Monday, June 19, 2017

Overview/General Build Log: Prometheus

Prometheus: an electropneumatic, magazine-fed, fully automatic, open-bolt, 12.7x36mm, high-velocity capable dart blaster development project. Or you may know it as an HPA engine design, or as one of those new "NIC" blasters that isn't technologically stuck in 1995.

This is Prometheus.

Where do I even begin? This is a very involved, expensive and technologically dense project that has been going for around a year. It is only within a month or so that I have returned to working on and designing superstock blasters and chipping away at the massive flywheel-related to do list. That big dark band in DZ's history as a blog was due to this happening. I have been dropping bits and pieces of the build around reddit and facebrick but have never disclosed anything about it in full detail until now.

The last I posted was this early Sketchup (I hate Sketchup, BTW) mockup of a very, very early concept. We can call this thing "Protometheus".

Some ideas did influence the actual Prometheus, mostly the construction of the receiver, but "Protometheus" had schematically nothing to do with what would become Prometheus. It used an operating concept I refer to as a closed-bolt FASOR ("FASOR" being a paintball-sourced acronym for Forward Air/Spring Operated Return). It is also known as the "Spudfiles engine" from use in component-based DIY semi-autos, and as the preferred engine concept of Spectre (everything from the original Red Baron configuration to Prometheus's big rival Deathblossom uses a flavor of it with assorted features added on) and the very earliest HPA guns (CaptainSlug ABP5K and ARR, and Doom FANG).

In the closed-bolt FASOR, or Spudfiles engine, the bolt is forced closed by a ram connected to the dump chamber, while a spring or airspring biases it in the opposite direction. Firing is initiated by shifting the main valve, usually through pilot control since the main valve is a QEV. When the dump chamber depressurizes, the spring moves the bolt open. Repressurizing then closes the bolt.

In the ZDSPB classifications, this family is described as unbalanced heldforward, spring bolt return, firing poppet controls bolt and includes the AirStar Nova, PGI Mayhem and ICE Epic paintball markers.

Now, there are a few major problems with this taxonomy regardless of how it is controlled. One, the bolt actuating pressure is the firing pressure. This leads to a lack of control over bolt force and speed. If you vary the velocity, the bolt force follows in lockstep. Two, following on from that, CBFs designed optimally for a certain operating pressure may not be able to cycle reliably at lower pressure. Three, is the inherent tradeoff created when the bolt itself is unbalanced i.e. the chamber pressure results in a large net force trying to blow the bolt open - in order to be completely kosher and not rely on fuzzy things like the bolt inertia and friction to prevent early bolt retraction and venting from the breech, the force from the ram must always be greater than the force the chamber pressure applies to the boltface, so the ram area must be greater than the diameter of the breech, plus the area required to overcome the spring force at the lowest operating pressure. The diameter of the breech is a given (because it is the caliber of the projectile). This again leads to ramifications on bolt force. For ultra-low operating pressure designs, it is workable, but for high pressure, say 150psi, not so much - the bolt closing force will end up as hugely excessive in order to have a self-locking bolt.

I designed "Protometheus" around a 3/4" bore ram in order to: use an off the shelf ram, counterbalance the 1/2" diameter bolt's blowback force, and account for the ~5 lbf return spring. This ram gave an available bolt force of 66 lbf at 150psi design pressure! Holy chopfest, let's not do that!

And the origin of the 150psi? I specifically wanted to explore higher OPs. Much of the nerf hobby's pneumatic designs have gravitated to lower pressures, 30-70psi. There is, in short, no technical reason for that. Off the shelf industrial components and pipe fittings are uniformly rated for operation at least 150psi. Higher pressures reduce the dump chamber volume required, have theoretically better efficiency in transferring energy to the projectile, are better for most regulators' performance, and reduce the proportionate impact of friction losses in piping.

I also did not want to build a semi-auto, like the original Protometheus was going to be. I'm an arms racer but more than that I don't personally like semi-auto as a control mode. Full auto is my thing, and electropneumatics held huge promise for fine control over timing parameters anyway, so early-on it was decided that I was going to build an electronically controlled gun. I didn't want to just convert a CBF to electro, I was looking for alternative operating concepts. At the same time, the engine designs I could explore were limited by the desire to avoid machining in this particular build.

