de Havilland Comet 4C

[hobofat]

Oh the drool. You make me wish I were sitting in a real one right now.

[Pirx]

I really love the way you create the graphics. It's fantastic to see how a developer has it's own way and stile when reproducing the reality, and the textures you build are awesome.

I now it will sound you repetitive but I really hope a soon release.

[OlaHaldor]

Wow!! This is just awesome! I love how the engines are fit into the wing like that. Sexy. Futuristic, in a 60's kind of way.

[guym-p]

Thanks! Comets were beautiful. Amazing that design started in 1946, and the first flight was in December 1949. The first Comet was a rather different aeroplane to the last one, though. It's only the tail that makes it look very dated.

By the way, if anyone is in London, I recommend going to the Science Museum and following the arrows to the Alan Turing exhibition (it's free of charge). In a glass case, they have a large section of the roof about 6-8 feet long, showing the fatigue-failure from the corner of the ADF window, and the first use of computers both for root cause analysis and for the calculation of stresses for the re-engineered (and reassuringly robust) Comet 4.

Edited August 6, 2012 by guym-p

[guym-p]

I've just finished tuning the performance of the Comet 4C for X-Plane 9.70. I've written an overview of results here: http://www.dh-aircraft.co.uk/news/files/1fb6133dab05132c84561db743ab5dfa-99.html

It's cause for a mini-celebration (at the very least, a cup of tea and some Marmite sandwiches), because it means the Comet model is now finished and ready to go for XP 9.70. XP 10 is another matter: I made some initial back-to-back comparisons this morning, and found that XP10 thrust is 11% short compared with the last iteration of XP 10. It could be a new setting in Plane Maker that I need to make use of properly, or it could be more serious. I shall start the changes necessary for XP10 next week ...

[deniasol]

I've just finished tuning the performance of the Comet 4C for X-Plane 9.70. I've written an overview of results here: http://www.dh-aircra...3ab5dfa-99.html

It's cause for a mini-celebration (at the very least, a cup of tea and some Marmite sandwiches), because it means the Comet model is now finished and ready to go for XP 9.70. XP 10 is another matter: I made some initial back-to-back comparisons this morning, and found that XP10 thrust is 11% short compared with the last iteration of XP 10. It could be a new setting in Plane Maker that I need to make use of properly, or it could be more serious. I shall start the changes necessary for XP10 next week ...

That's excellent news! Can't wait to get my hands on this file which promises to be one of the most exciting releases ever.

[guym-p]

I've taken some time this morning to gather a few screen shots and post them here:

 

http://www.dh-aircraft.co.uk/news/files/26afff15b5a355380e75d62b0917ca57-103.html

 

This is the first time I've had it back together and in any way photogenic since some rather major revisions.  The changes were precipitated by adding a periscopic sextant in the roof (which, actually, I see I forgot to post news about here) but, because the sextant required changes to internal and external models, UV maps, new shadow-bakes and textures, the changes were far-reaching.  While I was at it, I swept in some other improvements, too.

 

Apart from any truly unsightly aberrations, black triangles, gaps or thrashing I find during testing, that's it cosmetically.  I can now clear the decks and concentrate on programming.  More on that later.

 

Guy. 

[Nicola_M]

Wonderful.  This is going to be a "buyer" for me.

[hobofat]

I personally like the new feel to the cockpit (since I'm in a simulator might as well go back in time!). Your attention to VRAM budget and optimizing for online flying is a huge plus!

[guym-p]

Thanks!

 

And, yes: I will produce a version for X-Plane 10.  I have already taken an earlier alpha and modified it for X-Plane 10 to understand the differences.  The basic changes to enable the electrical, hydraulic and autopilot systems to work properly are very easy.  There are some small but important changes to the object files, mostly to do with how things are lit.  The Rolls-Royce Avon Engine plugin will need totally different thrust curves and an independent test programme to make sure they are accurate.  It will take time, but days or weeks rather than months.

 

The biggest nuisance will be the need to maintain development paths for two parallel products, which is why I have decided to develop the primary model for X-Plane 9.70, and only split them at the last possible moment.

 

Guy.

[steven winslow]

...I have decided to develop the primary model for X-Plane 9.70...

 

Thank you!  Thank you!  Thank you!  This makes me a very happy man!!!  I still do most of my flying in 9.7 even though I have V10 and keep it updated.  The changes to the views controls still irritates and frustrates me to no end.  9.7 for me is still a much more pleasing experience.

[guym-p]

I've finished the performance model at last!  It's taken ages because the full series of tests takes more than a day to complete, even with two computers running and climbs and descents running automatically, via plugins.  It wasn't always necessary to test everything, but I logged no fewer than 116 iterations, ranging from a simple digit change to overhauling a large chunk of code.  That's a lot of testing ...

 

I've written a report.  It's a dry read, but if you're into this kind of thing, you'll appreciate it:

http://dh-aircraft.co.uk/news/files/e744a904ffa2f49adffbe4c7791de9f2-105.html

 

Guy.

