Print

[Jerry Germann]

Some time ago, it was mentioned that a/some drawing/s of the patch installation method may be posted, I have a mechanical background and am interested in such things. I am not questioning the fit, I am mostly interested in the manner of patch installment. In this report;

In the artifact analysis report;

2-2-V-1 Sheet

The sheet was a comparatively large piece (23 inch x 19 inch) of 0.032 inch thick aluminum alloy, shown in figure 1, with a pronounced curvature across the short dimension. The sheet had four rows of evenly spaced 3/16 inch diameter rivet holes and one row of 5/32" holes along its long dimension. Measurements determined that the rivet rows were not parallel but rather showed a slight convergence. Nominal spacing between rows was about 4-1/4 inches at one end of the sheet and 1/8 to 1/4 inch closer at the other end. The skin around these holes was, in general, dimpled inward toward the concave side of the sheet suggesting that the sheet had been area loaded from the concave side while the rivets and underlying structure were intact. The remains of a solid brazier-head rivet were found in the hole denoted by arrow "R" in figure 1. The manufactured head of the rivet was on the convex side of the sheet and was marked with a single round dimple in the center of the head, as shown in the right center photograph of figure 1. The dimple usually signifies a 2117 aluminum alloy rivet. The length of the undeformed rivet shank (distance between the manufactured head and the formed head) indicated that the skin had previously been attached to an approximately 0.06 inch thick underlying member. The faint outline of 1/2" tall letters "AD" were noted on the convex side of the sheet in the circled area in figure 1A and are shown at 2X in figure lB.

0.06 seems to be the maximum stringer/underlying fixture thickness that this artifact could have been attached to; .....stringer thickness from the Idaho wreckage was also measured at 0.06

http://tighar.org/Projects/Earhart/Archives/Research/Bulletins/71_RiddleOf22V1/71_Riddle22V1.html


The procedure outlined below; shows the use of shims ( original skin thickness) to bring the patch flush as to avoid a concave indentation in the center of the panel when the rivets are bucked, and would seemingly eliminate voids near the patch ends making it difficult to buck up the rivets there.  It seems by the remaining rivet shank undeformed distance ,along with the Idaho stringer thickness information, that 2-2-V-1  seemingly wouldn't have room to have shims installed .....possibly, there are other repair methods that would allow one to attach the panel directly to the stringer and perform the repair ,in a timely manner, as a scab,.. Greg, Jeff, I spoke with you fellows about this before, but, would you have any drawings that would show a procedure by which that could be done?

The drawing in the attachment shows a typical installation regarding fuselage scab patch installations


A  patch installation procedure; page 99 ,.... pg 102 left hand side in this manual
http://www.avialogs.com/list/item/4128-structural-repair-manual-for-the-model-dc-3

[JNev]

320 line may be a smidge over 12.5" from the 307

Questions regarding Station Lines:
Jeff noted that the Wichita plane appears to be less than 13" between Sta. 320 and Sta.307 which if accurate seems odd to me.

"Stations" and "rivet lines" are not necessarily the same.  What I hoped to respectfully point out was that if one merely examines the tape measure in the picture, one may determine, simply, what the actual distances are between rivet lines.

Why wouldn’t they be 13” apart? The Station lines are 13” apart? Construction tolerance? Restoration of the plane?
What are the Station Lines supposed to be to? The center of the original form work for the Circumferential stiffeners?
The center of the structure or the center of the rivet lines?

All good questions.  You're getting there - rivet lines may be offset from a 'station' for any number of reasons - some of which you have touched on here / below.  "Stations" are more about a design reference point than they are physical stuff like rivet lines; a rivet line may lie right on a called station, or it may be off by fractions or more of an inch for some reason.  A 'Z' channel with a flange offset toward the forward end of the ship might explain what we see at the 'station 320' rivet line - rivets that apparently lie ahead of, not on and not behind, the station line.  Station 307 on the other hand may (I am not certain this is true) have a rivet line that is true - I think those fasteners go into a vertical channel (examining the interior of the Wichita bird should refresh on this).  Those are guesses as to how the rivet lie - all I can say with certainty is that the tape doesn't lie, and the tape tells of some different distances between fastener lines than we might assume were we depending on station numbers and math alone to determine, e.g. "320 - 307 = 13, so obviously it is thirteen inches from these rivets to these other rivets - WRONG: the tape tells us closer to 12 1/2", maybe 12 3/4".  Same from the skin edge to 307, e.g. "307 - 293 5/8 = 13 3/8" - so either the rivet line at 307, or the skin edge NEAR 293 5/8 is NOT on the station - because the actual distance from 307 to that skin edge is 12".

