Hi folks,
Our club is running its annual Build-the-Same-Kit competition and one of the nominated models was the BSG Viper. Luckily we got enough votes and there are now four of us building them.
I was going to do the new Mk. 1, but I found a cheap Mk. 2 so I thought I might try and build both.
I washed the sprues (long soak in warm soapy water, then clean with a toothbrush and rinse) since they felt greasy. They still feel not quite right, so I think I might give the sub-assemblies another wash before priming.
I've done a tape-assisted dry-fit of both kits, and the difference is not quite night and day, but certainly pronounced. From my limited experience with Moebius their CAD is quite good, but the tool making and moulding quality control is so-so. So once you get the parts cleaned up they fit, but cleaning up can be a pain. The Mk. 2 engineering is a bit wacky too - too many interlocking bits.
The Mk. 1 however is really good - great fit, much less clean-up required, and the way it slots together in paintable sub-assemblies is fantastic. It's also got a bit more presence than the Mk. 2 and better surface detail.
I also have the Paragrafix sets (not really essential but I wanted to try them) and a very nice resin Starbuck/Apollo figure from the original series.
Click for biggerYou can see how much bigger the Mk. 1 is compared to the new version.
I assembled the wings, weapons and undercarriage for the Mk 1. and dug out the cockpit for the lighting gels. Lots of chain drilling, scalpel carving and filing later:


Then I fitted the etched parts with 5 minute epoxy, the front panel first. They ended up a little bit high, but that was where they seemed to want to sit, so...

I filled and sanded the seams on the fin and started on the wings, but then got distracted and decided to get the engines together. This meant dealing with the clear insert - I know this is supposed to be a good idea but it seems weird to me - the insert has all this detail which should be opaque, and is basically impossible to paint, so the etch replacement is handy.
You still need the insert to get the etch in the right place and provide most of the engine tubes, so I opted for the brave-and-stupid approach and drilled out the backs of the nozzles with a power drill and a big spade bit

I shattered the supports but the tubes were OK, and after a good hour of sanding starting with 80 grit paper I ended up with this:

I sanded the flat backs of the tubes to about 2/3 thickness to allow for the thick brass, otherwise the whole assembly won't quite fit tidily over the engines.
Here are the bits that are going inside to light it up.

The microcontroller board is on the left - it's an $8 Adafruit product which is basically an Atmel ATTiny85 with a USB interface and voltage regulator. On the right there's a break-away strip of RGB LEDs, also from Adafruit although they call them "NeoPixels". These need a 3 wire daisy-chained bus and they're really easy to set up, with biggish solder pads on the backs of the PCB.
Here's the mostly-assembled lighting rig - I added a few more bits since I took this picture:


Caveat: This is a poor diagram of what may be a poor design, I'm not an expert! D1 is to allow USB power when plugged in since I was running the batteries down too much, and C1 is recommended by Adafruit to absorb in-rush current from long strings of pixels. Probably not needed here but I thought I might as well do it properly since it's being stuffed into a model where I would like it to remain.
I also forgot to add a resistor (I think I used 330 Ohm) from pin 2 to the first "pixel"'s data input. This was also recommended by the vendor.
The other LEDs are just driven by spare pins. It's probably better not to use pins 3 and 4 since on this board they're part of the USB interface and they seem to make it slightly less reliable. Not that it was very reliable on my PC anyway - it only connects every other time, although once connected I haven't had any problems.
<NERD WARNING>
I finished writing my software today - I like coding for these little machines, this has half the RAM (512 bytes!) of a ZX-81, which was my first computer. Although it does have about 5K of flash available for program code and constant data.
I did some experiments with writing custom code for the LED animations, but decided it would be better to write a keyframe animation system instead. That lets me define the animation colours and timings in data, and then have some glue code to decide when to trigger which animations, and tie all the results to the outputs. I also added a couple of basic signal generators for the pulsing and flickering stuff and used those to modulate the animation outputs in some cases. It's quite compact - a keyframe is 4 bytes (stored in flash) and the state required to run an animation is something like 6 bytes. All the maths is 8 and 16 bit integer for speed, accuracy but mainly reduced code size.
<END NERD STUFF>
And here's the result:
https://flic.kr/p/nKCpTo Sorry the video is so dark, otherwise the light was washing out the LED animation.
The viper cycles through four stages - "off" (flashing red LED), "start up", "run" and "shut down". The engines are lower right as I'm sure you guessed, the four white LEDs are the cockpit panel backlights, the lower single LED is for the main display backlights, and the two overly-bright red + green LEDs top right will be feeding single lights in the cockpit via fiber optics.
I thought having something flashing would be a good idea so that people would look at it long enough to see the sequence start if they arrived in the "off" part.
Now I need to paint and finish the cockpit so I can get the forward fuselage together.
Will