Kevin Horton said:This event highlights one big weakness with the classical DO-178B approach to software certification. . .
Ralph Kramden said:I have been on the F-22 software program. Twice. It was horrendous. It doesn't suprise me in the least that they had that problem. They have had many worse. For a while there, they were lucky if the flight software went for a whole 20 minutes without crashing. Nuff said.
REK
Kevin Horton said:Bottom line - you can only foresee the things that you foresee. No matter what standard the code is written to, there will be bugs that don't show up until it is in service, because no one foresaw this particular set of circumstance. Expect the unexpected. Have backups, or a fall back plan. Trust, but verify.
rickmellor said:Ralph, that's a little disappointing... I've always had this grand vision of life on defense contracts doing cutting edge military implementations and all the rigor and dedication that goes into it.-Rick
rickmellor said:I thought the DoD oriented methodologies (I don't know them) would manage out these problems.-Rick
SvingenB said:I would believe that some simple cosine tranformations somewhere in the unit would fix the stuff ?
SvingenB said:These kind of problems happends all the time in all technologies due to inexperienced engineers or too little knowledge of real life operation conditions.
...
This only means one thing; they cannot be trusted to operate satisfactory in a real life installations with the reliability and robustness you would expect.
This only means one thing; they cannot be trusted to operate satisfactory in a real life installations with the reliability and robustness you would expect.
I'm not sure you understood me correct. What I basically mean, with some other words, is that a general computer engineer (software/hardware) with no pilot experience is no better fit making an EFIS than a general mechanical engineer with no pilot experience is designing an aeroplane. A pilot, proffesional or amateur like most of us, with some basic degree in engineering would stand a much better chance of making a good product. The reason is requirement, specifications, robustness etc. To make a good and robust product you must be able to ask the correct questions, not merely make a solution according to some standard and a set of requirements made by others (this is basically what differentiate a good/excellent and an average/unexperienced engineer). I have worked in R&D for almost 20 years, so in all modesty, I believe i know a thing or two about this.nicolcarstens said:Sure: "trivial" transformation matrix. But I guess it was not part of the design requirement?! I would not say there is something "wrong" with the product ... as long as the designer made it clear that the customer is using it in a way it was not designed to be used! This could pass as a "bad" installation...
How did you get to this conclusion?? The technology is not the problem here, but the design team and/or the requirement specification and/or the installation, not so?
I wouldn't be so sure. I only fly VFR and keep my eyes outside of the cockpit most of the time (at least I try to ) In my requirements for instruments, readability is way up there. I want to be able at a short glance to see what the speed, alt and RPM is. So far have have not seen any EFIS that can compete with round analog gauges in terms of readability. It is not without reason that automotive instruments, although fully digital, have analog gauges and dials of the clock type.nicolcarstens said:Go look at that picture of the EFIS vs steam again (few posts back), and tell me "computer things" are bad and mechanical systems are good (with a straight face).
Kind Regards,
Nicol.
Engineer with MGL Avionics.
PS: ... don't confuse an EFIS with the "thing" on your desk... If computer HW and SW was designed application specific, the world would be a lot better place...
SvingenB said:Anyway, my reaction was that the above posts suggests that EFISes are (mostly??, sometimes??) made by people with little adequate training and engineering experience and zero piloting experience. Then I say that equipment made under such circumstances are way below the standard of equipment made under optimal circumstances. Then, given the fact that we are flying experimental aeroplanes with experimental (uncertified) equipment, can you by looking at the screen of the EFIS say that this one is good and that one is bad? No you can't. Therefore they cannot be trusted, and you would need a backup of some kind.
Well, SvingenB, after nearly 20 years, you seem to have missed noticing a few things about engineering. And, that is that most of the engineers who design advanced products cannot afford to actually own them. Many never even get to actually operate equipment that they have designed, or been part of the design.SvingenB said:... What I basically mean, with some other words, is that a general computer engineer (software/hardware) with no pilot experience is no better fit making an EFIS than a general mechanical engineer with no pilot experience is designing an aeroplane. A pilot, proffesional or amateur like most of us, with some basic degree in engineering would stand a much better chance of making a good product.
...
I have worked in R&D for almost 20 years, so in all modesty, I believe i know a thing or two about this.
Regarding proffesionality, an organization may very well be 100 times as proffesional as one single individual could ever be (Apollo program). Yet an organization may also be 100 times worse than even the worst individual (Challenger). I think you are stretching it a tiny bit too far. Not a single person ever flew with V1 and V2 either, yet their primitive "computers" guided them to their targets every time. The apollo program, one of the greatest achievement in human history, can hardly be compared to making an EFIS for experimental amateur built aeroplanes.thallock said:From your logic, I guess you would also say that the engineers who designed all of the hardware and software that sent the Astronauts to the moon could not have done a good job because none of them ever rode in a capsule. By that same reasoning, Burt Rutan was a poor choice as chief designer on Space Ship 1 because he never flew it. This logic is flawed, to say the least.
Tracy (programming since 1969, and know a few things about engineering ).
