PDA

View Full Version : Robust, affordable gun design



jwatte
08-02-2013, 02:20 PM
A robust gun design consists of a few things:
- a gearbox
- a barrel
- the spin and gasket for the barrel
- the hopup for the barrel
- hardware/brackets to mount it all together
- some solution for magazines
- electronics to manage the gun

Ideally, the electronics are more than just "on/off." Specifically, they should support a variety of available motor voltages (so, PWM,) and some sensor to make sure the breach stays open for long enough for the next BB to fall down.

I have ideas on all of these, and have started by combining separate sensor and control boards I'm using now into one, that could be screwed into the gun gearbox with shallow screws, to minimize cable length and properly position the sensor on the slide. Here is the first rev of the design for that board. Interface:
- VCC, digital fire trigger, GND as a three-way connector, doubling the Atmega programming connector
- optional "ready" indicator on that same connector
- logic is 3.3V if VCC is 3.3V, and 5.0V if VCC is 5.0V
- Motor power separate, with screw terminals

Arduino, Arbotix, CM-9, and most other robo controllers should be able to tell this to fire. The software in the Atmega is intended to drive the motor with PWM, and make a half-second pause with the breech open to allow time for the BB to drop, even if "Fire" goes on.

Btw:

Should it be single-fire (requiring "fire" to go low and high again for the next shot) or auto (re-trigger when ready as long as "fire" is high)?

What duty cycle should the gun control PWM actually use? I'm thinking 3/5, which will drive the Desert Eagle plastic AEG just fine between about 9V and 13V.

Should there be an option for serial control instead of digital line? It could probably recognize Dynamixel-style packets. Hard-code the ID to 254, and "fire" if the goal position is sent as > 512. Setting the baud rate would be needed, too, and might actually be pretty hard to get robust at 1/2 Mbit, given that the Tiny runs on the internal RC oscillator at 8 MHz, and doesn't have a dedicated USART (will need to decode the packet using the "USI" manual clocking mode.)
Seems like "high to fire" would be sufficient, although serial mode would allow software-defined tuning of things like rest timing, PWM, etc.


Looking for feedback on the general approach, as well as feedback on this particular board design.


4914491549164917

escott76
08-02-2013, 07:26 PM
Maybe a jumper, even a solder one to select auto or single? This type of board could see applications other than mech warfare and they might like to be able to pick either one.

jwatte
08-03-2013, 10:06 AM
My general thought is that, because there's a programming header, if you want non-stock, then change the firmware :-)

I was thinking more about "what would it look like if it were a Dynamixel TTL bus device?" The answer is that I'd need a Tiny84a instead of Tiny85, and a crystal for serial bus baud rate stability. And a regulator to run it off servo voltage instead of external logic. All in all, 3" wide instead of 2" if using through-hole components. SMD could make this more elegant, but also pretty much non-DIY for most people.

This version still has the separate gun power, should you want to use it without DXLs. However, if the idea is to make a starter kit, and that uses AX-12s, the non-DXL parts might be unnecessary.
Also: All those little diodes, zeners, and resistors that guard against user fumbles add significant surface area :-(

4918

Gertlex
08-03-2013, 11:35 AM
Problem with hard-coding to 254 ID for dynamixel approach is if you have two, nothing works, I think? Though I wonder if this can be worked around simply by not having the attiny ever respond to packets (i.e., I'm not sure of the full set of conditions in which duplicate IDs on a dynamixel bus cases problems).

I'm surprised the through-hole board sizes are so big. Guess I've had success with 1.4"x1.75" (or so) boards by virtue of skipping a lot of safety bits... Hasn't bitten me yet ;). I could probably layout a through-hole design in Inkscape, following your schematic and a list of of components.

jwatte
08-03-2013, 11:50 PM
The second design would support adjustable id just like a dxl servo.
The size comes from two things:
-protection resistors and diodes
-the board is designed to screw onto the flat top of the aeg with the sensor art the slide
If "something else" positions the board, about 3/8" width can be saved.

I'm almost at the point where I'm willing to sacrifice the screw terminal gun power, and just use dxl power. 4A is a little high during start-up, but perhaps an electrolytic will fix that.

jwatte
08-08-2013, 12:31 AM
So, if we assume that all mechs will have a Dynamixel-compatible bus (note that you can wire a Hovis or even just a UART as a write-only DXL bus) then we can also draw power from this bus. And if we assume anyone with a serial programmer is not dumb enough to keep the programmer connected while firing the gun, we can remove several protection resistors. We end up with the below design. (To do proper DXL bus, I need a Tiny84a instead of a Tiny85) It's 2.4" times 1.4" where the 0.4" on top are there just to allow screwing into the top of the gun slide to position the sensor robustly. This is for the $12 Desert Eagle plastic AEGs.

