Busy, Busy, Busy

The last few months have flown by - hard to believe it's nearly the end of May...

So, having part-written countless blog post drafts, I've not had the time (nor inclination I suppose) to finish them. Suffice to say, I probably should.

Having had my Birthday recently, I'm now inundated with lots of bits I've been wanting for ages and claiming was imperative for my 'making' needs. Highlights include tonnes of buckeye couplings, a couple of new locos, a Faller car system kit, a loco kit and some DCC gear.

I'm pushing forward with my model rail layout at the moment, with the awful weather I've not really bothered with the R/C vehicles, though I have sorted most of the electronics out for the projects I'm working on. What is mostly outstanding now, is time to design and 3D print the parts I need to complete those projects and some time to paint/finish them.

The rail layout is naturally growing and evolving far beyond the original plans, but it is moving forward. I'm working on the design at the moment and a little stuck with it, but over the next few weeks or so I should be laying track and roadways. I'll be visiting Miniature Wunderland in Hamburg as well very soon, so hopefully that'll also give me the inspirational shove I need to finish the design.

My workbench is also getting busy, with a number of model conversions on the schedule, in addition to the electronics and 3D print projects I'm working on. Pictures will start flowing shortly, I promise.

Finally, head over to www.makerbase.co.uk and check out what I've been working on since January - we'll be officially launching really, really soon, but in the meantime we're 'beta' launching with a smaller set of products. Sign up to the newsletter and you'll be in the running for our prize draw as well.

Okay, back to the workbench...

Building a 3D Printer

Much has been happening on the 3D printing front. A few delays in sourcing parts (as will be explained), but otherwise things are going smoothly.

After some research and help from the donor of the seed parts, I chose the Ecksbot as the basis for my first 3D printer. Even before completing the first, I've got plans for an OpenBeam-based printer with a larger print area and finer print resolution.

I initially thought the electronics aspect would be the cause of the delays - I suspected I'd get it built and it would gather dust as I desperately tried to source the electronics and wire the thing up. Not so (so far), the biggest delay has been on a couple of the mechanical parts.

From the outset, I wanted to source parts from good businesses with great products - a combination of British, German, Japanese and even US-made parts. My aim is to find as many manufacturers in the UK and EU to source my parts from for all projects. Of course, if I were to be an utter purest, none of my projects would come to fruition, so I have been forced to source from China for some bits (the electronics at the moment being the sole parts for the printer).

Seed Parts

Printed in British-made extruded PLA, all are the standard parts (the majority of the Ecksbot parts are already are based on the improved versions of the RepRap's parts) with the moving parts in orange and static in black.

Mechanical Hardware

The majority of fasteners are A2 stainless steel hex cap-heads some of the M3 are rounded heads, similarly the threaded rods are A2 stainless steel. The smooth rods have been the hardest to source, even at the time of writing, I've still not found a company able to supply them to me for a realistic price - not to mention the majority of suppliers have neither responded nor acknowledged me.

Hot-end

Looking around at the available hot-ends, it proved easier to get it straight from Eckertech especially since they engineer and test each one to high tolerances. Plus, their pre-sales support was helpful in selecting a nozzle size. I plumped for the standard 0.4mm nozzle, even though I will want finer printing in the future, it's best to start with a standard construction and work from there.

Electronics & Motors

Settling for a RAMPS 1.4 platform, PCB heatbed (with custom made aluminium heatsink) and a set of upgrades including the LCD panel and SD reader. Not much to say here until it's built up and I get to grips with the pros and likely cons of the RAMPS platform.

Once I can source the remaining parts, I'll be able to complete construction. I've built up as much as possible (working around the missing parts when the structure permits), so I'm keen to get these last few parts. The shopping list now; hobbed bolt, smooth rods, T2.5 belt and stepper motors. The hobbed bolt will also be coming from Eckertech, the belts and motors I just need to collect, leaving the smooth rods to procure.

Next update should hopefully see construction complete and the majority of electronics connected.

New plan. Build a 3D printer. Simple.

As a redux of a plan I had in May, I'm now back on track with building a 3D printer. Spurred on by the release of MakerBot's awesome Replicator 2, it was either saving up and waiting until later next year, or hacking my own together now. Being impatient, the latter seemed appropriate.

Part of my logic is that I can quickly solve problems and technical issues with projects I have outstanding, by virtue of custom-making the parts I need and hopefully producing a better finish all round.

