GeckoSub Mirage Evo - And Adventures in 3D Printing Speargun Parts

popgun pete said:

What I found strange when I first saw the Mares bulkhead is how the o-ring groove is at the front of the main body of the bulkhead.

Are not the partitioning bulkhead body "O" ring always positioned at the front of the bulkheads rather than at the rear?

Click to expand...

Yes, I think they actually are! I have just been living in old MIrage land for too long and thought all bulkheads were the same as those. The Mirage bulkhead has the o-ring seat adjacent to the air transfer bore. I guess they moved it forward decades ago for good reason.
Sorry about the confusion!

popgun pete said:

It occurs to me that when you have your "Mirage Evo" completed and working, note I say when rather than if, Mares may start taking some notice and will be kicking themselves for not pursuing a new "Mirage" on their own initiative. No doubt Chinese monitor these forums and may think of doing something themselves, which would be very ironic.

Click to expand...

Thanks for the trust.
Yes, I have been thinking the same exact thing about Mares. If they did it, I think it would be cool - the basic idea is ripe for an upgrade. Especially if it's "just" two new molds for the bulkhead and nose cone. It must come up from time to time - if they are some people left in the company who appreciates brand history and performance.

Another thing that could happen is that someone else takes the fundamental idea, upgrades it and starts a (small) production run. Perhaps LG or a smaller, more agile company like that. Possibly, the patents have all run out by now? Actually, with the whole "makers movement" happening and the cottage industries getting a reboost it could be anyone with a inclination and machinery to do so. Though if you wanna go into production and then supporting a gun like this after the sale is the big question.

Not so sure the Chinese would get in on, though. Perhaps still too much of a niche product for them? If they can't sell in the 100/0s, they often don't bother. As you have pointed out in other places, they do clone bandgun parts (or OEM factories sell out the backdoor or rebrand), but in airguns, I think we still have only seen that Asso(?) clone you found some years ago.

Nose Cone - Prototype V4
First off, don't get confused - you haven't really seen the first three versions, but I am thinking I will keep the naming here consistent with my CAD program. Especially since a lot of what I write here is also intended for my own reference down the line.

I have been thinking about what Pete said about massaging the numbers and going with as small a pumping barrel as I can - all in an effort to increase the thickness of the walls in the nose cone without having to shift the barrels upwards from their original position.
I made some changes explained in detail below but I still have to lift the barrels but just by 1mm. I think that's the best I can do for now.

This will be text heavy and I don't have many images to go with this post, but let's start with a pic anyway.
Outer o-ring grooves just "indented", so I know where machine the real ones. Inner bores are undersized by 0.5-1mm and will be machined, too.


For this proto nose, I will stick to my idea of keeping the "shoulder" of the pumping barrel at 13mm - this in turn will drive other dimensions. It's still a darling of mine because it allows me to make a "healthier" sealing design. I effectively get a deeper o-ring groove while retaining enough wall thickness in the barrel next to the bottom of the groove. This deeper groove reduces the risk of pinching the o-ring on insertion - which was a real issue on my original Mirage nose cones.
It looks like this in the CAD (I don't know how to to draw compressed o-rings but this design will have about 25% compression):



With that decision I loose a bit of space between the bore of the pumping barrel and the shooting barrel. But then I had a look at the Predathor and its bulkhead. And Salvimar uses a really small o-ring there, seated quite deep in the barrel. This means, the o-ring doesn't "stick out" much and thus, the surrounding bore is equally small. In practical terms, they use a 1.6mm o-ring (though I have Salvi barrels which shipped with a 1.5mm o-ring) and the o-ring bore in the bulkhead is just 18.25mm.
As mentioned, this is possible since the o-ring groove is fairly deep, deeper than any other I have seen:


