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

[Diving Gecko]

I gave this thought a go and it is probably the one place where a roller/invert-style arbalete is lacking. Chamber fishing is a no-no unless you want to destroy your shaft haha
Power regulation is still possible by loading an intermediate sharkfin like the "loading tab" sharkfin some rollers use halfway on their shafts but the kgf loaded on the shaft would still be 2x of the Mirage which I think kind of defeats the purpose. A possible solution would be a limited edition very short hybrid, something like a 40cm in bandgun length terms so overall maybe 60cm from stock to muzzle. Fully loaded this would still work like a pneumatic with ~55-60 kgf on it and shots up to 2,5m can be feasible I hope, which is more than a classic bandgun double its barrel size with two pairs of rubber. But to have a 40cm shoot that far would be rather special. So that's another project put into the list haha, thanks for that

Got it, that all makes sense.

Here's are some older pics from Thailand. Before my Mirage obsession it seems. The "big" gun is a Sporasub One Air 120 shooting a ~120cm shaft but it was a bit of a "weird" gun in that it has a detacheable handle but that means the shaft can't insert as "deep" into the gun as in classic oleos. In effect, the gun ends up longer for the same size shaft. It does however mean you get a handle with a very high grip. No power regulator, though. Also, it has a very chunky reservoir so the gun is actually heavy for its size. It's by far the gun with the least recoil that I have shot.

The small gun however is a classic Mares Sten in the Medi size. Also referred to as a Sten 70. This one is newer version, they have been making them since the 80s, I guess and there was another revamp a few years back but mostly likely only a facelift.

That's the homemade shooting line, btw. So, yes, I admit a weird setup - a float for shore diving whilst on a boat, a tiny gun but 30m shooting line with ~200kg dyneema inside, haha!

Anyhow, back to that little Sten, this one has a vacuum muzzle on it, of course, and a carbon tube but speaking of the 2.5m, I think I shot some small trevallies and reef fish pretty much at that distance. And that's on a classic one-stage-loading gun with perhaps 20-22 bar in it. Not sure how long the shaft on it is, maybe 75cm or so (I have a document with many measurements of my guns but forgot to add this particular one).

Another aside, I may actually be getting into more proper shore diving again, so I'm putting together a bit of a board again (this time a flat mini-SUP as that's what I could get).
If I really get into hunting in holes which might be smart on this island, having a short oleo would be the perfect tool for it and having a board would make it easy to carry a spare gun.

It would be better if it was 10-15cm shorter, but it was cheap and you wont be surprised thay I have already drawn up a micro catamaran, haha. Think of it as a much less draggy version of the ones shown here. Time, I need more time.... But for now, the inflatable mini-SUP will do. I have already glued in some D-rings for a backpack strap and waiting for some smaller ones so I can add attachment points for guns and gear.

 

[Diving Gecko]

A note here, wobbling can reduce energy but leave accuracy unaffected (it still shoots straight even if it nosedives sooner). But I have seen 8mm double roller shafts bend like bananas during firing and the shot was still fine. So either the losses are small to begin with for all types of shafts, or they are amplified on lighter/thinner shafts due to increased drag due to the anti-hydrodynamic bending. But 6,25mm shafts have been tried on rollers and invert rollers and they had very good results given their low mass, including in penetration which means they didn't bleed KE. So I'm inclined to believe that there aren't significant energy losses to begin with. Yes shafts wobble during flight but it's only the initial plastic deformation and particularly the acceleration curve due to jerk and recoil that bleeds the shaft's energy. The wobbling we see during flight seems to be much smaller than the initial one... which also bands the shaft against the guide. So this makes sense (ETs already guilty of this). If these factors are removed then the initial wobbling isn't that big and has negligible energy losses on shafts of 6,75mm+ diameters. However it is a very visible effect on camera, just like in bow/crossbow arrows. So it is expected to get blamed for factors far less visible or intuitive to the eye
Any initial stage vibration on pneumatics happens inside the chamber and thus meets much lower resistance. So there are reduced energy losses even there no matter how tiny. Energy conservation at the beginning of the shot is exponential to preserved energy at the distance we want to hit

Now, to the rest:
2x kgf of the Mirage would be around 56kgf but the load effort isn't technically 28kgf, to give you a clue. It just feels like as if it was 28kgf on the Mirage due to both hands being used as well as back muscles
From what you say I take that this is a very powerful all-around so maybe even 50kg would be feasible for even easier loading on a single chamber without pump. So this is an option I am keeping.
This same setup can produce a version with halved loading effort either using a pump or a very nice trick (very sparsely used from what I've seen since it can't be utilized well with rubber bands) that is like having a pump, without a pump. It produces a similar effect. Because it would be possible to just lighten up the initial version (2x Mirage) without a pump and operate with lower atm, I think this would be useful in the version with double kgf (4x Mirage) which can be done either using a pump or the same trick as before.
From your examples, both of these options (2x Mirage and 4x Mirage) sound like they'd have their respective audience. Both load much faster than an invert roller also

Any gun with a single pump would be bulkier and less maneuverable. The 6x Mirage version I thought of therefore seems pointless I think. It's better to try and beef it up, initially I thought of two shooting barrels and two pumps to get to that 8x. But 25 bar on two 25mm pistons placed side by side is a bit bulky for an airgun so maybe 25 bar on a single 35mm piston would be better assuming 2 pumps - one on each side. I prefer having two smaller pumps here that can be engaged in a single pull instead of a single larger one. 6x is the "weaker" version of this with a smaller kgf capacity. But this can be achieved in the previous 8x gun as well, with a lighter setup anyway
For reference Pete said that Velair has a 32mm piston and its body is a mostly compact cuttlefish

Some more good news is that because pumping is done by hand (both hands) you can do all of it on the boat and when jumping in the water all you'll have to load is the equivalent of a single 16mm band pair, maybe lower. So maybe this can work as an alternative to what you were thinking of earlier plus the line would be already wrapped since the shaft would be like in bandguns before they're loaded.

But that's enough analysis for now. I'm keeping the 2x (no pump), 4x (no pump for the sake of simplicity) and 6/8x (double pumps engaged simultaneously) and I'll try to make some designs that hopefully lead to a prototype if I can get the right people interested in it
Fun fact: just a week ago I had no idea that I'd ever get such an interest in pneumatics. And I've already known of Dreamair and Velair for years, yet the spark was because of this thread and the Mirage. So if this actually materializes to something functional, all the doggies will have you to blame

Might get to the finer math points in this later, might not, but all I can say is that it sounds intriguing
But I propose you seek out at least 1-2 other hardcore oleo guys as I haven't been super active the past two years, but when I have quoted numbers, I have tried to also make it clear what I can load and what I think others can. I would consider myself the average in terms of available power. I know some can do more. It's worth double checking.

And yes, having the shaft already on the gun, latched in the trigger on the boat is a pro over a Mirage. Other pros: Being able to chose from more shafts and triggers as well as having a clearer sightline. And more easily be able to design a handle with a very high grip.
Another tiny pro, the slip tip line is normally tugged under a band and I have yet to come up with a clever way to do it on the oleo (I will work on some type of little rubber cleat). I guess you can tug it under the dyneema on a hybrid, just as you would under a band on rubber gun.
A pro for the Mirage is that for the first time in history, it will be considered the less complicated gun, haha. At least compared to a hybrid.
Manufacturing these would be more involved/complicated whereas a Mirage can still be made out regular round tubes. It took a long time to get the Dreamair to market, but I do think the market is there.

Now, my dream Mirage would have a custom cuttlefish CF reservoir but it's trickier in one important way than the Dreamair's - because the Dreamair uses such large pistons, they only seem to run 8 bar of pressure which is surely a huge help in terms of building it strong enough. But the rest of the Dreamair stock is way more complicated - I haven't seen the inside, but the CF work looks great, integral handle, integral reel base, etc, etc. This part would be much easier for me, I only need the outer cuttlefish shape and tapered front for a bit more total volume and much less drag when tracking. Nothing is going on on the inside, the reservoir is just a shell (until I get to the built-in lever pump or reel mounts). It will all be about whether I can make it stand up to the pressures. Worst case, I may have to do some internal bracing but whether it would stay in place or rip itself from the inside walls I don't know. (If I was setting out to put these into production, a hydroformed alu reservoir would likely be the best option but each size needs it's own mold and they can not be cheap).
And then I really want to do a custom handle as I have found some very worthy improvements I can add there. Both are projects I hope to get to this year. The handle comes first as I can validate it on normal round tubes.

