eldavo

Very few chemicals (at least those available to ordinary folk) will act as a solvent to PLA. The stuff is pretty hard even though it melts at fairly low temperatures. I suspect you will struggle to get a reasonably smooth tyre or even a nicely concentric one using an extruder filament printer. You could machine it after printing but be very careful of it melting. I've managed to weld small drills and even my piercing saw to various bits of the stuff. ABS is not that hard to print with as long as you have a heated bed on your printer (capable of about 100C) and the parts are not too big. Warping of large flat parts is the biggest problem. Using a large skirt or raft may help. Cheers Dave

PLA is awkward stuff unfortunately. In my experience the only things that will stick it to anything are cynoacrylate (superglue) or epoxy adhesives. None of the normal plastic cements or solvents will touch it. Cheers Dave

Looks like fairly standard specs for a Prusa i3 design. Looks a bit more sturdy than some of the cheaper options. Cheers Dave

Beautiful work! Cheers Dave

The SD card doesn't need to be big. Generally before files for printers are fairly small, a few Mbytes for complex items. Cheers Dave

Clear material will be challenging. The cheaper fused filament printers at 100 micron resolution will not give you good results in N gauge without a large amount of post printing rework. IMHO you will need an SLA printer using resin to get anything with good detail. It won't be cheap. Cheers Dave

Lovely bit of work! Hopefully the faulty LEDs will pop out easily with a little pressure from the front. Bet you're glad you finished filing all those locking bars. Cheers Davd

DesignSpark Mechanical available free from RS components is my tool of choice. Similar in some respects to Fusion 360 but with less function and thus a bit simpler to use.? Cheers Dave

Stunning! Cheers Dave P.S. A factor is someone who buys and sells. It ain't a factory guvna, they was factors.

The brewery is an impressively sized building. Personally I'm not sure about the backscene though. I hate to say it but the scale of the houses just doesn't look right to me. The chimney tops are at the height of the station building! Cheers Dave

I ought to think about digging Cramdin out from under the pile of junk in my workshop sometime soon. Look forward to a couple of days playing trains. Cheers Dave

Nice work Mr. Doolish. Cheers Dave

Wish I'd found those boards a year ago, would have saved me a ton of work. We are actually using I2C to communicate between Arduino boards. We've built an Arduino nano based board that does the same thing. Cost more in components than the whole board on eBay. Will be buying a few of those I think. Cheers Dave

That is a scarily neat control desk. Are you sure you didn't work for Hammond organs in a past life? Cheers Dave

Surprised you managed to achieve that with the print from my cheapo printer. Bit of light bleed but encouraging. I really need to invest in a resin printer. Cheers Dave

Morning awl. Commiserations to those down on their luck. Only real prank I recall from school was when a few of us sneaked in before the start of school and removed all the square section bars that linked the door handles on all the classrooms. When the bell went every member of the school filed dutifully into their classrooms followed by the master who of course closed the door behind him ready to take registration. On hearing the 2nd bell of the day every class attempted to proceed to assembly but for some inexplicable reason none of the door handles worked. Took the caretaker quite a while to open all those doors. Just as well there wasn't a fire! I remember the "mind the gap" module down at the Camborne do. Drove some other mad! Cheers Dave

Don't think I've ever come across a chopper with high beams before!

An important historical record there Giles. All that subtle camera work stirred a few memories! Cheers Dave

Lovely work. Certainly the finest garden railway I have come across. If you can turn a plastic pate pot into something that looks like solid rusty steel you are very clearly a master modeller. Keep up the good work. Cheers Dave

Lovely modelling. Some blighter has had off with the colonel's golf bats! Cheers Dave

Making some good progress Roger. I suspect your problem with the third rail and catch point is because the latter is too close to the line it is protecting. Might be one of those compromises we have to live with in our compressed world. Cheers Dave

