JayW

Last I posted, I think I had declared victory on the digital surface of the top glass of the windshield: Well no. When attempting to make a solid from the surface, in order to 3D print a test part, I ran into all sorts of issues. Did you know that contoured 3-sided surfaces (note the piece is basically three triangle shaped surfaces connected by two blend radii) do not want to be offset to make a part with some thickness? The corners fold over themselves. Well I will not get into how I eventually got there, but it took a couple of days of banging my head against the wall, and re-watching a couple episodes of "Primary Surfacing" on YouTube. Here is a test part made from standard grey resin, shown with "Huge Thumbnails" from Imgur for your detailed inspection: OK. The main reason I printed off a test part was to see if the contour looked good. And it looks GREAT. Mission accomplished. It is .02 inch thick, which is what I want. I intend to try to print this part from clear resin, which is, based on some experience with a test canopy for the P-51D in 1/18 some of you may recall, going to be much more challenging. You can see where some of the supports were by the acne scars they left on the inside. So I will obviously have to rearrange my support config when I do the clear part some time in the future. If I fail, I am going to have to get somebody to vac form one. Not in my skill set... Well since I still had a big pool of liquid resin sitting in the tub on my printer, I decided to make some more parts - parts which do not require a still incomplete surface. How about main landing gear? Peter Castle aka Airscale made most of his undercarriage for Lope's Hope from 3D print parts. At least the wheels/tires/brakes. He recently sent me a huge data dump from his work on Lope's Hope, and included are some Rhino models. And included in that are models of his wheels and tires and brake housings. So what you are going to see here, are versions of what he sent me. Thanks a million Peter. Here, hot off the curing station: Now these are also test parts. Note one of the brake housings has a broken flange. Not enough care taken snipping off supports. And the outer wheel has a pretty serious blemish by the axle cap. That is Peter's part almost without modification. Magnificent. The brake housing and the inner wheel are my own, but using Peter's parts as a start. I will make a couple of minor changes to the support arrangement, and make some "production" parts pretty soon. But these test parts were mostly to check and see how round they are and if they will fit inside the tire. Based on my misadventures with the prop spinner, I didn't have a ton of confidence. A few shots of the parts in more detail: Note the interior stiffening ribs inside the inner wheel half. I wondered if they would print OK - They did and I am excited about that. Tires you say? I want my tires to have diamond tread patterns, and to have bulged flats, and hopefully "Good Year" on the sides. Like this: Yessir. So I am doing my own from the ground up. First though, I wanted to make another test part to see if my support arrangement is good, and if the tire and wheels fit together (meaning no out-of-round). Here is the tire test part, with a very pesky bulged flat: What do you think? When I did the Corsair, there were pages of discussion on how the flat should be. This flat is quite subtle; I am tempted to redo it with more bulge and/or more flat. But it is sooo tedious to model it in Rhino, I am more likely to just leave it. Accepting comments. I might tone down that ridge a bit, but it is needed for the tread pattern to run out onto. Next time you see a tire it will have the diamond tread pattern on it. BTW - I have seen umpteen pics of wartime aircraft including Mustangs, where the LH tire is not the same as the RH tire. Should I do that? I could.... So did the parts fit together? Yes they did! My test parts were largely a success, and I learned a couple of things. Next post you should see some tread pattern, and maybe the Good Year lettering and winged foot emblem on the sides. Never done that before. And, I hope to show progress on the surfacing of the radiator intake scoop under the wing. And, depending on which way the wind blows, I might try modeling the Malcolm hood. I have enough fuselage modeled where I could do that now: Or maybe the LG lower strut. I can barely contain myself. This is one exciting project at least for me it is. Until next time, take care and enjoy the beginning of Fall (or Spring for those of you on the bottom of the world).

Great work Pete! So I never knew just what that swinging arm on the Dauntless was for. Now I do. And, I now know the Helldiver has one (two actually) too. What an engineering nightmare this aircraft is. And, thinking on this (first time I did that), I guess a bomb could actually overtake the aircraft itself in a steep enough dive. Like near vertical. And with brakes out. I wonder if the navy found this out the hard way...

Thanks. Aircorps Library has the E&M manual, so that works for me. I will not be putting an engine in this thing, although it's tempting. In what form is that Merlin OH manual? I'd be interested in that.