The only real issue is whether this deep groove of only 15.7mm diameter is enough to weaken the barrel to such an extent that it might split some day...! Apparently Salvi doesn't think so but I will be running higher pressures so the impacts from the muzzle will be higher, too. The thing is, while Salvimar trust this solution for the rear end of the barrel at the bulkhead they still run a thick, 3mm ring with no groove at all at the nose. Perhaps they are afraid of the effect of any corrosion to the front groove which will be in contact with salt water? That said, I will try running the thinner 1.5mm seals in both places. If I really wanted to protect that groove from the ocean I could place a seal around the muzzle but that's overkill;-)

So, with inspiration from the Predathor's diminutive seal design, I have tweaked my numbers a bit and by going with a 1.5mm ring with a bit more compression (Salvimar runs only 17% compression on that seal), I can get the groove a little less deep at 16.15mm vs. the 15.7mm of Salvi's barrels..
It's something I am willing to try.

Besides wanting a specific design for the front of the pumping barrel, a few other things drove this decision. The minimum order for 10x12mm stainless steel tubes were a bit too high for this prototype. I did find a 10x12mm titanium tube but the vendor admitted the inside finish was not very smooth, but I could get a 10x13mm alu tube of known quality.

Here's a quick print I did just to get a sense of things (actually not V4, but V3; only the bores are slightly different). This time in a "lower resolution" = taller layers. You can probably tell it looks more "steppy":


Also, I realized holding it in the lathe for when working on the rear of it is going to be a challenge. The jaws wont like that front:


I used this nose to practice machining the outer o-ring grooves and they look alright.

For work holding I am going to go simple with a press fit mandrel inserted from the front. Even if I scratch the bore a little in that part of the nose cone, it's OK; there are no seals there anyways.
It will look like the alu mandrel inserted in the tool holder in this pic, but it wont be put in a tool holder but in the four-jaw chuck instead:


I could print a perfect split female mold of the front of the nose and use that as a jig - but it would take me a few days to learn how to make that part in CAD, so I wont.

The fit is great, I have had that dialled in for a while hence why I am putting a lot of effort on the sealing now. It was always the main issue with the originals, too. Got you on the anti-banana idea, too.
Don't know if I have time to make the switch for the first version of the gun, but will add it to the list of possible improvements - if the design allows.

Not much news from today - I am heading to North China for some work in the afternoon, but I did get one print done and will see if I can get more one done before I leave - the nose cones take about four hours and the bulkheads a little less than two. While the printer has run fine from day one, I don't leave it unattended for long spells. There are some nasty stories of people leaving a printer by itself only for it to decide to try to burn the garage or house down..

But I took Pete's advice and dunked one of the testers in some gun oil:



It's been there overnight but I will check on it after my work trip and report back.

In lesser news, I managed to get the supports dialled in a little better. The trick is for them not to fuse with the part itself but if they are offset by too much, they don't really offer much support and the layers above them, which they are supposed to support, will sag a lot.
Here's a slightly modified bulkhead I printed this morning:


The supports now come cleanly out with a firm pull:



I don't have a good pic of the surface they left behind but it's less saggy than last time around.

I have ordered a few more reamers and a smaller boring bar, so I should be good to start machining in anger in a few days.
Also, I have some small stainless balls for the check valve on the way.
I have mentioned this before, but for all this DIY stuff I really like being in China where I can get pretty much anything online in a matter of days - and often quite cheap, too.

Shopping
Some goodies have arrived over the last few days.

I have been reading a lot about SpearQ8's results using stiff dyneema as shooting line, so wanted to see if I could source something equally useful locally:


The thicker one is certainly stiff, but not sure the thinner one is stiff enough or will crimp well, but I will figure that out when time allows. Price wise, not much cheaper than buying the proven line from the US, actually. Not everything is cheap in China any longer - but as you can tell from the next few items, some things still are.