As for the CF cuttlefish reservoir, there are many ways to go about this for prototyping and small "production runs". 3D printed molds or wax or foam mandrels are some ways. My plan is to to make a shortened one first which still has the full belly as that's where the structure will see the most force - so basically the real design but shrunk length wise to around 40-50cm. And then that will get blown up to see if this is feasible at all. But it's also one of the reasons I want/need to go to a 15mm piston over the traditional 13mm. Both for the bigger ratio for loading but also to lower the overall pressure.

Writing this, makes me think I should go even bigger than 15mm... Historically, the adversity to big pistons come from before the time of vacuum muzzles where the space around the shaft inside the shooting barrel would be flooded with water that had to be pushed out during the shot. This is the main reason why 11mm pistons became a thing and it's why people used to enlarge the ports in the old muzzles to let water out more easily. And it's the whole reason why a vacuum muzzle is such an improvement on an oleo.
But people like to state that a large piston has more friction and the added mass of it could potentially lead to breakages of the piston of shock absorber. But I am fairly sure I can make a thicker pistons almost as light as most of the classic 13mm pistons even many of the supposedly optimized ones. Which only leaves friction, which might not be enough to mattter much, especially given the upsides.
So, perhaps the only real downside would be that I wouldn't be able to slim down the front as much as I'd like.

Anyhow, I'm happy that my meanderings helped spark your thinking and thanks for you doing the same for me

 

[Diving Gecko]

Losses With Depth - A Problem Or Not...?

An emerging meta I've seen is deep spearfishing with pneumatics because they have fewer losses than bands in increased depths

Yep, I think I saw Giacomo or one of the other very deep guys do a lot more hunting with a pneumatic since a few years ago.

And you are saying that bands also contract less efficiently with depth, right? I never thought of that.

I have however thought about the same in airguns, surprise, surprise, haha. The losses with depth have actually been used against the oleo perhaps as they are pretty easy to prove theoretically. Say, you shoot at 40m, that's 5bar of outside pressure pushing the shaft into the gun vs. 25bar in the gun trying to push it out - so you'd lose 20% right there.
Except, my math is wrong, since if an oleo has 25 bar of pressure in it, it actually has 26 bar (the gauge starts at 0, not counting the 1 bar of atmospheric pressure). And the atmosphere works on both sides of the shaft, inside and out, front and back, so can be eliminated leading to the loss being 4/25 = 16% and not 20%.
(At least that's what AI told me)

For my use case, most of the time, my max depth is 25m and on a 35bar Mirage, that would equate to a ~7% loss (or ~11% on a non-Mirage with 23 bar at the same depth).
Now, an 8bar Dreamair at 40m we might be losing 50%... or ~30% at 25m?. Unless somehow the piston is completely isolated from outside pressure, but I can't wrap my head around whether that would be possible. I would hope so, otherwise this is a bit worrying for the deep divers.

So, I guess this is an area where smaller pistons/higher pressures have substantially less losses than larger pistons...

 

[Diving Gecko]

Finishing The Non-Violent Mirage Valve
I finished the experimental Mirage valve today - just that front cap with a spring inside was missing:

It took longer than I thought as I broke some tools along the way and had to scrap the first attempt but it's been a long time since that happened so that's alright.

But mounted on the handle with the bulkhead it looks like this in the closed position which is the one for loading/pre-pumping:

In this position, we now have the knob at the end of the handle locked in the rear-most gate and the spring inside the cap is helping to push the gasket onto the lip in the bulkhead. The idea being that with just the first pump stroke we will already build a pressure differential to really push the valve shut, but we can't have it leaking until then.

Next pic shows the valve in the equalizing position. This is where air from in front of the bulkhead is now allowed to move through the internal channel of the power regulator rod, effectively bypassing the bigger cap that's still closed against the bulkhead:

You can also clearly see the spring here. Though I may need to machine a new cap and front dome. I think there's a risk the spring wont stay centered, it may somehow move and possibly knick the o-ring. So, I am considering adding a little flange or slot to the front face of the cap to keep the spring from moving about too much. But I will try it as it is first.

Finally is the shooting position where the power regulator is pushed fully forward (once the pressure has equalized enough to do so):

The cross area of the valve opening (taking into account the obstruction by the 4mm rod) would equate to a 9.7mm classic, circular bore. So, bigger than pretty much anything out there. And Zahar (and AI) say that above ~8mm (I think it was), we are supposedly not experiencing any throttling anylonger.

And yes, the valve is a weird shape and you may be worried it would turn and not seal properly but it's kissing the inner wall of the reservoir and the out erwall of the shooting barrel and the pumping barrel check valve, too - the idea is that those three will constrain the valve and keep it from turning. Here's a pic though I haven't made the pumping barrel valve yet, but it will go into the empty bore next to the valve:

Let's have a look at a few pics of the machining as there were 1-2 interesting things. The main challenge was always going to be that slot for the spring. I would have to use a 1mm endmill, one with extended flutes (making it even more fragile). So, first I drilled a bunch of holes with a small drill to remove as much material as possible and then I went in with the endmill. I had a failed attempt where I drilled 0.8mm holes and but the 1mm end mill broke during the slotting. So, I played it safe(r) the second time around and drilled 1.1mm holes but used the same 1mm end mill.

It worked but left a bit of a wavy wall. I could have cleaned that up with two contour passes, or I could have tried again with an undersized drill and even more conservative feeds & speeds for the 1mm end mill. But as I said, it worked and this is just a proof of concept anyways:

Afterwards, I milled an M3 thread into the middle bore. And off-camera, I then machined a 2.5mm cross hole that I can insert a small screw driver into when I need to assemble this thing. Next, on the lathe I cut a little brass holder so I could mount the part with an M3 screw in the newly cut thread:

And then back to the CNC to shape that front dome with a ball end mill:

That was the gist of how the machining was done.

Will it work, though?
I think the physics of it all should work. But I see a small handful of potential killers;

• The gasket may come off of the alu cap. It's just super glued in place. I think that would be solvable, though. As far as I recall, the Salvimar valve may not even use glue, perhaps just some double-sided tape. Anyways, I could upgrade the super glue to a marine adhesive like e.g. 3M5200 which is not to be trifled with
• Tomi pointed out that the gasket could possibly be cut by the lip on the bulkhead. If that happens, I think it's a fairly easy fix, too as I can make the lip more flat or blunt with some sanding and then change the CAD to reflect that for when I make the next bulkhead
• There's also the potential issue I mentioned where the spring might catch the o-ring and knick it. Changing to a captive o-ring inside the cap would solve that but so would the redesign with a little flange/slot at the front for the cap to keep the spring from moving sideways
• Finally, and related to the previous point, I worry the most about the general longevity and reliability of that same o-ring. If that wears out, the whole Mirage system goes with it. And it's a tiny o-ring, even smaller than the ones used in 1.5mm triggers which also means that it's more picky about getting the compression just right and more susceptible to leaks if tolerances in the machining or the o-ring itself are off

A possible solution to that last potential problem would be to change to a captive o-ring inside the cap instead. That would allow for a bigger o-ring to be used which would make things easier. But the downside of this design is that the front cross holes in the rod need to pass through the o-ring and of course, that might cut the o-ring, too. But if I keep the holes small, maybe go to three smaller ones rather than two larger and round off the edges of the holes nicely and polish them, too then perhaps it could work. It would look like this:

@tromic and @Zahar
Do you think it would be possible to make the above valve work, the one in the last drawing? Notice how the front cross holes in the rod have to pass through the o-ring - That's what I worry about. Do you think if the holes are small enough (smaller than in the drawing) and the edges rounded and polished that the o-ring could survive...? Perhaps some of the gun builders in the Black Sea has some experiences?

P.S.
I have mentioned that I wasn't sure if the ~3.5kgf needed to open the current valve by pushing against it with my thumb would be hard or not. So... I did what the kids do these days and asked AI. Two of them. "They" agree that it's very doable, especially when being able to counter against the grip. One called it a "moderate" force, the other said "clearly noticeable but easy". So, that sounds alright.
Also, the 3.5kgf is at a pressure differential of 35bar and hence, pretty much the worst case scenario.

I guess I will find out. Perhaps tomorrow or Monday when I will hopefully find time to make the new check valve for the pumping barrel. That's the last missing part of the puzzle but I ran out of time to make it today.

 

[rollerGR]

Losses With Depth - A Problem Or Not...?

Yep, I think I saw Giacomo or one of the other very deep guys do a lot more hunting with a pneumatic since a few years ago.

And you are saying that bands also contract less efficiently with depth, right? I never thought of that.