As was mentioned in earlier posts the pointwork in the fiddleyard is switched using servos controlled by Arduino microcontroller boards. Our intial servo control boards were based on the design of board I had used on my layout Waton and soldered up on Veroboard. After playing around with a couple of boards we ran into a few problems. First up we we encountered Arduinos not working, arbitrarily restarting or just plain doing odd things. After spending some time looking for software errors we eventually identified the problem as voltage drop. The layout has both 12 volt and 5 volt DC supplied from a PC power supply to all the baseboards. Initially we had supplied the servo control boards from the 5 volt supply but after checking we found that we were getting a voltage drop on some of the boards which are server by long cables. We didn't spot the problem sooner as some of the Arduino boards are more tolerant of supply voltage than others so the problems were intermittent. To solve this we rejigged the boards adding a 5 volt regulator and feeding them from the 12 volts supply. After getting the Arduinos stable the next problem became apparent, the I2C messages didnt always get through between the Arduinos. Again this was narrowed down to the length of the cables in the layout. The I2C protocol was originally conceived to link integrated circuits or modules less than a metre apart whereas some of ours are 7 or 8 metres apart. Luckily there are chips known as bus extenders that can solve the problem and they are cheap if you get them direct from China. Having got things basically working we than had a rethink about the way we were building the Arduino modules. Soldering up circuits on Veroboard takes a bit of practice and a fair bit of care if you want something reliable. As we needed 8 or 9 servo control modules it was decided to try etching up a custom circuit board. The circuit isn't complex so how hard could it be? The schematic for the boards was drawn using EagleCAD... The circuit has an Arduino Nano module at its core along with an LM7805 voltage regulator and the snappily named P82B715P bus extender chip. There is a bank of 6 switches which are used to set the I2C node address, a couple of pullup resistors for the local I2C bus, 4 LEDs and a bunch of connectors. The board is capable of controlling 6 servos and 4 logic outputs that provide either 5 or 0 volts. The direction of throw of the servos or logic level delivered by the logic outputs can be reversed and other settings tweaked using a push to make switch. From the schematic diagram a singled sided circuit board design was created again in EagleCAD... This board layout was printed onto glossy magazine paper (from the Gauge O Guild Gazette) and transferred to the copper side of some circuit board using a hot iron. This was then etched using Ferric Chloride and given a clean up before drilling the 100+ holes using a 1mm drill. The result is shown below... Component positioning info and connection legends were then added to the non-copper side of the board again using laser printing and toner transfer with a hot iron. This side of the board was then given a coat of clear laquer before the various components were eventually soldered on. A completed board is shown below... All looks pretty simple from this side! The 3-pin connections for the servos run down one side with 4 screw terminals for the logic outputs at the top. The Arduino module sits in the centre with the voltage regulator, address switches, bus extender chip and power connections to the left. To avoid having to fit a large expensive 5 volt regulator to cope with the potentially 6 amp demand from the servos the servos are fed direct from the layout's 5 volt supply while the Arduino and other circuitry are fed from the 12 volt supply via a small 5 volt regulator. The 6 pin power connector also hosts the 2 connections to the extended I2C bus. Hopefully these boards will prove to be reliable... Cheers Dave

That's an appalling piece of bodging! The wire looks to be enamelled so is insulated but a long loop and an unterminated piece is just asking for trouble. You might expect to see that on a prototype but not on a retail product. Cheers Dave

