Skip navigation

Updated 01/10/2015 to reflect current simulator installs

(TL;DR Dinking with computers may actually be finished for a while. “All statistical models are wrong, some are useful” — George E. P. Box.)

We’ve had a new, more powerful, computer running for a few months now. And briefly mentioned some of the overall system design just a few weeks ago.

The new computer, running an Intel processor at 4.8GHz, is a big jump in performance and is our primary simulation system. The previous computer, an AMD processor running at 4.0GHz, is in use as a development/testing system.

The new computer should last 4-5 years based on the past five years of flightsim history (more about this in a minute).

Left largely unsaid was that neither of these computers were in their final configuration as of a month ago. In the last 4 weeks, we have reformatted and re-installed operating system/applications/add-ons on both of these computers twice. That is correct, a total of 4 re-formats and re-installs in 4 weeks.

Our task(s) were a bit less daunting than the average excruiatingly painful re-install because we designed our disk drives such that data, operating system, and applications were separated. Of course, apps and add-ons had to be re-installed, but we did not lose all the individual settings configuration when re-formatting the operating system drives. Along the way, we devised a way to mass activate several gigabytes of installed scenery files in FSX, FSX-SE, and P3D, but that is another discussion.

We can finally say “final computer configuration” after all the R&D that has been going on since May of this year.

The primary simulation computer (which we will call YUREI), is running Windows 10 operating system and has 4 simulation engines installed. Starting out as Windows 8.1, it was removed from our Active Directory domain and re-installed as a clean Windows 8.1 OS so that we could get the Windows 10 upgrade license.

The simulation engines installed on YUREI are:
1) FSX boxed edition by Microsoft
2) Prepar3D v2.4 by Lockheed Martin Removed and refund requested
3) X-Plane 10.36 10.40 by Laminar Research (via Steam)
4) DCS-World 1.2 by Eagle Dynamics/The Fighter Collection

Interesting to note, 3 of these 4 simulators are due for signficant upgrades any day now.

The development/testing computer (which we will call KITSUNE) is running Windows 7 operating system and has 3 simulation engines installed. Also installed are a variety of SDKs and development tools such as Sketchup, AFX, Airport Design Editor, GIMP, etc. Starting out as Windows 7, we did a clean install of Windows 8.1 only to see that many of the development tools would not work in our security environment. As a result, we formatted and re-installed it as Windows 7 for now.

The simulation engines installed on KITSUNE are:
1) FSX-SE by Dovetail Games (via Steam)
2) X-Plane 10.36  10.40 by Laminar Research (via Steam)
3) Rise of Flight by 1C Game Studios/777 Studios (via Steam)

These past few days have mostly been about performance benchmarking in P3D, the newest engine in the hangar. So far, the results are very good.

Mentioned in a previous screed, it seems we are living in a multiple sim world. We have had some very good, challenging and interesting discussions on Facebook and over chat about this idea, its ramifications, and not surprisingly, the future of flightsim. This discussion is taking place in many forums as well.

One of the viewpoints we read used Google trends to make a business case that a particular flight simulator under development should not confine itself to flight simulation. Although the business decision to make a multi-purpose simulation engine is very good (and being done by some other teams), the manner in which Google trends was used was very clumsy and un-nuanced.

Check out this dramatic looking Google trends diagram:

TrendsExample1We are way past our prime

Not only is the “Interest over time” in “FSX” declining but look at the yellow “flightsim” line. After the viewer catches their breath, they might want to look at the dates along the bottom. Of course, FSX searches were more frequent from 2006-2009. One to three years is the prime lifespan for most gaming software. The fact that “FSX” is a more common search term than “flightsim” says more about the longevity and legacy of FSX than it says about the decline of flightsimming hobby. Note also, that the trend analysis has “FSX” more or less stable over the next 2-3 years.

Further to this point, removing “FSX” from the search terms creates this diagram:

TrendsExample2More detailed

The “interest” in “flightsim” continues to decline and is predicted to continue to decline. However, “interest” in specific flight simulation products is on the rise. This is not say that there are the same number of people buying flight simulator software today as there was in 2008. In fact, we would say there has been real decline in the total number of copies sold between 2009 and today.

However (and if these “interest” statements hold any validity), the number of copies to be sold in the next 2-3 years is on an upward trend. Further, users are more educated about the products as shown by the decline in the generic “flightsim” “interest”.

