Sentry Gun Project

[jackssmirkingrevenge]

hehe good one though I really like the Portal sounds to be honest, more creepy than gay.

I think I'll built another semi-valve prototype with the design concept above in the meantime, it will give me something to do while the postal service does its duty

[spudamine]

Instead of a solenoid valve you could just buy the solenoid bit and use that to hit your schrader but if you've already bought the servo you might as well just use that.
Of course if you're feeling sadistic you could make your own solenoid like I did, you could even pot it in epoxy

[jackssmirkingrevenge]

Instead of a solenoid valve you could just buy the solenoid bit and use that to hit your schrader but if you've already bought the servo you might as well just use that.

Also a possiblity but that would mean getting an extra power supply for the solenoid and I want to keep things simple. I'll try and get started on the valve tonight, been a long week and I haven't had time so far. I hope to start receiving bits next week in the meantime, woo.

USGF spotted these valves from McMaster:

46365K73
(Same as 46365K12)
Brass Sure-Fit Solenoid Valve Ruby Seals, NC, 1/4" NPT Female, 0.11 Cv, 24 VDC
In stock at $91.96 Each rated 1100 psi

7892K64
Solenoid Valve W/ Compression Tube Fitting Polyurethane Seal, 1/4" Tube OD, .02 Cv, 115 VAC
In stock at $89.40 Each rated 2000 psi

Both would fit the bill but I'm trying out the cheaper homemade option first.

I've been having some thoughts about calibration.

The software maker suggests attaching a laser to your gun so you can mark out calibration points:



The issue is that this is only useful where there is a backdrop to see the laser dot on, and even then the software only sees in 3D and has no concept of range, so really you can only properly calibrate the sentry for a given engagement range.

Another camera could be fitted on the gun itself with a cheapo scope or simply a set of crosshairs calibrated with the impact point in case the backdrop is non existent or too far away for a laser dot to appear on.

[jackssmirkingrevenge]

Update: I received the software and it's great, still fiddling around with the settings but first impressions are very good. Not much I can do before the servos arrive but nice to have a play in the meantime. Also, I leave it on when I'm not around so it can creep people out with the Portal sounds (for anyone considering this software, the sounds can be replaced with whichever mp3s you want) whenever it detects significant motion. I can also take snapshots in this instance, a useful feature.

I've been thinking about the gun itself in the meantime, I want something that can engage small fast moving targets, this means a high rate of fire is essential to ensure good hit probability. For once, a vortex block BBMG is actually ideal not too keen on filling the place with 0.177" BBs though, I was thinking of 3mm as I had used in my "micro" BBMG, but I figured I would go the whole hog and go for 1mm. Some of the benefits include:

- light weight (more than 70 times lighter per round than 0.177" BBs) which is a consideration when the chamber is turret mounted

- cheap, you can get 5000 for $14.99 from eBay sold as nail art decoration (so the more camp can get them in a variety of colours ) so they are inexpensive enough to be disposable

- with HPA they will reach high velocity and be devastating at close range, but it will taper off relatively quickly, like birdshot from a shotgun, limiting effective range and rendering them harmless if they should be fired into the air.

- A 1mm barrel has 20 times less flow than a 0.177" barrel, greatly limiting flow and therefore is much more economical with air.

Given the above consideration, a slightly open ball valve (say around 1/4") will provide more than enough flow to exceed the barrel flow, so I think the simples option is to have such a valve operated by the trigger servo, it doesn't have to move much so the power should be there. I've also ordered a remote line so the HPA tank can be separate from the turret.

I've ordered another cheapo webcam that will be permanently mounted to the gun, this way I can mount a simple crosshair that will greatly simplify calibrating the system. The idea is that I match target location on the crosshair to impact point at the average target range, say 10 metres, and after that I don't have to match any calibration points with actual shots, just line them up with the crosshairs.

I should start receiving parts this week, tired of playing with the MW2 turrets

[rp181]

I notice the "2fps" in the bottom corner.With this software, Im thinking fast moving targets are not feasible.

[POLAND_SPUD]

USGF spotted these valves from McMaster

I am sure you could find something similar on ebay. If you are that concerned with $ you could even buy some used ones

go for 1mm

hmmm good thinking. It's not something I'd try but it makes sense

[jackssmirkingrevenge]

I notice the "2fps" in the bottom corner.

I'm not sure what that refers to, it could be the rate at which images are considered (it has a setting where for example for every three camera frames it only considers one)

Well, I only want relatively fast movement tracking, after all servo and trigger response time are also a factor. I tried to see how it performed with objects in motion (yes, throwing shoes around one's room is science! ) and it only tracked them once they hit a hard surface and decelerated a bit. Still, good enough for my purposes, I'm not trying to build a Phalanx CIWS tracking suspersonic missiles

It's not something I'd try but it makes sense

It's the classic exponential rule that is ubiquitous in our hobby. Half the barrel diameter and you're using 25% of the air. Unfortunately there's a corresponding decrease in energy, but I did some fiddling around with GGDT, basically HPA + large barrel length:calibre ratio + 0.005 gram projectile = pathetic muzzle energy but some impressive velocity figures

[rp181]

I bet thats the rate which the image is retrieved from the camera. Just beware the maximum is 30 fps, with a lag free program.

[jackssmirkingrevenge]

We'll see, I would think the biggest challenge is to get the turret to rotate as fast as possible to match the software. I'm going to do my best to get the gun itself as light as possible so I can use the servos on direct drive as in this configuration.

The software seems to have no problem tracking small targets, have a look.

[inonickname]

The silencer looks like it will cause the barrel to bow a little. Perhaps omit it completely, recoil won't be too great and who cares about noise when you have an automated machine gun?

[jackssmirkingrevenge]

The diagram I posted is a basic layout, in practice it will look different. still having second thoughts about the valve, ideally a solenoid would be the neatest solution but I can't find one that can handle high pressures at a reasonably cheap price. I think I'll have a go at rebuilding this unit - 3mm porting will give 9 times the flow of a 1mm barrel so it should suffice for instant high pressure in the chamber - before making any further investments.

As I'm in an urban environment I want it to be as stealthy as possible, with such a small bore it's unlikely to make much noise anyway but a simple silencer made from say a section of PET bottle would not add much weight and it would certainly cut down on the muzzle blast.

[iknowmy3tables]

building your own solenoid valve is a possibility but I'm having trouble imagining a cheap and simple design

what about a high pressure poppet actuated by a low pressure piston, the electric valve for low pressure are more affordable

[jackssmirkingrevenge]

That would work but you're introducing an extra power supply for the solenoid valve as well as a low pressure air source, not all the added complexity might as well use a servo actuated ball valve.

[POLAND_SPUD]

@iknowmy3tables
that's probably not practical here as there are ready made solenoid valves that can handle high pressures. But I like the way you think

[jackssmirkingrevenge]

But I like the way you think

Says someone with a penchant for off-the-shelf parts I might be willing to turn to psycix's machining facilities for this one though, we'll see.

edit: I've ordered a couple of 1/8" and 1/4" ball valves, it's probably the best solution. They're not rated for the 800 psi I intend to use but I've used domestic 3/4" valves for that pressure without problems so I should be fine