Small Scale Military Headquarters

It's been quite some time since I reported the status of my V2 project. A few months ago, I was on track to begin offering kits for sale in November -- just in time to soak up a little Christmas money. Three things have intervened to delay my launch (!) until, hopefully, right after the first of the year.

First, a couple of major projects in my real-world work have developed such that the amount of time I have to devote to fun stuff has been significantly reduced. Second, my beta builder has had some real-life distractions that have kept him from getting started on his build. This is an important factor, since I don't feel like I can offer the kit until at least one person outside my mad scientist lab has built a rocket. Finally, in the last few days, as I noted on the Val thread, I've had some improvements in technique that really should be applied to the V2 and which, I'm hoping, will result in fairly dramatic improvements in some of the least satisfactory elements of the beta version I've shown before.

With that said, I have made progress on getting the kit to a first public release. First, I've got a fairly complete set of documentation to a beta level of completeness. I showed some preliminary versions of instruction graphics previously, but I've changed the approach for what I hope to finally release, doing the drawings in monochrome so they can be economically printed to be included with the kit. Here's an example of the what the final drawings will look like:





Second, there will be a full set of textual assembly and finishing notes included with the printed instructions. At the first draft level, these are running to 20+ pages. The length shouldn't be intimidating, though, because I'm trying to be as clear as I can in light of the fact that this will be the first kit of its kind and will involve a few tasks builders aren't used to.

Third, I've made, and will continue to make improvements before release of the first kits for sale in things that were documented in the original thread. For instance, a couple of things that have already been done are an improvement in the printing of the warhead/nosecone that solves much of the problem I was having with the printed contour of the last few millimeters at the tip, and a printed-in hole for mounting the thrust-direction vanes in the bottom of the fins, eliminating the need to drill a mounting hole for the vane shafts. As I type this I'm printing some experimental parts for what may be a major improvement in the structure of the fin (as mentioned on the Val thread) that will simplify construction and make the final product much better and easier to build.

Fourth, I think that a builder of moderate skill will be able to get a satisfactory result with three to four weekends of moderate effort, at most.

Finally, I'm pretty sure the price for the kit will be $150 + shipping. Shipping shouldn't be much, since the parts nest down into a compact package that is fairly light.

So . . . with that, one pic from the previous thread as a reminder of what a finished model will look like:

Very nice, I know you've probably already said, but what is the over height?

Forgot to mention the instructions look great, better than some other well known kits

ram04 wrote:Very nice, I know you've probably already said, but what is the over height?

Forgot to mention the instructions look great, better than some other well known kits

The instructions do look great !!! As well as the whole project.

I doubt if I will ever have top shelf instructions as that.

Heck some of the mods I have done never came with instructions

I think the price point is fair as well.

ram04 wrote:Very nice, I know you've probably already said, but what is the over height?

Forgot to mention the instructions look great, better than some other well known kits

It's 30.5 inches tall

Right now I'm thinking the instructions and notes will come with the kit in a 3-ring binder with a cover page bearing the unique serial number for the kit. So all you 1/18ers, get your low-number kit when they go on sale !!!

Very cool! I agree with pickel: the price is quite fair. Great job!

How many total parts are there, and how many require surface finishing? At one point early on, you estimated the weight of the model as comparable to a 21st Century P-47: is that still true, do you think, or have refinements changed the situation?

Have you gotten another printer for production purposes?

Love it...great price...can you make it so I can launch it at my neighbors house?

aae83 wrote:Very cool! I agree with pickel: the price is quite fair. Great job!

Thanks ... and to all the others who have given me positive feedback through the months as I've gone from "hey, maybe someone ought to try this 3D printing thing . . ." to "hey, maybe this will work . . ." to getting close to "please buy my stuff so I don't feel like a complete idiot . . ."

aae83 wrote:How many total parts are there, and how many require surface finishing?

Here's an exploded view of the current beta:



Because of the "straight on" view of this drawing, some parts are doubled up visually, but the total in this version is 61 parts. With changes I've already made or that look likely to be made before release, this will probably decrease to 53 parts.