I thought maybe that if one of the circumferential stiffeners with a “Z” shape was put in backwards it could change the distances between rivets.  But from the interior pictures of the Wichita Plane, both 320 and 307 stiffeners seem to be oriented the same, with rivet lines forward of the former.
I did notice that on some interior photos of other L-10 planes, the rivet lines seem to be forward of the former, but in one picture of AE’s plane the rivet lines seem to be aft of it. Do any of the available drawings show details of the Lav window area and reference the station lines with rivet lines?

You've explored it well and only an examination of the drawings and layout of these details could fully answer how the true stations lie relative to the rivet rows.  It simply occurred to me that what is prime for our need is to understand the physical landmarks we're talking about, and the actual measurements between them, etc. and I came to realize that assumptions about stations providing distances was not accurate.

Regarding the “ghost stiffener” near 307. If the patch installer didn’t want to remove existing rivets above the patch to splice it in, they may have spliced it to vertical part of the Z on the aft side instead. Another possibility is that the added horizontal stiffener was so close to the existing horizontal frame of the window, it made I difficult to attach it at all to the existing Circumferential stiffener.

I think the 'ghost stiffener' near 307 is an unintentional herring with no relationship to the Electra - to match it up rivet-line-wise at 307 requires the artifact to ride too far forward for it to remain a candidate, in my own view.

[JNev]

The point I am making is that we can see line 5, as well as lines 1-4, and other rivet lines I didn’t number as well, yet we don’t see any rivet lines on the patch, even when they would be within an inch or two of visible external rivet lines. The fact that we can’t see rivet lines on the patch while nearby external rivet lines are quite apparent is precisely the problem that is not consistent with the artifact. 

What is your theory for how the patch remained in place for roughly 21,000 miles in the air?

I would say at a minimum that it was riveted around the periphery. 

Once again I am plagued by trying to copy from the Purdue photo, so pardon please, but this is contrasted a bit that you might see more easily what I see and am trying to convey - a suggestion of rivet lines along the periphery of the patch.  They appear similar to other rivet 'trails' we're seeing in this photo.

I also added a color-toned photo but it doesn't save as well here as it looks on my machine for reason - I get tons of nice contrast out of it on my laptop before uploading it.  Not sure what I'm losing - but it's not too bad here.

[Ric Gillespie]

I would say at a minimum that it was riveted around the periphery. 

Now we're talking about the patch being just a plain, unbraced sheet rivet around the periphery?  Not one experienced mechanic we've talked to (including you until now) has suggested that anyone would be so foolish as to do that. 

The Miami Herald photo has far better resolution than the currently available copy of the Darwin photo. At least two rivet rivet lines on the patch that match the location of rivet lines on the artifact are discernible to the amateur eye in the Miami Herald photo.  Glickman, who is not an amateur, has found evidence of the other lines.  They too match the artifact.

[JNev]

All I noted was that I could see what appear to be rivets around the periphery; I simply don't see similar evidence in the mid-field.  Make of it what you will.

No, I would not cover that area without some bracing if I had a choice.  Nor, actually, would I really choose to brace it in the manner suggested by 2-2-V-1 but would prefer to match the original bracing, if possible.  But I wasn't there, either.  We were also looking at an artifact with holes across the center that might be related to such an effort - to explain same, if that artifact were indeed the cover.

I too have held that the Miami photo was compelling, but now having had a better study of the Purdue-Darwin ramp photo must say it might help if you and Jeff can help us see more clearly what you see in the Miami photo.

[Hal Beck]

The point I am making is that we can see line 5, as well as lines 1-4, and other rivet lines I didn’t number as well, yet we don’t see any rivet lines on the patch, even when they would be within an inch or two of visible external rivet lines. The fact that we can’t see rivet lines on the patch while nearby external rivet lines are quite apparent is precisely the problem that is not consistent with the artifact. 

