Sentry Gun Project

[POLAND_SPUD]

oh I see but that still means that accuracy is somewhat limited

[jackssmirkingrevenge]

Depends what you want. If you're trying to hit a small fast moving target using the motion detection software, accuracy is a bit moot. It's the same reason most modern fighters have gatling/revolver cannon, you want the largest volume of fire in the shortest possible time.

On the other hand, if you're firing on something slow/stationary, you might as well just get one of these, strap your launcher to it and aim manually.

[dewey-1]

Well, brand new servos aren't exactly expensive.

What I have in mind would be a high pressure BBMG (hit probability FTW ) so it wouldn't really be that heavy, basically barrel, valve and BB supply, the air can come from a remote line.

This project reminds me about the one you and I discussed via mail several years back!

Do you remember the mouse-trap design that consisted of a radio controlled CP-88 with laser and wireless video camera transmitting to my 46" LCD TV?

This kind of project will be very interesting for you JSR!

[JeffSinger]

Steppers are stupid for automated turrets. I'm so tired of people wanting to use steppers. steppers are great if you want to make a simple remote control turret but totally useless for an automated one.

Imagine you're writing the software for the thing. lets say you've gotten as far as getting it to know where the target is and where the turret is initially pointed. you say, ok, i'll just tell it to move x amount of steps, based on calibration, and it should be aimed right. well that works good and well if you're test environment retarded and slow but here's the problem; what happens when the turret is interrupted before it gets to where it wants to be and is told to move some other direction by a higher priority target being detected? now everything is off because it didn't make it, but it thought it did. that's why servos are superior and are used in the majority of newer industrial machines. they actually know where they are no matter what. they don't need to be zeroed out or any stupid sh*t like that. you simply tell them, "go to x position" and they do it. Also, there's no point in having a motor that can turn 360 degrees when the camera can only see 40-60 degrees.

in summary, steppers are dumb and servos are smart.

oh, and this guy right here:

you've got limited field of fire

If I understood correctly the camera is fixed to the base, so your field of fire is limited to what the camera can see. If the camera is moving around the software can't tell between a moving target and the moving background.

knows what he's talking about.

[Ragnarok]

oh, and this guy right here knows what he's talking about.

Nah, Jack doesn't know what he's talking about.

Most of the time we just ignore him and nod until he goes away.

[jackssmirkingrevenge]

This project reminds me about the one you and I discussed via mail several years back!

Haha YES! This would be autonomous though, it can be left to its own devices Technology eh?

If one were splashing out, a thermal imaging camera would be sure to track down the little blighters

[POLAND_SPUD]

that's why you use encoders...

what I am trying to say is that:
with servos you are limited to something like this -> http://www.thinkgeek.com/geektoys/cubegoodies/a1c2/

so it's both small and has short range as it would suck at longer ranges


but with step motors you can build something like this ->

[jackssmirkingrevenge]

Most of the time we just ignore him and nod until he goes away.

Hey, I'm not the spudfiles grandad

with servos you are limited to something like this so it's both small and has short range as it would suck at longer ranges

I think we have different visions of what this should look like. What I have in mind is something like the Samsung SGR-A1 - small calibre, full auto, fast firing, relatively short range. You seem to want something like the Telepresent Rapid Aiming Platform with a longer range large calibre semi auto (probably with a few spool valves thrown in ) - in this case, like the TRAP it would probably be a good idea to have a launcher mounted camera looking through a telescopic sight, on a platform that is manually controlled.

[Ragnarok]

Hey, I'm not the spudfiles grandad

Of course not. Now go back complaining about the youth of today, smoking your pipe or whatever it was you were doing. :tongue5:

[POLAND_SPUD]

@JSR
well to some extend you are right... I would build a universal platform that could be fitted with a lot of different guns


however, my point here is that step motors are better and could achieve better accuracy... just have a look at that Samsung SGR-A1... I am almost sure that it uses step motors as it has pretty good accuracy

the guy from paintballsentry.com uses servos as they make it slightly simpler but none of his videos show that thing shooting at ranges longer than 10 - 15 meters

I am not good when it comes to electronics but I think that his software could be used with step motors... you'd need one extra pcb board

[jackssmirkingrevenge]

Of course not. Now go back complaining about the youth of today, smoking your pipe or whatever it was you were doing. :tongue5:

Back in my day, they hadn't even invented electricity. We had to watch television by candlelight!

*ahem*

however, my point here is that step motors are better and could achieve better accuracy... just have a look at that Samsung SGR-A1... I am almost sure that it uses step motors as it has pretty good accuracy

It also fires at 3,000 feet per second through a rifled barrel not really spudgun territory, unless you're some insane Nordic fellow with a machine shop.

the guy from paintballsentry.com uses servos as they make it slightly simpler but none of his videos show that thing shooting at ranges longer than 10 - 15 meters

I live in an urban area. 10-15 metres is about the extent of the range I have so it doesn't bother me. Within that range, a 450-850 psi BBMG should be accurate and devastating enough... which sadly is bad news for whatever else is in the area you're trying to defend, it doesn't seem to care much for back stops... we could call it the DYI Sentry Gun

[Insomniac]

While servos are a bit less accurate than steppers, they are very quick and still relatively precise, and that's only using the feedback from the servo's inbuilt potentiometer. Perhaps with an infra-red laser on the gun, coupled with a camera that had been modified to see in the near IR spectrum (remove the IR blocking filter from the lens assembly), would allow the computer to constantly correct for any inaccuracies in the servo's centering. Alternately, a second camera mounted to the gun could help the computer know exactly where it was pointing.

[POLAND_SPUD]

@insomaniac
yeah that's the point... while his programme is pretty cool it has some flaws which make it more like a toy than something that could be used to build something more decent

[jackssmirkingrevenge]

it has some flaws which make it more like a toy than something that could be used to build something more decent

What exactly did you have in mind? Something that will fill your neighbour's greenhouse with holes while you sleep I was thinking of either a tripod mount, or have it ceiling mounted like this one:

[youtube][/youtube]

Even without a functioning gun, it would be pretty cool to have it follow people around the living room especially as the default sounds for the program are the cute but creepy effects from the Portal sentry turrets...

The more I think of it the more it makes sense, I think I'll be ordering the software.

[kenbo0422]

We used servos on our XY stages to achieve 1 micron accuracy at the place I previously worked. I guess its how you achieve your precision more so than the motor type itself. Steppers can be microstepped to achieve much higher precision than they were originally meant to have. The main reason for servos was to have feedback to make sure we were where we wanted to be, not necessarily for the precision placement.