In terms of parts requiring surface finishing, a better way of thinking about it than number of parts, is number of assemblies, since finishing happens after the major sections are assembled. Here's a drawing of the sections that stay separate in the assembled rocket:



The builder will have the option of finishing the kit with the fins attached, or leaving them detachable, so the answer is either 4 or 8 sections to be finished, depending on which option you choose.

aae83 wrote:At one point early on, you estimated the weight of the model as comparable to a 21st Century P-47: is that still true, do you think, or have refinements changed the situation?

I happen to have a P-47 on its gear right now, and I just did another "heft test": that still seems very close to being right.

aae83 wrote:Have you gotten another printer for production purposes?

No. My current plan is to buy a second machine out of the end-of-year profits distribution from my real-life business. Although the build of my second printer will likely go somewhat faster than the first one, I will still need to budget the better part of two weekends to put it together, calibrate it and get it working.

This will have a minor impact on my production rate once I start selling kits, since I won't take money until I have a full kit printed and ready to ship. As I've said before, my estimate is that, on average, I'll be able to produce one kit per week per machine. If I keep my original machine devoted to developing new kits, then the new machine will be able to produce 40-50 kits in a year. Right now, I can't imagine they'll be much more demand than that.

For what it's worth, my current DREAM of a SKETCH of a business model is to fully develop one kit and add one printer per year. As demand for earlier kits tails off, production for more recent kits could be run on more than one machine as my "bot farm" grows. Thus, in (much) later years, I'd be able to achieve production rates of more than one sold kit per week. Because "tooling" is entirely digital, there's no decay in the quality of older kits, so the entire catalog would be available indefinitely. But that's the far distant and very uncertain future.

The pics below document work to validate the new simpler, stronger fin design:





The first pic shows the test parts for the new 2-piece design in front of current beta parts for the 3-piece design. In the new design, the middle piece has been incorporated into the bottom piece, and the key-and-slot construction has been replaced by the round slot-and-peg method I developed on the Val project, using styrene tubing for the pegs.

The second pic shows a closer view of the new test pieces, the connection method, and the old piece that is going away in the new design.

The third pic shows the new design mounted on my "test mule" engine section with three other fins from other test series.

Some notes:

1. Because these were test pieces to simply validate the new concepts, they're relatively rough. While I'm at it, I've done some work to do a minor bit of refining on the shape, size and precise location of the flap actuator fairings on the bottom fin part. This needs further work, because the changes I made caused a "manifold error" in the part that I didn't catch before I printed this test piece.

2. In order to eliminate the middle part of the old design, I had to push the print-height for the bottom piece to the absolute limit -- the part ends just a couple of millimeters short of the printer's z-axis end-stop. To make this work, I skimped a little on the part's (should-be-patented) "Greg Burch Build Base." There's a slight loss of print accuracy as a result, but this is at the joint between the two fin parts in the new design, and will be completely masked by the more modest surface finishing now required by the improved design.

These test pieces completely validate the new round peg-and-slot part joining design, and they are a MAJOR improvement on the current beta, being stronger, having a more precise fit and eliminating one joint requiring major filling. So I will proceed to chuck the four full set of first-release fins I'd already printed, bring the new design to release level and, hopefully, remain on track for my "launch" target to begin selling kits just after the first of the coming year.

FIN FAB: With the peg-and-hole and 2-piece improvements now demonstrated, I've decided to see if I can incorporate the panel line techniques I've been developing on the early Val work into the now-delayed V-2 kit.

This pic shows the current state of experimentation:



On the left is a fresh print of the upper fin piece with panel lines as deep as I can make them (see notes below). On the right is the first full assembly of the two pieces with panel lines, joined, and sanded after a coat of green putty, but not yet with the putty scribed out of slightly shallower panel lines than the piece on the left.

Some notes:

1. When I first began tackling the V-2 I had no plan to do ANY panel lines. I eventually did do some panel detail on the engine section, but only by RAISING the few panels one sees on this part of the rocket in photos. Many V-2 kits have no panel detail on the fins, and I was content to meet that standard. But with the Val, I knew I'd have to come up with some way of addressing panel lines, and have shown to my own satisfaction that printing in a depressed line, and then using those lines to scribe the putty out at the surface finishing stage yields a pretty good result:



2. BUT, the paneled parts on the Val are all thicker than the fins on the V-2. The stabilizer thickness at the roots on the Val are roughly equal to the root thickness on the V-2 fins, but the latter taper toward their outer edges much more than the stab parts on the Val. In fact, getting a clean, structurally sound outer edge to the V-2 fins was an early challenge on this project because they taper to very close to the printer's resolution. I spent many weekends coming up with a solution to THAT problem.