Following on from young Richard's posts I'll try and give a bit of an overview of our emerging control system a.k.a "The Wiring". There is quite a bit of it! As Rich's pictures show we have a fairly conventional control panel for controlling the shunting operations around the creosote plant and sleeper works itself. This uses toggle switches to operate the tortoise motors and we have LEDs to give some sort of route feedback. There is a twist in that the reception and departure roads, known as The Elephant siding(s) are usually under the control of the works operators but can be given over to the mainline control panel for movements into and out of the yard. We are using Arduino micro controller boards extensively with all points in the fiddleyard being operated by servos controlled by Arduinos. We also have a few Arduinos controlling logical operations like switching the frog polarity of some of our more complex track configurations and interlocking between control stations. In the future our signals (and some animations!) will also be controlled by Arduinos. As we have all this computing power scattered about the layout it seemed appropriate to use it to cut down on some of the wiring. All the Arduinos communicate with each other using the I2C (normally pronounced I squared C) protocol which the Arduino supports natively and uses only 2 wires. With 40 something points in the fiddleyard this has reduced the number of wires needed to each baseboard significantly. The following is a simplified schematic of our control system... We are building two main control panels, one to control the sleeper works and one to control the fiddleyard and movement of trains on the main running lines. The fiddleyard control panel provides a route setting capability and contains an Arduino node that is the bus master for I2C communications and issues commands to all the other Arduino nodes. Anywhere else on the layout where we need to control things there are Arduino nodes mounted under the baseboards. In addition there is a slave Arduino node in the fiddleyard control panel that controls section isolation when we are running in DC mode and another one in the sleeper works control panel to handle interlocking and use of the reception/departure roads. The following is schematic of the fiddleyard and main running lines for the layout. The sleeper works is not included. I think this shows pretty clearly why we need some route setting smarts in the system as nobody could be relied upon to switch all the pointwork individually under exhibition conditions... The fiddleyard is separated into areas, down main (lines 1-4), up main (lines 5-8), freight operations (lines 9 and 10 with associated sidings) and the Romsey branch. There could be a fair few trains to oranise in there. Ill try to give an explanation of the complexities in the fiddleyard control box shown below... It does look a bit busy but it's not that complex honest! The vast majority of the mess is wiring between the 9 25-way D shell sockets and a bunch of tag strips just to keep things organised. At the top you can see the two black boxes of the NCE PowerPro system that was generously donated by Roger (Doolish). This is a 10 amp system so is far more than we will need as it will provide track power only. The output from the DCC system is fed through a pair of relays at the top right which serve to isolate it completely from the track feeds if the layout is run in DC mode. Switching between DC and DCC modes is via the pair of 9-way D shell sockets and plug in the lower right corner. There is a similar setup in the sleeper works control panel so that unless both control panels are set up for DCC we shouldn't get DCC power to the track and fry any stray DC only locos! When in DCC mode the track feeds to the fiddleyard are fed through the overload protection board (green thing below the righthand NCE box). There are 3 MERG overload protection boards in the sleeper works control stack that seperate the scenic sections into manageable districts (up, down and sleeper works). In the bottom left of the box is an Arduino node and a set of 8 relays... This set up allows us to isolate roads 1 to 8 when in DC mode. Each relay is defined as an accessory, just like all the points, and so can be switched as part of the route setting scheme via the I2C communication bus. In the top centre of the control box is the "master" Arduino node... This is the Arduino node that issues instructions to all the other Arduinos in the system. It has route setting logic in it that allows it to set up any of the 150+ possible routes around the layout and to check for conflicts. Currently it is capable of having up to 5 seperate routes active (assuming you can find 5 that don't conflict). It also has a couple more tricks up its sleeve. The master node also has a constant feed from the DCC system seperated out from the defined districts. The code on the Arduino monitors all DCC messages flowing and can respond to accessory decoder requests. A DCC message requesting a change to any of the points on the system will be picked up and converted into an I2C message to the Arduino node controlling that point, signal or any other device. This allows us to control any part of the layout through the DCC control handsets. Just as a final twist all the routes through the system are defined with a DCC accesory address so that you can set any route through the handsets as well. That's just as well as we don't yet have an actual control panel for the fiddleyard! Cheers Dave

To be a little more accurate, it used to be a very serious matter in the days of mainframes. I remember spending inordinate amounts of time ensuring that the software we produced (a company much related to HAL) which ran just about every bank, ATM network and insurance business didn't break the applications that ran on it. Then again the customers were paying thousands of $s each month for the privilege of using it! As to PC software and backward compatibility, there isn't any and that's not surprising as nobody pays for it. Have a good one all. Cheers Dave