I have however thought about the same in airguns, surprise, surprise, haha. The losses with depth have actually been used against the oleo perhaps as they are pretty easy to prove theoretically. Say, you shoot at 40m, that's 5bar of outside pressure pushing the shaft into the gun vs. 25bar in the gun trying to push it out - so you'd lose 20% right there.
Except, my math is wrong, since if an oleo has 25 bar of pressure in it, it actually has 26 bar (the gauge starts at 0, not counting the 1 bar of atmospheric pressure). And the atmosphere works on both sides of the shaft, inside and out, front and back, so can be eliminated leading to the loss being 4/25 = 16% and not 20%.
(At least that's what AI told me)

For my use case, most of the time, my max depth is 25m and on a 35bar Mirage, that would equate to a ~7% loss (or ~11% on a non-Mirage with 23 bar at the same depth).
Now, an 8bar Dreamair at 40m we might be losing 50%... or ~30% at 25m?. Unless somehow the piston is completely isolated from outside pressure, but I can't wrap my head around whether that would be possible. I would hope so, otherwise this is a bit worrying for the deep divers.

So, I guess this is an area where smaller pistons/higher pressures have substantially less losses than larger pistons...

Yes high pressures make bands lose a lot of energy, not entirely sure why as there are a few explanations given (drag slows them down, compression counteracts stretch ratio, cold water temperature and some other stuff)
So a 16mm might behave like a 14,5mm which is 20% energy loss or more hence why a lot of people use 4-band wooden guns for aspetto here and in Italy. Which can have the same power as a high pressure pneumatic but way worse tracking and recoil. Here, recordman Giankos has switched to pneumatics for a while now

From some math I found

Force = ΔP × Piston Area

Gun A: 30 bar internal
-Surface ΔP = 29 bar
-40 m ΔP = 25 bar → ~14% loss

Gun B: 14 bar internal
-Surface ΔP = 13 bar
-40 m ΔP = 9 bar → ~31% loss

However this is theoretically eliminated with a hybrid. See:

Another advantage of the "Dreamair" is the gun is unaffected by ambient pressure as the compressed air and sliding piston system is totally enclosed. It will have more frictional losses than a standard pneumatic speargun, but has the big advantage of loading from the rear.

As water never gets inside the gun the piston has pressure on one side and a near vacuum on the other in what is a sealed system, so it is completely unaffected by ambient pressure at depth. [This is from a different post]

Which is an unexpected bonus. I put two AIs to check this and both confirmed it, as long there are seals there are no losses due to pressure

Bit of a wild idea but wouldn't it be possible to have a secondary pump on the Mirage (or a single pump on a non-Mirage airgun) that is used to pump in-barrel water out? So instead of transferring air into the "Mirage chamber" before loading it pushes water out from a hole after loading, kind of like an inverted ballast design that some BW classic guns use

 

[Diving Gecko]

Yes high pressures make bands lose a lot of energy, not entirely sure why as there are a few explanations given (drag slows them down, compression counteracts stretch ratio, cold water temperature and some other stuff)
So a 16mm might behave like a 14,5mm which is 20% energy loss or more hence why a lot of people use 4-band wooden guns for aspetto here and in Italy. Which can have the same power as a high pressure pneumatic but way worse tracking and recoil.

Yep, I heard about the cold and I thought about the depth maybe making them not want to contract as freely.

Here, recordman Giankos has switched to pneumatics for a while now

Ooops, I mixed him and Giacomo up, haha.

From some math I found

Force = ΔP × Piston Area

Gun A: 30 bar internal
-Surface ΔP = 29 bar
-40 m ΔP = 25 bar → ~14% loss

Gun B: 14 bar internal
-Surface ΔP = 13 bar
-40 m ΔP = 9 bar → ~31% loss

The AI disagree on that math by 1bar, haha. The math you found wants to subtract 1bar from the gun pressure whereas when I consulted the AI, it did not. Either way, the numbers are close, as (A)I would get 13% and 29% in your examples.

However this is theoretically eliminated with a hybrid. See:


Which is an unexpected bonus. I put two AIs to check this and both confirmed it, as long there are seals there are no losses due to pressure

Not sure how the internals look but if you can keep the piston from having to act against the water, then I guess you wouldn't have these losses. The airgun's piston "feels" the force of the water through the shaft. Not sure how it works with the Dreamair/Velair. I will have to study whatever Pete dug up some more [Duh, I didn't even see the two quotes from him that you shared].

Bit of a wild idea but wouldn't it be possible to have a secondary pump on the Mirage (or a single pump on a non-Mirage airgun) that is used to pump in-barrel water out? So instead of transferring air into the "Mirage chamber" before loading it pushes water out from a hole after loading, kind of like an inverted ballast design that some BW classic guns use

Sorry, can't picture this one, either, hehe.
I mean, there's no water in the barrel in my Mirages. Well, there would have been in an original one because vacuum muzzles weren't invented back then - but even though I forgot to make it clear, when I say "airgun" in this thread I generally mean a vacuum muzzled airgun. I think I only ever shot my first airgun for a few days with the "wet barrel" before the parts arrived to turn it into a "dry barrel" gun. And I have never shot a gun without a vac muzzle since.

I said there's "no water" in a vacuum muzzled gun but that's not entirely true. There's a tad of water but only very little, only water amounting to the space between the front of the piston and the vacuum seal minus what gets expelled by the shaft when it gets inserted into the piston. That's the only water in the barrel in any (vacuum-muzzled) airguns. Some users even tip the gun upside down to have no water because they have had issues with salt water creeping behind the piston seals and into the internals with repeated shots.

 

[rollerGR]

Not sure how the internals look but if you can keep the piston from having to act against the water, then I guess you wouldn't have these losses. The airgun's piston "feels" the force of the water through the shaft. Not sure how it works with the Dreamair/Velair. I will have to study whatever Pete dug up some more.

(...)

I said there's "no water" in a vacuum muzzled gun, there's actually there's a tad of water - but only very little, only water amounting to the space between the front of the piston and the vacuum seal minus what gets expelled by the shaft when it gets inserted into the piston. That's the only water in the barrel in any (vacuum-muzzled) airguns. Some users even tip the gun upside down to have no water because they have had issues with salt water creeping behind the piston seals and into the internals with repeated shots.

I did get that response too, and asked it to clarify.

"The spear becomes a mechanical interface that allows external pressure to enter the barrel volume. The shaft itself is not magical – it is simply the movable boundary that allows external hydrostatic pressure to enter the system.

Now Compare to Dreamair / Velair​

Here the situation is fundamentally different.
Let’s restate the architecture:

• The pressure chamber is completely closed
• The piston is inside that chamber
• The spear is NOT connected to the piston
• The spear never enters the chamber
• Energy is transmitted via dyneema outside the barrel

Is there any path for pressure transmission?
Let’s trace possible routes:

Route 1: Through the spear​

In classic pneumatics:

• spear slides into the barrel
• pressure can follow it inside

In Dreamair:

• spear is entirely external
• it sits on a top rail
• it never contacts internal air

So:
This path does not exist.

Route 2: Through the dyneema line​

Could water pressure travel along the dyneema to the piston?
No.
Because:

• Dyneema is a tensile element only
• It transmits pull forces
• It does not transmit hydrostatic pressure

Water squeezing on the rope radially does not create axial pressure on the piston.

Route 3: Through the gun body​

External water compresses the gun body uniformly.
But that compression acts on:

• the outside of the pressure tube
• equally on all sides

It does not create differential force on the piston.
This is just like a scuba tank.

Therefore:​

In Dreamair/Velair:
There is no mechanical or fluid path from external water to the piston face.

Imagine two scenarios:

Scenario A – Classic Vacuum Pneumatic​

• You push the spear into the gun at 40 m
• External water is pushing with 5 bar
• That pressure acts on the back of the spear
• The spear compresses the air inside
• The piston “feels” that 5 bar

This is unavoidable in practice.

Scenario B – Dreamair​

• You load the spear externally
• The piston is already connected to dyneema
• No part of the loading process opens the pressure chamber
• No volume change occurs in the chamber

So ambient pressure has no entry point.

• Classic pneumatics:
  The spear acts as a moving interface that allows ambient pressure into the working volume, so the piston effectively feels depth.
• Dreamair/Velair:
  The spear is mechanically decoupled from the pressure chamber, so there is no way for ambient pressure to reach the piston.