This uses nice, thick traces -- 24 mil for signal, 66 mil for power -- so you might be able to make this using laser toner transfer. Although payint $30 for 3 at OSH Park seems easier :-) I could sell complete kits -- parts, headers, PCB, and pre-programmed MCU -- for perhaps $30.
We could go smaller by going for SMD devices -- then we could use an 8u2 MCU, which has a real UART, even, as well as USB interface -- but at that point, it's not so much "homebrew friendly" anymore. The software would still be openly programmable, of course.

49264927

Th232
08-08-2013, 01:06 AM
So, if we assume that all mechs will have a Dynamixel-compatible bus

Minor query: While I think the assumption of a Dynamixel-compatible bus is reasonable, what about people using RS-485 as opposed to TTL?

ArduTank
08-08-2013, 07:53 AM
Th232, what controller are you using? If you're using an ArbotiX, you should still be able to access the 3-pin dynamixel bus, even with the 4-pin bridge attached.

TXBDan
08-08-2013, 08:29 AM
I like the idea of using the dynamixel bus for trigger signal/command and power. Very clean, good idea.

I'd choose a smaller SMD version at a reasonable cost hit over the DIY through hole version.

escott76
08-08-2013, 08:48 AM
I like the idea of using the dynamixel bus for trigger signal/command and power. Very clean, good idea.

I'd choose a smaller SMD version at a reasonable cost hit over the DIY through hole version.
Except if it isn't clean, and gun noise gets put back into the dynamixel bus.

Th232
08-08-2013, 09:04 AM
Th232, what controller are you using? If you're using an ArbotiX, you should still be able to access the 3-pin dynamixel bus, even with the 4-pin bridge attached.

At the moment I'm using a USB2Dynamixel connected to my fitPC, which from my understanding doesn't allow both TTL and RS-485 at the same time. On that note though, regarding the Arbotix my impression was that using the RX bridge disabled TTL, as per the FAQ on the RX Bridge's page?

http://www.vanadiumlabs.com/rx.html

darkback2
08-08-2013, 11:30 AM
Th232 is correct you cannot access the TTL bus with the RX-bridge. For years I have been using AX-12 guts to drive my guns. I had a few ax-12s where the motor stopped working or some other thing went wrong and I was able to use the control board as a ttl gun interface. Seams to work great and your guns show up as an AX-12. You just have to switch it to wheel mode and make sure you don't back drive the guns. It seams that someone could make a similar board to the internals of an AX-12, actually I think someone already had a few years ago...but, it seams that it could be done reletively cheaply if it was done in bulk.

Others have used battle switches. If you are using an arbotix you have to put a regulator in there to keep from frying the switch, but you can run it off of the servo outs on the arbotix. I also think something similar could be made to this relatively cheaply.

Finally couldn't you use a relay coupled with a digital pin on the arbotix?

(Clearly I am not an electrical engineer.)

DB

Lupulus
08-08-2013, 12:13 PM
Along the lines of a battleswitch, I've been playing with a "Beefcake Relay Kit" from Sparkfun. It's $8 and is easily controlled from a digital pin on the arbotix.

jwatte
08-08-2013, 12:43 PM
gun noise gets put back into the dynamixel bus

The Dynamixels themselves are pretty good at doing that, too :-) The PCB design I put in has a back-EMF diode, a 100 nF close de-coupler, and a 10 uF entry de-coupler, which should make it at least no worse than the Dynamixel servos themselves.

Note that the problem is not "I have a digital out from my controller; how can I make the gun motor go?" That's trivial -- a $1 MOSFET will do that just fine.
The problem is "how do I build a robust gun that doesn't jam, doesn't mis-feed, and can be controlled by a variety of control mechanisms?"
If you're using a USB2Dynamixel or USB2Ax, you have no "digital pins" to drive with, for example. Hence, why the bus is attractive.
Also, this board has an optical sensor on it, and the board is intended to screw onto the gun, such that the sensor can detect the position of the slide. This allows the controller to sequence the gun to give enough time for a BB to drop into the breech, as well as providing "semi-automatic" operation where the rate of fire can be controlled.

