Merlin Engine Static Restoration Photo Album. Slow but Ongoing….

Funny thing about the printer, it can slightly alter the Z axis scale on different builds, which makes fitting these parts together more challenging. But I’ll get there.

You might be able to tweak the models a bit by slightly scaling it in the slicer program that you use.

AfterCooler Box Printed, Awaiting Gluing

Well Sportsfans, it’s taken a while, but I finally have all the printed parts for the aftercooler box. They printed reasonably well, but it will take some prodding, twisting, pushing and heavy clamping to align all the parts while the epoxy cures. Funny thing about the printer, it can slightly alter the Z axis scale on different builds, which makes fitting these parts together more challenging. But I’ll get there.

You can even see the name “Acme” on the bottom of the box, which would make Wile E. Coyote proud.

What you see here is about 20 parts. This does not include the header plate on the port side, or the header tank, which are in the post just above.

I’ve uploaded 8 new pics to the Flickr site, but I’ll show a couple here.

Tom.

Hi Minimans;

I typically thread with a coarse tap, then screw studs right into the holes. I just print holes that I consider pilot holes (size is not super-critical), then tap without any actually pilot drilling. Some of my colleagues use helicoil or brass inserts, so either works. Fine threads tend to rip out more easily. Being plastic, it’s a little too easy to damage threads, so easy does it on the nut tightening.

I almost have the full aftercooler box printed, although the pieces need a bit of tweaking to fit together for gluing. I’ll try to post some pics soon.

Tom.

Is the material strong enough to drill and tap threads or will you have to use inserts of some sort? Although expensive this is a very cool way of building what is fast becoming unobtainable!

Cool Tom

Time for an Update

Here’s a newly printed set of parts. You Merlin guys will recognize the Aftercooler Header Plate and Tank. I’ll just add one picture here, but there are more at this link:

https://www.flickr.com/photos/130143587@N02/albums

Look under “Tom’s Merlin Parts.” I also added a bunch of close-in pics to the “Vintage Wings Gatineau Canada” album.

Cheers, Tom.

Scotavia;

Most companies seem to charge about $25 per cubic inch of melted nylon material. I just checked, and there are 466 cubic inches in my full cylinder head. So that would mean roughly $11650 Canadian funds, and that scares even me! I strongly suspect this would be much higher in the UK, because everything seems to be a lot more costly overseas. If you were building in aluminum, not nylon, it would be much higher.

My organization doesn’t charge its research clients using this method. We use a combination of my labour hours to load the machine, and facility (unmanned) hours while the machine is chugging away, to arrive at the amount we charge. And again, I simply use volumes left over in the machine, once I have all the project parts loaded. It’s a way of getting at least some use out of unused powder, rather than choking fish with it in our rivers.

Tom.

A great insight into what is a fairly new process. I understand you cannot quote actual total cost as you explained however just to give some broad ideas can you pass on some estimates? I have been in conversation several times with rebuild enthusisasts and the matter of 3 d printing larger parts comes up. But no one can give figures and this leaves a question about the use of 3d.

Nice work Tom, it will look the business when its done.

Kind Regards Mike

A Few More Pics of the Head Parts

Just a few pictures of the 12 separate cylinder head parts. I’ve tried to capture a closeup of the “step” joint, or tongue and groove joint that allows parts to be epoxied together with real strength. The joints will need to be strong, as the head weighs 20 or so pounds. I’ve done my first glue joint and I’ll see how it looks in the morning.

Filling tiny cracks and sanding each surface to make it look like one solid piece will take time, but overall the joints worked out, despite a bit of part warping during the 3D printing. Shouldn’t look too bad, especially once all the steel studs are screwed in.

Cheers, Tom.

Hi Trak-Tor and Stuart;

Yes, it is kind of like Lego, very expensive Lego. I didn’t do a cost estimate on any of these parts. I justified building all parts by telling my management that if there was leftover room in any Selective Laser Sintering build, which we use to make research project parts, I’d add parts for myself, or for other guys in my shop. Lots of skeletons and automobile cell phone brackets for everyone.

To answer your question, I would honestly bet that these plastic parts would be as expensive as real Merlin parts today, or close to it. Especially large parts if I had to send these to a print bureau. The valve parts would be considerably cheaper than real valves and spring sets, just based on their small size. Real springs are stupidly expensive (Jack Roush Aviation).

That’s a nebulous answer, but really all I have at the moment. Tom.

Hey Tom, that really looks the business, it’s almost leggo (in a giants toy box kind of way) what does it cost to produce individual sections?

Wow! Really great work! :applause:

Replica Merlin Head Parts 3D Printed. Awaiting final touches….

Hello Group;

Sept 27 was my last post, so the months really tick by. That’s how long it has taken to print 12 parts of the left Merlin head. And a few valve sets.

The head is around 42 inches long, so no chance of printing it in one shot. At best, if I owned the printer, it would be 3 large parts. Doing it in 12 parts made it easier to add parts of the head into other project builds, when there was spare room. Fitting in a few valves and springs, here and there, is easy.

I have a lot of work to do to complete this head. I’ll meticulously clean each tongue-and-groove glue joint, and then glue all the parts. I’ll have to create some alignment jig to make sure all the parts are nice and straight as the glue cures. Then I’ll do some “bondo” work at each joint to conceal each crevice that is visible, so that it looks like it was made in 1 piece.