The piston cares only about differential pressure (internal vs external).
"

So the odds are looking good! Basically what Pete described
I'm not sure if it's exaggerating a bit on pneumatic losses or if we (or it) are missing something even though I explicitly asked it about ambient pressure on the shaft and it was adamant that it makes no difference (it does not experience net axial force from ambient pressure, the rod does not push in any direction by itself and sees equal hydrostatic pressures, cannot carry static pressure along its length because it is the water that is at ambient pressure instead etc).
But the chamber seems to be completely sealed off to any losses no matter how these might occur via transferring ambient pressure at depth to get to the piston. AI insists that vacuum pneumatics experience significantly fewer losses than wet barrels but over time the front of the piston ends up at or near ambient pressure, and hybrids experience the least possible (no pressure path to the piston face), and if you don't specifically tell it that the pneumatic has a vacuum muzzle it assumes it's a wet barrel by default. So this argument was true for old pneumatics but not newer vacuum ones as most bibliography is for older pneumatics. It gave this as well:
"Effective piston force = (Internal pressure – Ambient pressure) × Area"
and it estimated 25-35% losses for old airguns, 10-20% for vacuum muzzles and 0-1% for hybrids (mostly from friction)

Either way if this is built it will have to be tested at depth so we'll know for sure.

 

[Diving Gecko]

I did get that response too, and asked it to clarify.

"The spear becomes a mechanical interface that allows external pressure to enter the barrel volume. The shaft itself is not magical – it is simply the movable boundary that allows external hydrostatic pressure to enter the system.

Now Compare to Dreamair / Velair​

Here the situation is fundamentally different.
Let’s restate the architecture:

• The pressure chamber is completely closed
• The piston is inside that chamber
• The spear is NOT connected to the piston
• The spear never enters the chamber
• Energy is transmitted via dyneema outside the barrel

Is there any path for pressure transmission?
Let’s trace possible routes:

Route 1: Through the spear​

In classic pneumatics:

• spear slides into the barrel
• pressure can follow it inside

In Dreamair:

• spear is entirely external
• it sits on a top rail
• it never contacts internal air

So:
This path does not exist.

Route 2: Through the dyneema line​

Could water pressure travel along the dyneema to the piston?
No.
Because:

• Dyneema is a tensile element only
• It transmits pull forces
• It does not transmit hydrostatic pressure

Water squeezing on the rope radially does not create axial pressure on the piston.

Route 3: Through the gun body​

External water compresses the gun body uniformly.
But that compression acts on:

• the outside of the pressure tube
• equally on all sides

It does not create differential force on the piston.
This is just like a scuba tank.

Therefore:​

In Dreamair/Velair:
There is no mechanical or fluid path from external water to the piston face.

Imagine two scenarios:

Scenario A – Classic Vacuum Pneumatic​

• You push the spear into the gun at 40 m
• External water is pushing with 5 bar
• That pressure acts on the back of the spear
• The spear compresses the air inside
• The piston “feels” that 5 bar

This is unavoidable in practice.

Scenario B – Dreamair​

• You load the spear externally
• The piston is already connected to dyneema
• No part of the loading process opens the pressure chamber
• No volume change occurs in the chamber

So ambient pressure has no entry point.

• Classic pneumatics:
  The spear acts as a moving interface that allows ambient pressure into the working volume, so the piston effectively feels depth.
• Dreamair/Velair:
  The spear is mechanically decoupled from the pressure chamber, so there is no way for ambient pressure to reach the piston.

The piston cares only about differential pressure (internal vs external).
"

So the odds are looking good! Basically what Pete described
I'm not sure if it's exaggerating a bit on pneumatic losses or if we (or it) are missing something even though I explicitly asked it about ambient pressure on the shaft and it was adamant that it makes no difference (it does not experience net axial force from ambient pressure, the rod does not push in any direction by itself and sees equal hydrostatic pressures, cannot carry static pressure along its length because it is the water that is at ambient pressure instead etc).
But the chamber seems to be completely sealed off to any losses no matter how these might occur via transferring ambient pressure at depth to get to the piston. AI insists that vacuum pneumatics experience significantly fewer losses than wet barrels but over time the front of the piston ends up at or near ambient pressure, and hybrids experience the least possible (no pressure path to the piston face), and if you don't specifically tell it that the pneumatic has a vacuum muzzle it assumes it's a wet barrel by default. So this argument was true for old pneumatics but not newer vacuum ones as most bibliography is for older pneumatics. It gave this as well:
"Effective piston force = (Internal pressure – Ambient pressure) × Area"
and it estimated 25-35% losses for old airguns, 10-20% for vacuum muzzles and 0-1% for hybrids (mostly from friction)

Either way if this is built it will have to be tested at depth so we'll know for sure.

Got it. I knew the spear was decoupled, but I didn't know if the chamber was fully isolated - for some reaon I didn't see the two quotes by Pete that you shared, haha. As long as it is, then it's all good. Better in fact than an oleo as you explain

Now, tiny thing, but yes, normally 40m would equate to 5bar. The reason, I stated in my examples that it's 4bar is that AI insisted that we have 1bar atmosphere behind the shaft, too from when we loaded it on the surface so the AI said we can subtract 1bar. Why not do the same to the pressure in the tank? Supposedly that's already done by the gauge which starts at 0bar, so if it says 25bar on the gauge, it's actually 26bar absolute.

As for AI, on niche stuff like friction and losses for spearguns, it must have scrubbed the forums (I think you said that yourself earlier), it doesn't actually know. So, whatever conclusions it spits out - always with worrying certainty - is in cases like these probably based on what a bunch of spearos have speculated on. The ~20% has often been stated as the improvement from a vac muzzle, btw. My gut feeling is that spinning the pulleys - and in the case of the Velair, having to pull those pulleys through the water, plus the shaft friction to the gun stock is going to be more than 1%...?

Either way, the important thing is that the internals are isolated on the hybrid, otherwise your idea would be a bit dead in the deep waters

On another AI note and to do with my recent valve design.
Yesterday, I asked both Gemini and ChatGPT about how fast my new valve would let the air pass through/equalize the chambers. I gave it all the info needed, sizes of chambers, pressures and the basic geometry of the transfer port. Same exact prompts. ChatGPT said 3-4 secs (after doing what seemed like a lot of math), Gemini on the other hand said 0.11sec. Since 3-4 secs is more in line with what I would expect from experience, I wanted to challenge Gemini so I sent it the correspondance from ChatGPT and thought it would be "oh, sorry, I messed up". But instead it found math errors in ChatGPT's output.
ChatGPT had literally stated the right formulas but gotten the math x10 wrong. So, I went back to ChatGPT and it more or less said "yes, math is hard" (we are talking about a decimal error...). Anyways, it redid the math and got to the same very low time of ~0.12 sec to equalize across the valve.
Now, it will never be that fast but "we" went a bit deeper and it likely has to do with those 90 degree cross holes (air doesn't like sharp turns), the sharp edges of those same cross holes creating a ton of turbulence and possibly the two cross hole inlets fighting each other. Anyways, I don't need sub-1 sec times, but I did in the end gain a few pointers on how to perhaps smooth the flow out to make it a bit faster. I guess petrolheads who ported and polished their 2-stroke engines or the newer gen of them tinkering with the geometry of intake manifolds know all this.

 

[Diving Gecko]

The Non-Violent Mirage Valve Works! (Maybe...)
Today, I made the one-way valve for the end of the pumping barrel. This is the last part I need so that I can assemble the gun and test the new power regulator cap valve. On my other bulkhead, this check valve is built in. Not on this one and I think I remember why. It's because it would have gotten in the way of machining the little lip for the new valve.

The valve is a simple "Italian Ball Valve" - the same tech as is used in all our pump inlet valves though I have adjusted the internal dimensions a bit. It was made partly on the lathe, partly on the CNC:

The outer features were done on the lathe, so was the brass bushing and in the above pic, I am doing a test fit on the thread.
Btw, I have mentioned in the past how I only use insert tooling so I don't have to learn to sharpen my lathe tools. This is one of the reasons:

A pretty nice mirror finish on a simple facing cut where I didn't even try hard... The other reason is that the high rake inserts for alu work really well on small machines like this mini lathe. I even use them in hardened 17-4PH when I cut down my shafts and in titanium and stainless, too.

The internal geometry was done on the CNC, basically just three different bore sizes. One for the spring, a slightly larger one for the o-ring and a slightly larger one again for the threaded 7mm bushing. Just to try something different I made the bushing M7x0.7 instead of a more regular M7x1. It's basically just a difference in programming as a single point threadmill can do all the pitches up to it's nominal pitch (and I was using an M6x1 threadmilling cutter). And then on the lathe, since I now have the ELS, it was as simple as just setting the pitch to 0.7 there, too.

