davepallant

Thanks Ben,

That will be the next job - fitting the heads in the reply above the gantry onto the gantry. The wire has a solderable insulation so I am a little uncertain as to its strength pulling it around the sharp edges of the brass of the gantry., I'll need to thread the two lots of wiring up into the gantry along and down the leg. I'll probably leave it until fitting it onto the layout. I'm also wary about fitting the heads and then blowing up an led. I think I'll get the DCC decoder drive for the leds working and checked on the leds before fitting as well. I'm going to use a nano Arduino to drive them from DCC.

Dave

Next I needed modern signal ganties. Since I have Catenary I need some gantries for individual and double signals as well since the signals have to reach over the catenary in places. I looked at building my own and have purchased some girders and brass plate and brass netting but thought initially I wanted something less experimental to start with. I found the Traintronics N gauge kit and bought with the idea that I could cut them into individual single ended gantries as suggested on their website.

In the end the gantry is not long enough for me to get two of the lengths I wanted from it so I have gone for a full gantry over three tracks for two heads. The gantry is advertised as being for four tracks but I really dont think you would be able to do it with Peco track spacing. Inside of the two uprights there is barely 100mm and I allow 26mm or so for a track width so that I dont have to butcher Peco points for crossovers. For comparison N Brass OHLE portals cope with Peco track widths in the 3 way (87mm inside uprights) and 4 way types fine. Since I have three tracks on the mainline I am crossing the Traintronics gantry is nicely relaxed if a little wide.

It took about two hours of building. I didnt need to use bars for bending since the brass is nicely perforated. The instructions are a little vague in places. There is no mention of how or when to put the railing along the top of the gantry but I ended up tacking them on at the end. I didnt take any pictures of the gantry going together but here it is complete. This one has signals for the adjacent and third road in the three lines so needed a very long arm hence the use of the full gantry. I have several other places where I need part gantries so I'll be building at least two more of these kits. Traintronics suggest using the gantry as two parts but beware that there is only one ladder to the ground so you need to get some extra bits for a second ladder.

Now I've looked at the gantry more I do worry about the engineer climbing the ladder and finding the protection missing half way up. Not sure what I was thinking about using the two parts of the railing with a gap like that!

Dave

No link to N Brass or Traintronics just a happy customer.

My first signal gantry head in Frosted Ultra Detail seemed to leak light between the leds and I realised that I had given the signals above a coat of Halfords primer BEFORE wiring in the leds. By misting the FUD with primer it stops the light travelling between the leds. This should work with fitting leds into any slightly transparent material where you want to keep the light from seperate leds apart.

Now I have built another pair of 4 aspect heads ensuring that I prime the body before fitting the leds and it seems to keep the light from travelling through the body.

The leds are a little bright in the photos because I only had a coin cell lying around. I think I will tone the light output right down on the layout to help reduce bleed and to prevent them being distracting.

Dave

Bringing a fibre from under the board might be an idea but you would probably need 1mm diameter or so for the fibre. That did give me another idea though which is a single white led at the bottom of the feather and a triangular light pipe which narrows toward the end. That might reflect light into each of the 5 holes from the single led. That is similar to Ben's suggestion of a longer led.

Having said that I also have a double feather printed which would require four wires - feather 1, feather 2, common centre led and return... I'm in two minds whether I need that one.

Dave

Hi Dave,

 

For the catenary gantries and much of the infrastructure on Horseley Fields I just used Halford's primer grey, which is an acrylic paint (I think) and in my experience is pretty much 100% reliable to give a smooth, matt finish.

 

For some cable boxes etc I sometimes dry brush rail grey onto the primer to give the things a little texture.

 

cheers

 

Ben A.

Thanks Ben,

Just had a look at the gallery and Horseley Fields is a very nice layout! Nice atmosphere with lots of weeds and junk! I notice there are signals next to the Pretendolino on your gallery photo. Which manufacturer's signal is that if you don't mind me asking?

Thanks

Dave

edited for bad apostrophocation...

A feather. I had problems with one of the leds in the row, thought I had caught it and superglued the lot and then found the tip led is not working. I think the pad has split from the led. Since the FUD is quite translucent then not painting before fitting the white leds is the way to go here. This was painted after fitting the leds and was relatively easy to do. I have done the 'bad thing' and connected all of the leds in parallel. There was no other way really and we have had several signal feathers wired this way on Preston Club's Bee Lane for several years with no problems.

