Thread: How things are designed these days

Originally Posted by wheelinround

Having worked as a mechanic seeing machined bell housing line up holes out by 2 to 3 mm as well as input shafts out into flywheels by 1.5 which throughout my gear box badly.

Originally Posted by nearnexus

Yes, we have glaring examples of this in Toyotas that speed up to flat out in the USA, VW Golf that shut down at any old time, jet aircraft where thrust is reduced on stall warnings, the list goes on.

You've got to be really careful here guys - There are some issues that are genuine design issues where someone hasn't thought something through, but a lot of this sort of stuff is not a design issue as such but a failure of process. In the modern automotive industry in particular there are all sorts of processes in place to limit the possibility of something going wrong, so for some of this it may not be the original design engineer at fault but that the checks and balances are not there or incorrectly applied.

Normally if an automatic system goes wrong the safer alternative is to shut things down but some things you can't and then it is a matter of recognising that and altering the response. Trouble is, as things get more and more electronic finding a way to deal with all the fault possibilities get harder too. I'm not sure that there is a B-double sensor on the market yet though. Ray is right, in that the electronic's response was not appropriate for the situation but then again it could have been appropriate for all the fault situations that were reviewed.

As Pete will tell you though, a lot of pilot refresher training is to deal with things that are unlikely to go wrong but if they do will spoil your day. I don't know all the details of the failures mentioned, but sometimes incidents are compounded by the people involved doing the wrong thing. In this day and age people seem to want risk free existences without acknowledging that there is risk in everything and that electronics can only do so much in compensating for inappropriate judgement or behavior. When doing Failure Mode Effect Analysis (a common automotive risk mitigation tool) we were always told to analyse the possible not the improbable. Software failure would have been dealt with. Software failure + B double right up your bum? Unlikely to have been covered.

Anyway, a long handed way of saying that not everything that goes wrong is a design issue and it is not solely the original designer that is always at fault.

Michael

There's also the cases where an implement can be designed exceptionally well... but what makes it out of the assembly plant bears little resemblance.

One of my long departed uncles used to design hifi valve radios/amplifiers. As a young kid we had a cheap(ish) valve radio in the kitchen and one day Mum told me he'd designed the innards. Years later, I asked him why he never mentioned it (he was the type to mention all his past achievements at every family meeting, so his omission always seemed to me to be a bit odd, considering it was in plain view of everybody... ) and his reply went something like "I designed a good radio. The penny-pinchers pulled parts out of it until it stopped working, then put the last one back. That's what that thing is. It's not mine."

I suspect things haven't changed that much...

Originally Posted by Gammaboy

Castlemaine Rod shop bellhousing?

Toyota Crown was first I came across then Landcrusier there were others.

Originally Posted by Pete F

Can you provide specific examples of this?

He read it on the internet. It had to be true right?

Originally Posted by Ueee

He read it on the internet. It had to be true right?

I was referring to the crash into the forest (in France ?) of the (A380?) during a test flight where the software reduced power when the plane should have been climbing. Was widely publicised as a software fault. Everyone was killed. Some videos around showing it occur.

Originally Posted by nearnexus

I was referring to the crash into the forest (in France ?) of the (A380?) during a test flight where the software reduced power when the plane should have been climbing. Was widely publicised as a software fault. Everyone was killed. Some videos around showing it occur.

It's difficult to even determine what you may be referring to, that post being so devoid of any fact, but I presume you're talking about this incident?

https://www.youtube.com/watch?v=Y1FKAIrb0fQ

That being the case, it was an A320 and not an A380, the latter has in fact not had a hull loss (touch wood) to date. It was not a test flight, it was a fly-by at an airshow. The aircraft operated precisely as it should have operated and there was no "software fault" nor did the software "reduce power". Jet engines don't accelerate from low thrust settings very quickly, which is part of the reason we use very high drag on approach to ensure the engines are "spooled up" ready to go-around if needed. Not everyone was killed, indeed everyone survived the accident. Of all the passengers on board, only 3 were killed, and then only because they either couldn't or didn't evacuate the aircraft after it came to rest. It was however in France, yes.