Here are the parts, not much in it, really:

I forgot I wanted to shorten the bore for the spring as that's just dead space in the valve where we can't compress the air which makes the pumping a little bit less efficient. Well, next time. I think I will also order some smaller balls and reduce the bore to match them.
Also, the only reason the middle part of the valve is so long is because the new bulkhead is 8mm shorter then the one it replaces and I didn't want to make a new pumping barrel... I think I made this bulkhead shorter to save some weight and to shrink the rear volume a bit (less pump strokes needed, but less power for the low power shot).

Here it is, all greased up and with the o-rings installed, ready to go in the bulkhead:

I guess I could have deburred that slot with my little grinder. Well... Also, you can see the shape of the air transfer opening to the right of the valve:

And here's the assembled bulkhead with the new check valve and the new power regulator/Mirage valve:

It Works! But It Doesn't... Some Findings
I took the gun up to 30 bar and wouldn't you know, we have a leak... An internal one somewhere in the Mirage system (we will get back to that). But the leak was just about slow enough that I could do some test pumping and learn a few things:

• The violent slamming when opening is gone! That's the main take-away. No surprise but still nice to have it proven
• I was reminded that I prefer the direction of the Magnum type (same as Predathor) valve over the Mares type. The main reason is that when going from a high to a low power shot, you simply flick the regulator knob out of the gate and the pressure in the gun will "automatically" close the valve. On a Mares type valve, you would need to force it closed against the pressure (which in a Mirage starts taking noticable force)
• Pushing this valve open to equalize the pressure between the two chambers was not bad at all. Of course, this was helped by the leak resulting in a lower pressure differential but I think even with a fully working valve and a higher differential, it would still be very doable
• The air transfer did seem to happen really fast so it was not like I had to hold the valve open for any annoying or tiring amount of time. Maybe it took 1.5sec or so. That said, because of the leak, I think the pressure differential when I got to opening the valve had dropped to about 20bar (I had the pressure gauge inserted and it read ~10bar). But the way pressure and flow supposedly work is that the air moves a whole lot faster when the pressure is at its highest. So, say the pressure differential is 30bar instead of 20bar when I open the valve, it's not like it would take 1.5 times longer to equalize
• I have also stated that it would take ~3.5kgf to open the valve at max pressure, but I didn't account for the pressure in the gun also working in the same direction to push the power regulator "out of the handle". And I think those two forces could be additive and thus the force to open the valve is actually a bit more than twice as high as I have previously stated. But it was still very manageable
• I think I instinctively grab the topside of the handle or reservoir with my left hand as I push forward the knob with my right to get a bit more leverage and control. As this would happen as part of the loading routine and not in the middle of a stalk, this is totally fine

OK, Back To The Leak
Trouble shooting Mirage systems is always frustrating as there are so many potential leak paths.
For the valve:
1. The tiny o-ring inside the larger valve cap
2. The valve seal itself (the rubber sheet being pressed onto the alu lip)

Add the bulkhead:
3. The seal between the bulkhead and the shooting barrel (as I had to cut a new o-ring groove for the shorter bulkhead, this goes on the list, too)
4. The seal between the bulkhead and the reservoir (this is a brand new bulkhead, it's never been used though the o-ring groove should be made to proven dimensions)

The Pumping Barrel Check Valve (the one in the bulkhead):
The pumping barrel itself did not leak when used previously but the check valve at the rear of it is new. It's not the likely culprit but:
5. The seals between the new check valve and the new bulkhead could be leaking
6. The ball check valve itself could be leaking but for this to happen, air would also need to be getting back into the pumping barrel which it could if one of the following is leaking:
7. The seal between the barrel and the check valve (not so likely as the o-ring groove geometry is the same as previously used)
8. The Zahar valve (unlikely as it has been flawless so far)

Like I said, frustrating...! That said, the pumping barrel has proven itself leak free with the same Zahar valve and a check valve with identical o-ring groove design and geometry. I have designed for 25% compression on all these o-rings. I guess since they are static (except the one inside the valve cap) I could have gone to 30% but 25% should be fine .
The list is actually pretty much in the order of what I suspect is the likeliest culprit and I really think it would be between those first four.

So, glass half full/half empty at the same time. On one hand, I think this valve will work but on the other, chasing Mirage leaks kinda sucks...
By the way, the Magnum/Predathor valves also use the gasket idea, so I do hope it will work for me - if not, I have another idea in my notes that could possibly work instead. Using an o-ring instead of gasket material but it would still be a push-to-open Magnum style valve and of course, with the smaller internal valve, too.

Here's a reminder of the crescent shaped Magnum/Predathor valve next to my version:

 

[rollerGR]

Got it. I knew the spear was decoupled, but I didn't know if the chamber was fully isolated - for some reaon I didn't see the two quotes by Pete that you shared, haha. As long as it is, then it's all good. Better in fact than an oleo as you explain

Now, tiny thing, but yes, normally 40m would equate to 5bar. The reason, I stated in my examples that it's 4bar is that AI insisted that we have 1bar atmosphere behind the shaft, too from when we loaded it on the surface so the AI said we can subtract 1bar. Why not do the same to the pressure in the tank? Supposedly that's already done by the gauge which starts at 0bar, so if it says 25bar on the gauge, it's actually 26bar absolute.

As for AI, on niche stuff like friction and losses for spearguns, it must have scrubbed the forums (I think you said that yourself earlier), it doesn't actually know. So, whatever conclusions it spits out - always with worrying certainty - is in cases like these probably based on what a bunch of spearos have speculated on. The ~20% has often been stated as the improvement from a vac muzzle, btw. My gut feeling is that spinning the pulleys - and in the case of the Velair, having to pull those pulleys through the water, plus the shaft friction to the gun stock is going to be more than 1%...?

Either way, the important thing is that the internals are isolated on the hybrid, otherwise your idea would be a bit dead in the deep waters

On another AI note and to do with my recent valve design.
Yesterday, I asked both Gemini and ChatGPT about how fast my new valve would let the air pass through/equalize the chambers. I gave it all the info needed, sizes of chambers, pressures and the basic geometry of the transfer port. Same exact prompts. ChatGPT said 3-4 secs (after doing what seemed like a lot of math), Gemini on the other hand said 0.11sec. Since 3-4 secs is more in line with what I would expect from experience, I wanted to challenge Gemini so I sent it the correspondance from ChatGPT and thought it would be "oh, sorry, I messed up". But instead it found math errors in ChatGPT's output.
ChatGPT had literally stated the right formulas but gotten the math x10 wrong. So, I went back to ChatGPT and it more or less said "yes, math is hard" (we are talking about a decimal error...). Anyways, it redid the math and got to the same very low time of ~0.12 sec to equalize across the valve.
Now, it will never be that fast but "we" went a bit deeper and it likely has to do with those 90 degree cross holes (air doesn't like sharp turns), the sharp edges of those same cross holes creating a ton of turbulence and possibly the two cross hole inlets fighting each other. Anyways, I don't need sub-zero times, but I did in the end gain a few pointers on how to perhaps smooth the flow out to make it a bit faster. I guess petrolheads who ported and polished their 2-stroke engines or the newer gen of them tinkering with the geometry of intake manifolds know all this.

GPT is notoriously nerfed if you don't use a paid $$ subscription and isnt set to "long thought" or something like that (I don't remember the exact terminology, it's for premium plans). I use deepseek to double-check but mostly for types and physics principles (what can transmit what, differential pressure etc)
to save up energy/compute time LLMs which are not inherently trained on 3d diagram simulations of what we're asking it (which is 100/100 of LLMs as there's no dedicated spearfishing one) will just regurgitate whatever they find on blogs and sites. And are going to miss a lot, I've had both gpt and deepseek omit the very crucial details about shaft flopper(s) when asking it about possible energy loss sources that I've missed but if you bring it up they'll get into it ("This is a crucial detail and you're right to bring it up blah blah"). we go into it expecting a cranked up agent but in reality it has tunnel vision and must be spoonfed. Most people will ask it about cooking recipes so it's built with that in mind unless there's a very expensive monthly subscription at play haha

yes all free LLMs I've tried look up details online. What I posted was about 1/3 of the response, the rest was just quoting 3 forums (deeperblue and 2 others) so I just kept the part about differential pressure and whether the shaft's behavior in acting like a pathway or not to ambient pressure. I generally don't give it any math

Note that shaft friction on the guide (or through the seal) was not calculated, only in-barrel losses due to outside pressure making it in

 

[Diving Gecko]

An Alternative Non-Violent Valve
I stayed home today, needed a day off from the mess that's my workshop these days

But I CAD'ed up another version of the power regulator valve. Maybe a waste of time as I haven't yet worked on fixing the first version but I wanted to "see" the alternative roaming around in my head. I find that drawing things makes me understand them better and since I can't draw well in hand, CAD it is:

As you can tell, the gasket is gone. Part of the valve now protrudes into the bore with an o-ring.

