Boeing 737 Max8 Aircraft Crashes and Investigations [Part 6] 17

[Sparweb]

This post is the continuation from this series of previous threads:

thread815-445840
thread815-450258
thread815-452000
thread815-454283
thread815-457125

This topic is broken into multiple threads due to the length to be scrolled, and images to load, creating long load times for some users and devices. If you are NEW to this discussion, please read the above threads prior to posting, to avoid rehashing old discussions.

Thank you everyone for your interest! I have learned a lot from the discussion, too.

Some key references:
Ethiopian CAA preliminary report

Indonesian National Transportation Safety Committee preliminary report

A Boeing 737 Technical Site

Washington Post: When Will Boeing 737 Max Fly Again and More Questions

BBC: Boeing to temporarily halt 737 Max production in January

 

[IRstuff]

https://www.seattletimes.com/busine...-the-777x-handing-the-job-back-to-machinists/

Seems like it was not specifically an issue with dimensions.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Tomfh]

or what's mostly going to happen is they will use a 50mm spanner and grind a bawhair off one face. .

For real? I feel guilty enough using a 13mm socket on a 1/2 bolt for something that doesn't matter.

 

[Lou Scannon]

what's mostly going to happen is they will use a 50mm spanner and grind a bawhair off one face

Begs the question, one face of the fastener, or one face of the spanner? (I think I know the answer... the spanner will become 50.8mm/2" nominal, ready to start a rampage of rounding off legit 50mm fastener hexes.)

"Schiefgehen wird, was schiefgehen kann" - das Murphygesetz

 

[Alistair_Heaton]

It never is specifically IRstuff, its always some thing doesn't quiet work right, setup time per unit is vastly more than anticipated, run time is 1-2% more because the input deck is imperial sized and the machine is converting every single measurement.

Starting end of the hull I expect was more than acceptable but by the time it had done 2-3000 rivets in a line the tolerance will be out or averaging in the same direction as the error racks up. They will have tried to fix it by doing multiple position resets but there won't have been any reference points away from the planned start points. A position reset takes time and skilled manpower and they will have been chasing their tails constantly with constant monitoring required.

To be honest most Engineers in Europe/world won't know or have seen the issue. They just have never had access or need to have ever worked with imperial geared machine tools. Everything comes with a metric feed screw. Places like India will be well up to speed dealing with it but they have loads of old imperial machine tools and loads of skilled machinists to run them.

The example I got shown as a youngster was 4 machine threaded rods 2 meters long, two done on a metric lathe and two done on imperial lathe. The same metric thread on each and imperial thread on each. Afterwards you stick like threads next to each other matched at one end then see where they start diverging. Imperial feed gearing actually handles metric better than Metric handles imperial. Ie your further up the rod before you can visually see the divergence.

Modern Hobby lathes in UK you can get conversion kits between the two. And 1960/70's German/UK industrial machine tools you can also get conversion kits I really don't have a clue if you can get conversion kits for modern mass production machinery and if you can I suspect it significantly more to change than the old stuff. To convert old stuff its a complete strip down of the gear box and changing 2-3 gears and adding 1 if going from imperial to metric.

Tomfh this is my current standard present for technician mates in Europe... and the set I have in the house.

https://www.amazon.co.uk/gp/product/B01MV459FU

The imperial stubby's are very prized items, normal solution for them is to cut the shank in half of a long one which means they have to get their arm out to go from ring to open. These they can just reverse without doubling the risk of removing skin.

 

[VEBill]

An inch in the UK over the years.

1895 - 25.399978 mm
1922 - 25.399956 mm
1932 - 25.399950 mm
1947 - 25.399931 mm
1959 - 25.4 mm exactly

Ref. NIST The Gauge Block Handbook

Link might be something like:

https://www.nist.gov/document/mono180pdf

Easier to google the title.

Probably below the real world tolerances anyway.

 

[Alistair_Heaton]

VEbill from memory a long long time ago it was inch fractions like 3/64" that caused the most issues.

You basically had to reduce your tolerance by 2 to 3 significant places. If you didn't and used 3 significant places so say 3/64 is 0.0469" when you decimalise it instead of 0.45875. Your now 1.19126 instead of 1.190625‬ 0.000625 difference. Over 1000 threads or revolutions you into over 0.5mm rounding errors. Single measurements are not a problem. Its the worm drives and gearing which rack up the errors.

To be honest the addition of tolerances across a system and the resultant I never really got to grips with at Uni. I am sure that someone can generate the maths for it. Some imperial measurements it works fine with. If they have been created decimalised to begin with that also helps. I suppose you could help maters with rounding up and down alternatively or randomly with computer controlled machinery with metric gearing. But old school 1980's Texan Oil hardware which was made and designed using inch fractions I am not so sure it would solve things completely. Better to just make it using a imperial geared machine.

