US Defence Cuts Announced: F-35 program delayed AGAIN

[evangilder]

Thanks for the input, davparlr, I had a feeling things might be modular as a lot of things on the civilian side have gone to that for costs. Looking at projects that I have been involved with in the past, writing new code will always be quicker and cheaper in the long run than porting code to a new system. It sounds simple enough in the short term until you get chest deep into the project and find a bunch og unknowns, traps and god knows what else.

Having more programmers or developers doesn't always make things faster. Even with thorough design specs, how one interprets that and codes to it can sometimes lead to one piece not playing well with another, or providing contention for memory/processor, etc. Granted that having more eyes on a problem is not a bad thing, just that sometimes having too many engineers on something can lead to a mess.

Testing and QA cycles are also a big time resource. While I certainly see the need for thorough testing and "negative tests", that is probably where a big chunk of time is taken.

 

[ivanotter]

Evan, yes integration and testing is and will always be a huge task, especially in these environments.

My analogy is the development of the Airbus A380, which must have a similar degree of complecity (although different systems).

That did not take 15+ years, neither did the software development take billions.

The degree of safety, etc must be similar (maybe not identical), wherefore we should be able to compare.

After all, I cannot imagine trhe supplier of , say Storm Shadow not providing some API's for the software development and integration. That would enhance the sale I should think.

Or have I missed something?

 

[davebender]

Britain and several other nations spent a lot of money to develop the Eurofighter. To me it appears a pretty good aircraft.

Why didn't Britain fully support the Eurofighter program by designing a CV version and designing their new CV(s) to use that aircraft?

Too late now but Britain should have considered these things during the 1990s.

 

[Matt308]

Integration testing is important, but you cannot possibly test all the millions and millions of SLOC to V and V its solution space. This is why there is now such a huge emphasis on development processes and tool qualification. As a systems engineer, I would argue that the getting the high level requirements and the HW/SW derived requirements correct to begin with is likely more important than simple exhaustive test techniques. The military is finding itself in a real bind by wanting to use commercial off the shelf (COTS) HW/SW to minimize cost, but cannot ensure that this same COTS HW/SW does not contain additional functionality which cannot be assessed via traditional process assurance, and thus may introduce an unsafe condition. There is a whole industry rising up around this discussion and the answers are not readily available to choose a path amicable to everyone's business case.

Eurofighter? I think that folks are staying away from it because it is a platform that really will not have a viable air-gnd capability until Tranche III. Thus it is only an air superiority platform from the outset.

 

[ivanotter]

Without turning it into a discussion on software, there must be a reason why developers develop for Windows/DOS/NT and not for CP/M. Cost, functionality, market demand and support.

The key question is still in my mind: do we need a paradigm shift in development cycles or is it handed down from high altar that developments cannot be done faster?

I still like the liberty ship analogy: nobody could imagine that a ship could be built in days. When somebody did it, it had an impact on everything.

 

[davparlr]

Evan, yes integration and testing is and will always be a huge task, especially in these environments.

My analogy is the development of the Airbus A380, which must have a similar degree of complecity (although different systems).

That did not take 15+ years, neither did the software development take billions.

The degree of safety, etc must be similar (maybe not identical), wherefore we should be able to compare.

After all, I cannot imagine trhe supplier of , say Storm Shadow not providing some API's for the software development and integration. That would enhance the sale I should think.

Or have I missed something?

You are missing a lot. Comparing the A380 development to the F-35 development is like comparing Toyota building the Avalon to General Motors building the Volt. Airbus already had much integration experience in building the large, four engined A340, and I would be surprised if less than 60% of the integration effort was transportable to the A380, including the flight control programs, tailored of course. Very little integration is transferable to the F-35 from any previous programs.

The same can be said about the difference between commercial airliner and military fighters. I am going to make up a thumb rule, I'll call the "davparlr estimating factor" (there probably is a real one out there somewhere). The complexity of design, integration and test of a system is a geometric progression of its component subsystems. For example, the design, integration and test of a system that has twice the subsystem components of another system, will be four times the effort required for the other system. So lets compare commercial aircraft to a military aircraft, this list is not meant to be all inclusive but representative.

Commercial systems

Crew station
Engines
Fuel
Electrical
Hydraulic
Landing gear
Flight controls/w computers
Environmental
Navigation
Global Air Traffic Management
GPS
VOR
ILS
Communications
VHS or UHF (military)
Satellite
IFF
Radar
Weather

For a modern military fighter, all the above plus the following, I think these are typical

Radar
Air to Air

• Search
• Track (including multi-target)

Air to Ground

• Moving target indicator (MTI)
• Synthetic Aperture (SAR)
• Ground Map

Electro-optical target detection systems
IR target detection systems
Weapons
Air to Air

• Targeting
• Sidewinder
• AMRAAM

Air to Ground

• Targeting
• JDAM
• Air to Ground Missiles
• Dumb Bombs

Defensive systems (most of this stuff is classified and I don't know except it is complex}
Threat Detection

• Rf
• IR

Countermeasures
Rf

• Active
• Passive

IR

• Active
• Passive

Communications
Satellite Data Link
Local netted data link
Navigation
Automatic carrier landing systems (Navy)

Now, stick all that on a 30,000 lb airframe and make work from negative gs through 9+ gs, incredible vibrations from 20mm cannon fire and opening weapons doors at high subsonic to supersonic speeds, stress from landing on a carrier (Navy), et.al.