RS-485 is a challenge for bi-directional. An add-on circuit could easily be built (as there is space for an "aux" header on the PCB) that uses a driver chip. Another option is to do input-only, and just read the "A" line as "UART in" with a single transistor for level conversion.
Actually, the driver chip conversion wouldn't be very bad at all. Would make a neat little addition.

Deimos
08-08-2013, 03:08 PM
I think jamming / mis-feeding needs to be solved in the hopper, airsoft high-cap mags have a spring loaded sprocket looking thing that forces bbs from the reservoir into a single file channel that feeds into the gun. I think this could be accomplished with laser cut parts and a micro gearmotor. If there was a attiny or something babysitting the feeding mechanism it could still be operated with a simple interface (a single gpio/ttl daisychain).

Also, detecting a shot can be accomplished by looking for spikes in current draw.

darkback2
08-08-2013, 04:49 PM
Maybe we are working on this the wrong way round. Perhaps we should be looking for a robust hoppup/barrel solution for the AEG gearboxes and building the guns around that. Andrew and Cire have already done the cases and hoppers, though maybe a spring loaded or motor driven hopper would be a better solution. A while ago I designed a 3D printable hop-up, though a number of people seam to have made something work by coupling the hopup and barrel from a defender gun with an AEG gearbox. Gdubb also has a pretty good one that he designed also.

DB

DresnerRobotics
08-08-2013, 05:05 PM
Th232, what controller are you using? If you're using an ArbotiX, you should still be able to access the 3-pin dynamixel bus, even with the 4-pin bridge attached.

Please try to refrain from posting advice that you haven't researched or tried out yourself. Not a huge deal, but it improves the quality of our forums when we don't have people posting assumptions and conjecture. It's okay to be new to this stuff and I realize you're trying to help out, but I've seen a handful of posts even in the last month from you that weren't very helpful because there was no fact or experience backing what you were saying. If you'd like to contribute to a topic, slow down and make sure what you're posting has value to other people and that you're speaking from an informed position- be it through personal experience or research. If you're unsure of something, ask! You want to avoid speaking with a tone of authority in topics you're not that well versed in though.

So with that said, I won't give a criticism without trying to also provide some information behind it.

If both the TTL and RS-485 sides of the UART were available, we wouldn't have kept the design in such a way that covered up the TTL ports. Mike and I experimented with using both types of servos when the RX-bridge was prototyped, but if I recall neither of us were able to get anything workable. If my memory serves me, the idea was that we could theoretically get the TX side of the TTL UART working (so, we could talk to AX-12s but not get data bacK), but I believe the max485 chip (or whatever brand was used, havent looked at one in awhile) that was tied to the same TTL line was causing interference of some sort. This was about 3 years ago and it was a brief attempt, so I'm a bit foggy on the details. Nonetheless, both Fergs and I made an attempt and deemed it a lost cause, so I wouldn't offer that as advice.

DresnerRobotics
08-08-2013, 05:20 PM
Maybe we are working on this the wrong way round. Perhaps we should be looking for a robust hoppup/barrel solution for the AEG gearboxes and building the guns around that. Andrew and Cire have already done the cases and hoppers, though maybe a spring loaded or motor driven hopper would be a better solution. A while ago I designed a 3D printable hop-up, though a number of people seam to have made something work by coupling the hopup and barrel from a defender gun with an AEG gearbox. Gdubb also has a pretty good one that he designed also.

DB

I agree that it'd be great to have an open source gun driver. I know a few have been made, but I don't think anything that incorporated the AX/MX bus was ever brought to a finish line. Last I recall was Seth at Shepherdstown was working on something, but I think that disappeared along with him.

DB brings up a good point- in terms of the physical side of the gun build we should definitely focus on a solution that incorporates the plastic AEG gearboxes. Why you ask? Because they're readily available off the shelf, and they work great. Also, I have a direct source overseas from a competent manufacturer, so if we can get a full set of gun plans together I'll invest in getting kits made for the store. I'm going to go digging through my cad files and upload as many useful things as I can to our Github in the near future, pretty sure my AEG design is one of the better performing ones out there, save for R-Team's setup.

jwatte
08-08-2013, 08:25 PM
I think jamming / mis-feeding needs to be solved in the hopper

A main reason for not feeding on my guns was the cylinder cycling too quickly. The obvious solution to that is to run the cylinder back so it cocks and clears the breech, and then wait a little bit, and then fire.
The question is, how would you wait a little bit with the breech open?
Answer: Put an optical sensor that detects when the cylinder is pulled all the way back.
That's how this particular driver started -- I built optical sensor boards (with an opamp level detector/driver) to mount to the plastic gearbox for this reason, for Onyx.
Now, needing to build a simpler gun (as my wife is starting to put together a Phoenix quad,) I figure that a single board for both driving and detecting will be the simplest and the cheapest.
Hence, the single board with both sensor and driver on it.

