Aug 31 2025

Simulation of “Smoke in the Cockpit”

After finalizing an almost complete Boeing 737 Flight Simulator, I thought about going one step forward and trying to simulate emergency situations. Smoke in the Cockpit is one of them, and not a funny one (See for instance the fatal Swissair SR111 Accident, where a fire with the flight entertainment system generated heavy smoke and fire in the cockpit, led to many system failures and eventually rendered the plane uncontrollable)

I already had three original EROS oxygen masks, so next thing was to make them operational. The first idea was to buy an air compressor and distribute compressed air among the three masks. However, air compressors distributing breathing air (instead of e.g. supplying paint brush equipment) needs to be free of oil and would not simply fit inside the cockpit itself. A friend of mine pointed out to a much simpler solution: how about using a common diving air bottle with compressed air inside?

That solution actually worked, thanks to the friendly crew of the Tauchsport Käser AG (TSK) in Bern, which by coincidence sits quite close to my home. I now have a regular diving air bottle filled with regular air at 200 bar pressure, a first stage distributor, which regulates the pressure to 4.5 bar and three customized air pipes supplying the three EROS oxygen masks. The diving bottle fits well in the first officer corner of the cockpit and can be easily recharged at the TSK diving shop. The two original EROS smoke goggles also come in handy, of course. The air in the bottle should last for around 30 minutes if three people use their mask in regular (non-emergency) mode.

From left to right: Diving air bottle and pressure regulator, EROS oxygen mask, EROS smoke goggles

After fixing one last air leak at the oxygen mask container, a few minutes breathing through the mask demonstrated, that the setup works as expected. Now, let’s do the emergency situation simulation …

WAIT: Where is the smoke?`Here …

This is the little handheld smoke machine Lensgo Smoke B, actually intended for movie sets or smoky caipirinhas, but it is well able to fill a Boeing 737 cockpit in a few minutes.

So my 10 year old son and I did a full Smoke in the Cockpit test yesterday and it was a hit. After a regular take off from Zurich Runway 28 and SID VEBIT4W I climbed to FL150 while my young copilot tested the five point safety harness and made sure his oxygen mask did fit his head. After around 2 minutes at FL150 I have started the smoke machine, we both put on the oxygen masks and smoke goggles and I tried to follow the “smoke in the cockpit” non-normal procedure checklist in Chapter 8.8 of the Quick Reference Handbook. It became clear that the fire was not controllable and not in the cockpit itself. I initiated a rapid descent to around 7000 feet and programmed the ILS RWY 14 into the FMS. The situation became quite uncomfortable altough the air we breathed was clean. I also realized that I could not speak to my copilot due to the mask. Connecting all the microphones, headsets and cockpit speakers through the audio control panels is actually my next project, watch out.

The landing was uneventful, although we were a bit fast (170 kts and flaps 15), since I decided to take a shortcut to get the smoky bird down to earth as quickly as possible. Evacuation was started as soon as the aircraft halted on the runway. Was lots of fun! See yourself …

It took about 30 minutes to empty the cockpit from the smoke again and I did not find any remains of the smoke oil so far. So: to be repeated!

May 18 2025

Microsoft wants your kids to play games 15 hours a day!

Like many parents we use Microsoft Family Safety to limit screen time of our children. It turns out that the default screen time is …

15 HOURS PER DAY !!!!!

So when you register a Microsoft device (e.g. Windows computer or Xbox) for your 9 year old, the default screen time for your 9 year old is 15 hours. Actually, the default setting is: no control of screen time. Isn’t this hilarious?

Dear Microsoft, it is very nice that you have an App to limit screen time of our children. It is possibly a fact that many parents are overwhelmed with managing the opportunities and threats of today’s technology together with their kids and will not apply screen time at all.

However, wouldn’t it be a service to those parents who care about the health of their children to start with a default screen time of 1 hour upon registration?