It also means I can quietly add to my stack of projects, but claim they're just sample print pieces when they fail or languish.

I also joined DEMU, a great resource for information (and detailed technical data) on diesel and electric traction, with a great deal of information on the on-track plant equipment I find myself drawn to (probably being bright yellow and a definite appearance of function over form).

I'd already heard of DEMU, but seeing them at the Warley show on Saturday and some amazing samples of printed parts (a pair of incredibly-massively-amazingly EMU detailed bogies) was just the shove I needed to join. I plan to make good use of their drawings for accurate modelling and printing of vehicles I've been trying to find desperately.

I've started sourcing parts for the printer, so it's probably time to start writing up the build in more detail. Even at this stage, there seems to be a whole range of technical issues to contend with before I even get hands on with building it.

I'm not sure how a 3D printer will solve my problems, but I just know it will...

To quote the Underpants Gnomes;
1. Underpants
2. ?
3. Profit

UX - More Than Fluff

My day-to-day work tends to get split three ways between information science, information technology and control technology. But the one underlying thing that's common to all and massively neglected across all, is good User Experience or "UX" design.

Too often developers excuse horrific processes and user experiences to "getting things done," or commonly the lack of a user interface design is excused with a weak contrast of consumer and industrial - after all, you only design an interface for flashy consumer gimmicky software and instead you should dump your industrial and business interfaces on users from a-high (akin to a pigeon shitting on your shoulder).

Most of our control work is fixed within the SCADA and networking of plant systems, so more often than not, we're involved with designing interfaces that control large motors, pneumatic actuators and quite often high pressure steam. And yet the portfolio of products designed for this industry isn't frankly good enough to automate a kettle. As a result, we've had to work outside the box and draw from our years of experience building enterprise-grade systems (both hardware and software), to distill what's expected at the high end of IT and deliver it to the field (quite literally).

And yet it's a struggle to convince people that a rainbow colour screen with randomly placed information, weird animations that are linked to unrelated attributes and all running on a crapped out consumer Windoze PC is not the high of industrial automation.

It does frustrate me that more time and effort is put into the graphics and displays seen on films, then are ever put into the real-life counterparts. For what it's worth, we're trying to change this and introduce a level of user experience that matches the serious nature and importance of the systems involved.

For now, I'll leave you with these first class UX/UI designers:

• Mark Coleran
• Jayse Hansen

Rant over.

Winter, the Project Motivator

Now that winter is slowly and cruelly creeping upon us, my lackadaisical project efforts have started to give way to a bit of focus. Over the next few weeks, there should hopefully be some big progress (and write-ups) on my projects, that include:

- RC Platform (aka UAV/Rover)
- RC Hovercraft Rebuild
- RC Boat Rebuild (set back slightly due to new hull damage)
- 28mm Model Building/Painting
- 4mm Plastic & Card Models (1/76th aka OO)
- Train Control System (LCB)
- Train Control Panel (DCC/Traction)
- 4mm Layout (1/76th aka OO)

Plus a big sprinkling of vac-forming, CAD work, panel building, Arduino hackery and lots more fun stuff. Oh and I'll also be lusting after 3D printers and CNC laser cutters. And robots.

Stay tuned.

RC Dual-Motor Control

The mission; to develop an Arduino based PWM mixer that can control dual-motor powered vehicles.

Having bought two RC vehicles that rely upon dual-motor control, in the process of stripping out their original low-spec electronics, I was in desperate need for a mixer to take the steering and throttle channel inputs and mix them into left and right motor outputs.

Having searched for a solution, the closest I could find was a ludicrously expensive RC tank ESC, circa £300. So, with an Arduino to hand and a couple of servos, I started developing a simple mixing routine.

The input signals were PWM, or Pulse Width Modulated - a simple way of transmitting an analogue signal using digital pulses - supplied by the steering and throttle channels of the receiver. The corresponding outputs would also had to be PWM, in order to drive the ESC that the left and right motors are driven by.

The outputs are the easy part with help from the Arduino servo library. Most people have driven a servo from their Arduino, it's a nice simple piece of coding, no circuitry, just plug-in and go. But reading PWM turned out to be a bigger problem to tackle then I initially thought. The big mistake was assuming I could read PWM signals easily with the Arduino.

A couple of solutions are floating round, I think I found some on StackOverflow and others on robotics forums, where people were trying to integrate RC receivers with Arduino as a simple alternative to ZigBee.