Yeah I will second that. Now filling the gear bay with wet tissue - that sounds like a pretty good idea. I have seen gear bays filled with all kinds of stuff. Is this your preference?

Kevin - lofting is generating lines and surfaces for a boat or an aircraft. Word has it that at boat building houses way back when, when hulls were being designed, the lines were drawn up in "lofts" above the factory floor. From there, often tooling or molds were generated from the loft lines, for building parts. Back in the -70's "master plaster's" were made for producing skin panels or form tooling and such. These master plasters were made out of lots of metal plates trimmed to loft lines, placed in the proper relation to one another, and filled in with clay or plaster to make a surface. Hence the term, and it stuck. Even today when an aircraft surface is digitally defined, the surface is called the "loft". At least it was where I worked.

Uh oh - did you build one? I don't know. Many models are a bit unconvincing either because the tires were not weighted, or the paint job is too pristine, or perhaps the model isn't a Tamiya! I like models that look as if they were shrink-rayed. John's pretty good at that!!

Every Griffon Spit model I have ever seen is unconvincing. I have some expectation you will be breaking that mold.

That Griffon is positively bursting out of its engine cowl. Clearly it needed a large aircraft!

Last post I stated that I have surfacing to do. So that is what I have been doing. It turns out that I am not the only one to digitally create surfaces for vintage aircraft. Another fellow I was made aware of has painstakingly done this for several aircraft, where he had access to point data. I would use what he has done, and he offers it to the public, only it was done on a CAD program different from mine. So I will keep soldiering on. What I have so far: The windshield - boy the windshield. It was as hard as the smiley face, which I did not expect. The P-51B windshield surface has three components - a flat forward portion, a single curvature side portion (and its opposite), and a top compound curved portion. The flat portion was easy of course: Only one side was modeled; the other side is just a mirror opposite. The side glass, not easy at all. First, I used a fairly simple technique of lofting from edge curves, where I scaled the windshield assembly drawing (see above) for the lower curve and projected it onto the basic fuselage surface, used the station 113 loft (remember I have point data for that) for the aft edge. I don't even remember how I came up with the top edge. The forward edge coincides with the flat plate front glass. Four edges, and viola: But I really didn't know how accurate it would be. So I studied the engineering drawing for the laminated glass detail: What you see there is a ton of point data for a periphery, and a cross section that gives a single-curvature part. Back then, compound curved plexiglass was just being introduced, and manufacturers did single curvature wherever they could. There is precious little information on that drawing defining how to place it in space where it is supposed to be. Just enough for a supplier to make the part. So, to check how good or bad my side glass contour is, I digitized that point data into a surface: And then did the best I could to put it on the surface i created at the start of this process: The yellow is the digitized drawing of the side glass. The turquoise blue is my simplified 4-edge surface. You can see they do not perfectly match. BUT - the mismatches are small. And at 1/18 scale, very small indeed. So for reasons I will not further bore you with, I chose to go with the original blue surface. This exercise was but a check, and a successful one. Then it was on to the compound curved top glass. What a PITA that was, but I got something I think is pretty accurate: So now, I think I need to start in on the feared and dreaded wing-to-fuselage fillet fairings. Nothing there at the moment: But in order to do that, I need to finish the fuselage side panels, and the radiator scoop behind the wing. That radiator scoop might be the toughest challenge yet. I also have a tub of resin left over from my prop spinner efforts. Rather than take the trouble to pour it back into a bottle, I might make a couple more parts. Parts that do not need my unfinished surfaces. Hmmmm.... Stay tuned!

Oh my word John - you must be pretty darned pumped with that front end! I have not experienced the support ends being more brittle than the rest of the parts. FYI, 3D print parts go through two stages of cure. One is when the printing is actually being done. Each layer receives a several second cure to solidify the resin layer into the shape of that layer, allowing the next layer to be cured in top of the previous layer. But it is not a complete cure. After the part (and its supports) are finished printing, they are a bit soft and sticky (just a bit). Then it is cleaned of uncured resin with IPA, allowed to dry, and then it goes into a post-cure under UV light for a couple of minutes, where the finished part is dry and stronger. It can actually be done by letting the part sit in the sun a while, but I prefer a curing station where there is a rotating table in front of an array of UV light bulbs. I have one, and it was not very expensive. I suppose the thinner the part, the more curing it sees. And the support ends are quite small and thin. That might be what you are experiencing.