He read it on the internet. It had to be true right?

Indeed! ... or not even

I'm not trying to pick on anyone here, so before being too offended try to understand the point, that is firstly just because somebody says something on the net doesn't make it true, but in relation to this topic, this is exactly the type of instance where BS just gets perpetuated as fact. Assuming that was the incident being referred to, the engineers copped a lot flack after that incident for designing a "faulty" aircraft. It put Airbus' fly-by-wire project back years as a result (indeed there's still significant scepticism by many, despite it being used by Boeing now also). Once the investigation was complete, it transpired that the aircraft had performed precisely as it was intended to.

But hey, carry on.

Originally Posted by Michael G

You've got to be really careful here guys - There are some issues that are genuine design issues where someone hasn't thought something through, but a lot of this sort of stuff is not a design issue as such but a failure of process. In the modern automotive industry in particular there are all sorts of processes in place to limit the possibility of something going wrong, so for some of this it may not be the original design engineer at fault but that the checks and balances are not there or incorrectly applied.

Normally if an automatic system goes wrong the safer alternative is to shut things down but some things you can't and then it is a matter of recognising that and altering the response. Trouble is, as things get more and more electronic finding a way to deal with all the fault possibilities get harder too. I'm not sure that there is a B-double sensor on the market yet though. Ray is right, in that the electronic's response was not appropriate for the situation but then again it could have been appropriate for all the fault situations that were reviewed.

As Pete will tell you though, a lot of pilot refresher training is to deal with things that are unlikely to go wrong but if they do will spoil your day. I don't know all the details of the failures mentioned, but sometimes incidents are compounded by the people involved doing the wrong thing. In this day and age people seem to want risk free existences without acknowledging that there is risk in everything and that electronics can only do so much in compensating for inappropriate judgement or behavior. When doing Failure Mode Effect Analysis (a common automotive risk mitigation tool) we were always told to analyse the possible not the improbable. Software failure would have been dealt with. Software failure + B double right up your bum? Unlikely to have been covered.

Anyway, a long handed way of saying that not everything that goes wrong is a design issue and it is not solely the original designer that is always at fault.

Michael

Michael Having worked from chasis level through to putting number plates on of Car's, Trucks, Buses & Coaches, Ambulance bodies.
Initial pre-delivery checks at dealerships and repairs, restorations.
Knowing design teams no longer work together but on each individual bit then try marry them up to fit parts coming from around the world

.. I have seen some incredible B&^% up. Electrical wires worn through, burnt, inaccessible nuts, bolts, screws which should be easy for need of general service requirement. Oil filters, fuel filters, break lines, handles, key locks. Holdens HQ if you had 3 or more keys attached to ignition key an turned the steering wheel the wrong way the keys dropped out and locked the steering.

The time Nissan put a car/4x4 on the road with blessing of many gov depts even the NRMA gave it a bonzza write up.........it had no reversing lights.

I understand why people want/demand risk free design those who are designing and engineering and running the show go through Uni for years, they get paid what mere mortals make in a life time or take a life time to make. They are supposed to be the cream of the crop who have proved they can out do others.

I'll mention the government website on re-called goods which are dangerous, can and have killed.

Its simple to blame the fellows on the front line doing the assembly its how ever sad we see little of Journeymen these days fellows who have worked their way through from the ground up in all industry. These fellows can see more and avoid more disasters because they have been hands on.

Edited to add. I still have respect for good engineers and other proffesionals who take their job and do it well. For those who must stop and pass it on becuase they don't know the next faze or use the excuse "Its Not My Job" sorry at all levels from the ground up please let someone who can do the whole job take over.

Hi,
Every one seems to have forgotten, or is skating around the old chestnut

"Designed by a comity"

Regards

Who here has not not worked on that vehicle that was fitted with an engine where the engine oil filter,oil drain or some other part is almost inaccessible.

I know its a problem where the same engine or gear box ,or diff is used in a number of models. The computer yells (edit-should be tells) the engineer or designer that there is clearance.

What the computer does not know is what clearance humans need to use said space ,but are not equipped with multiple elbow joints to fit into impossible spaces.