Regarding detecting "firing" by detecting current draw, that has two problems:
First, I'd need a current sensor. A power resistor with sufficient dissipation yet still usable for ADC detection is at least as expensive as an optical detector. Such a resistor->ADC setup will also have trouble with PWM, and when driving the 6V gearbox motor, I'd prefer to use PWM to driving it "straight."
Second, the current draw actually changes right *after* the position I want to detect. The motor will draw current while cocking the gun, which is fine, but it will stop right as it releases the cylinder/spring, which is too late to stop. The best I could do would be to time how long it takes from beginning to end of cocking motion, and then aim for 90% or so of that time, which would be fiddly and fragile.

Yes, the breech/feeder will need to also help avoid jamming. I'll get there, and I'd be happy with some help. First, we need to figure out what barrel/bucking to use, and then perhaps look at the hop-up. The cheapest short after-market barrel I've found is a MadBull for $13. Given that the Defender gun is only $20, perhaps recommending just cutting one of those up is actually the better option -- although it may have a weight disadvantage.

Deimos
08-08-2013, 09:57 PM
Regarding detecting "firing" by detecting current draw, that has two problems:
First, I'd need a current sensor. A power resistor with sufficient dissipation yet still usable for ADC detection is at least as expensive as an optical detector. Such a resistor->ADC setup will also have trouble with PWM, and when driving the 6V gearbox motor, I'd prefer to use PWM to driving it "straight."
Second, the current draw actually changes right *after* the position I want to detect. The motor will draw current while cocking the gun, which is fine, but it will stop right as it releases the cylinder/spring, which is too late to stop. The best I could do would be to time how long it takes from beginning to end of cocking motion, and then aim for 90% or so of that time, which would be fiddly and fragile.


I use pololu motor driver boards with a built in current sensor, so I hadn't considered the implications of building one from scratch. As for the timing, I obviously failed to think that one through :p. I think you're right about looking for a more reliable solution though. A simple solution might be to put in a mechanical switch that is depressed when the breach is opened (placement could be tricky though). If we figured out how to get something like that to work it would make things quite a bit simpler and cheaper.

jwatte
08-08-2013, 10:23 PM
A limit switch is more expensive than an optical sensor, too :-)
The TCRT 1000 is about 80 cents. And it's contactless, so inherently will last longer than a mechanical limit switch.

kamondelious
08-08-2013, 11:03 PM
jes1510's dual gun motor driver is still a great option.

http://forums.trossenrobotics.com/showthread.php?3907-Dual-Gun-motor-driver

(http://forums.trossenrobotics.com/showthread.php?3907-Dual-Gun-motor-driver)

Deimos
08-08-2013, 11:27 PM
I see that i was confusing myself, as I am unfamiliar with optical sensors and did not attempt to use my google powers to remedy my ignorance. Please excuse my laziness :p.

A main reason for not feeding on my guns was the cylinder cycling too quickly. The obvious solution to that is to run the cylinder back so it cocks and clears the breech, and then wait a little bit, and then fire.
It should be noted that applying force to bbs when the breach opens eliminates problems feeding, this is why high quality airsoft guns can still be reliable when fired at upwards of 15 bbs per second. I doubt, however, that applying aforementioned constant force with a gearmotor would result in a happy robot (would need to turn on motor right when the breach is opening). This solution also helps prevent jams in the hopper itself, as the motor acts as an agitator.

Of course, rate of fire is not such a large issue, so just leaving the breach open for an extended duration would be equally effective.

As for actually contributing to this, I do have some CAD drawings of some hopper components that I am going to be testing out for my mech, I'd be happy to share if they work out! Otherwise, please disregard my fanciful ramblings and speculation and continue searching for practical solutions :veryhappy:.

jwatte
08-09-2013, 11:00 AM
The main problem right now is finding an affordable source of bucking.
6mm piping from McMaster can serve as a good enough short barrel (although Airsoft die-hards will now extoll the virtues of precision ground 6.01mm barrels...)
But, I need the bucking to mate with the cylinder, and to spin the BB (so, a bit of protrusion into the barrel on one side.)
Typical sources charge $5 for a single sleeve.
I guess I could become a Mad Bull distributor and buy 50 at a time...