I am sorry, that was not intended at all. But You didn't disagree with the notion that some (a lot in fact) of the software made for aviation is substandard within any type of quality measurement.Rainier Lamers said:Granted, your comments are likely valid for the big guys, but certainly not for us - in fact they hurt slightly.
...
...
Rainier
CEO MGL Avionics
Very possible. If so: sincere apologies. I hope we can still walk away as friends after this "debate"?SvingenB said:I'm not sure you understood me correct.
Where did you find this??SvingenB said:Anyway, my reaction was that the above posts suggests that EFISes are (mostly??, sometimes??) made by people with little adequate training and engineering experience and zero piloting experience.
thallock said:Tracy (programming since 1969, and know a few things about engineering ).
Furter up there when referring to the F-22 and the posts following.nicolcarstens said:Where did you find this??
SvingenB said:It seems to me that you MGL persons are reading things between the lines I have written that simply is not there.
It does not always take much for a small misunderstanding to blow up. Especially if you are not listening carefully to each person.SvingenB said:No problems Nicol I guess I just write in confusing ways.
thallock said:Kent,
Yeah, I have some, but I am hanging on to it in case the Museum of Computer Sciences happens to loose their last sample. Same as my SOL16. You never know when those things will be worth something, again. You may want to stash a few floppies, just so you can show your grand kids in a few decades.
Tracy.
kentb said:Oh yea, back when floppies were floppy. I through away all my 8 inch ones. I should have saved some to show off. Oh well I still a couple of boxes of 5 inch ones.
Kent
Ralph Kramden said:Floppies? Hehe... I remember programming PDP-8s with the toggle switches on the front panel. Magnetic core memory, baby! I was a power user with a thermal paper teletypewriter with a 1200 baud acoustic modem!
--REK
Ironflight said:Atari 400....with a cassette drive.... Good times....
Paul
Ironflight said:Atari 400....with a cassette drive....programmed it in machine language . . .
Ralph Kramden said:You know, getting this thread slightly back on track, I bet I could program one of the Ataris to be an EFIS. I still have the old programming manuals ("De Re Atari" or some such). I would be able to wire up and program some sort of interface to the sensors out of its parallel port. It would be a cool novelty...
--REK
Now you are bringing back memories! Only we used to use a 300 baud acuoustic modem and 8" floppies for storage. of course that was on a CPM system and I wrote an entire payroll system in FORTRAN. Now, as a CTO they don't let me touch that stuff any more and for good reason.Ralph Kramden said:Floppies? Hehe... I remember programming PDP-8s with the toggle switches on the front panel. Magnetic core memory, baby! I was a power user with a thermal paper teletypewriter with a 1200 baud acoustic modem hacking into Bell Labs and Fermilab in the 70's!
--REK
Rainier Lamers said:. . .perhaps the 3D terrain is going to be a little much at the high frame rates (it's mostly assembler on a 200 Mhz ARM9 RISC running directly from single cycle TCM (tightly coupled memory - 5ns cycles !)).
. . .
N941WR said:Now you are bringing back memories! Only we used to use a 300 baud acuoustic modem and 8" floppies for storage. of course that was on a CPM system and I wrote an entire payroll system in FORTRAN. Now, as a CTO they don't let me touch that stuff any more and for good reason.
The funny thing is everyone on this list (in the states) gets the results of some FORTRAN code I wrote in the early 80's every sunday. Can you guess what that might be? (Talk about drift!)
Ralph Kramden said:Floppies? Hehe... I remember programming PDP-8s with the toggle switches on the front panel. Magnetic core memory, baby! I was a power user with a thermal paper teletypewriter with a 1200 baud acoustic modem hacking into Bell Labs and Fermilab in the 70's!
--REK
I think that was an L or 1 and that keyboard was called a typewriter.Ralph Kramden said:I have memories of IBM terminals, keypunch machines and cards and some guy in an office whose job it was to load the cards into the reader for you and then later tear off the paper from the line printer and put it into a slot in the wall for you. I seem to remember that on some keyboards you didn't have either a "1" key or an "i" key - I can't remember - you were supposed to use "i's" when you needed a one.
--REK
N941WR said:. . .On my desk at home I have punch card and a metal card plate as a soviner from days gone by.
Ok, I'm feeling VERY old!
Ralph Kramden said:And then sometime later Heathkit came out with it's line of kit built computers...
Ralph Kramden said:Can I come and work for you?
--REK
Rainier Lamers said:Yes, but you won't like it.
Our R&D is run along the lines of a Roman slave galley.
A set of drums is used to keep the pace and a whip is used as motivation. Payment is in the form of stale bread and thin soup.
Employee of the month gets to sleep right side up - everybody else sleeps hanging by their feet...
Nah, not that bad...
Rainier
CEO MGL Avionics
Now let's not blame the thread for that! We know your dark CP/M secrets!Rainier Lamers said:My guys here are following this thread and are starting to look at me funny.
I wrote space invaders for a TRS-80 color (4 colors including black !).
Ralph Kramden said:Do you remember Commodore 64 books that explains some of the unimplemented /unintended /undocumented opcodes that their processor had and suggesting cute ways these undocumented opcodes could be used?
