EEVblog #1267 - How To Control a 747 Flight Simulator

How do you control the thousands of switches, sensors, actuators, bulbs and ARINC 429 bus devices in the world's most expensive 747 flight simulator?
Rod Redwin from Simulator Solutions explains the design on the I/O solution for such a mammoth task.

Support the EEVblog through Patreon!

Buy anything through that link and Dave gets a commission at no cost to you.

Donate With Bitcoin & Other Crypto Currencies!

[flymypg]

I once worked at a small (and extremely ambitious) aviation instrument company, and most of our toughest problems weren't on the development side, but were on the test side. For every hour spent developing hardware and software, three hours were spent in system testing. And that excludes component-level testing for hardware and software, which is part of development. Beyond that, an equal amount of testing is needed for the separate certification effort, where the testing is focused on more than just the instrument, but on its integration with the aircraft as a whole (including Type Certification).

If you've ever wondered why FAA-certified flight instruments costs 10x-100x the cost of uncertified gear, that's the reason: 90%-99% of the cost is in the testing and certification. It's not just the paperwork (which is merely the proof it was done): It's the endless hours of patient and thorough testing and analysis, and the racks of equipment needed to make that happen (which also needs to be tested and calibrated).

Every engineer was also expected to be a testing expert, and (unlike Boeing) we worked closely with our DERs to ensure we were doing all the right things the right way. One of the greatest difficulties was simulating real-world inputs, and we put extensive effort into developing our test infrastructure. Some signals could be modeled mathematically and simulated in software, but others had to start from recordings of real-world systems in flight.

As Rod explained so well, there is a vast multitude of ways signals are encoded within aircraft (in both the analog and digital domains) using a chaotic mess of "unstandardized standards", many examples of which exist merely to link two or three specific instruments. Even things one would expect to be "bog standard", such as GNSS links, have a perverse plethora of permutations.

So, not surprisingly, there is a significant market for test equipment able to generate and receive/analyze these specialized signals. Many of these are no longer made, others are built-to-order at exorbitant cost, and some (a surprising number) simply don't work for their intended use. Which means we generally had to make our own simulation and acquisition equipment.

We took a brute-force approach. For example, rather than directly digitize the sinusoidal multi-phase synchro/servo signals, we simply bolted high-resolution digital quadrature encoders and geared stepper motors to the shafts. We made fantastic use of Ebay to get old equipment we'd re-engineer to become test equipment. We even got some WW2-era test systems whose signaling methods are still used today!

We had to verify our test systems were faithful to the real-world environment, which meant sampling real signals in flight. Being of the brute-force mindset, rather than build specialized signal recording systems, we took multi-channel GHz-bandwidth o'scopes with extremely deep trace depths into the aircraft. We also became expert at making Hall-effect sensors to non-invasively measure current-mode signals within operating aircraft.

We then flew the aircraft to its limits, both in terms of maneuvers, but also the environment, including flying into storms. Of course, almost nobody would rent you their airplane for such testing. Even NASA wouldn't let us piggy-back on their storm-chaser planes. So we had to purchase, operate and maintain our own small fleet of test aircraft. Those were, by far, our most expensive pieces of test equipment! Yet another reason why FAA-certified instruments cost so much. This was for a company with under 100 employees, including the staff for our in-house manufacturing lines.

When we took our cockpit instruments and sensors to trade shows, we also took some of our simulator equipment to drive them. At one very large aviation trade show we received many more inquiries about our simulator equipment than for our flight instruments! Those requests came from other instrument makers, and we realized we had a significant competitive advantage in our test systems, else they wouldn't have asked about them. From then on, our test gear was kept in locked and sealed racks that were never opened in the presence of others.

I'm wondering if Rod and Simulator Solutions have explored the aircraft instrument test market? Unlike the flight instruments themselves, the test equipment hardware/software is not FAA-certified at all, other than ensuring the signals they generate meet the applicable aviation signal standards (more of a specialized calibration/verification than a certification), and that the test equipment will not damage the instruments to which they are connected (per industry-standard safety and EMI/EMC specs).