Deimos
08-09-2013, 11:46 AM
The spin on the bb is probably irrelevant at such short ranges, as it is meant to counteract drop off at long ranges. I believe those barrels have a hole for the hop up near the base, but if you just left it out it would be easier to use 6mm piping.

Lupulus
08-09-2013, 12:20 PM
The spin on the bb is probably irrelevant at such short ranges, as it is meant to counteract drop off at long ranges. I believe those barrels have a hole for the hop up near the base, but if you just left it out it would be easier to use 6mm piping.

Spin (on a bb, and especially on cylindrical bullets) also helps it to go straight, period. Basically, it acts like a gyroscope to prevent minor wobbles that turn into tumbling and veering off course. At 2ft this is probably irrelevant, but it would be interesting to test whether spin makes a difference for shots fired the length of the arena.

YoMarchYo
08-09-2013, 12:52 PM
Spin (on a bb, and especially on cylindrical bullets) also helps it to go straight, period. Basically, it acts like a gyroscope to prevent minor wobbles that turn into tumbling and veering off course. At 2ft this is probably irrelevant, but it would be interesting to test whether spin makes a difference for shots fired the length of the arena.

I agree with Deimos. The only spin that's advatageous to bb's is backspin; to provide a bit of lift and counteract drop off. Your best bet for stability at short ranges like these would probably be to just buy the heavier bb's and let inertia do the work.

Deimos
08-09-2013, 12:54 PM
Spin (on a bb, and especially on cylindrical bullets) also helps it to go straight, period.

My research on the internets has led me to believe that a small amount of spin is imparted to the bb by the barrel itself, although at this point I will have to do some testing.

ArduTank
08-09-2013, 01:01 PM
Deimos, rifles have rifling on the inside of the barrel to spin the bullet along the axis of travel. Airsoft gun barrels are smooth on the inside (most, at least) and so, any spin can be considered random in direction due to it being imparted by contact with the barrel or aerodynamic effects.

Deimos
08-09-2013, 01:09 PM
See lupulus' post for why spin in a random direction could still be helpful, as I said much testing will be done shortly ;)... speculation only gets one so far.

ArduTank
08-09-2013, 01:12 PM
Deimos, the only spin that will straighten a flight path is along the axis of movement. If, say the bb is spinning clockwise along the y-axis, it would go right due to it causing an airfoil-like effect. when you have spin along random axis at random speeds, your bbs go everywhere.

Lupulus is talking about two different kinds of spin: BBs(Backspin) and Bullets(rotation around axis of movement(roll))

jwatte
08-09-2013, 01:27 PM
So, even without the hopup spin, I still need bucking to let the BB in and mate with the cylinder nozzle.

I was thinking of punching a hole in a latex balloon and wrapping it around a bit of tubing; we'll see how that works out...

Deimos
08-09-2013, 01:58 PM
when you have spin along random axis at random speeds, your bbs go everywhere.
I would assume that such an extremely minute amount of spin might help to stabilize the bbs, while not being nearly enough to influence the flight path at such short distances?

Alternatively, use madbull barrels, and a preexisting hop up solution? There is a small chance you might be able to find something compatible with the parts you are planning to use somewhere like here: http://www.airsoftgi.com/index.php?cPath=745_31_170

Deimos
08-09-2013, 04:23 PM
I know Jwatte would like to focus on more important things than the hop up, but I just though I would post my findings.

I tested accuracy under the following conditions:
- Defender style gun
- 0.2g bbs
- range of 20'
- gun clamped in a vise for consistency
- about 10 or so shots per target (I think)
- hop up was removed for second target

I noticed little to no difference in grouping size (i was going to use the targets, but then realized i was testing consistency :p)
4930

Additionally, Ardutank, I don't misunderstand you, I simply disagree :D. If you have evidence to support your claim, I'd be more than happy to be proven wrong!

(not sure if pic size is good, I'm pretty sure its right though)

- by hop up i mean the small protrusion into the barrel that puts backspin on the bb
- - curses self for mixing up homophones

CogswellCogs
08-09-2013, 05:19 PM
Incidentally, I chrono'ed the Defender gun. It clocked 65 FPS with a .2g BB. I also checked the cut-down mini-AEG that Boojum uses. It was 83 FPS.