This is where brainstorming led:

FIRE CoreFinn's improved ram/exhaust, the bolt-sequenced piloted valve operating concept upon which the Prometheus blaster is based.

So, let's turn the CBF on its head. The CBF was designed in an era of single-shots. It is an outgrowth of single-shot airguns that adapts them to have a breech and a bolt and self-load. So instead of starting with the main valve and building the rest of the cycle in compromised form around the main valve shift, let's start with the bolt and use the bolt as the initiating event. We need a small, double-acting ram to cycle the bolt and a solenoid valve to put that under electronic control.

Then how do we tie in the main valve? There are a number of directions to go with that, but the main two are to separately control it in software (dual-solenoid), or if we want only open-bolt operation, to sequence it mechanically to the bolt. The latter is appealing for simplicity of control. Since we're using components, the main valve is a QEV, and what we are going to hit with the bolt is its controlling three-way valve.

Finally, this should be dual-regulated so that the firing and bolt ram don't impinge on each other's tuning, and furthermore, dual-independent with two regulators operated in parallel, so that there is no the arbitrary requirement that one pressure be higher than the other. That's OK for paintball, because velocity range is narrow. Here, wide velocity range is good.

Bam. We now have something vaguely like an Ego with a pilot-operated main valve. That is the FIRE core. Let's build a blaster.

There's no good way to transition into this build log. The construction of the device will become apparent.

Testing layouts: Clippard MJVO-3 with 90 degree fitting inside 2" square, 1/8" wall 6061 aluminum tubing, which is what the Prometheus receiver is made of.

The 3-way and a firing assembly with non final dump chamber. The QEV is a Humphrey Products SQE-2, 1/4" port.

Starting to figure out the dimensions. Note the construction concept: the internal components are mounted into the receiver by transverse trunnions.

Made a bolt tip from brass tubing. This did not survive into the final build, since the brass sealing ring was not compatible with the aluminum breech/barrel and caused galling.

Some more work on the trunnions, now mounting the three-way and the ram.

Basic structure test-fit using a short offcut. Also, the barrel stock is 5/8" OD, .527" bore 6061 aluminum tubing.

Here is where the layout really materializes:

Rear trunnion with firing assembly and ram.

Center trunnion with 3-way.

Front trunnion is just a barrel support and has a duplicate at the front of the handguard section. This makes 4.

Now I have finished up most of the trunnions, cut the receiver out of that hunk of tubing, and started fitting. This is a difficult concept to understand, but the internal components load into the bottom of the receiver. You are looking up at the bottom of the blaster.

This is toward the rear of the blaster. Final firing assembly.

Now that you see the dart in the breech, is it any clearer?

Lots going on here, rear to front:

The Mac 43 solenoid valve (43A-AAA-RFEJ-0BL, 5V 4.7W coil, M3 threaded ports, 5 way/2 position i.e. "4-way") is visible where it will be mounted.

The rear trunnion carries the firing assembly, which will later mate to the powertube, and the ram. All trunnions are built-up from PVC, except the single aluminum layer on the rear trunnion to provide sufficient strength to the ram mounting flange.

The center trunnion carries the Clippard MJVO-3 three way valve. Holes have been drilled for 1/4" tubing runs for the firing circuit.

Then the breech with a dart in it.

A view into the bottom of the receiver showing all that stuff again.

Magazine interaction with other components.

Powertube and bolt body shown in position. I will explain the powertube/bolt structure later.

Bolt tip visible halfway closed in breech. Note that breech is cut into the barrel as a single part and has an integral feed guide, and that the flimsiness of the feed guide area is negated by the center trunnion supporting the rear of the breech.

Bolt closed. 3-way, check valve, and tube fittings visible above breech in the very top of the receiver.

Bolt open.