[spyderMN]

I spent 20 years as a software engineer in the simulation industry (I am intimately familiar with the AH-64D helicopter sims the US, UK and other govt/military use as well as the F35). I primary worked in the visual system side, but also worked with the hardware and interface side of things. It is super impressive to see a smaller group of individuals take a project like this on. I can tell you that the civil aviation simulators had huge budgets with 100s if not a thousand or more people working to develop these solutions. Will be fun to see how this all turns out when finished.

[spencebarton]

I'm a retired United Airlines Pilot and definitely spent many hours in that cockpit and simulators at United's training center which were indistinguishable from the real thing. It's sad to lean that that beautiful airplane was cut up. BUT... it's a good use of surplus parts.



Boeing never "over-engineered" anything. They used appropriate redundancy, strength and reliability. Until recently the greatest airplane manufacturer, ever.

[andymouse]

"what constitutes an Arduino ?" the guy sounded a bit embarrassed...he didn't need to be, I've heard it alot from people who have large super complicated projects, and when you say " so it's got an arduino in it ?" they rapidly tell you about their firmware and how much its NOT Arduino...I think its super cool that it has off the shelf jelly bean MCU's with now almost de facto footprints at its heart..go 328PB go !...It looks very much like an Arduino because it IS an Arduino !....great project and what seems like a passionate bunch! I would very much like to see more please Dave.

[olivialambert]

God I'm jealous. Some people have the most amazing setups. Combat sim pits would possibly be interesting for EEVBlog viewers too. They tend to get a bunch of real world components, but of course not every component is available and the interface to the computer is different so they have to design their own back end. I recall seeing one guy get a real RWR, essentially an electron beam display. But for some strange reason the control plates were mounted at 45 degrees, and of course didn't come with instructions for the connections, so he had to work it out and design the controller to write symbols and letters where and when the computer asked for them. It was a great read. I've spent a few thousand on my setup, but I'm not even on the board compared to what a lot of people have managed to build and the work they have done.

[martinda]

I had watched the vids on your other channel, and was blown away with the amazing engineering and effort these guys put in. Wonderful - and something I have dreamt of since I was six...Hmmmm?

[stclairstclair]

I've never thought about traveling to Australia until I got into EEVBlog!

[nickkapirnas]

I love these, I've seen a lot of these 1:1 scale sim cockpits and the attention to detail is astonoshing! I personally have built a 1:1 scale virtual pinball machine that's a simulator of the real thing, and though it's not quite a sim cockpit, something about recreating a real word thing as accurately as possible running a PC based sim is just awesome to me, along with how passionate people are when they do stuff like this.

[Spookieham]

Great video Dave - can't wait for the next one! Great work by this team, very impressive. EEVblog at its best.

[Digital-Dan]

This is the kind of thing I always wanted to know about these systems. Nice to see they're building their own version of a CAN bus. First thing like that I ever encountered was when we used a variation of Ethernet inside a Xerox copier in about 1978.

[christopherj]

wow he's put a lot of work and effort into those stacks

[mikemike]

Those nifty gray keyable connectors are Wago series 733, with an ultracompact 2.5mm pin pitch. I've been looking for connectors like these for a long time.

[Sixta]

So basically, it is just an arduino, bunch of shift registers, bunch of random mosfets. Unprotected inputs, unprotected outputs, no controlled switching slew rates, no ESD countermeasures, no EMI compliance etc. I see a lot of work there to be done. But I stay being corrected!

[BlackEpyon]

LOL! Dave, who walks into a cockpit and says, "Don't turn it on, take it apart"?

[billh]

Very cool, I was going to build an L1011 cockpit and bought all the Autopilot controls for it. Ultimately I became an airline pilots and abandoned the idea. Still have the L1011 autopilot controls. Today I fly 747-400's for a living.

[Knight]

Amazing project; can't wait for part 2 - soon as you can Dave!

[Vidicon]

That's a very cool project. Is one of the leds on the pcb labeled L1 ?

[tomgeorge]

Great video Dave, looking forward to CH2.
I have opened a topic on the Arduino Forum, highlighting this video, hope it gets you some more well deserved subscribers.

[MetallicBlade]

They should bring the 747's back.


Cool job this guys been doing.

[peterandersen]

Great project. But I still doesn't understand: who makes the simulation software that runs the hole thing including the displays?? like airdatacomputer, sensors, stallwarnings, radios, autopilot etc. etc. all these thing. That is extremely big and complicated project to replica a aircraft in software.