Just some data for reference.

jwatte
08-09-2013, 06:51 PM
gun clamped in a vice

What did you use? Cigars? Whiskey? Loose women? :-)

My eyes want to fool me into thinking the left one is more consistent, but it's probably not by much.

ntopper
08-10-2013, 12:15 PM
As far ah hoppers go, has anyone played with the idea of adding a servo driven paddle that un-jams BB's at the bottleneck? like paintball hoppers? Or is that crazy?
4932

Deimos
08-10-2013, 12:59 PM
An agitator like that would definitely help prevent jams in the hopper, but I think one of the reasons problems feeding occur with gravity hoppers is that unlike paintball guns, airsoft guns tend to shake the hopper every time they fire (massive spring being released). I find that this can cause bbs to actually bounce up when the breach is opened, and not fall in. Granted, as a gravity hopper painball hoppers are as good as it gets, so an agitator like that would go very nicely with Jwatte's plan to leave the breach open for an extended period of time.

Short version: Force feed hoppers are most helpful with a high rate of fire, otherwise an agitator should do the trick.

Janis
08-10-2013, 03:56 PM
Guys,
I was going to ask this question a long ago...
You are discussing jams and solutions a lot and I have seen quite a few of, I would say, crazy designs for feeders, hopups, unjammers here in the forums. And as a rookie I was a little afraid of how to deal with it... So, we decided not to deal with it:)

In my experience with HD2 we had worked with 2 stock defender guns each fired more than 1000 times and not a single jam.
Can I ask what exactly is a jammed gun and do we know exactly why it happens?
1) two bullets somehow jam at the entry into hopup preventing others to move?
2) It is because fire rate is too fast so bullets just cannot keep up with the speed?
3) What happens when gun is jammed? motor does not spin? motor is spinning but gun not firing?
The only case we experienced something like jam was when there was not enough voltage or the motor controller was too weak for motor to start when the spring was almost fully loaded just before release.
In my case the hopups are not modified in any way and BBs are gravity fed. Fire rate is 8 BBs/sec.
There are few little tricks which HD2 uses:
- Fire in short bursts 2-10 bullets, not more. In competition it was 2 shots per trigger only.
- Use amperage to count rounds/stop fire. (there is a noticeable drop in current in piston releasing phase)
- Looks like the shaking of gun while firing is enough to avoid any jamming.
- The guns are always aimed slightly downwards as our robot is rather high.
Can it be that HD2 is just lucky and it's first jammed gun is waiting for it just around the corner?

darkback2
08-10-2013, 06:01 PM
Janis,

I think you bring up a pretty good point which is that a cheap reliable gun already exists...It is the defender. I think we are looking to develope a gun based around the AEG gearboxes because they do, when they work, provide an advantage both interms of reliability of triggering a hit when the target pannel is struck and the rate of fire being slightly higher. That said, the AEG gearboxes aren't easy to get working properly because they are designed to be spring fed. So, jamming...

I have briefly flirted with using AEG gearboxes from time to time and have found the major difficulty to stem from getting BBs to fall into the breach at the right time while still maintaining a reasonable rate of fire. Gdubb solved this by using a servo wound spring to push BBs down a tube. This while a workable solution added a lot of weight to his mech and required quite a bit of serious machine work on his part.

When I have tried to gravity feed the AEG gearboxes a couple of things have happened to me.

1) BBs don't make it completely into the breach before it closes and get sliced in half. Bits of plastic sprey everywhere and sometimes get caught up in the guns mechanism stopping it cold.

2) BBs bounce out of the breach before it closes causing the gun to repeatedly dry fire accomplishing nothing.

3) Multiple BBs get into the breach and block up the barrel jamming up against eachother and then roll out when the gun tilts forward

4) The rubber gasket that is supposed to stop the BBs from just rolling out of the barrel and let the air pressure build up behind the bbs gets shredded and clogs up the barrel.

For all of these reasons my best AEG gearbox based gun shot about 1 in three shots. My worst never really worked at all. I have since switched back to using defenders which seam to work just fine for me.

All of this is just my own personal opinion. Sorry if I am sounding preachy.

DB

jwatte
08-10-2013, 07:49 PM
The main thing I have against defenders is that they have the mechanism molded into the shell, and thus they are bigger and heavier than I'd like. The stock magazine is also kind-of dumb, as it's small capacity, and in my experience actually does jam.
When I talk about jams, I generally mean that BBs wedge themselves somewhere in the magazine->breach chain, and will need force and/or vibration to get un-stuck.
When I talk about firing speed, that's a "dry fire" problem where the BB gets pushed back out of the breech, rather than into the barrel, presumably because it hasn't had enough time to settle in the breach.
I have not seen BBs sliced in half, or the bucking (gasket) shredded. I guess I'm lucky there :-)