Extreme closeup. Yuk.

Here is the feather held against a signal. I don't have any individual signals with feathers on my layout - all of the feathers are on gantry heads - so I haven't glued it on. I'm not sure there's room for even one more wire down the post anyway!

Great job Dave! Will you be making them available to buy anywhere?

Hi. I've just been wiring a feather and not having a lot of fun what with the FUD melting and surprisingly, the white leds melting! I dont think it would be worthwhile me making runs of these things since I dont own a lot of the equipment I'm using. I would need to get a good soldering iron and a microscope to do enough to make it worthwhile and I dont have a lot of spare time. I can make the bare items available to buy on Shapeways but they will have to come with a warning that you need bionic eyesight and a fine tip iron!

Hmmm. Just been looking at a few source photos now and it looks like the box on the back of the lamps is often grey? Also my Railmatch Rail Grey metalwork is looking a little bright for modern infrastructure grey so any hints on an alternative acrylic colour off the shelf for signals, gantries and catenary posts would be useful!

My reference for signal lamp colours came from a Marl document concerning modern led signals. The Marl colours and the led colours I used are:

Green   Marl 500nm    Leds 520nm   True Green slightly less blue than the real thing

Yellow   Marl 593nm     Leds 589nm   Looks slightly less red than the real thing

Red      Marl 633nm    Leds 625nm

Dave

I've got the first wired with leds and a coat of paint and is as good as can be expected. I have cleaned this one for rough edges before painting and just brushed on acrylic over the spray undercoat. I think the photos above are quite harsh on the surface - these are better. The acrylic has filled in a lot of the surface and the matt colours help disguise it as well. I havent bothered putting any liquid glass in for lenses. I'd never get it in and it would probably be invisible.

I havent bothered putting the small panel on the back of the lamp box since the wiring fills the available space quite nicely.

Obligatory scale pic with a one penny (UK).

Straight on view

All of the leds one at a time.

I'm using the diode setting on a multimeter to power the leds one at a time but they are looking about right in terms brightness. The range of leds was chosen to get the green frequency as low as possible. The ones I used from CPC were:

Dave

It was printed in Frosted Ultra Detail. It was the only material that would get down to 0.3mm thick panels. All of the items based around the signals I have sent off for printing so far have been rejected for some part which has crept under the 0.3mm structural limit. Now I have seen the effect I am happier beefing up certain parts of the models to be thicker but I dont think I'll be able to go to the slightly cheaper Frosted Detail. As it was the feathers order arrived with a lot of bits in a bag. It must have kept them busy finding all the bits that belonged to me. I think I went a little too thin on the links to the sprues.

I still dont have my seperate order of 4 aspect heads for gantry use. Again slight thickness problem on the links to the sprues. I'll make another order once I have something else to get to share the postage.

Dave

A next step on my layout was to plan the signalling. Since I have a modern image mainline layout I decided to go for 4 aspect colour light signals. There are a few off the shelf N gauge 4 aspect signals available and also a couple of kits but I thought I would try 3D printing which had the possibility of being relatively cheap... if it worked. I can always fall back on ready to run. The software I have used is Sketchup. I drew up a 4 aspect head, a 4 aspect single signal on a pole and some feathers and a shunting signal. The first ones from Shapeways have now arrived. The seperate heads were rejected and need to be reprinted. The single head on a pole looks quite good. I have blown it over with grey primer and it looks a little dotty and I've got a couple of hairs in the paint but not too bad. Another coat of the right colours should help. The signal is pretty accurate with the red at just over 3m from the 'ground' and the top of the signal at 4.5m. The whole is just about 30mm from board surface to the top of the signal head.

I have started wiring one of these. It uses 0805 surface mount leds and roadrunner wire which is like fine transformer wire but has a coating which melts at soldering iron temperature. The pole is big enough for 5 wires (4 plus return) down a 0.8mm hole. I had expected to have to wire the leds outside of the head and then slide the lot in hoping that the common was the right length between. I made the first that way and three of the leds worked but not the red. I repeated but first gluing the leds into the head. With a fine tipped iron I could work around the safety bar and solder each wire in individually. I also have to say that I'm using a microscope since my eyesight is not up to focusing at this size! Again the red led does not work. It looks like the red Multicomp 0805 leds I got from CPC have the cathode symbol round the wrong way.... 