The visually looking at 4 rods of machine thread produced to educate the young pups about the dangers of mixing and matching made a lot more sense to me. You could see the thread was different and even if blind you could feel it.

 

[LionelHutz]

What does gearing have to do with CNC machines? All CNC machines pretty much have a tolerance limit based on the movement screw and stepper ability to move each motion a certain minimum amount accurately. You can list this value in imperial or metric but it's still the same physical limitation of the machine.

 

[moon161]

https://www.seattletimes.com/business/boeing-secretly-testing-new-automation-for-777x/

According to a presentation reviewed by The Seattle Times at the time, a phased plan was presented to Alan Mulally, then chief executive of Boeing Commercial Airplanes, in September 2003 under the code name 777 HAL (for Higher Assembly Level).

You'd think people would have learned about mixing automation and the HAL acronmym by now.

 

[waross]

The answer may be here:
in·com·men·su·ra·ble
/ˌinkəˈmens(ə)rəb(ə)l,ˌinkəˈmen(t)SH(ə)rəb(ə)l/
Learn to pronounce
adjective
adjective: incommensurable

1.
not able to be judged by the same standard as something; having no common standard of measurement.
"the two types of science are incommensurable"
2.
Mathematics
(of numbers) in a ratio that cannot be expressed as a ratio of integers.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[dik]

That's why they made 6 pt. and 12 pt. sockets...


Dik

 

[FacEngrPE]

Here is a discussion about metric Metric Transposing Gears

http://www.conradhoffman.com/metricthreading.htm

Changing lead screws is unnecessary, and an exact conversion is possible, but not always practical as the 127 x 100 transposing gears will not fit in all lathes. Allstair's experience with mismatched threads is one I contended with when I made parfs - millwright training. If precision is needed, you need to be careful about temperature.

My experience is that usually tolerance fitup problems are usually related more to engineers not understanding how to correctly describe requires tolerances, than they are to the measuring systems used. Sometimes shops do not get the measuring part correct.

Geometric Dimensioning and Tolerancing (GD&T) was supposed to help us with this, ASME Y14.5 Geometric Dimensions and Tolerances (GD&T), ISO TC 10 Technical product documentation, and ISO/TC 213 Dimensional and geometrical product specifications and verification.

Getting machine tool specifications and commissioning plans correct could also play a part in Bowing"s riveting robot troubles. If you don't require what you need you rarely get it.

As to the use of inch fractions, we are supposed to use appropriate measuring tools. While it is possible for a very skilled machines to fit up a bearing race with only calipers, I never was one of them. I always use micrometers for tight tolerances. My set is equipped with a vernier that clams to measure to 1/1000 inch. Scatter on measuring a standard bar for me with one of them is about 3/1000 inch.

Scatter on a measurement with a 12 inch vernier caliper is a bit larger.

Trying to tease out 1/128 inch out of a scale marked in 1/64 yields inconsistent results.

Fred

Fred

 

[waross]

Sorry. I guess that after the last change in the definition of an inch the systems are not incommensurable, just inconveniently awkward.
Given the management culture at Boeing, the tolerance issues may be management issues rather than engineering issues.
As far as machining metric threads on an Imperial machine:
The inch was made commensurable with the meter in the 1959 re-definition.
It probably took some time for the new definition to make its way to the construction of new lathes.

An inch in the UK over the years.

1895 - 25.399978 mm
1922 - 25.399956 mm
1932 - 25.399950 mm
1947 - 25.399931 mm
1959 - 25.4 mm exactly

Any lathe built to post 1959 standards should be able to cut perfect metric threads with the proper gearing.
It may have been difficult to build a lathe to 1959 standards when all the existing machines were built to earlier inch standards.
However, a metric machine with the proper gear ratios would easily build to exact post 1959 standards.
Two suggestions Alistair.
1> The Imperial lathe was built to pre 1959 standards.
2> A measurement would show that the Imperial thread cut on the metric lathe was actually the correct pitch. The thread cut on the Imperial lathe would be correct to some pre 1959 standard.

Of course there are at least 346 people who are no longer affected by all this.
Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Alistair_Heaton]

That agrees with my knowledge that metric on imperial causes less issues than imperial on metric. From memory of the rods there was 50cm difference between the divergence of the imperial V metric machine. I can only presume the gear ratio's give much less options for getting things right when the machine is metric.

As with all these things the skill of the operator is the icing on the cake. Unfortunately your skill set is becoming a rare breed FacEngPE even more so that you seem to be a PE that has time on the tools.

