The HT16K33 isn't too difficult to find, and can be picked up quite cheaply on AliExpress if you are prepared to wait. The DLG2416 displays are quite expensive from typical vendors, however hunting around on ebay etc might result in more affordable sources. They have a fixed character set which is unfortunate, however they remain useable and pleasing alternative to the 16 segment displays which the actual UFC requires. If didn't manage to find suitable 16 segment displays of a physical size to work with this project. The comms channel indicators are 16 segment displays (be careful not to buy 14 segment) but unfortunately I was unable to find displays which were not slanted - for me a minor issue.
- The ESP32 boards come in all sorts of shapes and sizes, and my controller board needs a particular type of 38-pin board and I found to my cost that there were several boards with different widths. I would expect these boards to be the correct size.
- The green 16 segment LED displays now (2020) seem to be very hard to acquire.
- R2 and R5 on the controller board should only be installed as a last resort of the I2C is not stable without them
- The PCF8575 is a PCF8575PWR (TSSOP)
- D1 and D2 are BAT54CDW and not BAT54TW
I would love to have a full Simpit for every aircraft – who wouldn’t – but space / time / money all conspire against this. I personally feel that there is more immersion from the sensation of real switches, knobs etc, but compromises are part of life.
I wanted a mix of virtual cockpit and some real devices so I wrote a native F/A-18C interface for Helios, and I started a project to create a hardware Up Front Controller. This is what I am presenting here. While the 3D printed UFC enclosure is very rough, hopefully you’ll get the idea of what I am doing. The UFC just has power going to it, and it has the potential to drive future parts of the cockpit via i2c (I am aware of the limitations and potential problems with the approach of using i2c, but it seems reliable in my set up). The UFC is controlled by an ESP32 board to get me the direct pins and interrupts that I needed for the encoders. Because there is only a single power connecter to the UFC, it means that it can be removed from the set up extremely quickly. Likewise, the Thrustmaster Cougar MFDs are also very easy to remove as they simply hang from the top of the screen, primarily to allow hiding of the USB cables. I rotated the rocker switches and 3D printed the hangers so the MFDs sit over the top corners of the 23 inch Dell touch screen.
The UFC itself is pretty simple using DLG2416 displays (I like the dots). Unfortunately these do not have any programmable characters, but the result is OK IMHO. I would have preferred 16 and 7 segment displays, but I could not find suitable devices. The Comm Channel displays are 16 segment LED displays driven by the same HT16K33 that reads the keys and switches. Other things I would change if I ever revised the UFC design would be to replace the 2 x volumes and the brightness knob with potentiometers and put an i2c ADC onto the board – but I am happy enough with the encoders in the current design that I’m not planning to do this.
I’m going to put a third Cougar MFD in for the AMPCD, but I’m still considering how I can do this while keeping the mounting discrete.
With two inidividuals (that I know of) successfully recreating this project, I chose to revisit
the facial. I purchased a small SLA 3D printer which allowed me to print smaller key caps, and
I was able to paint and use my 2.5W LED Laser to label these. I also created new comm channel
selectors, as well as more harmonious knobs (ie not Mil Spec accurate) for the UFC.
My LED laser will not touch acrylic other than black, so the facia was cut from that material and
white polystyrene sheet was then glued to the surface, and painted black. The LED laser was then
used to create the lettering. While all of this is an improvement, I have to confess to not being
very happy with this. This is in part why the GitHub project has not been updated with these changes.
Are there going to be further revisions to this project? Well maybe. I am considering a V3, primarily to see if I can improve the visuals of the displays, and to see if I can gain some level of crossover with the AV-8B which has a similar UFC. DLG2416's are expensive and difficult to get hold of. 16 segment LED displays are also out of fashion, and green ones seem extremely rare. I have created Eagle drawings for discrete 0603 / 0402 LED displays which can be charlieplexed and might be coaxed into looking like the 16-segment displays used in the real UFC. I just need to finalise these plans, and pluck up the courage to use a PCBA service to build a prototype (I have neither the patience nor the skill to solder the volume to LEDs needed). If I get around to getting a display prototyped AND I can work out how to make it look fit for purpose, then I will almost certainly begin some work on a V3 of this UFC.
Controlling the very large number of LEDs will also be interesting, but I have been doing some prototyping and it does look possible.
I started out many years ago with the A-10C and Loz’s profile for Helios which was a a great place to start. After a while, I built a real caution panel based on the Holtek HT16K33 which was driven by an Arduino Pro Micro which communicated via a serial connection to a c# program which talked to DCS. Over time, the A-10C parts grew a bit and I had CMSC, CMSP, UFC, NMSP and a few other panels.
The EOS project was a little too complex for my needs, and while DCS-BIOS is a fantastic option, for a number of reasons, I have chosen to stick with my own technology.
Fast forward a few years, and the Pro Micro and c# program were replaced by an ESP8266 with more memory so that it had the capacity to do all of the parsing needed and the wifi communication allowed direct connection to the DCS exports.lua. This was later replaced by an ESP32 board.
When the Hornet came out, I decided that I again wanted to have a hybrid Helios cockpit, but I have always taken the view that offloading as much work from my DCS PC so that it can concentrate resources on DCS, was a good thing. Since I always have access to under-utilised PCs, I wanted the virtual cockpit to be on a remote machine like I had with the A-10C. Unfortunately this ruled out Capt Zeen’s Hornet profile because it uses keyboard commands which appear on the wrong PC, and the use of the A-10C Helios interface requires a lot of processing in the exports.lua on the DCS machine.
I’m happy to make 3D parts visible on Fusion 360 on request. Schematics etc are Eagle format.