Aichi D3A Val

[gburch]

Has anyone ever made a 1/32 prebuilt of this plane?

If not, what's the largest scale kit that's readily available?

I can't imagine why I'm asking I must be crazy ...

Think there'd be interest in a 1/18 Val?

[Jay]

YES, YES, YES, GOD YES!!!!! WHAT EVER YOU ARE DOING, I AM IN!

[gburch]

YES, YES, YES, GOD YES!!!!! WHAT EVER YOU ARE DOING, I AM IN!

You might want to curb yer enthusiasm there, cowpoke. By the time I start selling V-2 kits (did I just say that?) I will have been working on that project for ten months or so. Imagine how long a Val would take ...

[gburch]

... but somehow, this got into my computer:



... some crazy person must have been using my CAD tools ...

[Dauntless]

To my knowledge there have not even been any 1:32 model kits of the Val or any of the IJN dive bombers, or torpedo bombers either. Admiral did however do a Kate prototype in 1:32 that never made it out of the gate.
All the 1:32 model kits are Japanese fighters.

A Val would be cool in 1:18, or anything WWII Japanese besides a Zero.

Have you ever thought about starting something a little less challenging like a Rufe float, pylons, struts, fuselage rear fin kit? (maybe a dolly) Though from what I've read the tail rudder is larger, which might be a problem.
Plenty of Zeroes around to modify. I'll bet you could sell a bunch of em.

Member VMF214 did a couple custom Rufes that were simply amazing (the pics are gone or I'd link you to em)
I don't think there was an offer for any kits to sell though.

[Jnewboy]

I would love one! or a Kate, or an Oscar or a George or anything...

[snake]

Marushin do this in 1/48 heavy diecast.

Not sure about the kits, though.

[gburch]

Marushin do this in 1/48 heavy diecast.

Not sure about the kits, though.

Have you seen it? Do you think it's accurate?

[gburch]

So ... here's my thoughts about the Val as a potential next 3D printed project. It's a RELATIVELY simple shape and a manageable size. It also has fixed gear, which would save that challenge for a later project. It's also essentially a complete impossibility that anyone else would undertake it.

I'm still recovering from shelving the X-15 project because of the rumor that Pegasus might do an X-15 -- even the hint that someone else might do something is enough to scare me off from a project -- given that we're talking hundreds of hours of effort to get to a finished prototype ... I'm just never going to risk that.

The Val also has a relatively simple canopy. Which leads to the challenges. I see four main challenges:

Wings: They're much broader at the root than the biggest dimension of my printer, which means I'll have to split a big portion of the wing into more pieces than I'd like. Balancing "printability" against "buildability" and strength will be a big challenge.

Canopy: I suppose I'm going to have to start from zero on vacu-forming. Picking a relatively simple canopy shape and design speaks in favor of the Val, though.

Prop: I have some ideas, but that's all. If I'm going to make stuff pre-jet age, I'll need to master this eventually ... it might as well be this.

Panel detail: Coming up with a solution to this is a must. Again, it might as well be the Val as anything else. I have two or three ideas about how to explore solutions to this challenge. At the rate I progress, I'll be getting to that in a few months ...

As to why I don't try something smaller and simpler ... well, I'm crazy

[gburch]

This is the easy and fun part:

[pickelhaube]

G Burch,

Let me tell you there is nothing easy about aircraft .

Because you have rounded surfaces. I truly thought that the Horten I am working on would be done in 4 weeks tops.

It has taken me 2 months of hands on to get it this far.

Taking about 4 months to do 2 months worth of work.

The easy part of my build was the rough in wings and center section. I think that took only about 4-6 days. The rest spent on details and finish ( sanding filling )

All of the work goes into details and finishing.

Your printer leaves a course surface which means that you will have more sanding and furnishing.

If you could refine your bot to make smooth surfaces with details it would make things easier. Like the more pricier bots.

Otherwise this project may end up having you pull all of your hair out.