3. This presents some serious challenges to printing in indented panel lines on the V-2 fins, especially the line that runs about .75 cm inboard from the fin edge, where the part is already very thin. Suffice it to say that striking a balance between ending up with a workable panel line indentation on both sides of the fin and having a part that can even be printed, much less one that will be structurally strong enough to be finished (including sanding) and hold the weight of the rocket ... has been a bear.

4. Right now, I THINK this is a solvable problem. If it's not, I have a good, solid "smooth fin" design in the can. If it is, I have time to include a panel-lined fin in the final "commercial" version of the kit that will go on sale after the first of the year.

BACK TO THE DRAWING BOARD: With the time I've got available now (which is not much, with real-world work heating up a lot from now until mid-December), I've taken the opportunity of having 2-3 months until my self-imposed deadline to start selling kits to explore a radical redesign of the entire engine section and fin assemblies. The pics below show progress on this project, with notes following:





This redesign represents coming around in a complete circle to an approach I used in my very first demonstration build, as well as a fundamental change in approach to how the rocket-body parts for the engine section are fabbed. In my original designs, I included the fin fairings as part of the body parts. When I switched over to the large tab-and-slot fin/body joining method, I incorporated the fairings into the fin assembly parts. Here, with the promise of the stronger and more precise peg-and-hole joining method I'm perfecting, I'm exploring going back to the original concept. Also, for lots of reasons, I made the choice when I switched fin designs to go to a body design for the engine section that had four separate outer parts, and three separate internal joining parts. With more experience dealing with overhangs and large, long prints, I've decided to explore reducing the engine section body parts to just two, and incorporating, if possible the internal joining elements directly into the external body part print.

The first pic shows the lower of the two new body parts vertical on it's build platform, The second pic shows the same part tilted to catch shadows to highlight an issue with this approach. These pics were taken when the part was fresh off the printer, and there's been no sanding or other preliminary finishing work.

The downside to how I'm approaching this is that the sharp convex turn from the body to the fairing produces some artifacts in the smoothness of the external surface. You can see that the worst of these issues are limited to the top third of the body shell, where the shell is thinnest, to reduce print time (more about which below). If I stick with this method, this lower engine section part will be the roughest of any of the parts in the kit. Given that the entire surface needs to be puttied, though, this MAY be acceptable -- something I won't know until I do a pass at surface finishing on these parts.

Among the many benefits of this approach is that the exhaust fairings on the bottom sides of the part come out much better, with no need for major clean-up to address outer edge issues that resulted from the fact that they were overhangs supported by sacrificial material. Also, with fewer transitions from one body part to the other, and with more surface continuity without the large indentations of the bigger slots, the over-all contour of the part will require much less third- and fourth-pass sanding to get a smooth contour, a problem that was inherent in the previous beta design.

The last pic shows the status of the redesign project as of this morning. Four of the new two-part fins with panel lines have been printed, and are mated together only with dry-fit of the styrene pegs. No sanding or surface work has been done on them -- they look just as they do when separated from their build bases. The line where they mate to the body is an indented panel line. The lower engine section part has had one 5- to 10-minute workover with a rough grit sanding sponge and a minute or so of "shaving" with an xacto to remove some "strings" on the external surface that occur in the printing process when the printer makes the convex external detour to print the body/fairing contour.

At the stage presented in the last pic, this redesign project looks promising enough to support the substantial effort that will go into bringing it to "final production" level. A major issue is the upper body part. Incorporating the fin fairings into it, accommodating the panel details in it, and incorporating what were previously separate internal "joining collars" are major challenges. The first attempt at doing this is currently printing. And therein lies a major challenge: That print will take 15 minutes shy of FIVE HOURS.

Previously, the longest print I've done on anything that might be a production part has been just a little over two hours. Up to now, I'd considered such prints to be very long and "risky" in the sense that, if there was a hardware or design failure that didn't manifest itself until near the end of the print . . . I was unhappy with the result.

But two things make me willing to risk such a long print. First, the printer is significantly more reliable now than it was previously. Since cranking it up for the first time in February of this year, Makerbot has made some important design upgrades, all of which I've incorporated into my machine now. The printer control software, stepper-motor driver for the extruder, the extruder heater, the plastic reel-and-feed mechanism and the build platform surface and wiring are all much more robust than they were in the first version of the printer I was running at the beginning of the year. Blown prints due to hardware failures are very, very rare now.