All of this is relevant to me (yes, switching editorial perspective) primarily because of the scarcity of flight time in my schedule. I want to continue development of some scenery products but frankly, I am a pilot first. The results of my little survey and data such as these trends tell me where to focus my efforts. I will confess to some interest in iconoclasm of opinion based on the un-mensurated.

The increase in the number and quality of flight simulation platforms we are seeing is a direct result of the increase of desktop computing performance AND the development of 3d rendering engines that are separate and distinct software products from the simulation product. There are many parallels to this evolution in the history of computing. Beginning first with in the 1980’s with the transition from centralized mainframes to distributed desktop computers. Prior to the relatively low cost/high power of the desktop PC, only a large corporation could afford to build the hardware and thus they also developed the operating systems. In a natural fit, they also developed the applications. This is the model that created the giants of IBM, DEC, etc.

With the development of the desktop PC also came a much more radical and significant implication: the user or 3rd party developer could develop an application without being the company who built the operating system or hardware. Just ask Bill Gates how important this shift to accessible application development has been.

The mathematics of flight physics, the mathematics of graphical rendering, and the mathematics of digital terrain modeling have all been very well known since the 1950’s and 1960’s. Prior to less than a decade ago, computing power was the major limitation to the development of a flight simulator application.

Microsoft was able to develop the (then) pinnacle of flight simulation because they not only tackled the flight physics problems, they also created the DirectX graphics API (application program interface) standard. Direct X is what allowed a developer to easily make a mathematical calculation communicate with graphics hardware to display a rendered pixel. By making DirectX an open standard, not only was Microsoft able to develop graphics intensive applications, but other software developers could as well (if they programmed on Windows platform…after all, Bill G was no business dummy). X-Plane, as an alternative, uses another popular graphics API called OpenGL. While DirectX is only on Windows, OpenGL applications can be implemented on Linux and Apple. This is why X-Plane is able to run those other operating systems.

The rise of both PC and console gaming created a new software market for full blown graphics rendering programs. These rendering programs are the packaged and polished mathematics that can take an XYZ coordinate and translate it for DirectX. DirectX then can talk to any graphics card that understands the language to produce the rendered pixel. With this advance, a game or flight simulation designer could focus most of their effort on calculating in-game physics and action and less time on calculating a rendered pixel. AeroFly is a natural result of this evolution. This concept is taken to the next level by developers such as Outerra, who have created a complete world building platform that can be adapted to many simulation uses.

The trend to be highlighted is that application developers have increasingly been able to focus on the end use of the application instead of all the “nuts” and “bolts” to get it to display correctly. This trend being a direct result of software developers (with Microsoft leading this position) opening their toolboxes to other developers.

All software platforms can be grossly divided in to open and closed application systems. An extreme closed application system would be like the old IBM mainframes…IBM controlled the entire hardware, OS, and application ecosystem. An extreme open application system is more like the Intel/Windows application ecosystem. In terms of operating systems, no matter what your personal opinion might be, Apple is only 8% of total PC market share whereas Windows (in various flavors) is 88%. This spectacular difference in market share is not because of the relative age of the platforms, there were both created and first marketed with the same period. The difference is because you cannot build a computer running Apple OS without buying the hardware and OS bundled together from Apple. Where once Apple held first place in the mobile market with their innovative iPhone, for the last 2 years, Android is outselling Apple and has risen to 51% percent of total market share. This trend, again, as a result of the required bundling of Apple hardware and software.

This open/closed story is being played out in our (not so) little world of flight simulation. Microsoft flight simulator and derivatives are the most open flight simulation platform right now with X-Plane a possible tie. The open nature of ESP/FSX engine has allowed Lockheed Martin to focus on core platform and future oriented developments, instead of using money/time/people to work on content. LM has also extended the open nature of the engine by including bathymetry and improving the mission/scenario sub-engines.

X-Plane has a bright future and we are starting to see major 3rd party developer efforts. Laminar’s platform is quite good and at a base level is a much more modern product than ESP/FSX core engine. The tools Laminar provides to 3rd party developers are much more mature and feature rich than the hodge podge ESP/FSX/P3D SDK. There is a growing market share for X-Plane within the FSX community, with more and more FSX users “dual simming” with X-Plane.