I’ll post a few more pics of the individual 12 pieces, for those who are interested, and some decent closeups. Here is the address to my Flickr site;

https://www.flickr.com/photos/130143587@N02/albums/with/72157650467086187

More later, Tom.

:applause: Thanks and well done :). Something like this could be used in the future to explain the basics of how things were built and work alongside a static original.

Hi Trumper;

The density of the laser-sintered nylon plastic is the same as water, so yes, a lot lighter than aluminum or steel. Still, when you have a thick-walled full part like a cylinder block, she’s pretty weighty.

Most of the parts I have are original Merlin, so it’ll still weigh 1000 pounds or so.

Cheers, Tom.

It really is coming on ,very impressive.I don’t understand a thing about cad or 3d printing so this looks really good .I am assuming it will be considerably lighter than an original engine.

3D Scanned Merlin Parts. A look at the Process

Hi;

Just wanted to share a bit about the process of how some replica Merlin parts are made, at least by me. I took the 3D scanner from work and visited Vintage Wings in Gatineau Quebec last Saturday, and came away with some very usable scan data. My goal was to capture as many of the parts that I don’t have as I could, with hopes of reverse engineering them into a clean CAD model. I would then take the models, cut them into printable sized pieces, print, glue them into one part, paint them, then mount them onto my ongoing project.

I need quite a few parts, such as the aftercooler box, valve covers, and many others. Scanning took longer than I’d hoped, so I limited my attention to those items. It’s always better to have any item to be scanned, sitting on a bench in front of you, rather than scanning it in-situ. This is because you can roll the item into several orientations to effectively scan all surfaces and capture as much near-watertight data as possible. Scanning the aftercooler box while it’s on the Merlin was useful, but not perfect, as you can see from the pictures below. The partial, sort of shredded looking box is actually a merge of 3 scans, and unfortunately I’m still missing the entire bottom surface, and most of the front. Oh well, it’s a decent start.

The valve covers worked out better. They are flat on the bottom, and even mounted on the Merlin, stick out prominently enough that I captured the entire upper surface (almost). That way, I don’t have to guess at reconstructing much of the scan data to make it watertight and usable for reverse engineering or 3D printing. The grid-like things and the goofy bumps that are visible on the top surface of the valve cover, are both parts of the net that I had to drape over the valve cover as I scanned it. This net has a grid-like pattern of reflective targets which the scanner must see in order to acquire its scan data. There are many different scanners, but our hand held scanner relies on this 3D grid of “stepping stones” to understand what it’s seeing.

After scanning, I can “switch off” the target dots, and just work on cleaning up the ugly, rough scan data. The 2 valve cover images are both ends of the left valve cover, and the first thing that comes to mind is, “Where the hell do I start?” The extra peripheral junk is cleaned away. Then the 2 clean pieces that I want are carefully aligned and merged into one semi-ugly scanned model. From there, it’s a case of removing bumps, fixing holes, and eventually getting to the point where it looks finished, like the green valve cover. Since it’s an stl file (called STereo Lithography) I can actually print the item right from this format, or I can choose to use a reverse engineering software to make it into a really clean, crisp model. The softwares I use to reverse engineer the stl scan data, are Geomagic, and the user-friendly SpaceClaim. Sometimes good old Rhino, although it’s not really a proper rev-eng tool. Rhino is usually good for simpler parts, where I import the 3D scan data as an actual “part” and then build shapes that mimic the scan model. I weld those parts together until I have a big, often complex assembly. That’s how I modeled the entire V-Drive, cylinder skirt and head. Only now am I starting to use proper reverse engineering software to make solid models from scan data, and it’s been incredibly helpful. The advantage of having good software is massive time savings, in preparing the scan data to become printable 3D models. The cylinder head has taken well over a year, and is not quite ready to print. Still more details to add. The valve cover, admittedly much simpler, has taken 2 days of spare time to go from hideous to almost printable. I’ll add the Rolls-Royce label, because even P-51 owners with Packard Merlins strive to find the valve covers that say “Rolls-Royce.” I would !

So, bored with details? Have a look at the images and enjoy. Cheers from Canuckville, Tom.

Hi QldSpitty, yes I thought you were doing some 3D printing. You offer parts, do you not? Keep up the good work !

Creaking Door, I have considered keeping the core plugs out, then gluing them in place, just to make it look more realistic. However, you guessed it right, in that I’ll have to cut the head into 6 or 7 pieces to print, based on its size. That easily allows me to air-blast the unsintered powder out of the coolant areas in the head, although some of those nooks and crannies are pretty tough to reach, even with 100 PSI.

Referring to the real castings, I agree that they are works of art. The tiny passages in the head must have been very tough to make, or at least the cores must have been challenging. I think sand cores were designed to partly collapse when heated, but still, it’s amazing how these parts were made. So many internal areas that could go wrong.

Well, more soon, I hope. Tom.

I also have to hollow out this solid model, but with modern software, that’s not too hard. I want the inside coolant passages to look like the real thing…

More impressive work!

How will you get the unused powder out if you’ve modelled the heads with the core-plugs in place?

If I understand your 3D-printing process correctly (and I possibly don’t) won’t the coolant-passages be full of powder when the cylinder-head is finished; or are you going to have to split the cylinder-heads into sections to fit into your 3D-printer?

I’ve got a small section, (almost) a single cylinder, of crash-damaged Merlin. The coolant-passages are something to behold (although crude by modern standards) but I’ll never understand how the ‘sand cores’ fail to crack the heads as the aluminium cools or how the ‘sand’ is retrieved through those tiny core plugs!