What is your theory for how the patch remained in place for roughly 21,000 miles in the air?

Well, judging from the Darwin photo, and that is the matter we are discussing, I would not conclude that the patch remained in place for 21,000 miles thanks to four horizontal rivet lines in the positions we see them on 22v1.

I agree with Jeff that the riveting was around the periphery; looking at the Darwin photo at the Purdue web site I do see indication of riveting just inside of the patch edges.

[Greg Daspit]

Regarding the rivet line just below the top edge of the patch in the Miami Ramp photo. I see that one clearly as well as one other one, but it was not as clear. I thought one possible reason the top one was so much more evident, was that it was the bigger 5/32” row of rivets. Of course the angle of lighting is another reason. If the darker top row was the 5/32" row, then the orientation of 2-2-V-1 would be tab up. It seemed possible to me, that a tab up orientation also fit with a previous study drawing I did to explain a possible scenario for the direction of the tears in 2-2-V-1(away from center).

[Bill Lucas]

Just as an FYI:

Stations are simply a measurement along an axis that runs down the centerline of the fuselage (front to rear) from
some origin point but not necessarily the nose of the aircraft (although it's been said here that for this aircraft station zero is the nose).
On some modern aircraft station zero could be some imaginary point out in front of the nose. It depends on how the production tooling is set up.

A bulkhead or any other structure for that matter at say station 320 means it's 320 inches from the origin which is in our case the nose.
Bulkheads generally being SQUARE to the centerline axis. (so bulkhead 320 would measure 320 inches at any spot on that bulkhead.)

Don't forget: any station measurement is a measurement from the origin point that is PARALLEL to the centerline axis of the fuselage.
That's why you could have rivets that measure 10 inches apart on an actual piece but their station measurements could be say 9 1/2 inches
apart (their in/out or "Butt-line" measurements would be different of course because that part would be installed at an angle to the centerline axis)

Waterline measurements (up/down) are perpendicular to the centerline axis mentioned above and also of course have an origin point.

Butt-line measurements are the in/out measurements that are also perpendicular to the centerline axis and are expressed as LBL (left butt-line)
and RBL (right butt-line). These are the measurements taken from the aircraft centerline axis.

In the 3 dimensional cad world that I'm part of .... an aircraft structure is defined based on an X, Y & Z axis. These axes are all square to each other. This X, Y & Z axis
never moves and is permanent. The Z axis forms the centerline of the fuselage (stations), the Y axis is for waterline measurements and the X axis is for
the Butt-line measurements.

I hope my explanation didn't confuse anyone.

Bill

[JNev]

Very good, Bill.

'Station Zero' is referred to as the 'datum'.  It can be anywhere the manufacturer designates.  Often that is something like "the forward face of bulkhead thus and so" - but canalso be an imaginary point so many inches forwaed of the tip of the nose, etc.  point being it is the datum by which all stations are measured, as you described so well.

Thanks for describing lateral stations (butt lines) and waterlines.  It is an old system, but very good.

[Ted G Campbell]

Why doesn't anyone concentrate on the artifact's lower double row rivet line and see that there is an "off set" rivet that has to be taken into account when trying to line up the patch to pictures of AE's airplane?
I have tried to point out that if this "off set" rivet aligns with any photos we've got, there is very little argument it's from AE's plane.
Ted Campbell

[Diego Vásquez]

Ric -

   One part of the fitment issue that has long puzzled me is the Glickman overlay (attached).  You have suggested that the Miami patch was approximately 24 7⁄8  inches in length.  The artifact was measured at 24 3⁄8  inches, leaving them only ½ inch apart in length.   One would think that the Glickman overlay would therefore show the two objects to be very close to the same size.  Yet the size difference between the two objects in the Glickman overlay is very large, obviously much greater than ½ inch. 

   Using the artifact as a known scale at 24 3⁄8 inches, the patch in the Glickman overlay appears to be well over five inches longer than the artifact (5.36” longer when I do the scaling measurements and the math).  This suggests that the patch would have to be almost 30 inches long, which we know can’t be true.  It would seem that the patch layer would have to be shrunk down by about five inches in order for the overlay to accurately reflect the posited ½ inch difference between the two objects.

   How can the large size disparity that we see in the Glickman overlay be reconciled with the idea that the objects are only ½ inch apart in length? 