DCS finally figured out that they absolutely must cooperate with 3rd party developers to expand market share. Although they are still struggling with the complexity of their platform. DCS is currently the “Apple” of our flightsim universe. An innovative, high dollar product that is still very closed and, despite their passionate fans, a relatively small market share compared to ESP/FSX/P3D. There is potential for much growth in DCS. But right now, they are making way more money selling the occassional new content to existing users than selling to new users.

Dovetail Games has proven there is a young and hungry market for general flight simulation. Reputably 200,000 copies sold so far, the Steam and Facebook communities for FSX-SE are filled with new folks learning about freeware and payware add-ons. Dovetail has promised a new flight simulator platform, rumored to be built on the MS-Flight engine. Many hard core FSX pundits have derided this MS-Flight heritage but I suspect these were the same folks that postulated no good reason to purchase P3D because it was (originally) just FSX in a new interface. The real challenge for Dovetail to outsell itself (FSX-SE), is to make the new platform very developer accessible. Dovetail had some issues with their add-on content paradigm early in the life of Train Simulator but seemed to have opened a bit. They will have to invite 3rd party development in a big way to get market share with a new product. IMHO the biggest issue with MS-Flight was poor vision and marketing from Microsoft. The MS-Flight core engine was, 3 years ago, closer to what P3D is now than their common ancestor FSX.

The wildcards in this arena of flight simulation platforms are not systems like AeroFly which will never amount to more than a very very, very small, probably first time buyer, market (sorry to burst your bubble AirDailyX). The true wildcards will be platforms like TitanIM, built on top of Outerra, that are being dual marketed for commercial and personal use.

As long as Lockheed Martin continues to sell to the consumer “educational” market, it will be difficult to them to fall to second place in market share.

For now, I would tell anyone who would listen: Fly mostly what your friends are flying but have at least one other sim. Finally, be prepared to buy a third sim.

FSX-SEJudging by the number of internet hits

A long time ago, in a galaxy far, far away, we uploaded a spreadsheet based HOTAS programming reference.

There have been a lot of changes to our systems since then, one in particular being that we use Tripod’s SuperScript for programming our Thrustmaster HOTAS for use with the VRS Superbug.

Here is how we have the HOTAS set up as of today:

  WarthoThrottleBaseSuperscriptuchi WarthogThottleSuperscriptuchi WarthogStickSuperscriptcuhi ThrustmasterMFDSuperscriptuchi

And, for your personal use, a graphical HOTAS programming reference that can work even if you are not using SuperScript.

Enjoy.

As a result of our participation in a few forum discussions concerning FSX performance, the website is getting a good bit of traffic for the information about we how analyzed and tweaked our system(s). And, a couple folks who were helped by our scars from work on Windows 8.1 have made some very kind and appreciative comments. Perhaps the time is right to discuss in more detail our latest system, including the software load.

The basic hardware specs for this computer are in the “Details” link at the top of the page. The best purchase decision we made was to go with a motherboard that did “one click” overclocking. The complexity of overclocking our system from 4.0 GHz to 4.7 GHz was limited to one drop down menu in the BIOS configuration. Literally, one decision, one setting to overclock almost 20%. We highly recommend this feature (and the ASRock motherboard). The second best purchase decision was a sealed system water cooler for the CPU. No mess, no fuss, easy assembly with an added bonus of much quieter sound levels from the box. The baseline performance of the system with almost no add-ons was a nice surprise.

YureiInProgressIn progress

From the earliest beginnings of this journey, we envisioned running the simulation system as a network of 2 or more computers. When the simpit build is awakened from hibernation, it will undoubtedly have an embedded computer running as a mate to our current simulation desktop. For now, the simulation-desktop-before computer (referenced in “Details” as the “Development/Testing Workstation”) is only being used as it’s moniker implies; no concurrent processes with the new workstation.

There are 5 separate hard drives in this workstation. We are not just using the machine for FSX, but also X-Plane, DCS World, and Prepar3d (to come a little later this year). It made sense to distribute the software loads across multiple drives. More significantly, we have a drive dedicated for the installs of all the ancillary programs, applications, and utilities for the various flightsims. This “AddOns” drive allows to install the programs outside of the default “c:\program files” structure to reduce the need for admin rights. There is also a side benefit to performance by having the active apps on a separate disk spindle from the operating system. The “AddOns” drive has a folder “AddonScenery” where all the custom add on scenery for FSX/P3D is installed. More on this a little later.  Finally, there is the “Workspace” drive. This drive is an “everything” else storage area. It is shared across the network so we can trade files between the 2 computers.