[Nicola_M]

Wow, that's some attention to detail.

[steven winslow]

Impressive!  But most definitely a dry read for this math-challenged guy!

[guym-p]

Thanks!

 

For an airliner, I believe that these things need to be right.  At the other extreme, something purely aerobatic must "feel" right, and let the fuel consumption do what it likes.  Obviously, best of all is to have one's cake and eat it, but if there is a conflict in some aspect of the flight model, then an airliner (and probably a bomber) would be biassed towards fuel consumption and range, a Pitts Special towards pure handling, and a fighter somewhere in between.

 

It's also a challenge.  I had some correspondence with a commercial simulator builder who is convinced FSX is "better" because simulator training is largely procedural, and the model must perform to the numbers.  Of course, FSX does that very well.  I wanted to see whether that level of statistical accuracy is also possible in X-Plane — and it is.

 

Have I therefore "wrecked" the Comet, or turned it into something dull and unresponsive?  No — I don't think so.  Thrust, drag and SFC were already in the approximate range.  All I've done is to give them a nudge to make them closer still.  Nothing has been tuned out or toned down.  "Blade element theory" is still fully engaged, the way the Comet handles is still pure X-Plane, and free to respond to its wind and weather.

 

Guy.

[scubajuan_new]

Very impressive work! I just discover this development and will certainly follow closely until it's release.

[Nicola_M]

How are we doing, Guy?  Hope it's not keeping you overworked. 

[guym-p]

Sorry for the lack of news.  I'm deep into C programming at the moment so there's not been anything complete or demonstrable to report.  I'm really pleased with the statistical performance in XP10.21.  Performance in XP9.70 is good, too, but it all seems to "click" in 10.

 

Now I'm working on systems plug-ins.  You may remember from earlier news articles that there are 400 individual steps necessary to get a Comet Series 4 from absolute stone-cold start to take off and climb-out.  Without any plugin whatsoever, just using built-in X-Plane datarefs, I could complete 135 items on the official check list — as demonstrated in the YouTube videos I published quite a long time ago.

 

When you read the real check list, you quickly realise that a large number of items are about communications with aircraft and ground crew administration, which are not simulated in any way in X-Plane.  Also a number of visual checks where there is no failure mode in X-Plane, therefore checking them is irrelevant.  So I will never achieve a ratio of 400:400.  Even so, it's an interesting ratio.

 

The motivation behind the plug-in is not to increase that number, although that will happen as a matter of course.  It is to make those things that do work function better.  For example, in the YouTube video, I had used a lot of generic light switch datarefs and sliders in order to be able to move levers to satisfy an item on the check list, but the levers didn't actually do anything.  Now they will.  I appreciate that this will be of no interest to some users, therefore there will also be a cheat button, effectively skipping to the end of the check list.

 

Progress is steady.  I've finished the flight instruments — debugging them at the moment.  The DME radios are now completely separate from the NAV radios, and the distance function works over a longer range than the standard X-Plane instruments.  All radios require not only the relevant electrical bus to be live, but also the inverters and radio power switches at the navigator's station; the navigator's station has its own independent ADF radios (ADF 3 and ADF 4).  That only describes a limited area of the whole plug-in, but it gives an idea of what I've been up to recently.  The next steps are to sort out the electrical supply, and the fuel and hydraulic systems.

 

At this stage, I am not going to replicate historical navigation systems like LORAN or DECCA.  That's a big project that's outside the scope of the Comet model.  It requires the infrastructure of transmitters and their correct geographical locations as well as the simulation on board an aircraft.  I don't have that information, so it would require a fair bit of research and correspondence before any work could actually begin.  I have been discussing this with other members of the UK X-Plane Development Team: we're all interested in aircraft from this period, and it would be great to have this equipment available, but we haven't begun to assess the scale of the task.

 

I am also in quite a busy period with my "real" work, and will be on assignment abroad for the whole second half of September, which will cause something of a hiatus for the Comet project — but I'm still very much on the case!

 

Guy.

[Orcair]

Great news! Looking forward to her 

[guym-p]

It's been a while since I reported anything new.  Cosmetically, the Comet hasn't changed, so I have no juicy screen shots to show off, just continuing work in code.  Today, however, I had written enough and compiled enough to see various branches of the project working together.  Nothing comprehensive, but it demonstrates the level of detail I'm working towards:

 

http://youtu.be/ORzg_Q2dfl4

 

Guy.

[ilias.tselios]

Great job!

[Airbus]

Excellent work as usual Guy!

[hobofat]

Absolutely outstanding work. And everything is so crisp and legible.

[guym-p]

Thanks!

 

I've just finished the fuel system plug-in.  Everything needs testing properly, of course, but the code is finished, a few changes or additions to the model have been made, and it's working well.  A fuller write-up is here:

 

http://www.dh-aircraft.co.uk/news/files/fb38443eed8229868c16b7817042f947-107.html

 

The job list is getting smaller ...

 

Guy.

[Claude D]

The job list is getting smaller ...

Guy.

 

 

Good news

slowly but surely !

great work !