   Wouldn’t changing the size of either object layer in order to properly reflect their posited sizes also destroy the alignment of the rivet lines seen by Mr. Glickman? 


Diego

[Ron Lyons]

We know the patch we have pictures of was lost out there somewhere.  Fact.  Maybe it was 200 miles away, or maybe it was right there on the beach.  Almost everyone is in agreement on that.  The plane went down, the patch was lost with the plane.

We know of no other piece of sheet metal missing in the area off of a plane, and certainly know of no other piece of sheet metal that's the same thickness, and that's within 1 inch of the size we're GUESSING the patch was.

Be back later, have to go sharpen my razor.

Sincerely,

William of Ockham
 

[JNev]

Some time ago, it was mentioned that a/some drawing/s of the patch installation method may be posted, I have a mechanical background and am interested in such things. I am not questioning the fit, I am mostly interested in the manner of patch installment. In this report;

In the artifact analysis report;

2-2-V-1 Sheet

The sheet was a comparatively large piece (23 inch x 19 inch) of 0.032 inch thick aluminum alloy, shown in figure 1, with a pronounced curvature across the short dimension. ...
0.06 seems to be the maximum stringer/underlying fixture thickness that this artifact could have been attached to; .....stringer thickness from the Idaho wreckage was also measured at 0.06

http://tighar.org/Projects/Earhart/Archives/Research/Bulletins/71_RiddleOf22V1/71_Riddle22V1.html


The procedure outlined below; shows the use of shims ( original skin thickness) to bring the patch flush as to avoid a concave indentation in the center of the panel when the rivets are bucked, and would seemingly eliminate voids near the patch ends making it difficult to buck up the rivets there.  It seems by the remaining rivet shank undeformed distance ,along with the Idaho stringer thickness information, that 2-2-V-1  seemingly wouldn't have room to have shims installed .....possibly, there are other repair methods that would allow one to attach the panel directly to the stringer and perform the repair...

The drawing in the attachment shows a typical installation regarding fuselage scab patch installations


A  patch installation procedure; page 99 ,.... pg 102 left hand side in this manual
http://www.avialogs.com/list/item/4128-structural-repair-manual-for-the-model-dc-3

Jerry,

Thanks for this, but it actually highlights a problem with calling the nav window covering a "patch" in the first place: we have no idea, other than by conjecture, that the "covering" was a "patch" in the structural sense at all. 

It may well have been nothing more than an expedient "cover" that was simply applied over the existing modified structure for the purpose of closing off the window opening.

And for all that has been wondered about the adequacy of the window mod and the "big hole" it created in the overal structure, we really don't have a big obvious reason to question the strength of the bird with that "hole" in it.  The primary bending forces in the fuselage with a heavy fuel load, for example, would be absorbed by the upper and lower surfaces - compression above, and tension below.  This puts the window (similar to other windows) in something approaching a more neutral zone in terms of stress. 

The side skins behave more like diagonal bracing during vertical bending; in those terms it is not clear to me that the missing material, where the window was, is that significant.  My personal view is that had some event resulted in damage leading to a last minute alarm and reinforcement of this area, more would be evident.  It has been speculated that seam damage can be seen at the lower end of the 293 5/8, but I believe we're only seeing ordinary shadowing at the skin lap there; in any case, were real damage there, I would expect to see some diagonal wrinkles throughout the area, not isolated joint distress as has been suggested about that lap joint shadow.

Consider too just what was - or, if you prefer, what was not removed to create the window opening: approximately a 23x17 inch section of skin, involving a 1x23 inch approximate lap joint with a u channel stiffener, and another 23 inch long section of u channel stiffener.  No major structural members like heavy stringers, etc.  We can't see what was added behind the outer skin except for the telling rivets near the top of the coaming at the forward end which suggests some internal addition for reinforcement.  It is reasonable, however, that given the thought we see in this mod by that one tell-tale feature that the need for some reinforcement may well have been addressed.  Finally, the window - although larger than the ordinary cabin windows, is still a similar installation: a look at how those openings were reinforced is not particularly remarkable.  It appears to me that the side skins of the L10 were well calculated to tolerate windows without a great deal of fuss.

