sign up .

August 13, 2020, 12:46:46 am

Login with username, password and session length
35 Guests, 1 User
  • Total Members: 3961
  • Latest: avvial
  • Total Posts: 55573
  • Total Topics: 7418
  • Online Today: 36
  • Online Ever: 582
  • (January 22, 2020, 08:44:01 am)
Users Online
Users: 1
Guests: 35
Total: 36



Will Depart in...



Collins CTL head units

Started by ame, October 24, 2019, 08:48:45 pm

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.



I have been having a go at building various things for a friend's flight simulator. Now that I have started I want to keep going. I have several projects to post, but I might as well start with the most recent.

I realise that the simulated elements don't actually have to work, but they have to appear to work, thus lending verisimilitude to the cockpit experience. What I am trying to do is build things that look right, given the limitations of DIY construction, do not use too many exotic parts, and could be replicated by someone else.

So, the most recent thing I have been building is the Collins control head, as found in a King Air, for example.

There are usually a cluster of six control modules: two COM; two NAV; one ADF; and one ATC. All of the four types of modules have a similar appearance, comprising of two 5-digit 7-segment displays and indicators, a momentary toggle switch, a selector switch, an audio volume control, two rotary encoders, and three small pushbuttons.

I have found several photos of the modules on the web and used them to estimate the size of various parts, and guess what components I could use to make something that looks and works the same. Although I prefer to work in metric I realise that a lot of avionics are designed using English measurements, so I have tried rounding measurements off to fractions of an inch, usually some whole number of sixteenths or thirty-seconds. I then convert back to metric for CAD. Often, popular sizes jump out, but I realise they are ever-so-slightly not the same. For example, 3mm and 1/8", 6mm and 1/4". I live in New Zealand, so a lot of stock parts are metric, but their history is obvious. Aluminium sheet is available in 1.6mm thick, which is about 1/16". MDF, plywood, and other materials are available in multiples of 300mm, which is nominally 12".

In general, my workflow is:

1) Find photos, drawings, manuals, etc. on the web and hand-draw a sketch and optionally a dimensioned and annotated drawing.
2) Use the hand-drawn sketch and notes to make a CAD drawing (I use QCAD for 2D CAD) with different layers for various parts of the final model.
3) Use the CAD drawing to make 3D models (I use OpenSCAD for 3D modelling) or to cut parts from stock (acrylic sheet, aluminium extrusion, paper, etc.)

When all the parts are ready they can be assembled into the final model.

This model looked like I could cut it from 6mm thick opal acrylic sheet, stick on a 3mm thick transparent orange acrylic as a display filter and stick a 3mm black acrylic layer on the back. I tried this but realised cutting and drilling the parts would be time-consuming, plus it would be difficult to get the accuracy between layers.

So, I decided to 3D print something that looked like the complete outside shape. This has the benefit that I can leave apertures and spaces in exact locations to add components later. Then I don't need a PCB (probably). I want the panel to be backlit, so I have left recesses for opal acrylic which will diffuse the light from LEDs below. The decal will be a sandwich of laser-printed paper and laser-printed transparency. The black toner will block the light, and the white paper will show the lettering and artwork.

So far I have the overall shape, with locations for fasteners and components. I have an orange filter, and I have ordered some yellow/green 7-segment displays to go behind it. I have 3D printed some knobs, and I have some ideas how I can do concentric rotary encoders and backlighting.

I intend to drive the whole thing with a single HT16K33 I2C display driver chip, and I have already written some Python code for this to turn LEDs on and off, and fetch the state of buttons. There will be one chip per module, with each chip at a different address on the same I2C bus.

The attached photos show the model printed in white and coloured with a Sharpie. Now that I have a few parts to play with I can see if everything is going to work together, or if I need to revise anything in particular.

Thanks for reading.
You cannot see attachments on this board.You cannot see attachments on this board.


Ok. Here's what I am trying to achieve. Yes, it's fake. It's the same model sprayed satin black, with a laser-printed decal. I intend to have a second layer on a laser transparency sheet, printed in reverse and glued onto the paper. That will give me an extra layer of black masking for the backlighting, and will protect the decal from wear.

I don't have the 7-segment modules yet. Or the rotary encoders. Or a number of things. But when they arrive I can carry on.

For the backlighting I intend to cut pieces of 3mm opal acrylic and insert them into the model. White LEDs shine from the back, and the decal is stuck on the front. Light should shine through the white areas of the decal and be blocked by the black toner areas.

It might be even easier to leave 'holes' in the body of the 3D model and print solid blocks of PLA the same size which drop in and act as diffusers. That way, no cutting. The body will be printed in black, and the diffusers in white or transparent. We'll see.

As for the rotary encoders, I don't know if they will work with the HT16K33, because it has built-in debouncing. It might just quash all of the encoder signals.

Thanks for reading.
You cannot see attachments on this board.


I 3D-printed a 'scaffold' to hold LEDs in place behind the front panel. I still have only a single layer of laser print as a decal, so there's a lot of extra light coming through.

I've also found a switch for the frequency controls which is not a rotary encoder. It is a rotary pulse switch, which acts like two independent buttons. One is repeatedly activated when the knob is rotated one way, and the other when the knob is rotated the other way.

Anyway, they're on order too now, so this project will probably be on hold for a while until parts get here.
You cannot see attachments on this board.


I designed an assembly for the back, where two cheap rotary switches can be used instead of an expensive dual-shaft concentric rotary encoder.

The idea is not mine, there are several examples on the web, but it's customised to fit into this model.

I'm waiting for some acrylic tubes which I will use as actuators for the switches. I will use 6mm tube with 3mm hole for the outer switch, and 3mm rod for the inner switch. I will use 8mm tube with 6mm hole as a coupler from the 6mm rod to the outside of the switch itself (which is also diameter 6mm). You could do it with brass or aluminium, but I'm hoping to back-light the knobs using the acrylic as a light pipe, even though they are not backlit on the original unit.

Generally this arrangement is used for rotary encoders, but I am worried that the chip I am using will debounce the signal from the rotary encoder until it disappears entirely. I will test this, but I have also ordered these:

They are Alps SRBM1L0800 rotary pulse switches, which I have mentioned above. They require no special decoding and will appear (electrically) just like a pair of pushbuttons.

You cannot see attachments on this board.

Currently two rotary encoders are installed. The round flange to the right is for a volume potentiometer which will be driven by a shaft through the centre of the on/off/selector switch.

Like the Website ?
Support and Click Below to Donate