Powertube (background) and bolt body (foreground): Powertube is built-up from brass tubing. Outer layer is 15/32" K&S tubing. Bolt body is 1/2" K&S tubing. Sealing ring on tip is still non-final brass one.

The powertube is fixed into the rear trunnion. The 7/16" K&S inlet tube at the back passes through and mates to a modified 1/4" NPT threaded brass fitting on the QEV.

Working on the bolt attachment lug to connect the bolt body to the ram. This steel/polycarbonate composite lug shown in the following images was destroyed on testing and is not used in the final build.

Looking into the stock end of the receiver at the back end of the QEV firing assembly and the ram.

QEV mated to powertube.

DPAR3.0: Dual-Pressure SLP Remote Air Rig, which will run this blaster. This setup and all its components will be covered in a separate post.

Started working on external covers and furniture: rear lower cover, main section, covering the dump chamber.

Stock base fabrication.

Lower receiver assembly baseplate, with magwell cutout.

Lots of stuff now drilled and tapped, stock finished.


Mag release detail: Lever pivots about a clevis on the left side of the magwell (I'm righthand) and is to be biased toward the mag by a spring in the grip frame later on

Mag inserted and locked.

More incremental changes and completion.

Pivot side of mag release is shown. This mag release design is one of my favorite things I invented in the course of this project. I will definitely use the concept later on. It isn't ambi in this form, but I don't build your guns, I build mine.

Breech and lower baseplate.

Closeup assembled magwell to baseplate.

As it stands.

Designing a grip frame. Note, space constraints imposed by other components make the trigger be cramped a bit far back and the trigger guard cavity a bit short.

Cutting, fitting components and laminating.

One step earlier mocking it up to the blaster.

Trigger. Clear PC, I has it sitting around.

Grip panels, unshaped, with fasteners and countersinks.

And bolted on. 2 extra screws are for the microswitch

This was where it came together for the first time.

This is the new bolt lug, handmade from a hunk of brass stock to replace that plastic bullshit from earlier that got wrecked.

New bolt once the new lug is soldered on and the polycarbonate anti-seize tip is installed, 5-40 tap drill and tap set, and a Bimba 0072-D ram which is what this thing runs. I had a used one that leaked and was replaced.

Rear trunnion with powertube, ram, and new bolt.

With barrel.

Assembling and testing the new bolt hardware.

Barrel length cut to the final 16.5" and QD base elbows and Foster Series 2 plugs visible. These are clamped between the lower rear cover and the bridge/dump chamber bracket inside the receiver. Just off the shelf pipe fittings and some clever design.

Foster Series 2 one way shutoff quick disconnect fittings are a sort of standard in airsoft HPA and are great. I recommend them.

Inside of grip frame with the switch.

Testing electronics: FCU shown unshrinkwrapped, and 3S 850mAh cylindrical cell battery. Yes, I built a tiny little Monolith pack.

This is the FCU:

It's a prototype, so it is built on perfboard and the microcontroller is a Sparkfun ProMicro (Arduclone) development board.

Notable features of the Prometheus FCU are that it powers the 5V solenoid properly from a switchmode voltage regulator (Recom R78 series), uses a three-wire interface for the switch (using a double throw switch and two inputs to achieve noise rejection without having to low-pass filter the input like normal SPST debounce strategies and introduce a trigger delay), does ROF (off-delay) and dwell adjustments with analog trimpot inputs like the Wolverine Airsoft SMP units, and... is full auto only.

The omission of selective fire is very much intentional. I don't care about fancy modes. All I ever use is full auto whether I am firing single shots or not. Yes, it is super easy to program select fire with something like this. I don't care. I just won't be bothered.

The assembled blaster, missing only a top rail, top rail riser, and flash hider.

Starting on paint and final assembly.

Rail riser.

Receiver with some cosmetic changes.

Striped up

And we're there! Here it is, end of the road.

4.1kg fully dressed. It's somewhat like Thor's hammer to people who don't work freight team when they aren't nerfing. I don't think it feels heavy. It is nicely balanced and wieldy.

Part 2 coming soon: Firing demos.

1 comment:

  1. Flat out stunning work, worth the waiting on this blog.