I hope I did not burst your bubble but THINGS ARE NEVER AS EASY AS WE FIRST THINK.

I know.

[Jay]

When I was studying design and building models for prototypes my tutor gave me some good advice about projected time management. He said work out how long you think you'll need to build something, and then triple that amount to get a projected actual time.

[gburch]

G Burch,

Let me tell you there is nothing easy about aircraft .

Because you have rounded surfaces. I truly thought that the Horten I am working on would be done in 4 weeks tops.

It has taken me 2 months of hands on to get it this far.

Taking about 4 months to do 2 months worth of work.

The easy part of my build was the rough in wings and center section. I think that took only about 4-6 days. The rest spent on details and finish ( sanding filling )

All of the work goes into details and finishing.

-- snip --

Otherwise this project may end up having you pull all of your hair out.

I hope I did not burst your bubble but THINGS ARE NEVER AS EASY AS WE FIRST THINK.

I know.

When I was studying design and building models for prototypes my tutor gave me some good advice about projected time management. He said work out how long you think you'll need to build something, and then triple that amount to get a projected actual time.

When I said "fun/easy" I was referring to the computer modeling. I've been doing that for over a decade now and I've gotten half-way decent at it. Once I get into the groove, I can whip out a pretty decent mesh from good 3-views and reference photos in a reasonable amount of time.

As far as time projections, once I got the hang of translating CADed parts into plastic, I ended up making a pretty good projection of how long the V2 project would take me. I'm roughly eight months into that project now, and I'm just about where I thought I'd be back last March or so.

Your printer leaves a course surface which means that you will have more sanding and furnishing.

If you could refine your bot to make smooth surfaces with details it would make things easier. Like the more pricier bots.

Otherwise this project may end up having you pull all of your hair out.

I'm sure what little hair I have left will all be gone by the time I either make something I like or give up. But getting a finer resolution printer is a matter of another order of magnitude of cost -- between $10k and $20k, instead of between $1k and $2k. I can't see doing that. The whole point of what I'm doing is to see how far I can go with true "personal printer" technology.

With the V2 project, I've established that I can create a finished product that satisfies ME, with a reasonable amount of effort, which is what I set out to do. The beauty of what I'm doing is that I'd be doing it for myself, anyway, no matter what. If anyone else thinks the results are valuable given the extra finishing effort involved, that's just icing on the cake.

[hworth18]

Marushin do this in 1/48 heavy diecast.

Not sure about the kits, though.

Have you seen it? Do you think it's accurate?

I have both Marushin Vals (D3A1 and 2 versions) and I think they are both superb, but they are also pricey. You can get the D3A1 for $130 from Aikens but the D3A2 is tough to find and typically go for a premium price. I sold my extra D3A2 on ebay a few months ago for $369..

[snake]

$369!

That must have put a smile on your face Harry.

Ironically I got my Val from Harry about a year ago for about $75 if I recall. Added a couple of Marushin Zero's since then.

Know a couple of collectors who love their Marushin's, and are willing to spend big bucks for them.

[Dauntless]

....damn $369!

Hasegawa made a 1:48 kit also:
http://hsfeatures.com/features04/d3a1ym_1.htm

[Jay]

Fujimi did a 1/48th one as well.

[plasticboy]

Love to see Aichi D3A Val in 1/18! Great project indeed!!

[gburch]

Mesh this morning, very much WIP:



I haven't been able to find nearly as much reference material on the web as I did with the A4-V2. I'm piecing things together from many fragmentary sources ...

[Jay]

Another avenue you could try is the various downloadable flight sim models (Combat Flight Sim, add on aircraft for example). I'm not sure what file types are used though so unsure how these would translate to a usable solid cad file.

[gburch]

Another avenue you could try is the various downloadable flight sim models (Combat Flight Sim, add on aircraft for example). I'm not sure what file types are used though so unsure how these would translate to a usable solid cad file.