Moving on - Stainless Steel balls in various sizes for the check valve I need to make (less than four bucks incl. shipping):


Also, the alu tubes for the pumping barrel came in. I got two pieces of 10x13x1200mm and two of 10x14x1200mm and a piece of brass for the check valve for USD 12:


I needed some more reamers, so here's an assorted bunch:


The shop had a cheap version and a six times more expensive one for each size. I went cheap and I don't know what tolerances they hold - but they are not far off, def fine for my needs. About USD 1.5 each.

But then I realized, I might be able to actually get boring bars small enough to cut the 9/10mm bores for the air transfer port and the check valve bushing. I also got some more alu-cutting inserts:


The boring bar to the right is my old one. I think the smallest it can do is ø12mm. With the smallest of the two others (on the far left) I can go down to ø8.2mm. So, the reamers wont really be needed after all.
The "02" on the lower blue box refers to the radius of the insert (I think it is 0.2mm) which is half of the other box which I normally buy. This should make them more "pointy" and give me an even better finish. It will also make the tip more fragile but for the 3D printed plastic parts that's not an issue.


Only 15 bucks for 12 inserts and two boring bars. I need one more tool holder and that will cost about the same.

Filling In and Something About Bananas...
The nose cones I've printed so far have been printed as solid parts.
But as you can see in this pic, there's a way to save some weight and perhaps not loose too much strength:


Now, this zig zag pattern is not something I have painstakingly modelled in the CAD software. I am not at that level at all, yet. But all 3D slicers - that's the software that prepares the CAD files for the printer - can do this. Most 3D CAD models come into the slicer as a solid model and this feature is basically a material and time saving setting. It's called "infill" and is incredibly easy to adjust - most slicers just operate with a percentage for the infill, which determines how dense the infill is. Combine this with the number of walls for the inner and outer perimeters and top and bottom layers and you have a pretty easy, but powerful way of playing with the balance between weight and structural strength.

But I digress so let's move on to what I was doing with that bulkhead tester in the first place. You might notice it doesn't have a sleeve protruding from the shooting barrel bore as some of the earlier versions did. I put the sleeve in to move the shooting barrel o-ring groove away from the bore for the pumping barrel. This way, I could keep the maximum wall thickness between the two bores in that critical area. But I wanted to test if, with some careful machining, I could get away with not having the sleeve.
And it just about works. Here's the part after both bores have been machined to final dimensions:


I have about 0.9-1.0mm of wall thickness between the bores and under the pumping barrel. That might just work. I feel it depends on how well the parts have been printed - how well the plastic has fused. But shining the beam from a torch light through the thin walls makes it look promising.
There's one issue, though. There's very little room for a proper chamfering of the edges of these bores and that's another thing I found really aids in not pinching o-rings on assembly. But if I wanted the simplest nose cone possible, then this is it. It also makes holding the part in the lathe a whole lot easier.

I feel I got the bores spot on in terms of compression of the o-rings. It's just about where it's difficult to insert but not too bad:


As has become custom for me, I fit two o-rings whenever I can.

(BTW, that's my preferred sealing design for the front of the pumping barrel. 13mm OD (shoulder), 10.9mm D of the groove and the o-rings are 13 x 1.5mm (OD x W). About 25% compression with these numbers).

Now, since it seems I can actually place the shooting barrel o-ring groove just next door to the pumping barrel one, this opens the door to Pete's suggestion of adding length to the lower part of the nose cone to help counter the banana'ing forces.
I fooled around with it in CAD just now, and it could end up looking something like this:


I might give it a go, actually. But next up, I think it's time to sort out the check valve for the bulkhead. Need to make that pretty bulletproof.

Thanks Pete,
It has crossed my mind on other occasions that in a complete redesign of a Mirage, you could use the pumping barrel to the same effect. It would have to be threaded and somehow locked in place at both ends but perhaps the bulkhead will then have an issue. But I still have crazy plans for an even more advanced Mirage down the line which may not even have this issue. Either different shape of the reservoir or a different pumping system. But that's for much later.