Yep! I'll agree with Wheelingaround and further add that the engineer./designer should be made to work on the prototype changing out those consumable parts just to understand what the end user has to deal with.

Grahame

Grahame I can assure you that computers don't "yell" at anyone, they're a tool to accomplish a task. I can also assure you, that any mechanical engineer worth his/her salt working on a design such as you've used as an example would know full well just how much clearance is needed by a human hand.

I think what some people missed, is the emphasis that engineering is all about compromises. That's not to say there isn't bad engineering, there's certainly no shortage of that, but often it's a case where there hasn't been enough (or any, with some copy attempts) engineering input into the final product. Just because you found a oil filter difficult to access doesn't mean that it's "bad" engineering. Maybe one of the design constraints was the volume and specific dimensions of the engine bay made it difficult to impossible to locate it anywhere else.

I think the problem here is the engineer designs an engine to be an engine, but the body designer is given a 3D silhouette of what has to fit in the engine bay that he designs. By the time the fopar comes to light all the tooling has been set up and deadlines are past. If the problem gets far enough up the management chain and it's a firm that cares, they may design a special tool for the standard tool kit.
Regards

Originally Posted by Pete F

Grahame I can assure you that computers don't "yell" at anyone, they're a tool to accomplish a task. I can also assure you, that any mechanical engineer worth his/her salt working on a design such as you've used as an example would know full well just how much clearance is needed by a human hand.

I think what some people missed, is the emphasis that engineering is all about compromises. That's not to say there isn't bad engineering, there's certainly no shortage of that, but often it's a case where there hasn't been enough (or any, with some copy attempts) engineering input into the final product. Just because you found a oil filter difficult to access doesn't mean that it's "bad" engineering. Maybe one of the design constraints was the volume and specific dimensions of the engine bay made it difficult to impossible to locate it anywhere else.

I think that it's more the case that the engine was designed for a totally different vehicle or even application, and to save costs doing some minor redesign, like changing the location of a distributor or oil filter, the old design is just used as is. In the new application, there may not be access points that were in the original application, so what should be a simple job becomes much more difficult than it should be. While I have no first hand experience of it I seem to remember a report of a Rolls Royce limousine which required the removal of one of the front mudguards, in order to change one of the spark plugs. If true presumably RR thought that if you could afford the car, then regular servicing would not be an issue, and no owner would have to worry about it.
I have had to change hydraulic hoses on a Cat 988 loader by going head first down from an engine hatch, to get to the end hydraulic coupling, and then being hauled out by the ankles when job done. The 988 was a mongrel to work on compared to the larger 992 which in general had much better access.
It seems to me that many manufacturers know that most will only find the bad design points after they have purchased it, and by then you are stuck with your purchase. If you actually knew beforehand you would at least have had the chance to make a different decision if you thought it important enough.

It's hard to be be critical when you don't know what the constraints that the designers were working under. Cost constraints are always a factor in the auto industry. I don't know where serviceability ranks in the order of constraint priorities, but I'm sure that cost and functionality comes first. No point being able to change the oil filter easily if it's so expensive you can't sell it.

That Air France 296 A320 crash was the subject of one of my favourite shows, "Aircrash investigation" Although I notice the youtube video is from the same show under a different name "Mayday", must be aired in some other country under a different title.

Mayday - S09E03 - Pilot Vs Plane - Video Dailymotion

Moral of the story, don't do a 30' slow speed flyby in a 40' forest..

Ray

I think Ray is quite right, while I'm just as guilty as the next guy as swearing at the guys who designed something (hey my last car was Italian!), in all fairness I think the engineers themselves cop flack they don't always deserve. We typically don't know he constraints the engineers operated under. Maybe they needed to fit engine ABC into engine compartment XYZ, the latter's dimensions provided courtesy of the marketing department? As the final consumer we're rarely privy to what the constraints were at the time, hence what compromises needed to be made.

I guess if you extend the definition of "engineering" to mean the whole design process by all departments, criticism is often well deserved, but I do often think the poor old engineers cop a lot more criticism than they often should.