You also have to remember that a 777 central hull is some 50 meters long. And they do it in halfs. And one robot each side. So they must be putting in some quarter of a million rivets and 2-3000 per row lengths ways on the easy rows without windows. If the movement between rivets was 20mm then I suspect they wouldn't have had any problems but I really don't expect Boeing specified a nice number to decimalise between rivets which matches a metric value. Then your into combining tolerances to get the resultant tolerances between the drawing and production. I think out of 120 of us on my degree only 4 people that stuff clicked with. And two of them are now academics in Maths with connections to engineering. And none of them you would trust with a bread knife never mind a screw driver. The tolerance of the machine stops being the limiting factor. It may very well be within the specs for single jobs and/or movements. Over 50 meters and 2000 placements maybe not. I suppose you could design to ensure nice numbers for metric production. But it would be easier to just design metric in the first place which I suspect is the reason why the USA automotive industry has gone metric.

I don't have a clue how they moved the entire thing or how they locate.

The 3/fractions I think cause issues because they are hit the worst by rounding errors when things are decimalised which they will have to be when going through to the making the input deck or processing the input deck later. But your into a situation that you have variable deviations depending on the size required.

Anyway the KISS solution is if its a metric design you make it on a metric machine. And if its imperial you make it on an imperial machine.

The ability to source local high quality manual measuring devices is not to be ignored either. You just can't get them in imperial most places in the world. They will use metric and convert to imperial yet another injection into the tolerance equation. End result on the final production floor hand tools come out to get the parts to fit..... I seem to remember reading about that being an issue with Boeing, all that expensive cold reaming for increased fatigue resistance out the window to keep hitting production targets.

 

[Alistair_Heaton]

To be honest waross the issue with the threads was a bit more involve due to it being a big beast of a thread on a steam engine restoration being the final goal. Now that you mention it I think we were lucky and the engine would have been built using the 1895 inch definition. Anyway attempts were made using a metric geared lathe on small pieces and the end result wasn't very good with several of us having a go including a couple of done it all their lives hairy bum machinists. After the final 5kg lump of metal for a small piece had been flung out the shed door followed by a load of Doric swearing we tried to source a proper lathe and a bottle of whisky secured one from an oil company workshop but it was too small for the big thread. After success with the imperial lathe we sourced a conversion kit from Denmark for the original German metric lathe and it worked, I think it was 1965 built. The initial thoughts were that they wouldn't have made it to very tight tolerances in the first place so we would get away with it. How wrong we were. I converted it back after they had finished.

The oil stuff I had no say over, they had been bitten before with it all and they weren't going to risk it again. I believe a poisoned well in Gabon was involved. I was doing the FEA side of things so after the induction where I was shown the 4 rods I didn't have anything to do with final production drawings or production. Everything internal was metric it was just anything that could conceivably be connecting to USA produced hardware they were strict on. There was two sets of drawings done one with imperial sizes which was more marketing than anything else. And the working ones. I don't have a clue how they do things these days. This was early 90's. And being Aberdeen born and educated in the period the American food shop was open in the 70's/80's and the place was full of Americans by the 90's they had pretty much all left. Les the OIM on piper alpha who was killed on the boat under it was my Scout Master.

 

[IRstuff]

It probably takes a grand total of 10 microseconds to do the math calculation for each hole; it's not like they're using the 1802 processor used on Apollo, at least, I hope not. My cell phone's GPS processor, by, itself, is doing dozens of large matrix inversions along with the conversions from XYZ to lat/long/alt every second while I'm driving.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Alistair_Heaton]

How many movements are involved with each riveting action? How many measurements?

Just had a search and SpaceX and Starliner are both metric built. We shall see if the fresh sheet designs are metric from Boeing. I suspect it would be suicide not to if the bulk of your market is metric.

https://www.seattletimes.com/busine...d-fix-a-manufacturing-defect-on-all-737-maxs/

Anyway more news on Boeing hand tooling the engine pods on the max to get them to fit.... You have to admit there seems to be a lot of issues with sizing and requiring reworking cropping up.

Boeing getting a waver on the wiring looms being outside certification are getting very remote.

 

[waross]

I'm up against the Seattle Times pay-wall.
I found this:[URL unfurl="true"]https://www.flightglobal.com/safety/faa-requires-lightning-protection-inspections-of-737-max-prior-to-flight/136914.article[/url]

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[LionelHutz]

Seriously, WTF does a geared thread cutting lathe have to do with the failed automated riveting system?

And then, what does the failed automated riveting sytem have to do with the 737 crashes?

AT the end of the day, I would bet money the failure was due to that systems supplier or integrator simply failing to do what it promised Boeing it could do.

 

[dik]

AH: " Now that you mention it I think we were lucky and the engine would have been built using the 1895 inch definition." other than the change in mm to in being fixed at 25.4, I wasn't aware there were other unit changes over time. I used to have the conversion of mm to in to about 14 decimal places. It was significant on a project about 10-15 years ago using the old conversion... I checked the location of a new building and based on the reference several kilometers away, the difference in location was a few feet. Confirmed by a real surveyor...


Dik

 

[rb1957]

a new thought … do the stored planes have engines installed ? does this prolonged storage affect airplane systems ??

another day in paradise, or is paradise one day closer ?