The leds I have used have a front face. I'll see how that looks and if necessary try a drop of liquid glazing in each led hole. Photos with leds on next week.

Dave

Would still contend that the power question isn't really that much of an issue.

I wasn;t suggesting that low power was an issue for train control. My thought was that the reduction in power meant that Bluetooth was becoming more appealing for all markets. More markets, more chips sold and the price drops. Now it's appealing as a low cost system for controlling trains.

 I believe that we should say as a group that we do not need or want a dcc-incompatible system. 

"Down with that sort of thing!".

What do you mean 'we'?! I don't want to say any such thing! I am just finishing DCCing my N gauge layout and even after spending that money and expecting to pay much more in DCC chipping another 50 or so locos over the next year I say 'bring it on'! It really doesnt matter what the messages over the Bluetooth link are and they might even be open source (which it sounds like Bachmann/Bluerail will do) the important thing is that for people who just want to run two or more locos at the same time on their home layout this is going to be easier and probably cheaper than putting together a DCC system. The features I expect it to have straight out of the box such as full length and descriptive loco names and maybe the ability to take photos of the locos with your phone and drop them into the app make it that much easier than selecting using four digit codes or one of a limited number of buttons. I've just spent the weekend typing four digit codes into Lenz controllers and the numbers cannot always be simply derived from the numbers printed on the sides of the locos because of duplication in whatever system is used (and whether different people in a club use the same system of reducing 5 digit numbers to 4) .

 

There is no advantage over DCC. DCC already has "direct connection to software driven computing devices". There are already a number of software packages from free open source to closed, proprietary and expensive. I know of, probably half a dozen, home brew systems connecting computers directly to the layout.

 

I know that these facilities are available on some combination of controllers and software but I really wouldn't wish these systems requiring PCs and customised software (I'm thinking JMRI because I am using it) on a newcomer to railway modelling.

Mounting a high volume, low cost chip onto a low volume PCB is likely to easily end up costing an order of magnitude more than mounting those same components onto a high volume PCB.

 

And that's a professional estimate, not an amateur one.

The only difference is that you are changing from a high volume low cost chip - the $0.50 processor on my N gauge decoder - to a high volume low cost Bluetooth chip with processor built in $1.00 in tens of thousands. They are both already being used in the millions and that brings the price down for everybody. I don't think you need to be worried about the cost of the Bluetooth implementation because it sounds like this system should rapidly create a new market for Bachmann, maybe cannibalise some of their DCC sales and end up with Bachmann using more of the Bluetooth parts than vanilla DCC parts anyway.

And who is to say that with the DCC being one way of providing the power to the decoder that the extra link for the processor to read the DCC data from the track (one or two resistors should do it) doesn't allow Bluetooth decoders to be both Bluetooth and DCC compatible? Most are already DCC and DC compatible and you can turn DC off if you want.

I am somewhat sceptical about the low power claims. The electronics themselves (e.g. the processors used in smart phones) are always advancing in terms of lower power but when it comes to radio communications there are fundamental laws of physics, related to bandwidth and range, that cannot be circumvented (we aren't quite there yet with the processors).

 

The lowest power bluetooth devices save power by transmitting very short packets and not transmitting continuously. A train control handset need to be "always on" sending a continuous stream of packets, and there will be little or no benefit in battery life compared to earlier Bluetooth or WiFi implementations.

My first 2.4GHz product was 8 years ago now and we are still using that Atmel ATR2406 even though newer parts have come out since. It uses an almost identical frequency hopping system to the Bluetooth of the time and similar power consumption to the Bluetooth alternative from Chipcon. Until recently the improvements in power consumption were not worth the work of rewriting the code.. It was stuck at 4dBm transmit power (maybe 100m line of sight) and it uses over 40mA in transmit and more importantly over 50mA in receive. Use short messages and you certainly get far better average current but there are times when you have to sit in receive waiting for a sync message from the master and the underlying receive current eats into the battery.

A more recent part like the CC2540 which is now Texas Instruments has switchable output power so you can drop to 0dBm (maybe 30m line of sight) and transmit and receive baseline are around 18mA for both.