YureiDrivesDisk overview

You will note that none of the drives are super huge in storage space. Distributing the data like this across multiple disk spindles not only increases overall application/system performance, it also reduces the probability and effects of any single drive failure. We have been working with disk drive technology for over 40 years, beginning in 1974 with removable 640 byte platters on an IBM 1130. The one absolute from this arranged marriage of 4 decades is: All hard drives fail, be ready. A secondary empirical observation from working with literally thousands of computers is: More spindles are better than bigger drives.

Both of the active sim computers are backed up each day to our file server  (a current backup is part of the “Be Ready” axiom). Our file server has 6 hard drives arranged into 3x RAID1 volumes. With this configuration, any single drive failure failure on the server will not create a problem. In fact, the server ran from July 18 until August 10 with a drive failure on the main data volume with no ill effect.

The entire server is backed up every night to a portable disk. The portable backup is what will make it out of any potential house fire or hurricane event.

Server cabinetThe Moroccan theme paint job on our custom server cabinet

No joke, this new computer really hurt our pocketbook and sim budget. We are on very tight purchase restriction until next summer as a result. Still, it is a long term investment and we are hoping to get 5 years from it running the various sims.

Stay tuned for details of our software configuration.

WARNING: These documents are published for flight simulation only and are not to be used for any real world flight activities.

KNID VFR Terminal Flight Patterns

R-2508 Complex Airspace Briefing

KNFL NAS Fallon Range Users Manual

KNGP NAS Kingsville Course Rules
KNGP NAS Kingsville Course Rules Slideshow
KNGP NAS Kingsville Additional Course Rules Slideshow

KNIP NAS Jacksonville JNFC Course Rules

KNMM NAS Meridian Course Rules

KNQX NAS Key West Course Rules

TW-5 Fixed Wing Standard Operating Procedures for North Whiting Field (including KNSE, KNPA, and outlying fields)
TW-5 T-6B Checklist Guide
TW-5 Rotary-Wing Operating Procedures Manual

KNUW NAS Whidbey Island Course Rules

KNZY NAS North Island Course Rules Slideshow

A good bit of this post may sound negative. The limitations in the sim, procedurally, technically, and with the various add-ons, significantly hinders high fidelity simulation of a US Navy aviator’s training. These limitations apply at a very basic level even before any OPSEC considerations.

Still, for someone willing to invest the time, the Navy curriculum is quite good at training an aviator.

For our purposes, we originally adapted the T-6 PRIMARY curriculum for use with the Iris PC-21. At that time there was no T-6B. With the current Iris T-6B (developed with the help of real world Navy PRIMARY pilots), a very faithful PRIMARY curriculum could be developed.

While attempting to use the existing T-45C model to adapt the INTERMEDIATE JET curriculum, the more we studied the PAT PUBS, the more disenchanted we became with trying to use that model in a faithful way. Very basic characteristics covered in the PUBS, such as fuel flow numbers, did not match the sim. Reading ahead into STRIKE, the discrepancies seemed even greater when trying to adapt WEPS. About 30 hours of flight time and 20 hours of study later, we just weren’t feeling it.

Long story short, we have been flying the VRS F/A-18E in an adaption of INTERMEDIATE JET for about 70 hours flight time and 40 hours of study. There are several obvious issues with learning to fly a 22 ton jet in a manner designed for a 7 ton jet, but, overall, we are pleased with the training. Adapting the Rhino NATOPS numbers and capabilities to the curriculum has been easier than trying to do the training with a sim that did not match many numbers at all. We still fly the T-45C a good bit, but the flying is for a more casual experience rather than serious training.

Somewhere down near the bottom of this post is a link to the coursework and maneuver requirements we have adapted for our training. Some notes, in no particular order or sense:

1) The tally requirements in this curriculum are WAY less than half what the real world nuggets are required to do.

2) You are going to have to read a good bit to use the curriculum. At the bottom of each section are the PAT PUBS that are required reading. Of course, you should know the NATOPS for your aircraft before you start.

3) With the exception of FCLP, any aircraft could be used to fly the adapted curriculum events. We have done PA’s just for fun in the DHC-1 Chipmunk for example. The turn, S, and landing attitude patterns are part of real world training for almost all aircraft (or should be).