May 08 2025

Bye Bye USB, Welcome Ethernet

I’ve just removed and replaced the last OpenCockpits Hardware, the Forward Overhead Panel, which was connected via USB to the flight simulator. Apart from the extensive wiring needed, USB devices by OpenCockpits were quite difficult to use because a) their communication protocol is not “Open” as the name misleadingly suggests, and b) the OpenCockpits USB devices have no serial number, thus cannot be distinguished from each other. This last point made is rather difficult to run them reliably, since accessing two servo cards (e.g. one driving the fuel temperature servo and the other driving the cabin pressure servo) was always a lottery. We don’t want to fly by luck, anyway.

Here are some pictures of the “old” Forward Overhead Panel. Nicely built, but too much wiring and occupying 3 USB ports. The last image shows all the hardware panels which were needed.

The new world is called “Ethernet” and Teensy and consists of a universal Teensy firmware which listens to the flight simulator and allows to be on-the-fly programmed for any attached inputs / outputs or daughter boards via the I2C bus. The new Forward Overhead Panel has the following hardware structure:

  • Teensy 4.1 Controller with Ethernet and I2C interface
  • 19 MCP23017 16 Input / Output boards via I2C
  • 2 PCA9685 PWM / Servo Output boards via I2C
  • 2 HT16K33 7 Segment LED boards via I2C

The new Panel also features 4 Solenoid Switches:

  • Yaw Damper Switch
  • Engine 1 Start Switch
  • Engine 2 Start Switch
  • Wing Ice Switch

The New Panel also features 16 latched Switches like in the real aircraft and it is mounted on a original B737-300 overhead frame, plus includes three original LED-retrofitted panels and a few original and LED-retrofitted Korry annunciators. Beautiful stuff, see here:

As you can see from the last image, the amount of wiring has not gone to 0. However, each panel can be detached without removing the overhead since each panel only has a 4 wire I2C connector, one Interrupt wire and a 12V backlighting connector. This could be optimized by a generalized panel interface (using e.g. a 6 wire plug).

Flight test is underway, but the new Forward Overhead Panel looks stable so far …

Teensy Code as part of the xpcockpits framework available on Github here.

Mar 10 2025

How to convert a Boeing Lightplate to LED Backlighting

Boeing 737 classic aircraft use cockpit panels which are backlighted by incandescent bulbs. They are driven by 5V (not 28V like all other bulbs, so please don’t burn them, it happened to me). The bulbs are connected in parallel, they are directly soldered onto a PCB and fit through holes into the semi-transparent backplate.

Unfortunately, after 30+ years those incandescent bulbs are not that effective any more and the backlight is rather dim. They also draw quite some power as you can imagine (more than 1A per lightplate).

I have now fitted 12V LED strips (using 120 LED’s per meter) in pairs of 3 LED’s to the lightplate and thus had to drill another two holes per bulb hole:

The result is fascinating. A more even distributed backlighting which is much brighter and is dimmable by a PWM mosfet driver from a common microcontroller PWM capable GPIO output. The current drawn by the lightplate went down from 1 A to 0.2 A:

You can for instance use the Joy-IT COM-MOSFET driver to dim the backlight from software. It accepts 3.3V or 5.0V logic inputs.

Mar 08 2025

How to transform a Korry 318 annunciator from a 28V to a 5V device

Korry annunciators are used in many larger aircraft. Boeing often uses (or: used, e.g. in the Boeing 737 classic series) the Korry 318 annunciator. It consists of two 28V white light bulbs (Type 387 T1 3/4 Midget Flanged). The annunciator can be pressed, activating a switch to test the annunciator’s light bulbs. The light bulbs can be exchanged easily by pulling the front part of the annunciator from the respective panel … no screw driver needed.

The use of incandescent lamps in annunciator produces a warm light, but 28V is usually not what we use in cockpit building. The annunciator also draws around 80 mA, which is quite a lot.

Flaps-to-Approach has demonstrated how to dissasemble those annunciators and how they work. I will now show how to turn the annuciator into a regular 5V device that you can connect to your GPIO pin of any modern microcontroller.