I think this gives some idea of the complexity of the modern fighter aircraft.

Hmm, more rambling

 

[ivanotter]

You are starting to convince me.

Yes i see there is a magnitude of complexity.

Let us take it one step further, though:

How many new systems will the F-35 carry, things never been used before? Known systems must have been integrated on other platforms; hence API's and code must be available (The cannon is not new, is it?)

So, how much would be ported from previous platforms?

What I am really aiming at is: by using more 21 century software tools and technology and integration technology, can the entire development phase be cut dramatically?

I am also seeing, thanks to this forum, that the "raw" engineering efforts are becoming less problematic as flight engineering knowledge is becoming more deep and sophisticated (let's forget the broken spar. it is not fair to harp on it I think. Bad engineering can be found in many places)

The software integration among sub-systems will then be key to any successful fighter programme.

If software development takes years and billions due to "old" technology, and ADA is "only" 40+ years old, will a quantum leap be for any country to utilise modern tools?

It has been mentioned that the CPU's in most flight environments are a bit old, the communication, based on MIL-STD bus, instead of TCP/IP, is slow and difficult to interface to. etc, etc.

I am not so hot on the "what-if" scenarios as they tend to get carried away, but i could imagine a technology demonstrator built using commercial software. Linux and iPhone interfaces, etc etc. Going overboard in IT technology, really. Can it be done? will it work? will it be safe (I don't believe in 'blue screen" in the middle of a dog fight either. difficult to re-boot or debugging, really).

What if China should use modern integration software on their new fighter? Or India, having millions of PhD programmers?

Am I totally off on a tangent? Should i accept defeat and go for conventional thinking that developments must be 15+ years as a minimum?


Ivan

 

[FLYBOYJ]

Last night I had a talk with my buddy who was on the X-35 program and has been on and off the current program. According to him the media is making the problems bigger than what they are. He says there have been over runs and delays but many of those have been driven by the government due to changes in the program induced by the agencies purchasing the aircraft. On the F-35B, as he puts it, "every nut and bolt is being scrutinized after every flight" so the flight testing is slow and tedious. As my friend puts it "there are more things that are right and that have exceeded expectations than are wrong with this plane." He thinks the media is clouding the waters to make everyone believe that all versions are in trouble.

From a LMCO employee - take it for what its worth....

 

[buffnut453]

Known systems must have been integrated on other platforms; hence API's and code must be available (The cannon is not new, is it?)

Ivanotter,

You are broadly right but there isn't one set of code for one thing and another set of code for something else. Coding has to be ultra-efficient, partly to facilitate safety testing and partly because onboard processors are always several generations behind the power available on the latest generations on the street. This means that code is reused extensively but perhaps not in the way you think. For example, I know of one aircraft where a customer decided to change the engines from those originally specified. Everything was going fine until they tested the software and it was found that the weapons controls had been affected...yes, by changing the engines. There are multiple feedback loops in control systems that can cause all sorts of problems if components send out signals that are not expected by the rest of the system.

Another factor is the need for (apparently) instantaneous response within operational flying constraints. You don't want to be in an aircraft where there is a noticeable lag between control selection and function of that subsystem or where a combination of switch inputs "blue screens" the system. This need for high throughput but with extreme levels of reliability, low power consumption and resistance to radiation (not something you find in your everyday laptop) promotes extensive use of Field Programmable Gate Array processors that are configured for specific roles within the onboard systems of which they are part. This certainly was the case when I did my flight trials training a few years back and I doubt things have changed very much.

If it was easy, anyone could do it...

Cheers,
Mark

 

[ivanotter]

Mark,

Yes you are spot-on. It it was easy everyone would do it.

Do you foresee a quantum leap in this area? What if somebody surprises us all and actually doing it? India has a huge IT sector, China could maybe?

Would this then be the next bigger achievement in a/c development rather than "raw" engineering and fancy paint? (I am building a 1/72 Raptor and the sheen is a curse I think)

Comments on this tangent?

Ivan

 

[Glider]

Last night I had a talk with my buddy who was on the X-35 program and has been on and off the current program. According to him the media is making the problems bigger than what they are. He says there have been over runs and delays but many of those have been driven by the government due to changes in the program induced by the agencies purchasing the aircraft.

From a LMCO employee - take it for what its worth....