I even hoped to use that same approach on other projects myself, so was set on getting it to work. In fact, it all sort of started coming together nicely - the RX is even powered off the Arduino's 5V bus, so it keeps things quite nice and simple.

But the code just keeps getting more complex.

In short, it's possible, but a lot of code is needed, and the Arduino can't do much else. From all accounts from others working on similar code, it struggles to read and write two PWM signals. Bugger. I might come back to it sometime, but it's not a nice reliable design at present.

So back to the drawing board.

The final breakthrough came in the form of a £4 electronic mixer. Better known as a V-Tail Mixer, these are designed for simple fixed wing aircraft, that utilise a pair of servos to operate bother rudder and aileron - the mixer takes the values of both and mixes them proportionally. Bingo. Perhaps this is the RC communities biggest secret. Perhaps I've found a genius alternative use of the mixer. Perhaps everybody does this. Who knows? I found jack shit on my travels, so I've come to the conclusion that I made a ground-breaking discovery.

The mixer is perfect - it takes the input values and passes them through mixed. It doesn't do anything else to the signal, so the resulting output seems to work with everything including uni- and bi-directional ESC. And they're £4. Tiny bit cheaper than an Arduino. And less effort.

Perfect. Let's rewind this. The perfect solution, that would fit virtually every use of dual-motor powered vehicles, whether they be boats, tanks, hovercraft - whatever. It costs £4. And yet, it's being sold and advertised solely for people with a V-Tail whose radio transmitter lacks an onboard V-Tail mixer setting. Great.

But all is not lost... In my struggle to find a solution I found the OpenServo project. Quite genius really. Simply, the project's main hardware development is an I2C based controller board, with virtually the same footprint as the regular PWM controllers found in servos. There's a lot in such a small package - including temperature monitoring, advanced programming, braking and a massive voltage range.

Coupled to the I2C bus on the Arduino, I have high hopes of advanced control for robotics (and anything else I can wedge it into). The I2C bus seems to be more and more popular, with it being used for a wide range of applications these days. So, not really of use in driving speed controllers for this mini-project, but definitely of interest for the future...

Productive Fortnight

Having a 4-day weekend is a definite recipe for hobby productivity...

Over the Jubilee weekend/week I managed to push forward a number of projects, including; the OO scale card modelling, the RC dual-motor controller and a couple of new projects.

One of the new projects is Arduino based and makes use of the output channels to control scale lighting on my forthcoming OO layout. The idea is to simulate a day in scale speed and loop through a list of lighting circuits which have start/stop times - as the time is reached the lighting circuit is either faded up (PWM outputs) or flickered on (in the style of fluorescent street lights). I want to avoid building too many circuits (555 timer etc), as the programmable nature of the system gives me the flexibility to tweak and change the lighting as desired (without having to fix on/off times etc with resistors/capacitors).

The main code function is a loop, so with a scaled pause and incremental counter, I can approximate a 24 hour day in whatever time base I want. I could use a potentiometer to dial-in a scale, or I could fix the scale in code. I'll define a number of lighting types, which will define either an analogue/PWM control or a digital control of the circuit - the PWM will allow for fading and the digital for simulating fluorescent and regular tungsten. I can also go ape-shit on the lighting with backlit billboards, building signage etc and funky lighting effects.

My plan is to also use an interrupt channel in order to reset or tweak the playback (perhaps skip an hour etc), that'll enable the loop to be broken and the interrupt code to be executed before the loop resumes.

No code written, but lots of notes and ideas on scrappy bits of paper. When I get a few minutes I'll start sketching out the code. The main issue is devising a template way of driving the lights - I'm thinking of using LEDs for most things, so each channel will need to power a handful of mini-circuits.

This started off the back of me wanting to simulate welding flicker - something I still want to model and may perhaps execute with this approach (probably a little overkill for an Arduino however). Another project on the list; just what I need!

The card models have come a long way - I've virtually finished the low-relief workshop and moved straight into building the concrete bridge. With that finished (sans a few bits that need to be assembled when I know the exact layout of the bridge) I moved onto two new ones; the inspection pit and the small goods store. I've modified both slightly, as I wanted two inspection pits out of the one kit and the small goods store didn't work too well with the type of card I'm using, so some extra glue joints and cover caps are needed there.

The dual-motor controller is an entirely different matter and deserves it's own post...

On a side-note, I did get to see the Queen.