There are two types of Korry 318 annunciators:

Type 1: ground seeking: common power pin and controlled by ground pin

Type 2: power seeking: common ground pin and controlled by power pin

The first step is to replace the lamps with LED’s. This is possible through the company ledlightbulb.net which offers T 1 3/4 midget flanged bulbs with LED’s instead of incandescent lamps:

https://www.ledlightbulb.net/index.php?main_page=product_info&products_id=2031

Please choose the white color LED with 5V dimmable configuration

For Type 1 Korrys please add in the comment field: “center pin negative”

For Type 2 Korrys please add in the comment field: “center pin positive”

You will get the LED replacements as shown below (left are two inandescent bulbs and right are two LED replacements. Put those LED’s into the Korry annunciator and test the circuit using a 5V power supply.

The next task is to build a circuit so that the LED Korry switch can be connected to your microcontroller. The Korry switch LED’s are connected in parallel, require 5V driving power and still draw around 20 mA, which may be too much for your GPIO pin of your microcontroller.

The following schematic is suitable for Type ground seeking Korrys:

The following schematic is suitable for the Type 2 power seeking Korry:

The circuits can be easily put on a small board right next to the Korrys, see here:

Mar 27 2024

Flight Simulator 1.0

Not many words needed. I’m happy to show you some pictures of my first flight yesterday on 26 March 2024 with the newly rebuilt Flight Simulator Home Cockpit, now featuring a second flight management computer and the newly built aft overhead panel, and, and, and … also many thanks for the VACC (VATSIM Controllers in Switzerland) staff for this greate Geneva to Zurich coverage!

Mar 27 2024

Back to Step 1

After having almost completely mounted the Sim in a newly built wooden house including a curved 210 degree screen and three projectors, a small metal part of the mechanical rudder pedals broke. Since the B737 sim is mounted in a real cockpit enclosure, there is no way to access the rudder pedal from the front. This meant unmounting the pedestal, throttle quadrant, seats and the main instrument panel MIP first …

The good thing about such an exercise: each time I have to take apart the Cockpit things can be improved and the rebuilding process becomes easier. The broken part was found, but the supplier of this otherwise great dual linked rudder pedal assembly, SimuJabs from Spain, unfortunately went out of business.

I was very glad to find a nearby wood and metal workshop in 3008 Bern called LU DE WIG, who created a reinforced replica of the broken part in s short time (piece is now painted in black instead of green):

Rebuilding the cockpit (start in front and work your way backwards was a breeze (only around five hours in total):

Aug 16 2023

Project Winedigger

The whole thing has actually started in 2007 at my bachelor’s party with friends when we buried three rather expensive bottles for 10 years about 1 m deep in a forest in central Switzerland. Guess what: the wine has aged perfectly under really adverse conditions since the bottles were subject to seasonal temperature changes of 10 degrees Celsius or more and were experiencing changes in soil humidity from completely dry to soaked in ground water.

So: What about those who say that wine can only be preserved with constant climatic conditions?

There have been several instances during the last years where either I or a colleague did bury a fine bottle of wine, geo-tagged the site and motivated others to find it. Preferrably on Elba Island. The idea ha ripened for building a geocaching initiative where you can find “wine caches” buried underground. The idea would be that you could get a nice outdoor experience and have a good drink with wine you would have never bought yourself. The rules for such a game would need to be created. The main rule would be that you have to re-dig a wine of your choice once you have emptied a cache. The main problem with this game is that current geo-caching initiatives do not (geocaching.com or opencaching.de) do not allow to hide alcoholic beverages since they are open for everybody. The project “winedigger” (let’s call it this way for the moment, there is no such website to date) would therefore have to be an adults-only (18+ year old) game.

Please get back to me if you would want to take the initiative and start the project.

Wine-Cache No. 1: Verzasca Valley Switzerland

Located in the beautiful Verzasca Valley in southern Switzerland, this typical (but not most expensive) Merlot grape wine was buried by me on 8 August 2023.

46.346050 °N, 8.791470 °E 897 m.asl. (Precision: 3-5 m)

Here are some pictures for you to guess what to expect. Note that the bottle is located in around 70-80 cm depth and you may have trouble to get to the place in spring during snow melt and after heavy showers. You need a shovel to dig since you may be off by 1-2 m.