I hae not been invovled in the development of any aircraft system but have been the Senior Project Manager on a number of IT Projects and I am willing to bet anything that the above as highlighted is a major part of the problem.

Projct Creep as its called, is always a huge factor. I wanted this and now I want that, or can you add this to it, or it was spec'd missing out a couple of 'features' to get the numbers down and now we need it. The F35 and F22 couldn't share data together or talk to each other and now we think its a good idea. Or the client didn't check that what the developer was building actually met their requirement.
What the people in power who make these requests forget is that every time you add or change something everything has to be retested resulting in a massive duplication of effort.

The list of examples available some of which I have been invovled with and others I have read about is unending.

 

[FLYBOYJ]

I hae not been invovled in the development of any aircraft system but have been the Senior Project Manager on a number of IT Projects and I am willing to bet anything that the above as highlighted is a major part of the problem.

Projct Creep as its called, is always a huge factor. I wanted this and now I want that, or can you add this to it, or it was spec'd missing out a couple of 'features' to get the numbers down and now we need it. The F35 and F22 couldn't share data together or talk to each other and now we think its a good idea. Or the client didn't check that what the developer was building actually met their requirement.
What the people in power who make these requests forget is that every time you add or change something everything has to be retested resulting in a massive duplication of effort. The list of examples available some of which I have been invovled with and others I have read about is unending.

I seen this on an "epic scale" when I worked on the B-2.

 

[ivanotter]

I used to be in IT, so yes, I can vouch for those things.

Building state-of-the-art engineering, fancy paint and basing it all on 80186 chips coded in ADA (which came with the ark) simply does not really make sense.

Sure, it must be safe and tested and working, but how many developers are maintaining ADA? Just because it is old (ancient, really) does not imply it works.

Does anyone want to guess on development times if it was all based on up-to-date state-of-the-art IT technology? never mind flying it, but as a demonstrator?

ivan

 

[jamierd]

I heard all the problems started when the test pilot couldnt reach the ashtray because he was left handed.so then they swapped it over but he couldnt reach the PS3 because the ashtray was in the way then the pop out mug holder got stuck and all hell broke loose.after this the Marine Corp found out about the PS3 and complained coz they wanted XBox live in theirs.then Mrs Burgess responsible for sewing the upholstery went on maternity leave so the whole thing is on hold till she delivers her little bundle of joy

 

[ivanotter]

very good, very...

Sad, but it could be so true. Reminds me of a new mini computer the company I worked for received from nobody knows. No OS, no drivers, nothing. We had to write it all from scratch. disks, tape, card, printers, screens, comms, all of it. Developing an OS from new is character building, and then someone comes along and asks for: "can we also just use this new screen please? the client likes it"? and this before TCP/IP, API's, Windows, CP/M. all in assembler. Thank you.

Yes, i can see it

 

[Catch22]

I know this isn't really relevant to all of the discussion going on in this thread, so I apologize.

If it is indeed the Marines' version causing the issues, would they not be able to release the other two independently from it? I know the CAF is supposed to have theirs by 2020, and they are needed by then as our CF-18s are on their last legs as is.

 

[FLYBOYJ]

I know this isn't really relevant to all of the discussion going on in this thread, so I apologize.

If it is indeed the Marines' version causing the issues, would they not be able to release the other two independently from it? I know the CAF is supposed to have theirs by 2020, and they are needed by then as our CF-18s are on their last legs as is.

YES! But the media and those against the F-35 program in a whole are lumping all 3 versions into one basket. One thing LMCO is not mentioning are the issues driving the cost over runs like government induced changes and a vendor base that his also partically directed by the government. I'm told some of the foreign suppliers are on their butts and there's little or nothing LMCO can do about it.

 

[Catch22]

YES! But the media and those against the F-35 program in a whole are lumping all 3 versions into one basket. One thing LMCO is not mentioning are the issues driving the cost over runs like government induced changes and a vendor base that his also partically directed by the government. I'm told some of the foreign suppliers are on their butts and there's little or nothing LMCO can do about it.

Thank you sir! That's what I suspected, but didn't know if there were other possible contributing factors that I'm just not aware of.

 

[FLYBOYJ]

Thank you sir! That's what I suspected, but didn't know if there were other possible contributing factors that I'm just not aware of.

The F-35 will always carry the reputation of being a money pit, thanks to the media. They did the same thing to the F-15 back in the early 1980s. The only thing that worked in the F-15's favor was the Soviet threat at the time silenced many of the critics. What the media and skeptical public miss is, yes these toys are expensive, but they're going to be around for the next 40 or 50 years. Think about how cost effective the B-52 has been!

 

[Catch22]

Yeah, there just isn't the aircraft turnover anymore. Upgrades sure, and that also will add costs, but that theoretically won't be until the reasonably distant future. If in 30 years the F-35 is still nowhere near retirement, the media will be saying that they were a great bargain.