Second, I've gotten a lot better at designing in a way that WORKS in the printer. The constant, quick feedback loop between design and fabrication I've mentioned before on this subject has resulted in me being much less likely to design into a part something the printer can't faithfully print in a way that works in the end.

All of which means I'm willing to at least see if a near-five-hour print makes sense. If it does, the kit will be much easier to build and have fewer, better-fitting parts. The downside may be that the occasional failed mega-print could mean that my per-kit production time might grow from one to two weeks, since the long prints will only be something I can do 2-4 times per week -- on the weekends. Printing all the other parts fit into before-and-after-work time slots during the week.

Very nice. It makes a lot more sense to put the locating tabs on the main hull.
Good thinking!

Too much tech stuff .


My head is spinning.

WOW!

Experiment Number 75 bazzilion and 2: Partial failure. Notes below the pics:





Pic 1: About 30 minutes into the top engine section mega-print, I heard the bad cracking sound of delaminating plastic in layers high enough above the heated build surface and low enough below the current extrusion layers to experience significant cooling, the bane of prints that are both long and thick. Figuring this was time to try a trick I'd read about on a Makerbot forum, I draped a towel over the machine to trap heat, and crafted a makeshift mount for a small hair dryer to try to raise the temperature in the build volume.

This had the effect of minimizing, but not entirely eliminating delamination. There's still way too much differential cooling in such a situation between the inside and outside of a very thick-shelled part like the design of the upper engine section I was printing. There are also some potentially bad effects for the printer, as the heat being pumped into the machine and trapped by the towel heats up EVERYTHING, including the printer's components. By the end of the print, I'd pushed the temp of the somewhat complicated connector to the heated build platform past it's capacity, and some of the more delicate leads separated. Fortunately, I had a spare and, when the print was finished, I replaced it in 5 minutes of minor robot surgery.

Pic 2: Although it had some deep cracks, the upper engine body part held together long enough to get a full test fit with all the new components and the last upper body parts I brought to completion. Here it sits in the rocket factory, now invaded with some Val parts in the foreground, and a stash of release-level finished tank section parts behind.

Pic 3: The flaws in the new part were too great for even minor handling, though, and within a few more minutes of fiddling to refine the fit of the fins, had cracked into two pieces at the deepest delamination -- not surprisingly where the greatest contour deviations occurred.

Conclusion: A much more sophisticated heating system for a part like this MIGHT make it practical. I can imagine developing a system of piped hot air that could be directed more precisely at the part, that wouldn't overheat the whole internal volume of the printer chamber. Without an even more sophisticated temperature sensing and control feedback system, though, successful implementation of such a system would be very time consuming and would carry the risk of thermally overstressing one or more key printer components before an acceptable balance of temperature was achieved, an expensive and risky proposition.

More promising is redesign of the upper engine part with a thinner shell. This will mean that at least one additional part in the finished kit for part mating will have to be added. I've done a preliminary thinner-shell design and compiled it for a print time that is almost exactly one hour shorter than the the failed part. That's still almost four hours. I don't know when I'll have time to try that, though, as I'll be completely absorbed with real-world work soon.

Finally, it might be necessary to break the upper part into two. This would still be an improvement over the current beta. If I had to bet, I'd say that's the most likely solution.

Having just finished the first of two "surges" in my real-world work, I've been able to devote a few hours to getting back to the V-2 revisions documented above. Hopefully I'll post some pics tomorrow or the next day. For now, the complete redesign of the fins and engine section continue to proceed on a path that looks like it can end with a final result that will be a major improvement on the previous beta, and that I'll be able to stay on course for first sale of kits some time in January.

The pics make it look a lot better, and the detail seems more pronounced.
Looking forward to more info my friend!

Progress . . . slow progress (notes after the pics):





Pic 1: The most recently-completed upper engine section in the printer, a few seconds from the last plastic being laid down.
Pic 2: The assembled components of all the most recent versions of engine section and fin parts.
Pic 3: A fin separated from the rocket body to show the attachment system.

