A lot of cockpit work has been going on simultaneously this summer: the new Cessna GPS printed circuit boards and switches with integrated LEDs have arrived.
But also the universal B737 COM/NAV/ADF PCBs, so I’ve been working on both in parallel.
And because I was curious what the planned backlighting would look like I’ve also been working on the matching panels.
It was a perfect dry summer to do the painting of the panels outside.
For the panels I’ve tested 5 mm opal plexiglass with 80% transparency and painting them with spray paint from Action: two layers of grey primer and two top layers of anthracite.
I learned that for a good result it’s very important to spray thin layers. It’s tempting to keep spraying until it looks completely covered but that often leads to paint runs.
The result doesn’t look too bad, but I’d like to have the texts more white with the backlight off, so I’m going to do some more experiments with a.o. polystyrene material.
For the dual rotary encoder I originally planned to use some second-hand knobs, but the outside diameter of the inner knob was a bit too large. I tried to make it smaller but it appeared quite difficult to do this so that it fit exactly centered in the larger knob.
At conrad.nl I found a nice series of affordable 14 and 20 mm diameter knobs of the brand Mentor that are sold by multiple shops and that perfectly fit together. Unfortunately they’re only available in black (except for the cap which is also available in grey), so I’ll probably try to paint them grey.
As the dry period went on and on I thought I might as well start painting the throttle quadrant panels, which was quite some work because they had to be completely disassembled and after the painting assembled again.
Some of the panels had not been wired yet, so I have been finishing that job too.
Also the throttle handles had to be disassembled for painting so this was a good occasion to polish the aluminium reverser handlers whose surface had been damaged a bit during bending. For this I used a polish set that can be used with a drilling machine. A small MDF construction was created to hold the drilling machine during the polishing work.
Considering the large number of connections in the throttle quadrant I’ve started to design a PCB to connect everything.
While working on the schematic I realized that using an OpenCockpits Mastercard will become problematic: the common GND connections of the inputs are multiplexed: every block of 9 inputs has its own multiplexed GND which may not be connected to the other GND pins. Even if I would divide the extra logic in the throttle quadrant in separate GND domains this would mean that also the clutches will be connected to a multiplexed GND, which doesn’t seem a good idea.
Using an Arduino Mega with direct connections and ArdSimX looks feasible. It also makes the wiring easier: by embedding the Mega 2560 Shield PCB in the design the Arduino can easily be piggy-backed at the bottom of the throttle quadrant PCB and the throttle quadrant can be connected to the PC by only a single USB cable.
To make room for the Arduino at the bottom I’ve created somewhat higher versions of the throttle quadrant panels 1 and 11.
In my original design of the control circuit for the electromagnetic clutches the limit switches would send a logic ‘1’ when closed, but as other switches use a common GND it’s easier to invert the limit switches polarity so that multiple switches can be connected with a single GND wire. This means that the 3-input OR gates in the clutch control circuit have to be replaced by 3-input NAND gates and the software set and reset signals must be inverted so that sending a logic 0 will activate and deactivate the clutch.
I found some nice cable clips for organizing the wires in the limited space between the throttle quadrant panels so that they don’t get damaged by the moving parts.
I’ve been thinking how to better synchronize the trim indicators at the left and right side of the throttle quadrant now that they are controlled by a single servo. Initially I used a single 3mm threaded rod, but this bends too much. For a new attempt I’m using three rods, two of which go through the servo arm and small custom-made spacers are used to connect the three rods. Unfortunately no nuts can be used on the two rods that go through the servo arm because the distance is too short for that.
To enable small adjustments of the indicator needles I’m trying these rods (type 007, 85mm) instead of the original custom-made metal strips.