Hi John - welcome to the world of 3D printed parts. Quite the aftermarket kit you have there! And such a coincidence that you get a spinner with some of the same issues I had doing my own spinner in 1/18 scale for the P-51. See my new build thread I started. I was able to avoid the messy aft edge by printing my equivalent piece upside down, leaving me with a mess to deal with on the nose instead. One must pick one's poison. As I stated earlier, the Achilles heel of 3D print parts is the many supports. So extra sharp clippers, and great care is about all you can do. That and CA or putty.... The good part designer will try to place supports such that the final product is not too marred up. I am encouraged by the thin trailing edges of the rudder. In my experience, anything less than .015 inch thick or so becomes risky with 3D print, as brittle as the material tends to be. This looks to be some project. Following.

Maxim - I just googled 1/24th LEM. Holy Moley man! That's AWESOME! Is that yours? All I can tell is that some guy from down unda did it.

Well Chuck - you got the old dog correct! I grew old on this very long project, but what better way to grow old? This stuff is really fun. Thank you for the fine complements, especially coming from one of the best modelers I have ever met, eh?

I am not familiar with that publication. Do you think it helpful?

A little scratch building huh John? Right when I start wondering "Who are these Border guys?", an error crops up. No - this kit is extraordinary, and of course in the hands of the master, well it couldn't be in better hands.

Ralph!! I did not know. I have contacted Hugh to see if he can help me with the -BC inboard wing leading edge ordinates - the only data thus far that I don't have. Thanks. As for the prop - get ready for a whirlwind. I wish I had access to actual design data, but Aircorps Library does not have (at least I couldn't find it). What they do have though is an overhaul manual that contains geometry data. Go to the "Props" section, click on "Hamilton Standard", then go down to the document "Hamilton Standard Blade Repair Manual", part no. 130, revision date Oct - 1947, page count 204. On page 10 of the document is a table for various blade designs. Go to blade design number 6523A - this is the basic prop blade for the P-51BCD. Note the shank size is "D". You will find the shank size dimensions on pages 36 and 37. Note also in the remarks column "6487A with AMS 4130 alloy". It is that 6487A blade that is in the geometry tables (not the 6523A). Also note that geometry data is listed as pages 155-160. And also page 193. But that is wrong - actually it is page 189, not page 193. Always look for data for blade 6487A. I would encourage you to read the whole manual (except the geometry data that doesn't apply to our blade type). That way you will better understand the terminology. That is what I had to do to make sense of it all. All this data is for the basic blade - nothing for the cuff. For that I obtained some information from the P-51 SIG back when I could access that website (I no longer can for some reason). Let me send you a note with that data. Soon.

Thank you Ray - of course!

Look at the difference in the inboard leading edges! Changing that from the -B would have been a big deal. So there had to be an important reason to do it, and I am betting it was a Center-of-gravity versus Center-of-lift thing.

Thanks Thomaz!! Pretty sure everyone enjoyed that read. So, part of why I am so enamored by P-51's is that my late uncle Henry Rudolph was a Mustang driver and served in the same fighter group as Dick Turner - 354th FG. However he was in the 353rd fighter squadron whereas Turner was in the 356th. Lt. Rudolph flew several aircraft, one being a P-51B (wish I knew the serial number - I would make it my subject if it had the hood) and more than one P-51D's. One of the last ones was P-51D-20NA serial number 44-63685, buzz letters FT-J. Here it is, supposedly, in March of 1945, after "coming to grief" in a training accident resulting in the death of another pilot (not Hank). I have learned that one cannot hang his hat on the buzz letters to identify aircraft, as they were used more than once on different aircraft after the original one was lost or retired. Who knows if this is really "Sissy Mana". But the six kills align with his record. Also, there is some confusion as to which aircraft carried the name "Sissy Mana" (which was named after his sister - my mother). It is believed by some that it was 44-63885 as opposed to 44-63865. And 63885 survived the war and is actually a restored flying aircraft today painted up as "Tempus Fugit". Some of you know already I have been hunting far and wide for pictures of the nose art of "Sissy Mana" should any exist. Were I to find it, I would immediately switch over to a -D model and use it as my subject for this project, and once finished present it to my still living mother with her nick name painted on the cowl panels, should she (and I) live so long! Late in his life, Uncle Hank described some of his adventures with the 353rd. I believe he did not enjoy speaking of his victories, but he freely described his harrowing escapes. Just as Col. Turner describes, Hank also survived a compressibility dive early in his combat days with the group, recovering from the dive with an altitude margin that brings back bad dreams decades later. He also claimed he was shot down in a protracted one-on-one dogfight that devolved into a low altitude duel over occupied Belgium (or Holland or France?), survived the belly landing, evaded a strafing run by the victorious German pilot, evaded capture by the Gestapo, and was later delivered by the underground to his unit. But curiously there seems to be no record of it. Some other exciting stories too, some perhaps embellished a bit by a sentimental and a bit senile old man. What a way to spend one's twenties.

