotely, for example, is something I'd need to consider. Bearing in mind time constraints the easiest thing is to have them come on when the signal is powered, anyway there's plenty of refinements that could be done which will have to be left out at this stage because of time constraints - I'm thinking limit/detection switches on the servos so that we can report a true "state" of the signal and prevent conflicting instructions, as sometimes what you ask a servo to do, 99.9% of the time happens, sometimes, for whatever reason, doesn't. Then again there could be problems with the detection so I'm going for the keep it simple route to start with.

You say a typical address would be "Signal 7 arm 1". As you know, I'm an electronics dunce, but isn't that two pieces of information, would it be simpler to just call it "Arm 17" or whatever, and ignore the post?

This is when the world of signalling and programming collide. Even a signal has a physical description containing two components - eg Arm 17 is "outer home", "distant" - how does the signalman know? Well it's probably on a plate attached to the lever and on the diagram. "Arm 17" is a mapping. The bus has to have each "device" ie signalpost, as a unique address - therefore Arm 17 is really device 10 coil 1 (for example) - how these things are mapped is just down to how we program it. It does point to some kind of plan as to what each signals "computer" address is. The "signalbox" will have to be programmed with this in mind.

[Richard Phillips]

So since last post I've spoken on the phone to Jonny and much to my relief he's doing all the hard work and sorting out all the locking - it's cut the problems in more than half. I've got a master and two slaves doing my bidding (and video to prove it!) but I've run out of components.

Jonny sent me the track/signal plan and we're not driving as many signals off each signalbox as I thought, so that makes life easier too. He sent me some voltage regulators and an Arduino Nano. These boards are a few mm shorter than the micro's and there's plenty of cheap copies available. Sadly though, no joy with the shared USB serial port on these boards, which leaves me with the Micro. These aren't as cheap, at about 20 pounds each - CPC do them at 15.54 each for quantities above 10... but obviously that's 170 pounds + postage to shell out. I have two already, so I make our requirement 11 more as there are 13 posts? The reason I'm looking at boards without headers is because I've changed my mind on mounting on a PCB - it seems to be additional cost whereas we could just solder up the boards directly - I'm thinking of dropping the DIP switch to set the address and just change the Slave ID for the bus when each signal has it's program uploaded.

Has anyone got a cheap source for the servos? I have 2 of the type Dave used, and could do with a few more for testing.

I have to think now about fault detection and making it all a bit more robust - if the master/signalbox doesn't see a signal/slave for a while it just gives up on it to stop it slowing down the whole bus. The master will need a lot more programming as I'll probably try and incorporate indicators for "failed" signals and a feedback on the position of the arms. One slight issue is the board I'm using as a master doesn't want to work off of a battery, I'm hoping our 12volt supply will give it a bit more umph. Also I'm pondering things like sending a degree to which I want the servo to move rather than an on/off instruction and whether to use one register for all arms on a post to speed things up, but that's just implementation details!

[Alan Wood]

I don't know about anyone else but I don't understand much of this tech jargon . But I'm sure you and jonney know what your doing and as long as there working a bit before Shildon I'm sure we will all be happy , from what I see here we all have our work cut out if we want the signals done for testing and painting by March , if any of you need any work done please ask the rest of us I'm sure there are plenty of helpers out there .

[Jonny Ward]

The last two weeks have been very productive.

The Block Bells are complete and tested (Much to the Domestic Authority's delight, though the Trainee Engineer thoroughly enjoyed the testing session).




The inards of which were getting a little cramped for space...




They don't look too bad alongside the Tokenless Block Instruments I made a couple of years ago either...



Most of the bits for the lever frames are either in the workshop or in the post on their way here. The only tricky bit left is making the Key Token Instruments. I've made a start on them tonight, and I'm hoping they wont take much more than a week.

The drawings for this lot are more or less there, I should be able to start putting the frames together too.

[Dave Noble]

Nice work Jonny, I don't understand them, but nice work.

Also thanks for the wiring diagrams, I don't understand them either!

On a more positive note, I've sent for some more servos, and will send some on to you Richard when they arrive. Expecting some saddle etc patterns on Tuesday, they can then go off to the casters. I will have a go at casting some balance lever brackets this week.

Anyone fancy a go at working out how to make the loops on the ends of the operating wires, known as Standard Post-Office Make Offs?

[Keith1500]

Dave

Can you clarify what loops you mean?
Are theses the ones used by the pole wires? If so I have done this work before on full size telegraph poles.... Or more likely do you mean the actual operating wires leading to the signal arm from the servos. This is the seven strands which are wound around the wire each in turn, in which case I have also done this work.

Also...

My thoughts are along the lines of Alan here that time is getting a bit short and perhaps a test of the system so far would be prudent with the view of enlisting help where required. Personally signalling fascinates ME but i am not an expert but happy to help if required. Of course the other benefit of testing is a spot of training. These people can then train others at Shildon once things are up and running.....would this not be reasonable?

Keith