The first pic shows the current state of work on the engine section body parts. At a print time just under four hours, each iteration for this part basically takes a whole day during the work week, if I'm lucky. Working at that slow pace, I have nevertheless had some fairly significant success in designing elements into the part to combat delamination from differential cooling, and also to greatly minimize the bad outer contour issues I was getting previously.

The same is true for the lower engine section part. It takes a little less time to print (right around 3 hours) because it's a little shorter and there's less plastic in it, over-all. For both parts, I've gotten to a stage where I'm now optimizing at a very fine level to try to remove plastic from the part little by little to a level that still addresses the structural and thermal issues and avoids the "external divot" problem I was having, but will decrease print time as much as possible. I'd say I'm three or four design-print cycles away from reaching the best possible balance of those factors, but that total print time for the two engine section body parts will still be over 6 hours.

The second pic shows a rough fit of all the parts. No surface sanding at all has been done, and the only other shaving and sanding work has been to get parts to fit. I expect some of the misfitted parts visible in this pic will be resolved in a final version.

The tolerances on this over-all design are closer than on the last beta to go to completion, so there's probably going to be somewhat more test-fitting and sanding and minor x-acto shaving involved in getting a good fit, even on a final release version. But the final result will be much better in terms of final fit.

The last pic shows some structure that's key to the new design. Splitting the fin from the body at a panel line has involved some rather more precise CAD-ing than I've done previously on this project, but the payoff is that a moderately skilled modeler will be able to assemble the kit so that there is no visible line between the fins and body at all.

In sum, progress toward the new design continues and I feel I'm on track for this to be the pathway to what will make it into the final kit I offer for sale.

Incredible.
That's looking a lot better! Well done there.

Putty-ing around (notes after the pics):




A weekend of long and frustrating experimentation has gotten me to a level where I have a set of parts I'm willing to invest the time in finishing. The first pic shows the new engine section parts glued together and puttied; the second, with puttied fins attached by their pegs.

At this point I've got an upper engine section body part that LOOKS great, but takes almost four hours to print and has produced some new challenges in fabrication. I pushed almost two pounds of plastic through the printer over the weekend, and count 12 failed experiments lying around the mad scientist lab. I've encountered some pretty heartbreaking failures along the way -- in more than one instance over three hours into test prints. To make a long story short, I'm working on a hypothesis that the extreme length of these prints is causing the stage bearings to heat beyond the level they've ever been to before, which can result in positional errors near the end of the process. I'm trying to control for as many variables as I can to test this hypothesis, but it's a painstaking process.

Meanwhile, as you can see, I'm trying out a new surface finishing technique: a full, heavy coat of Squadron green putty as the first step. I've been using this technique on smaller test pieces along the way, and think it may be a net improvement in surface finishing by putting all the mess in one step. I'll do some sanding and scribing tonight or tomorrow morning to see how it turns out.

That will be lots of sanding

Hmmm... Is it possible to "pause" the print halfway through to allow the piece to cool a bit?

pickelhaube wrote:That will be lots of sanding

It's the nature of the beast. There's going to be sanding with this fabrication process. Doing it this way with a heavy coat of putty up front just gets most of it over with early in the finishing process, or so it's been with the small parts I've been using this technique on.

Bar wrote:Hmmm... Is it possible to "pause" the print halfway through to allow the piece to cool a bit?

Not really. Yes, the print process can be paused, but the print head (which is hot) will be stationary just a millimeter or so above the part when you pause the print process. If it stays there for more than a few moments, it will cause Bad Thermal Stuff to the plastic it's positioned over. So far I haven't figured out a way to get the machine to remember exactly what position everything's in and return to it if I move the stages during a pause.

But, meanwhile, I think I may have figured out the conditions in which I'm getting a successful print of the Really Big Part. All three of the successful prints I've done have been from a cold start of the machine after it's been asleep for a long time. The failures were all begun after the machine had been running a while, i.e. the bearings were already somewhat hot before the print began. If the solution to the problem is just that this part needs to be printed first thing in the morning, that's definitely an acceptable way forward, as I won't be printing more than one of these parts per week, and I'll almost always have at least one day a week in which I can run the printer for four hours straight in the morning.

Hmmmm.
Well in that case, can you do the print in a cooling jacket in a cold room, or set up some household cooling fans blowing cold air onto it?

mmmmm so the ideal setup would be to have a warm print bed area, but have the actual print hardware/electronics kept cool?