--REK
jdmunzell said:Wow...what an interesting thread! I'm surprised however that nobody has mentioned something else pretty basic!
Comparing the safety aspects of glass vs round dials and IMC flying and the like can suddenly take a back seat when the one single engine powering our RVs decides to go on holiday!
Perhaps this is where ones' situational awareness (SA) can help the most. Knowing at the push of a button ( ie. glass technology) where any quick diversion airport is located relative to your nose of the aircraft could make a big difference in the outcome of a engine out emergency. Not that tracking your progress on your sectional/enroute chart wouldn't do almost the same thing....like we did in the [/I]old days.
Just to stir the pot a little here... I sincerely hope that when RVers are flying around in IMC, that you guys are doing so in weather that at least gives you a fighting chance at living! if you suddenly turn into a glider!
Iv'e done more than my share of flying point A to point B in a single engine airplane with no autopilot and routinely shooting an ILS to 200 and 1/2. I'm so glad I survived that phase of my career, and choose now to not ever have to do that again. ....at least not to plan on it anyway, and that's KEY!
Think about it! You're flying along in cruise in IMC ( or VMC on top, doesn't matter), and suddenly your engine injests a valve or two and comes to a complete stop!! ....Period!.....engines' done!....toast!.....not coming back! If the weather below you is at minimums or worse.....and you lose that engine for good.... my friends,....you....are.....screwed!!! end of story period!
That's all. Just a little food for thought. Ya'll be safe out there now, ya' here!
Mixed points. First the need to cross check does not negate the usefulness of Glass. A 6-pak by its very nature demands cross checking. I'd have to ask you, have you flown with glass or IFR rated?Ralph Kramden said:It seems to me all the cross-checking you need to do in an experimental glass panel system sort of negates its usefulness. Do you trust it or don't you, and my point is that you ought to be able to trust it with your life. You shouldn't have to be crosschecking with another system - the system you are flying should be doing that for you.
First NOTHING stands on its own, especially if you can avoid it. Airliners do have 2 or 3 separate "glass systems" working together, but they STILL have backup instruments and two pilots backing each other up. The single pilot is the real danger in IFR ops. You're confusing common sense and higher levels of redundancy with an indictment of glass reliability. Glass allows you to add more backup you don't get with a 6-pak alone. Yes the 6-pak divides it up into 6 instruments. That's a true point, but saying "glass is not perfect" so it's not safe is not honest or a good reason to deny all the other advantages, IMHO. No one thinks one instrument system is a good idea. Rent a mighty C-172 with a 6-pak, there's little redundancy w/ a Vac failure. Many have died on partial panel.OK, ok, we are talking about RVs flying most of the time in daylight VFR. The glass panels out there will do just fine, I am sure. Have fun and watch the pretty pictures. And the way most of the non-professional pilots are trained we would be expecting to be ready to switch over to a partial panel at any moment in IFR flight. So the need for super-reliability in the RV is not there quite as much as in the airliner. But if it's an instrument and you are looking at it for flight decision making it really (*really*) needs to be trustworthy standing all on it's own.
Bingo! We're very equipment dependant. All pilots, even the intrepid C-172 pilot shooting an ILS to mins with one vac pump and one crankshaft is a trusting dude. I've studied risk management, CRM - crew resource management. It gets to a point of philosophy, you've got to admit flying can kill you. Quote:But you know who has to have had guts? Those F-111 or B-1 aircrews that have/had the terrain following defensive systems. Imagine how much they are trusting the system to function correctly without their intervention.
Cool you will be very happy and I safe. Going w/ electric gyro or vac?What am I putting in my RV? A classy looking six pack and some sort of color GPS. If I had an extra $10,000 (and I don't!) I'd like to get a Garmin 430 and the ADS-B enabled transponder.
There's no need to justify you choice, but consider cost of your full TSO'ed analog/mechanical flt instruments. A TSO'ed airspeed is about $650, china knock off about $150. Once you add up all the instruments, you'll have about (guessing) ++$6k-$8k in a top line 6-pak analog flight instruments. You won't use cheap stuff, right. You can buy two Dynon's, with pitot/static instrument backup and T&B/coordinator for less. Add a nice wing leveler you have something that's safe & redundant, without spinning wheels & gears. Aerobatics & mechanical gyros, plus IFR don't mix IMHO. Cheap instruments are garbage and you will throw them away eventually.Hope I am not rambling too much. For me there is the simple joyful hobby of flying a simple airplane and then there are the technical aspects of designing cutting edge avionics for airliners. Sometimes its hard for me to separate the two approaches to flying. RK
I don't think many clear thinking pilots will routinely fly IMC in a single engined aircraft - it would be just a matter of time until something goes badly wrong.
With most "better" EFIS's you may have a better chance compared to the six pack as a good EFIS can give you a 3D terrain view - this can be more helpful in establishing situational and spacial awareness if the terrain view is suitably coupled to the horizon sensors and thus approximates what you would be seeing out of the window.
Cheap electric gyros don't empress me either. The good ones start at well over $2,000 to $4,000.
Ralph Kramden said:George -
Are glass panels safe? I don't know, probably, but I sure wish they were a few more steps along, technology-wise.
--Ralph