Antonio I have read this more than once over the years. Yet, I see no indication that the basic wing contours are any different between the two. I wish I had the -B wing ordinate drawing (I only have the -D) so I could do a point-by-point compare. Instead I did some scaling from wing rib drawings to see if ribs are deeper on the -D. And they do not appear to be so. They are the same. I don't know how 3 fifties got crammed in there versus slanted 4 fifties, there it is. So with the trimmed down rear fuselage on the -D, if skin, frame, and stringer gages were the same, I estimate about 40 pounds less airframe weight for that part of the D model. Everything else being the same that results in some amount of nose-down pitching moment compared to the B-model. Perhaps that's a big deal for the aircraft's balance?? Were that the case, that nose-down pitching moment would need to be offset. Extending the inboard leading edge would provide a nose-up pitching moment. Just spitballing here....

I have all manner of good news to report. First the bad news. Recall that I have been trying to make spinner halves that are essentially solid, with holes for the prop roots and prop shaft. Like this: Also recall that I have been having maddening deformity issues especially at the point of the initial print layers - a stretching of sorts. Also out-of-roundness. Ugly. Well, I tried a couple more times with revised and enhanced supports - hours of printing. Still failures. So I tried a completely different approach - thin shelled spinner halves with a separate prop hub (since the spinner halves would no longer be able to support the prop roots). And, bulkheads for mounting the spinner to the hub. BTW, this is how the actual spinner is on P-51's: It was suggested by Maxim61 and Peter (Airscale) that the aft half of the spinner be printed directly on the build plate rather than supported in mid air by a hundred supports (perhaps that would address the deformity issues). So I tried that: What a beautiful part. The shape is perfect. The challenge would be to free it from its sacrificial base without fracturing it. And I fractured it. But, it was repairable with thin CA, so no harm no foul. And here are the other new concept parts, fresh off the printer waiting for the post-curing station (the simplified hub and two bulkheads): Here is the collection of useable production parts, post cured and ready to go, for the very first component of this P-51 build: Assembled: Victory. Gone are the deformity issues. I am a happy camper, although it disturbs me how much pain and suffering it took to get these parts right. Hope in the future I don't have such trouble, but I am learning. Wish I knew what color the spinner will be - then I would just finish it. Note that these prop and spinner parts are common to the B/C and the D. So in theory I do not have to decide what I will do yet. In theory.... Kevin suggested that I could cover the little pimple marks from support removal by dabbing on uncured resin, and then curing it with a UV pen. Well Kevin - sunshine works too. Thanks for that excellent suggestion - look at this: That nose cone was a bit ratty with pimples and now it is practically perfect. A coat of primer will hide any remaining (very tiny) imperfections. BTW - that big hole at the forward end (.125 inch dia) is going to be filled with plastic rod and sanded to shape to simulate the rubber plug on the actual aircraft. Meanwhile - continued surfacing in Rhino 7. More of the engine cowl area, and importantly - the B-model specific inboard extended wing leading edge. Pictures: The P-51 expert knows that the -B/C wing has a different (significantly smaller) inboard leading edge extension than the -D. I had ordinate point data for the -D, but nothing on the -B/C. So I did some intensive study of drawings (wing rib cross-sections, etc), and came up with a damn good approximation. Interesting how it droops going inboard....yup - it really does. If any of you aerodynamicists can explain why the -D has that over-extended leading edge, please explain! I cannot figure it out... Look at the difference: I believe y'all have seen the last of 3D printed parts for a while, unless I just get sick and tired of the surfacing project. I have very tough surfaces to come - wing fillet fairings, the radiator scoop, the fin and h/stab fairings, and the cockpit enclosure (including BTW the Malcolm hood). At this point, I have run out of excuses not to do a -B/C. Unless something unforeseen happens, I think the decision has been made. Speaking of the Malcolm hood, look what Mike McM got for me: I have some reason to believe that fishbowl cross section for the hood is accurate. It looks to be a simple circle. Who'da thunk. So that takes a bunch of guess work out. I like! Don't go away as I continue to surface!