In case you get lost, here follow two pictures of larger stone formations in the vicinity. Actually, the bottle is located just in the middle of those two stone formations …

Wine-Cache No. 2: Elba Island, Italy

If you ever visit Elba, don’t miss out the southern part of Lacona. It offers fine food, beaches without big buildings in your back and … a wine cache. Also don’t miss out the restaurant Bagni Orano. You can enjoy fine fish dishes while massaging your feet in the sand.

42.759740 °N, 10.307220 °E 2 m.asl. (Precision: 3 m)

We also have included a few kid’s toys. But be aware: the cache is burried around 1 m deep in the sand. You will need a good shovel (an avalanche shovel is a helpful tool for sand digging) and you may need to get rid of daytime spectators before digging. Sunset or sunrise is a good time for wine cache digging. If you need a hint: the sixt pole shall it be and not more than 9 steps you shall walk towards the sea.

Wine-Cache No. 3: Pfynwald Wallis in Switzerland

The Salgesch Village in the Canton Wallis in Switzerland is known for its superb wines. If you ever happen to be there, please don’t miss out the Gregor Kuonen Wine Cellar for tasting a few of the finest wines of southern Switzerland. A few km south of Salgesch on the southern side of the Rhone river lies the Pfynwald nature reserve. I have burried a fine bottle of Gregor Kuonen red wine for you. Notably not in the nature reserve itself but in a agriculturally used grazing field just at the border to the forest. Depth is only around 20 cm. Have fun!

46.298690 °N, 7.583840 °E 560 m.asl. (Precision: 3-5 m)

Good luck with finding! And please let me know if you have replaced it with a bottle of your choice. And even more importantly, please leave me a note on the wine tasting ceremony!

Nov 13 2022

A little update

As I concentrate on programming and hardware work you are not getting too much news here from the project. You can always visit and see yourself.

What has happened:

  • Completely moved all hardware from my attic to the real cockpit shell
  • Wired 3 headsets and got them working with online flying on VATSIM
  • MIP almost complete (missing second FMC and Lower Display Unit
  • All Glass Displays working through XPOpenGC and Rasperry PI’s
  • Temporarily installed two small displays for visuals (I would say the Sim is currently well fitted for IFR)
  • Heating and Cooling working but not yet connected to the Sim
  • Done a few test flights and continuing to improve systems for 100% reality.

What will follow next:

  • Development of three projector set up
  • Development of a wooden house around the cockpit to hold the curved screen and the projectors
  • Re-Engineering of Center Pedestal (rip off OpenCockpits and install SISMO hardware)
  • Acquiring a few reserve Rasperry PI 4B and 3B computers (they are not available right now)
  • Starting to build the Aft Overhead Panel
  • Re-Engineering of the Fwd Overhead Panel (rip off OpenCockpits and install SISMo hardware)
  • Complete installation of Audio / Intercom system controlled by Center Pedestal ACP and Aft Overhead ACP
  • Testing and introducing Teensy hardware in order to later replace SISMO hardware

A few current images follow here … feel free to ask about any specific technical details.

Feb 20 2022

How to move Heavy Stuff without Wheels

I wanted to rotate the cockpit shell horizontally by 90 degrees (you may have read the article where we already had rotated it by 90 degrees vertically). The thing weighs around 600 kg and wheels are not an option when rotating in a circle.

The idea was: build skis of wooden plates and use frying butter to lubricate and get around static friction of the wood plates. I had no idea if it would work. This was simply a cheap try. The expensive solution would have been to rent a crane (been there done that, see here).

Fortunately it worked out thanks to the help of my neighbors (no names here in this public blog, but many many thanks again for their force and help!). Using frying butter it rotated with medium force needed … like on butter!

Take 1 (Turning)

Take 2 (Moving)

And … in order to build a house around the cockpit to fit the projectors and screens in a waterproof environment, the whole cockpit was moved away from the platform today … many thanks to the whole family for helping!