In a cross section cut, it looks like this:

It's still a "push to open" valve which will eliminate the slamming but using an o-ring should make it a lot easier to get to seal reliably.

Besides getting rid of the lip, the air transfer bore itself needed a different profile as the concave curves on the earlier version wouldn't work with an o-ring. So, now everything that goes into the bore is convex (still some concave curves on the front part of the valve to help index it against the pumping and shooting barrels):

Whilst the assymetry of the bore could make it look like the rod is not centered, the center of pressure on the main cap is really close to the center of the rod (off by 0.15mm). The idea being that when I close the valve it will hopefully not be off canter. Also, some nice 30 degree chamfers should help guide the valve into the bore.

The unrestricted cross area still equates to a +10mm circular bore (I have deducted the rod). As long as we don't start speculating about how various shapes, features, bends, etc impact the flow, but allow ourselves to only consider the cross area, then a 10mm bore is still larger than anything on any classic oleo (apart from not having a bulkhead of course, hehe). E.g. it's about a 56% increase over the Predathor bulkhead transfer ports and more than double that of a Sten. For reference, I measured the Predathor ports to be equavalent to a ~8.1mm circular bore but given that it's made up by four smaller ports it may even be more restrictive than an simple 7mm straight bore as found on e.g. Stens and Assos (Mares upped it to 9mm on this handle and I guess on the later Force, too).

Pros

• More reliable sealing than a gasket.
• Also, the internal o-ring is now a bit bigger (for the same size rod) which should also make it a bit more reliable. That is, if those transverse holes don't chew up the o-ring when they pass it...
• Now that the lip is gone, machining wise there's no problem with making the pumping barrel check valve part of the bulkhead.
• It wont really need a spring but if I am wrapping my head around forces correctly, if I do have a spring, at least that will reduce the force needed to open the valve. It might be worth it testing some quite powerful springs but that would be for a later time.

Cons

• This valve may not shut fully by itself when flipping the knob from max power setting to low power as a gasket valve would. The reason is that the o-ring needs to be compressed to allow the valve to fully enter into the bore. So, it may need a little pull at the very end.
• Because of the protrusion, this new valve also needs a bit more travel of the power regulator rod otherwise the rear part of the valve may sit a bit too close to the bore in the fully open position. This may (or may not) restrict the flow it so would be nice to be able to push it a bit more forwards. I only have 16mm between the two gates in this handle which is a few mm too little (Not an issue with the previous valve which is shorter overall). On my custom handle - which I hope to make this year - I would increase that to ~20mm which should give me plenty of space to play around with for various valve designs. It's funny how having a CNC and CAD now means a few mm is considered "a lot".

 

[Diving Gecko]

Yes, yes, yes - It Works!
I was kinda dreading having to troubleshoot the internal leak in the new Mirage valve system because there are so many ways that system can fail but this one was easy to spot and as we will see in a bit, @tromic was right

It was the lip on my bulkhead that had immediately cut into the gasket:

About 1/3 of the circumference had been cut clean through.
As for the corrosion on that domed front, I will get back to that.

But on the bright side this is pretty much the best I could hope for. An obvious failure point and hopefully the only one.

Let me remind you of the shape of that bulkhead lip:

Tomi saw this many months back when I first showed this bulkhead and voiced his concern that it might be too sharp. I think I said if it was, I could try to sand it down. So, that's what I did. I did this on a very flat surface and did my best to keep the sanding parallel to the front face of the bulkhead. I simply placed most of the bulkhead off of the sandpaper and shimmed it a bit with some other pieces of paper. Worked pretty well. Took it down to about 0.35mm.

I also have a little flexible Scotchbrite type sanding wheel that works pretty well so I used that a bit after the flat sanding:

Here it is before some final polishing and cleaning:

A lot less offending and closer to how blunt the Salvimar flanges are:

One thing that probably didn't help was that the supposedly "high quality nitrile sheet" that I used the last time around was actually really "brittle" for a lack of a better word. If I pulled on it, I could make it tear pretty easily but in a weird raggedy kind of way. Like small chuncks wanted to come out it. This time, I changed to using some motorcycle inner tube (it feels like a better material):

I was impatient so I just used superglue to stick the rubber to the valve cap and then cut it to size with an Exacto knife. Last time, I actually cut the gaskets on the CNC.
Also, the superglue might not be the best. I actually wanted to use 3M 5200, but that would mean I would have to wait at least a day to start testing and superglue is fast and easy. Lets see how it holds up.

I also took a bit of time to polish the domed fastener in the lathe before I did a test assembly:

The reason it looked so bad was because of this:

A few days back, I had it in the ultrasonic cleaner but at some point I added the galvanized trigger and that's supposedly a stupid move... The galvanization reacts with alu in the bath and I think it took it out on the dome since that was made from some no name 6061'ish rod I had around whereas most of the other parts are made in marine grade alu with higher corrosion resistance.
Anyhow, now I know and I really should use the ultrasonic cleaner more often. Just with more "material discipline".

Also, that little gap between the front dome and the cap made me shave off ~1mm of the length of the regulator rod before I assembled it all again and tested it:

The Result
It works! I took it to around 28 bar and did 7-8 pumping strokes which dropped the pressure to about 4-6 bar. I kept the valve closed and then I tested again after around 15 mins and the pressure was still the same - so no internal leaks and the Mirage system works!
I did four rounds of this but I didn't let it sit for much more than a minute or two under the higher pressure differential except for that one time.
Also, the time to equalize the chambers when pushing the valve open was about 2.2secs. That's shorter than I would have thought given how tiny the internal bore is that the air has to travel through. There's a bit of hissing and pulsing but I also forgot to round off the cross hole edges. Will do it next time I take the gun apart. I also want to cut down the OD of the shooting barrel and add threadlock to the power regulator parts if the valve survives the first few days of real world use.

One more thing, I had thought that locking the power regulator knob into the rear gate and trimming its overall length precisely so that it would pull the seal onto the lip would be important. But it turns out that locking it in is not even needed. The air pressure trying to push the 4mm rod out of the handle is enough to push the gasket onto the lip tightly enough that it doesn't leak during the beginning of the pumping.

Now, the one thing I don't yet know is the reliabilty of this gasket and valve. I will just have to use it and then we will eventually know.

I got lucky in the sense that it was obvious right away what had failed and the I didn't have to go chasing a leak that was "hiding".
And I am particular happy that the tiny o-ring inside the valve cap was fine and the compression on it seems fine, too as proven by not having any leaks. My check valve is fine, too and all the other o-rings grooves must have been done correctly, too

 

[Diving Gecko]

Bummer...
Also I have received some titanium tubes, just short ones as I wanted to check if they would be smooth enough to make shooting barrels out of.

The outside is a bit rough but that would have been ok if the inside was not. But turns out we are not even close to something useable:

This would take more than just polishing to make work. So, at least for now, this vendor is a dud.
I have one more set coming in and also some stainless tubes.

I also got some small pieces of 17-4PH plate. I hope the spot rust is stuff transferred from the saw onto this plate as 17-4PH really shouldn't rust:

The plan was/is to make some shaft sliders out of this. Yes, CNC cut out of "billet" 17-4PH but only out of necessity as I can't weld. So, I can't weld the line lugs onto sliders made out of rod/tube. I need it to be integral, hence the plate.
Yep, yep, I know Tomi and others have designed and made "lug-less" sliders where you whip the line in an undercut area of the slider. I may have to go down that path - but I want to try the plate approach first. Not looking forward to machining this though. You get spoiled when you cut alu...

The problem with 17-4PH is that I can't buy it prehardened and it needs heat treating to get the most out of it. So, later this year, I may get a small secondhand heat treating oven as I have some other parts I want to make in 17-4PH, too. Besides sliders, shaft tangs of course, but also tangs for pistons, slip tips and custom trigger sears.
The sear could be made in 440C which I can buy prehardened but I think 440C should probably only be used inside the gun, not surrounded by salt water.