Yeah it took long enough didn't it? Thanks for following all this time. Well by all means, tune in to this:

Oh, I just noticed the cross section pics with the blue fishbowl!! Awesome, very helpful! Thanks Mike.

Looking awesome Pete.

Alright - these test parts are very important. This model is supposed to be primarily made from 3D print parts, so as to assure great accuracy of shape. But what I am finding is that my parts have inaccuracies if not outright flaws. Much of this was unexpected, and could be operator error. By that I mean I am not printing parts off as well as other real experts could - due to the resin I use, or the machine settings I have created (exposure time, lifting speeds, layer thickness, aliasing), or how I am orienting and supporting the part during the printing process. But it also could be I am expecting too much of the 3D printing process. I am asking for some serious exactness. Take a look at my prop spinner parts - I already learned from test parts, and have applied some lessons learned. Also Alain suggested I print the nose upside down, which I wanted to try. Recall the test parts: Terrible fit, wavy edges. Part of the terrible fit, BTW, are from some subsequently discovered small but nonetheless serious clashes with the prop roots. I fixed that for the "production" parts. Those wavy deformed edges are serious: If I do not get a handle on that, this whole project is going to be a no-go. Peter (Airscale) has been having similar issues with his 3D print work and is working on it in parallel. Somehow that has to be addressed - all my parts are going to have edges of course. The "production" parts: Note there are two nose halves and one aft half. The nose half to the right was supported in a traditional way, tilted about 45 deg with nose up, so that all the supports were on a portion of the aft edge. Compared to the test part which was supported the same way, I only increased the number of supports to try to get rid of the local waviness. It worked in that there is no waviness. You can see remnants of those supports along the edge. The other nose part was printed upside down at a slight tilt, where all the supports were on the nose surface itself. That ought to leave the aft edges pristine. You can see the little marks the supports made on the outside surface after snipping them off and sanding smooth, just off the "pointy end". The third part of course is the aft half of the spinner. To my great dismay, the traditionally supported nose half has a large deformity in the region where it was supported: No waviness, but a giant mismatch with the aft half of the spinner. This part goes in the trash can. Why oh why did this happen? Well, clearly it was stretched and deformed during the printing process. That corner was the first to be printed. And as more and more layers are then added, the part gets heavier and its weight is pulling on those first small layers. Also, after each layer is baked on, the build plate is lifted to allow resin to flow back underneath it so the next layer can be created. There is some stiction force associated with that motion, which also pulls on all the layers previously laid down. One or both of those phenomena might be causing this. Perhaps if I spun it 45 degrees so that the first layers are not coincident with the prop root cutout? Then the first layers should be more robust. I dunno... Or, maybe my exposure time (3.5 seconds per layer) is insufficient to give a good hard material? Again, dunno. That is the recommended exposure time for this stuff at 5 microns layer thickness. The aft half, which was oriented and supported in a similar manner, suffered the same fate although less severe. It is, however, a bit out-of-round as a result. That part will also go in the trash can. Then there is the other nose half, which was printed upside down: Now THAT'S what I am talking about. Thanks Alain - good idea! Good edges, decent match up with the aft half. That is a useable part. It has a bit of a rash from the supports: But some putty or CA will cover that up. So I think I am about 80% of the way there. A few more shots that seem to indicate a more hopeful situation: The prop blades are just laid in there loose for now. So, I have a useable spinner front half, and an un-usable aft half although quite close to being OK. Next post I hope to show a useable aft half which is perfectly round at its aft edge (currently it a bit off), and not wavy or deformed. And then I can declare victory and move on with some assurance that this 3D print process will actually work OK. Bear with me, some may find this a bit painful to follow. But it is here where I can determine whether or not this project will succeed. Stay tuned.