There's a few points there. First, although the authors of some flight sim models don't claim copyright, many do, and I don't need that headache.

Second, and most important, video game models are (or should be) carefully optimized to have as few polygons as possible. I've actually made some sim game 3D models, and there's an extensive bag of tricks in that world to cut down on how complex the models are because, as the polygon count goes up, rendering speed goes down. A lot of surface detail is taken care of with surface texture images and "bump maps" which render faster than actual mesh detail. As I have translated my 3D modeling skills into the brave new world of designing for printing, I'm having to "unlearn" a lot of those tricks.

An example of this is "roundness." For something like a rocket body (on the V-2) or fuselage (on the Val), in a videogame mesh, a designer would likely go with, at most a 36-sided polygon in cross section (and in many cases, less). You don't notice this in a flight sim (much) because you're either not viewing the mesh from a close enough point of view or, these days, some real-time face-smoothing takes place with the game's rendering engine (a very computationally-intensive operation that only the latest and greatest video cards can handle -- or so it was when I was active in that realm a few years ago).

But in the world of CAM, what you see in the CAD program is what you get from the printer or other CNC machine. So, for example, I'm using a 96-sided polygon for the cross-sections of the V2 and the Val. A 36-sided cross-section would look and feel heavily faceted in any large-scale model. The near-100 figure I'm using is a good compromise between managing the CAD process and producing optimally rounded shapes. You can see in the V2 build pics that, even fresh off the printer, the faceting isn't noticeable. By the time you finish, sand and paint, you get a perfectly smooth surface.

The same definitely holds true for things like panel details. In flight sims and other computer games, things like that are handled 99% of the time with surface texture images and bump maps.

But, finally, the whole point of what I'm doing on a personal level is to develop processes for achieving the whole thing from beginning to end with the tools available to a single hobbyist with reasonably-priced tools. I want to do it all myself, so I can see if it's possible to go from just historical blueprints and reference photos to a completed large-scale model. I'm convinced that one of these days, we'll be in a different world with this hobby -- no longer at the mercy of poorly-capitalized manufacturers always living on the edge of bankruptcy and Chinese sweat shops. The only way this will happen is if the real enthusiasts have the most design and manufacturing power in their own hands and under their own control.

... so I'm doing it all myself ...

[gburch]

... speaking of which:



Because they're the most striking feature of the Val, I worked on the wings early, and as a separate mesh. My original idea was to marry them to the fuselage mesh. Wrong. As Matt pointed out above, the complex curves on aircraft are a real bear, whether in bits or atoms. The transition from the fuse to the wings at the root on the Val (and most other WWII aircraft) is a COMPLEX set of curves. I wasted at least a week of drafting time trying to stitch the wings I'd made smoothly to the root fillet. The result looked like the 3D version of a small child's crayon drawing.

So I sucked it up and am redoing the wings as a real "loft" from the fuse mesh. Just as painstaking, but the result looks much better. In the pic above, I've got the root fillet partially contoured and the wing lofted out to the place where the dihedral begins. Outboard of that is the old wing.

Meanwhile, if there are any "Val scholars" on the forum, here's a question: Am I correct that one of the differences between the D3A1 (Pearl Harbor version) and the D3A2 (post-1942 version) is that the earlier plane had a shorter span horizontal stabilizer?

It looks like that, a recontoured canopy and a big spinner were the main visible differences (although there may have been a slight recontouring of the fuselage) between the earlier and later models. If I persist with this project, I'm definitely going to do the D3A1 (at least first), although the lack of a spinner will make the prop even more of a challenge ...

[gburch]

All major airframe elements roughed in:



Now comes the work of refining the major contours.

[gburch]

First pass refining the canopy:



... those curves are a b1tch!

[gburch]

Another round of refining the canopy:



it's a little hard to see, but the basic contours are now established in a way that further work will -- hopefully -- make the pieces of the canopy "manufacturable," pretty close to scale and movable in the right way ...