You talked about getting the fit right in another post. If I can get the the nose to be have incredibly little wiggle room that should, of course, help, too. I am close but perhaps there's another 0.05mm I can add to the diameter;-). I just need to make sure the plastic doesn't swell - or shrink - the least bit when subjected to the oil. Later today, I'll check my test piece which have been living in oil for a few days now.

"Hey, What If...?"
I was supposed to be on the lathe today working on the check valve bushing between the bulkhead and the pumping barrel. But I guess I felt like spending time in front of the computer instead...:


I know, I know, not much different than the previous versions. Except, I did add the peg for the the pumping barrel. For a few reasons, I didn't want it there in the first place. One is that because of the way the layers are laid down, it has less strength across the peg and I am worried it might snap off if I am silly enough to loose grip of the pumping barrel while it is plugged onto the peg. But then I thought, perhaps before I start machining the check valve, let's give it a try at modelling it just for kicks:


There a few unknowns, such as will the o-ring in the valve stay put when the air rushes past it or can it move enough to get skewed. Judging from e.g. the pump inlet valves, it doesn't seem to be a real world problem.

But then, I thought...: "hey, what if..." and a little while later this is what I had drawn up:


Yeah, it's getting brave and advanced now;-). But if it works, it would be amazing. The "sleeve seal" in the original Mirages actually works well, but if I can make this on the printer, it's one less thing I need to turn on the lathe.
The printer needs to do a good job on the threads and it would be cool if the bores don't have to be touched up on the lathe. The one from the inside will be hard to reach - I am not sure I can get a long end mill or reamer of the right size easily. But come to think about it, since it is a static seal - which I actually don't want to move, it might be a good idea to keep the ridging. I'll start that way at least.

The part is on the printer now. I will not print the whole bulkhead - just the front half with the valves, so I can see what needs to be adjusted before printing a real one.

Yes, I have decided to leave that one-way valve out. I have never had it open - have just counted pre-loading strokes and it's worked fine. But I have had a leak from a very flattened seal in that bore. So, leaving it out all together means there's one less thing to worry about

Bulkhead - V8 Proto With Check Valves
As mentioned above, I printed the front part of the new bulkhead with built-in check valves.
The air flows through nicely, so the printer did a good job on the bend in the channel between the two bores.

And the threads and threaded bushings printed well, too as this little video shows:

(Don't mind those little blobs on the lower threads, they came right off - it was some weird setting in the slicer software. On another bushing, they were not there).

Now, I needed a way to verify that the check valves actually work. So, I fashioned some small test pieces. One version looks like this:


And with a few others:


They mount on a Salvi pump for the sole reason that the only Mares pump I have at home is completely busted - the threaded front part shoots out with almost every pump stroke:


I mentioned in a previous post that, while I didn't think it would be, it could be an issue that the air would blow the o-ring around and that it would not seat back properly and thus leak. I speculated it wouldn't happen as it doesn't in the pump inlet valves. But it does happen in this valve. I am fairly sure it is because the spring in inlet valves is very strong, whereas it is not in my check valves.
I then checked the original Mirage bulkhead - which of course I should have done first - and it does have a 'captive seal' design (at least I think that's what it's called) where the o-ring has a shoulder in front of it and a bushing behind it. So, it can't really move. The spring in that check valve is much softer - so I think that's why the o-ring needs to be kept in place by the shoulders.

I changed the design a bit:



And this version works. It releases easily in the direction of the flow, blocks firmly in the other and I can't get the o-ring to skew:
So, at least I think I have a check valve design that works now.

But more on what works and what doesn't in a bit...