 

[Diving Gecko]

Why Not A Face Mounted O-ring?
In terms of how to seal the bigger cap in this push-to-open valve we obviously have the simple gasket and I also showed the "radially" mounted o-ring that inserts into the bore. Think of that one as a regular piston o-ring method but there's one last idea I have also drawn up. A captive trepane groove:

The groove doesn't have to be in the cap, it could also be cut into the front face of the bulkhead.

Advantages

• No risk of the bulkhead lip cutting the gasket
• Changing the o-ring would be a lot easier than having to disassemble the whole valve to change a gasket
• No need to cut custom gaskets
• It's easy to design a groove to a target compression
• And like the gasket type, this valve would "self seal" when going from high to low power - no need to pull the valve shut like the "radial" design that goes into the bore might need

Disadvantages

• Machining wise it would be tricky as the tools would be super small, fragile and not cheap
• This design takes up the most space. For even a thin 1.5mm o-ring it needs about a 3mm flange width. More would be better. But with the space constraints of this handle it means that the transfer bore becomes smaller and the total cross area ends up below the target I set.

So, whilst I think this type of seal could be more reliable and easier to service, I don't think I will try it on this particular handle. But I might give it a go on a future custom handle where I can arrange things a little bit more freely.

 

[Diving Gecko]

A Short Video Showing The Valve Equalization
Here's a video from yesterday after I had sanded down the lip on the bulkhead which had cut the previous gasket.


Filming with the phone in one hand makes this harder to demonstrate - it would go a lot more smoothly in the water.

Btw, I forgot one small disadvantage to the "push to open" face valves vs. the Mares piston type valves: When disassembling gun, you need to unscrew and pull off the valve cap before you can pull the barrel out of the handle since the valve holds the bulkhead in place. Same on a Predathor. Not a big deal once you have a reliable working gun that wont be taken apart often.

But on the positive side, I also forgot to mention that they work as poppet valves and release if there's over-pressure in the rear chamber. So, effectively they do two jobs - safety valve and power regulator. The Mares style power regulator can't do this, which is why Mares guns have a separate safety valve in their bulkheads.
(I am 100% sure Pete has explained this years ago in other threads).

 

[Diving Gecko]

According To The Title This Thread Is Supposed To Be About 3D Printing...
Finally I have an upgrade to my first bed slinger printer that I gave up on as it was slow, finicky and blew a motherboard (a nice replacement Klipper board, even).

It's a bit of a Bambu clone but everything in the cheaper range is these days for good reason:

And yes, I buy a fair bit of stuff for my workshop (though the printer is at home), but this one, like almost everything else is used. I can source for the used market in China and when you are not in a hurry there are good deals to be made. This one was ~40% of the price of a new one, had very low hours on it and should last me a good while. Some of the other stuff I have sourced has been crazy steals and I still haven't had a dud.
More importantly it's enclosed, can be made to work directly with OrcaSlicer and I will add a chamber heater pretty soon.

The very first thing I printed on it was a practical part for my rotisserie - a pulley bracket:

It's not perfect but it's close enough. This is straight out of the box after having been shipped twice across China and then loaded into a sea freight container and I ran no input shaping or anything else other than auto-bed leveling. And compared to my old printer, I am very happy. This was just PLA.
I later printed the same part in ABS-GF which is what I want to do a lot more off. E.g. I think the first grips for my custom handle will be 3D-printed and ABS is pretty light and should be sandable if need be.

On the top is the old bracket in PLA, the bottom is the new ABS-GF. The old bracket was printed on the old printer with a larger nozzle to make up for how slow I have to run it. The new printer has much smaller nozzle so has to print less wide and less tall layers but the print times are about the same but the new printer can run way faster. (And yep, the new one is not perfect, but I think that can be fixed with some belt tuning and input shaping.)

I think even back when this thread started in 2018, I said that if you were considering printers, then do it. Now, I think that's even more true. Printers these days are so much more capable. Thanks to Bambu everyone has had to up their game so a lot of the printers are much, much easier to use and print way better than was the case just 3-4 years back.

Another purchase arrived today, too:

It's a four-stage PCB airgun pump (not the underwater type). These go up to insane pressures:

They used to be USD 300-400. That was when the US/European companies were the only ones who made(?)/sold them (they were possibly already made in China back then). Now, they have dropped to USD 60-100 as the Chinese sell them directly.
QC is possibly lacking on some of them. But when this one finally showed up for 1/4 of the price of a new, cheap Chinese one, I had to get it. I have been wanting one for a few years by now. For two reasons.
I don't want to pump airguns with it, but if time ever allows I would love to make a lightweight two-stage oleo pump. So, a pump that has both larger and smaller pistons than the 10mm pumps we all have. A two-stage pump should allow you to move more volume until the pressure goes up at which time, the second stage (smaller piston) takes over which allows easier pumping than the 10mm pumps. This would be for traveling with a Mirage.
So, the first reason to buy this was for inspiration. To take it apart and figure out how they are made so I can make my own.

The second reason is that I can use it to "blow up" my CF tube samples. They go to as much pressure as a scuba tank so if I just fill the tube with water I shouldn't have to even pump that much to get up to some high enough pressure to validate the strength of the tubes. And this means, I no longer have to rent scuba tanks and drive around the island with them on my scooter which is a bit of a hassle.

Also, my new custom CF tubes came in for the Mirage. In short, I think they work - I am off to the pool now so will report back later

P.S. For The Printing Nerds
I may resurrect the old I3 Mega as I have a another board somewhere that I could pop in with some work (a bit of rewiring, new connectors, new config, etc). The reason for doing this is that I may want to print some vacuum seals in TPU and I already have a pretty nice direct drive extruder on that printer and could keep the filament path short. It also has a simple spool holder with ball bearings and TPU prints slowly anyways so that printer would be fine speed wise.
I think I will buy a roll of soft TPU and see if I can make it work in the new printer (outside of the AMS) and if not, the old pony may get brought back to life.

 

[Diving Gecko]

Finishing Up The Mirage (For Now)
Yesterday, this arrived:

A small batch of four custom carbon tubes. Matte on the outside and I had asked the factory to use a woven fabric on the inside where they normally all use unidirectional (UD). My thinking is that UD is too fragile to deal with the twisting of nose cones and handles we sometimes do when (dis)assembling our guns. Also, almost all of my tubes over the years have arrived almost a bit too dry on the inside, so on top of asking for the 3K on the inside I asked if they could make it more resin rich. There are ways to do this even when using prepreg but the factory said they couldn't.

Anyhow, the outside looks alright. Not as good as last time, some more pinholes this time:

But the inside 3K I think is an improvement:

But even more pinholes on the inside. And this next backlit pic makes it look worse than it is in real life:

But I have seen that bigger kind of void on the bottom of the pic on many of my tubes. I speculate it happens when they pull the mandrels out of the tubes which I think takes a lot of force. Luckily, they are rarely at the very ends.

I had hoped I could use these tubes as they were but I felt I still needed to resin coat the ends. In the meantime, I have also gotten some coating epoxy which shouldn't fisheye. First test was great:

Coating epoxies have additives in them to lower the surface tension and make it flow better so it levels better and doesn't bead up as easily.
Yet, when I did a test run on an off cut of the new tubes, we were back in fisheye land...:

I wiped that off, cleaned the tube even better and tried again and whilst it wasn't perfect I gave it a go on the real tube for the Mirage.
Cleaning regiment was soapy water, sanding, brake cleaner (!) and alcohol in that order. I speculate that despite the voids in the inner wall, the factory uses a hefty release agent that's just really hard to clean off.

Here's the proper tube on the rotisserie:

Still not perfect but I could actually smooth it over with a latex gloved finger and it almost stayed as a film.

Once cured, I did a final sanding with 400-800-1500 sandpaper and then chamfered the edges by hand, too.
But I do the inside sanding with a mandrel now to reduce the risk of not sanding evenly. I count strokes, say 5, then turn the tube, sand 5 times again, turn, rinse and repeat.

Messy, I know but I made this mini bench specifically for sanding jobs. It has a CNC cut flat plastic plate which comes in handy when I need things to be sanded flat.
The Predathor bulkhead in the background is my go/no-go gauge

Next up of course was to assemble the gun with the new CF tube and leak test it in a bucket. It worked. Zero bubbles.
And since this tube is longer than the alu one, I could get rid of that delrin spacer I have had to have in front of the nose cone:

(That blue safety is upsetting me now...)

The circular scratches on the muzzle is from when I stupidly haven't fully inserted the custom wrench and it slips when turning it. The shaft protector is my own design, it has a captive o-ring for friction so the protector stays put and a small vent hole that ends in the channel for the little line. Works a treat.