I dont even need to refer to new parts on their way later this summer since I've just found a new part (new to me) that would be perfect for fitting in a loco and that It's the ST BlueNRG part which has a baseline of 8mA at 0dBm in transmit and 7mA in receive. Also it is available in a chip that is 2.7mm by 2.6mm! That should definitely go in an N gauge 8 pin decoder! http://www.st.com/web/en/resource/technical/document/datasheet/DM00092683.pdf  Since it contains a processor as well it would drop into the processor space on most decoders but would also need an antenna which could possibly be a copper track around the edge of the board or possibly a short piece of insulated wire, maybe 20mm, that would need to be routed inside the loco body from the decoder.

As anyone found out the cost of a the blue tooth smart, chip is likely to be. After looking on the web I haven't been able to get a price. Which could be the making of this or even the thing that kills it of before it gets going.

The company I work for is developing new Bluetooth Low Energy devices. The power consumption for the latest chips has dropped enormously so is now even better for use in handheld devices, watches, health sensors etc. The popularity of those devices increases now that they dont need charging every night so the numbers sold are going up and the price is coming down. We have just had a rep in from one of the major chip manufacturers talking about prices less than a dollar for Bluetooth Low Energy chips (buying 10s of thousands or more in a year) and that includes a built in processor capable of decoding the messages and outputting to a motor driver. That puts it in the same ballpark as a DCC decoder which might have a slightly cheaper processor but not by much. The track can now power the loco by DCC, a constant DC power supply or even a standard DC track controller turned up full. You would be able to take your Bluetooth loco which might cost £10 more than the DC loco and run it on pretty much anyones layout controlling it with your phone or tablet.  I'm thinking that the big gain for someone starting out is being able to power your own medium sized layout with a £20 power supply and not have the initial DCC cliff of a £200 5A DCC track booster to afford!

Use of Bluetooth is quite well defined and, given that the interface is embedded in the chip the messages and addressing should be very standard and therefore easily backward engineered. The messages over the air have to be sent to a certain standard for the system to work and for it to be called Bluetooth and use the Bluetooth chipset. The only risk to open source could be the manufacturer deciding to encrypt the data content of the messages so that it is impossible to work out how the messages have to be encoded to send the correct message to control the loco. Given that manufacturers don't currently do that for DCC and in many cases publish their proprietary control bus message specs so that groups like JMRI can easily interact with their equipment does suggest to me that train control will continue to be open.

An interesting thought is that DCC currently enters a layout through a single track  booster. It is difficult to use multiple track boosters on a layout except when they are fed the same command bus signal. Multiple controllers are handled by connecting them with one of several command buses like LocoNet and XpressNet. However, given a way of assigning which control device speaks to which loco or accessory then the air is the Bluetooth bus and any device could talk to any loco simultaneously with multiple manufacturers controllers and devices all working at the same time. If you want a dedicated Bluetooth points controller panel for your Bluetooth accessory decoders then you dont need to worry about whether your favourite command bus has a specific controller that will do the controlling job, you should be able to use any.

Dave

Yes!

Well done Mike and Ben, you should be proud of yourselves for having done this work and got us to the position where Jason can say that the UK market deserves the Pendolino!

Thanks Jason and Rapido for effectively fronting us the last quarter or so of the total money to get us from our deposits to the final product.

This is really good news. Thank you all!

Dave

Many thanks for everyone on this thread. I've read through it over the last few days and have just gone for a Silhouette Cameo on Yolo.   I've probably spent a lot of my discount on a spare blade and several sheets of iron on materials for my daughter to try out but I think it is money well spent. I'm going to be using it initially for Scalescenes cutting and then progress to custom buildings.

Interestingly I saw a US labelled company selling the basic Cameo 2 for £181 on Amazon.co.uk but I presume it would arrive with a bill for the VAT and the shippers charges for charging me the VAT. I may be wrong. If we were in the US we would currently be able to get the Cameo 2 starter kit with spare blade, spare cutting mat and glitter pens(!) for $270 which is £179!

Hi,

The polarity of the frog relays always reverts to the 'off' position when the power comes back on. The point is left where it was when the power went off. So for whatever reason the power went off, any point that had been set 'on' before the power went off will have the frog the wrong way. The only way to get the point and frog to be in sync when the power comes on is to either set all points to 'off' before powering off or cycle the points that could be wrong when the power comes on.

For now I've carried on using the Seep PM1 auxiliary switches where I need to power frogs and in a couple of places used Frog juicers which I already had. 