4) As we advance thru the curriculum and update the adapted course work, we will update the link below to the latest version. Also, updating these notes as needed.

5) Real world nuggets use standard IFR flight plans to ingress and egress the working areas. Below is a link for a couple that can be used at Lemoore. The package includes jet logs, flight plans, and MUs.

Here is link: INTJET/STRK curriculum v1.2

Here is link to the stereo routes package: Stereo Routes Package v1.0

PS, the BI portion in real world is all done under an IFR hood (except for the FCLP I added). So you will need full IFR conditions, an IFR panel view, or a hood simulation similar to what is in the Virtavia T-34C. Those wacky nuggets are required to do this BI from the back seat even before FAM. All your comments, suggestions, and corrections are welcome. We are off to fill in our MIFs.

Kneeboard parking diagrams for popular (and good) KNID and KNYL sceneries:

KNID parking diagram

KNYL parking diagram

NRR_TJRV parking diagram

And a modified CNATRA Single Engine Jet Log flight planning form:

Jet Log v5 

Enjoy.

Although we have never taken the controls of a real world aircraft, we have spent way more than a few hours in small aircraft as passenger, navigator, and/or cameraman. The specific aircraft being Cessna C-337, Cessna 172 and Piper Archer. Very little of this experience was in IMC conditions but a lot of it was in windy conditions.

FSX, even with addons such as OpusFSI and Accufeel, has never, for me, replicated the real world airplane/atmosphere interaction. The wind in sim is usually too consistent and stable. While the turbulence is more like what you feel in a large jetliner. The atmosphere is a fluid and the motion of a real world aircraft is more like a fish in water than a dart throw in still air.

The opinions of multiple real world pilots who have flown more than one PC flight simulator have guided us to make this X-Plane investment in time and money. Agree or disagree, it’s all good. But, should you be wondering, this is why we are exploring X-Plane for aerobatics and crosswind approaches.

After about 10 hours in X-Plane, it does indeed, for us, do a radically better job of simulating that aircraft/atmosphere interaction. We will be spending some significant time in X-Plane over the next few months, doing, for us, what X-Plane does very well. Big spoiler….not going to replace the FSX family of sims as our main ride anytime soon.

So how much money were we willing to place on this bet before we flew the first hour in X-Plane? Well, under $150 US got us 3 high quality aircraft and about 10 payware quality airfields.

X1$1Our X-Plane world at this time

One of the barriers preventing some folks from trying X-Plane is its initial purchase cost. Even at full retail price for X-Plane Global, it is still competitively priced compared to, say, Prepar3d when you consider real world weather and an absolutely superior view system are builtin to X-Plane. On sale, it is a bargain.

X1$2Addons to FSX that are baked into X-Plane

There are, perhaps, other thoughts about X-Plane that we may record as we use the sim more. If you have anything specific you would like to suggest, ask, or explore…you know how to get hold of us.

March 21, 2012 was the first flight I recorded in FSX. The software had been installed for about 4 weeks prior to that, while I figured out the software menus, graphics settings, and many matters unrelated to flight training. The Iris T-6A was installed as my first airplane because that was the current US Navy trainer.

That first recorded flight was from JFK in New York city to Sikorsky Memorial in Bridgeport, Connecticut. The route and lesson came from this book: Learning to Fly with Flight Simulator. Several other of my early recorded flights were also from this guide.

KBOSFSX on a dual core Pentium IV computer by the way

Since March, 2012, I have not only recorded my flight time, but also recorded my other flightsim hobby related time. The difference between anecdote and science is the difference between “noticing” and “measuring”.

In round numbers, it breaks down like this:

Fight time            450 hours
Pit Building          340 hours
Systems work      150 hours
Social media        300 hours (desk only, does not include mobile device time)
Scenery design    520 hours
Total                   1760 hours since March, 2012

That is about 11 hours each week spent on this hobby. But only about 2.75 hours of flight time per week. And about 20% more time over 3 years spent on scenery design than on actual flying. In truth, all of the scenery design work occurred in the last 2 years. If pro-rated accordingly, I’ve actually spent a little over 5 hours per week in scenery design (with a corresponding decrease in flight time).

FlighttimeMay2015A primary characteristic of success is focus

Worth noting, that in 3 years time, a real world Navy nugget can go from 25 hours private flying time to FRS in an F/A-18 Super Hornet.