All of which simply means that trying to rationalize the covering as a structural patch may be an overreach in conjecture; we may simply be seeing a cover with little, if any, features in terms of beef-up.  All we can really do is see what the visual record can tell us and should probably use care to avoid forcing an answer to make the artifact fit the picture we might tend to create ourselves.

Merry Christmas to all -

[JNev]

We know the patch we have pictures of was lost out there somewhere.  Fact.  Maybe it was 200 miles away, or maybe it was right there on the beach.  Almost everyone is in agreement on that.  The plane went down, the patch was lost with the plane.

We know of no other piece of sheet metal missing in the area off of a plane, and certainly know of no other piece of sheet metal that's the same thickness, and that's within 1 inch of the size we're GUESSING the patch was.

Be back later, have to go sharpen my razor.

Sincerely,

William of Ockham
 

That's where this kind of started out, and some of us did a great deal to try to eliminate other possible sources.  In fact I believe I was an early trouble-maker who thought he saw potential for this artifact to be a possible fit for the nav window, once it was realized that the window was covered expediently in Miami - so my heart is with you.

And true - to this day we don't know of another 'direct fit' source out there. 

It has even been suggested that 2-2-V-1 stands until someone can identify another source.  I understand the sentiment and once felt very strongly the same way, but the problem is we still have an obvious migration possibility.  We have seen it with the B-24 navigator's bookcase, and apparently with other still-unexplained shards of metal that have turned up: the islanders had a knack for getting that stuff, God only knows where from.

There's a lot of hair on that hide - we looked at what we could see, and now the uppers of a PBY have been examined.  We looked in bomb bays, wheel wells, flooring and interior bulkheads where we could get at them, and of course all the accessible outers in Dayton on every type we could think of.  Did we miss some crucial floor section, or a bulkhead hidden from our view?

Having wanted to see the double #5 row along the 'bottom' of the artifact as relating, it now seems a stretch when I look at that light skin lap on the L10, for instance: who would take a whole string of #3 rivets up to #5 in a light skin joint like that?  We have to consider the possibility now that 2-2-V-1 might relate to an installation where it covered a large area of light structure, but picked-up a row or two of heavier members where those #5 rivets were - not an easy match on the Electra.

So slash away by all means, Razor - but it is a large and strange world. 

[Jerry Germann]

Jerry,

Thanks for this, but it actually highlights a problem with calling the nav window covering a "patch" in the first place: we have no idea, other than by conjecture, that the "covering" was a "patch" in the structural sense at all. 

It may well have been nothing more than an expedient "cover" that was simply applied over the existing modified structure for the purpose of closing off the window opening.

All of which simply means that trying to rationalize the covering as a structural patch may be an overreach in conjecture; we may simply be seeing a cover with little, if any, features in terms of beef-up.  All we can really do is see what the visual record can tell us and should probably use care to avoid forcing an answer to make the artifact fit the picture we might tend to create ourselves.

Merry Christmas to all -

Thanks Jeff,
              The mechanical thinking portion of my mind keeps coming back to the installation process; It seems 2-2-V-1 was laid out rather well, regarding rivet pitch and rivet line spacing, seems some time was spent in it's fabrication. My thought is , how would it be installed as a cover/patch ( structural or otherwise), and include all the features of a standard repair procedure, regarding scab patch applications.
           It seems most repair methods include the use of shims the thickness of the surrounding original skin to bring the underlaying layers flush to the outermost surface, so as to allow one to solid buck the entire rivet line along the repair. Absence of them would seem to create a concave indentation in the center of the panel ( the original skin thickness) , in this case 0.025, ......granted, half the thickness of a dime, so looking down the fuselage line it may not be very noticable....however, when working to the outside edges of the patch, the void between patch/stringer ( caused by the entrapment of the original skin) would seemingly prevent a rivet there from being bucked tight, without  some skin dimpling or possible tearing.
It also seems the problem of forming the sheet to a desirable shape near the top of the fuselage may be further compounded by this indentation.I am basing this idea on the thought that the added stringer thickness was 0.060. My reasoning would become moot if added stringer thickness was 0.035 ??  ..( kinda thin) and a 0.025 shim was added. Your thoughts?  Below , I tried to draw my view of this; ( definetly not as good an example as greg's ).

Black = stringer
Maroon = patch/cover
Gray = original skin