I'm using the decoders on N gauge and I've realised that to guard against any frog relay being the wrong way, for any reason, I would need to cut the two rails between the frog and the blades on every one of my points! Using the Seep switch set up properly the switch can 'augment' the blades by paralleling the blade contacts. (On OO that is the standard way of DCCing a Peco point anyway as the point is designed with the cuttable wire in the rails between frog and blades.)

Dave

A great model! It has clean sharp sides unlike some of the resin kits that by design have to have a little soft blobby feel to make them mouldable. Those photos make it look sharp enough to be OO!

Dave

Great Pics Jon,

 

The second one is really striking and makes the pendo seem very at home in its surroundings!

The second photograph also demonstrates to anyone worried about modelling the catenary that the supports are few and far between and the wires are effectively invisible. Plenty of great layouts out there with no wires, just a few supports.

My Amazon sale Railroad Mallard has now arrived:

REF01-90355

R2784X-50-979

My Amazon sale Railroad Tornado is

REF01-91200

R3060-11-852

I'm surprised that there is no date code anywhere on these things and I would expect some sort of batch tracing to be going on. Looking at the numbers following the factory code they seem to be incrementing over time so my guess is that the five digit numbers currently beginning 9 are the purchase order number from Hornby or maybe a delivery batch number either of which can be tracked to a date.

Sainsbury's car park and Preston North End (its north end)

:

Didnt enjoy DCCing this one bit. Serves me right for trying to set it up "on the main". I bought one of these second hand DCC fitted. The power car ran like a dog and all of the lights were on on the driving cars. It really does feel like you are trashing these coaches to get them open. I ended up pulling the inner buffers out first and then running my nails along the sides of the coaches but it is a bit fraught since if you dont do both sides at the same time the other side seems to pop back in. I also had the problem with the motor end of the power car not coming out easily. I think the pcbs get caught above the window mouldings. I am surprised that the body side paint is not ruined but the paint seems to have survived.

It turned out that there were no decoders in the driving cars so a couple of Bachmann decoders later I had the driving car lights flickering all of the time and the decoders in the driving cars buzzing quietly. None of the decoders would program easily and I couldnt use the section my layout I'm working on as a programming track because all of the stationary decoders complained by rattling their relays! In the end I set up a seperate piece of track as a programming track. I still had problems with the lights flickering and it wasn't until I cleaned the wheels carefully that I could finally program the decoders properly. I also had to pull off each bogie in turn and clean the contacts on top of the bogie towers and the contacts in the bottom of the body. Then the lights were on solid and the train would crawl at a slow speed. I finally had the rear lights on permanently one end. That was cleared by swopping the two end bodies - there must have been a slight misalignment fixed by disassembling and reassembling.

No mention of Eames of Reading yet? There was also a model shop in the Oxford Road. And was it Platform 5(?) in Wimbledon?

I definitely remember the Platform 1(?) shop in Wimbledon on The Broadway. That was where I bought most of my early model railway stuff back in the early 1970s. Shelves full of second hand stuff all along the right hand side of the shop and then the counter down the left hand side. I remember the long counter was made of glass with a model railway inside. Also they used to do A4 printed price lists for their secondhand stock with enormous numbers of items all priced up. I didn't have a lot of money as a kid back then so I bought a load of secondhand rubbish!

My other haunt was Beatties at Lewisham but that was probably for new track..

Dave

The unit does not necessarily know which way round you have got the point set to switch so if it switched the frog output between the two DCC inputs then the chance is it will be wrong 50% of the time. You need to be able to set which way round the DCC feed into the frog relay is to match the rest of the track around the point.

There is however a bigger problem than that with the ADS-8fx decoder and that is that it seems to forget the position of the frog relays if the DCC cuts out for whatever reason. Once the DCC comes back up the point will stay in the same position but the frog relays all drop back to the 'off' position. This means that any of your points set to the on position when the DCC went off will have their frogs set to the wrong polarity. The only way out of the problem is to cycle all of the points to off and then set them again when the power comes back up. Difficult, slow and noisy!

The long term way out of this problem is to run a second DCC bus around the layout and have a separate power section for the track and accessories so that if the DCC is shorted then the ADS-8fxs are still powered but this is precisely what the instructions tell you that you do not have to do!

I've given up using the frog relays for frog powering for now and have fallen back on the Seep solenoid switch contacts. At least I know they always follow the point position.

Dave