Houston, We Have a Problem...
But first, say hello to my Apollo:


And then allow me to backtrack. It was a beautiful day in the apartment-yet-again-turned-workshop and I was beginning to see the end to this magnificent quest. As the printer performed its stepper motor symphony, I was chucking happily along testing and refining my check valve design when I thought a glass of water was in order. After all, you do see things more clearly in water sometimes...:


You see, those bubbles shouldn't be there. It's not that the valve is leaking - it's the whole 3D printed body of wonder that is leaking... The stuff doesn't hold air in! Shine a bright torch through its thinnest walls and all looks great, but on a microscopic level this is Swiss cheese.
And just like that, my mission has hit a potentially catastrophic speed bump. At the very least, it's the mother of all Mirage-related setbacks. (And yes, why did I not do this test weeks ago...).

In a strange mix of denial and logically looking to verify this shocking discovery I drew and printed the Apollo nose cone. I needed something real, something with a proper resemblance to the Mirage parts I was making. The easiest and fastest was a nose cone for a Predathor.
Took a few hours before Apollo loaded up at 20 bar crashed heavily in the kitchen sink:

The air just comes out of everywhere in the nose cone. It is now 100% certain that the Moon nose is made of Green Swiss cheese...

Now, my mind has been racing to come up with possible solutions.
With a ton of luck, one of these could potential work:

• "Anneal" the parts (Some people anneal their PLA part, but I don’t have harbor high hopes this will work)
• Print in other materials such as PETG, ABS, HIPS or polycarbonate (I have PETG, so will try that)
• Coat parts with epoxy (Probably won’t work as the leak paths are everywhere and 30 bar of pressure would likely lift a surface coat right off)
• Impregnate parts under pressure with a very thin epoxy or PU (Could work, but I would have to build a pressure chamber)
• Print the parts with SLA process (I don't have an SLA printer so would need to outsource this)
• Simplify the parts and cut them on the lathe in Delrin/POM
• Use printer to make female molds and make the parts in either carbon fiber or a pourable polyurethane

Stay tuned, something's gotta work out.

Rescue Mission - Annealing
I mentioned other people have "annealed" their PLA prints with varying success. It does make the part stronger and stiffer but dimensions change, too.
I baked a handful left over parts - some bulkheads, some of the check valve testers and a few of the calibration parts. All of the parts shrink along the X and Y and grow in the Z direction.

The parts do come out a whole lot stronger but while the dimension changes on the simple parts can be handle in the CAD but adjusting for it the problem is that the more complex parts change in complex ways. I think the change happens in relation to the amount of material in any given spot in a part.
E.g. the bulkheads shrunk a lot more at the end facing the handle where there's less material.
My feeling is that it is going to be very hard to design for these changes in the complex parts. You could possibly insert jigs to help keep the dimensions during the bake and I am considering that.
But one more test is in order before I go down that route. The whole idea in my case is to see if the annealing would close the small gaps and stop the leaks.

For the red PLA which I have used so far, it still leaked - I tested on one of the check valve tester which had been baked.

So, I think annealing PLA, even if I could control the dimension changes, doesn't help one bit in terms of stopping the leaks.

Rescue Mission - Different Materials (PETG)
I have some PETG from two different manufacturers and originally, I wanted to print in PETG as it resist higher temperatures better. But PLA is so much easier to print, so that's what I settled for. I just have to keep the gun away from strong sun...

So, on to checking if the PETG suffers less from leaks. So, I made a quick test design - just a long cylinder with a pump inlet valve at the end of it.
In the kitchen sink test both the green and silver PETG failed. I think it might even leak more than the PLA...

BUT let's anneal the PETG:



Nah, didn't help much, if at all..
But then I thought, what if I raise the temp a little but then this happened:


I was about to chuck the part in the bin when I thought I might as well test it in the sink and it does hold air completely...:


The silver PETG also went banana shaped when I raised the temps a bit, but kept it under what I had baked the green stuff at - but the silver part still leaked:


So the problem is that the point where the material fuses fully seems to be above where it can hold its shape, so it's kind of a worthless exercise.

The conclusion is that PETG in both printed and annealed form wont solve the problem.
The search continues...