I am still not sure what happened to those two old CF tubes that I couldn't get to seal. Some day, I will cut them down some more for some shorter guns and maybe it will work then. My only guess is that I messed up and made them crack when chucking them in the lathe. I still think you could chuck them very carefully but of course some custom softjaws would be the real trick.

 

[Diving Gecko]

Finishing Up The Mirage (Pool Testing A Bit)
I put 32 bar in the gun after having swapped out the alu reservoir for the new CF one. Which btw saved me about 120g as well as gaining me some buoyancy. I was about to turn down the OD of the shooting barrel to save some more weight but I wanted to test the gun in the pool as I hadn't yet tried the new valve.

Good thing I didn't turn down the OD because... the gun floats... It floats!

I even brought a scale so I could weigh how negative it would still be... but it floats!

I built a Mirage which floats - WITH a reel. That's crazy. I mean, it's not the size of the shorter, original 80, but still.

And yes, the gun is not loaded in those two first pics, the shaft is outside the reservoir and thus displaces a bit of water and adding some buoancy, but it floats when loaded, too...:

This is nuts. And it was my dream, but I thought the reel would kill it. This means I can probably make it neutral with a 7.5mm shaft by turning down the shooting barrel OD and possibly get very close even with an 8mm shaft.

Anyhow, the pool session was weird. I am a better builder than target shooter these days, but that's probably given that I have only been in the water for a few days in 2025.
But here's a pic of the penetration from 6m:

But the testing was weird, the gun or I shoot left. By a fair bit and I need to investigate that but the power is there. I had changed to mono for this test and the whole idea was to do a back to back test between mono and dyneema but I couldn't even hit the target after I had switched the line. Besides the left shift, I am almost certain the dyneema makes the gun shoot high. I have no idea why but all my shots went completely over the target and then I ran out of time since the pool was needed by someone else.

My action cam mount is wonky so the angle is horrible but will get that figured out soon (I am using a custom thingamajig, no surprise there but it needs more work). But here's two clips showing 6m shots (apologies for the colorful language as they say in broadcasting...):


Except for a few CM, they have the same penetration. Maybe it has to do with how firmly you hold the gun when taking the shot.

BTW, I double checked the length of my "underwater tape measure" afterwards because I was impressed by the penetration and was wondering if the line had shrunk. But it hadn't. 6m was indeed 6m

Also, the new valve is fine so far, the one where I sanded down the sharp lip. No internal leaks after the handfull of test cycles in the workshop and ~13 shots in the pool. Of course, that's not a measure of durability but it's an alright start.
And I can actually open the valve in one go now. No need to hold it in the intermediate position waiting for the equalization. I guess it's because most of the air transfer happens very early on when the pressure is the highest - which means the force on the valve drops a lot right away and thus I can push it all the way forward without having to wait for the last bit of air to move.

 

[tromic]

Bummer...
Also I have received some titanium tubes, just short ones as I wanted to check if they would be smooth enough to make shooting barrels out of. View attachment 61471

The outside is a bit rough but that would have been ok if the inside was not. But turns out we are not even close to something useable:
View attachment 61472
This would take more than just polishing to make work. So, at least for now, this vendor is a dud.
I have one more set coming in and also some stainless tubes.

I also got some small pieces of 17-4PH plate. I hope the spot rust is stuff transferred from the saw onto this plate as 17-4PH really shouldn't rust:View attachment 61474View attachment 61473

The plan was/is to make some shaft sliders out of this. Yes, CNC cut out of "billet" 17-4PH but only out of necessity as I can't weld. So, I can't weld the line lugs onto sliders made out of rod/tube. I need it to be integral, hence the plate.
Yep, yep, I know Tomi and others have designed and made "lug-less" sliders where you whip the line in an undercut area of the slider. I may have to go down that path - but I want to try the plate approach first. Not looking forward to machining this though. You get spoiled when you cut alu...

The problem with 17-4PH is that I can't buy it prehardened and it needs heat treating to get the most out of it. So, later this year, I may get a small secondhand heat treating oven as I have some other parts I want to make in 17-4PH, too. Besides sliders, shaft tangs of course, but also tangs for pistons, slip tips and custom trigger sears.
The sear could be made in 440C which I can buy prehardened but I think 440C should probably only be used inside the gun, not surrounded by salt water.

Maybe this might be interesting to you, regarding line slide. I just find this on Italian forum. One of my version of this is Salvimar-Tomba slider: youtube]iRg3rxIwDH8 This one at the image also has something like a slider, not just a ring... Very important detail is how the spectra is attached.

I motivi come dicono gli chalo e gabriellum possono essere diversi a volte e anche solo l aletta detto questo io per ovviare alla precisione e per ottenere il massimo dell idrodinamicita uso le aste gost con tessitura tessile consigliatami da Emanuele Zara lui e anni che pesca cosi zero attriti velocita e precisione ho fatto vedere un video a daduz dove sparo ad una spigola a circa oltre 4 metri con un 80 cm precisi che mi sono costruito con un serbatoio skorpion modificato e canna da 12 a18 ATM e asta da 6,5 adesso provero la sette che secondo me dovrebbe andare anche meglio per quello che voglio ottenere . L asta free non mi e mai piaciuta e se appunto l'hai montata per pensare di essere piu preciso e veloce Credimi che la soluzione che sto usando e che usa Zara e il top allego la foto del tiro dove a steso 3 passate di filo del 1.30
***
The reasons, as Chalo and Gabriellum say, can be different at times, and even just the fin. That said, to overcome the need for precision and to obtain maximum hydrodynamics, I use Gost rods with textile weave recommended to me by Emanuele Zara. He has been fishing like this for years, with zero friction, speed, and precision. I showed Daduz a video where he shot a sea bass at about 4 meters with a precise 80 cm that I built myself with a modified Skorpion tank and a 12-18 ATM rod and a 6.5 shaft. Now I will try the seven, which in my opinion should be even better for what I want to achieve. I have never liked the free shaft, and if you have mounted it in the belief that it will be more precise and faster, believe me, the solution I am using is that Zara uses and the top. I attach the photo of the shot where I laid out 3 passes of 1.30 line.
***

 

[Diving Gecko]

Maybe this might be interesting to you, regarding line slide. I just find this on Italian forum. One of my version of this is Salvimar-Tomba slider: "" This one at the image also has something like a slider, not just a ring... Very important detail is how the spectra is attached.
View attachment 61509

I motivi come dicono gli chalo e gabriellum possono essere diversi a volte e anche solo l aletta detto questo io per ovviare alla precisione e per ottenere il massimo dell idrodinamicita uso le aste gost con tessitura tessile consigliatami da Emanuele Zara lui e anni che pesca cosi zero attriti velocita e precisione ho fatto vedere un video a daduz dove sparo ad una spigola a circa oltre 4 metri con un 80 cm precisi che mi sono costruito con un serbatoio skorpion modificato e canna da 12 a18 ATM e asta da 6,5 adesso provero la sette che secondo me dovrebbe andare anche meglio per quello che voglio ottenere . L asta free non mi e mai piaciuta e se appunto l'hai montata per pensare di essere piu preciso e veloce Credimi che la soluzione che sto usando e che usa Zara e il top allego la foto del tiro dove a steso 3 passate di filo del 1.30
***
The reasons, as Chalo and Gabriellum say, can be different at times, and even just the fin. That said, to overcome the need for precision and to obtain maximum hydrodynamics, I use Gost rods with textile weave recommended to me by Emanuele Zara. He has been fishing like this for years, with zero friction, speed, and precision. I showed Daduz a video where he shot a sea bass at about 4 meters with a precise 80 cm that I built myself with a modified Skorpion tank and a 12-18 ATM rod and a 6.5 shaft. Now I will try the seven, which in my opinion should be even better for what I want to achieve. I have never liked the free shaft, and if you have mounted it in the belief that it will be more precise and faster, believe me, the solution I am using is that Zara uses and the top. I attach the photo of the shot where I laid out 3 passes of 1.30 line.
***

Thanks Tomi,
Will look into this
Years back I actually experimented with splicing thin dyneema and made some splicing tools for the purpose. I think it was for shooting line, too but I think at the time the issue I ran into was that the supposedly best dyneema for shooting line is prestretched quited hard in the machines and then has a PVC coating applied. The idea is that we want a stiff dyneema as the soft braid tangles so easily. But it makes it very hard to splice, I found it impossible in the thinner sizes if my memory serves me.
The soft braid is generally much easier to splice. Not to mention, the braid doesn't have a protective sleeve which the stiff dyneema does. Anyhow, I think it could be worth revisiting the idea