Things to ponder with how much pressure you need.

[athomas]

If a valve output causes the level of thrust to be decreased by some means such as restricted flow or turbulence then the efficiency is reduced. The matrix does a better job than the shocker of maintaining initial higher pressure behind the ball and and then allowing the pressure to drop at just the right time to reduce the waste air flow out the barrel.

The chamber size vs pressure is critical. The barrel length is also a larger factor in lower pressure guns because the pressure is maintained for a longer period in order to allow slower acceleration to the desired fps setting.

A short barrel would be extremely inefficient on a low pressure gun. It takes a certain amout of torque to accelerate a ball to the desired speed. If the pressure is lower the rate of acceleration is lower. Therefore it takes the ball longer to reach the desired speed. In a short barrel, the distance is not there to allow that to happen. Therefore the pressure must be increased to allow a greater acceleration. The greater pressure combined with the larger chamber of lower pressure guns = larger amount of air used. Most isn't utilized and just pours out the barrel after the ball has left.

There is so much to consider. One thing that hasn't been discussed much is the effect of friction in most of the equations. Here is where the lower pressure guns gain back some of there inefficiencies. Typically, the effect of friction is a function of the square of the velocity. Since lower pressure guns accelerate the ball at a slower rate, the force of the friction is not as great in the early stages of the barrel as it is in higher pressure guns that achieve greater velocity earlier in the barrel.

Hope this helps.

[AGD]

You guys are doing so well with this I am going to jump in and share some data that I never let out in public. Here is the pressure profile of the Angel that you can use to determine accelleration rates. Virtually all paintgun profiles look like this except for the Matrix and Mag which take longer to get up to pressure. The Y scale is pressure and the X scale is time in tenths of milliseconds so 10 counts = 1 ms.

As you can see the pressure and therefore the accel. is not linear. The barrel was a dye 14" with a sensor at the end. To make matters worse the vent holes in the barrel let out the pressure before it gets to the end.

AGD

[bjjb99]

Tom Kaye wrote:
> You guys are doing so well with this I am going to jump in and share
> some data that I never let out in public. Here is the pressure
> profile of the Angel that you can use to determine accelleration
> rates. Virtually all paintgun profiles look like this except for
> the Matrix and Mag which take longer to get up to pressure. The Y
> scale is pressure and the X scale is time in tenths of milliseconds
> so 10 counts = 1 ms.
>
> As you can see the pressure and therefore the accel. is not linear.
> The barrel was a dye 14" with a sensor at the end. To make matters
> worse the vent holes in the barrel let out the pressure before it
> gets to the end.

That's a fascinating plot! I think it's wonderful that you're willing to share data like this. I've taken this plot and extracted the pressure/time coordinates as best I could (Adobe Photoshop plus some elbow grease), entered them into an Excel spreadsheet, and calculated acceleration, velocity, and ball position as a function of time.

However, I've run into a bit of a snag.

In order to accurately determine acceleration, we need to know the mass of the nylon balls that were being shot. When I used an average paintball mass (about 3.3 grams), I ended up with a final velocity of around 213 fps and a total distance traveled of 9 inches from the ball's starting location before we get to the "ball exits the barrel" marker on the plot. This indicates to me that the mass of the nylon ball used is less than 3.3 grams. Based on the plot provided, I'd estimate the nylon ball mass to be around 2.3 grams, which seems low to me. Perhaps the resolution of the plot is just not enough to perform an accurate acceleration/velocity/distance calculation.

Out of curiosity, how do you measure pressure as the ball is travelling down the barrel? I'm guessing you've got a port drilled into the barrel right near the breech, with a pressure transducer mounted there; outputs from the transducer go to either a digital oscilloscope or a digital acquisition board in a computer. I would like to know what kind of transducer you're using, since the only ones I've seen available (Omega.com) require pretty large ports (>>0.125 inch diameter) to mount them. I don't want to put a 1/4 NPT hole into the side of a barrel and then try to fire paint.

BJJB

[AGD]

BJ,

I don't like to give away the details of the sensors we use but we do drill a hole right at the barrel breach interface to feed the sensor.

The nylon balls are about 3 grams but where you are probably running into problems is in the pressure numbers. The sensor is dynamic so it's the delta of the curve shown. It actually starts about 5-10 psi negative so you have to add that back into your numbers. So that makes the peak pressure about 95-100 psi. Run those numbers and it should bring up your velocity.

We do use a super high speed DAC board.

AGD

[bjjb99]

Plot resolution and pressure sensors

Tom Kaye wrote:
> The nylon balls are about 3 grams but where you are probably running
> into problems is in the pressure numbers. The sensor is dynamic so
> it's the delta of the curve shown. It actually starts about 5-10
> psi negative so you have to add that back into your numbers. So that
> makes the peak pressure about 95-100 psi. Run those numbers and it
> should bring up your velocity.

When I first looked at the plot, the curve looked like it started below zero pressure. I figured it required baselining, so I went ahead and did so prior to calculating any accelerations. The peak pressure I obtained at that time was 94.6 psi, give or take a few hundredths.

Using the plot, I came up with a scale factor of 0.4484 psi per pixel for the vertical axis, and 0.02188 milliseconds per pixel for the time axis. These are based on the 0 to 100 psi scale taking up 223 pixels, and the 4300 to 4400 time scale (10 ms) taking up 457 pixels. I used the lowest point to the left of the pressure spike to represent zero pressure (baselining), and scaled everything accordingly. My results for a 3 gram ball end up with the ball reaching the "exit point" marked on the plot after 10 inches of travel and a final velocity of 233 fps.

What I think has happened here is the plot just lacks sufficient resolution to take numbers from it and get an accurate result. Even if I can't get 300 fps out of my calculations resulting from that plot, it's still a very interesting look into the dynamic world behind a paintball as it flies down the barrel.

You mentioned that the pressure sensor is "dynamic so it's the delta of the curve shown." Does this mean that the sensor is measuring delta-P as a function of time, and the curve shown is basically the integration of that delta-P from time zero to time T? If this is the case, what caused you to use a delta-P sensor rather than an absolute-P sensor? Do delta-P sensors have a faster response time? I imagine that you probably need on the order of microsecond resolution for accurate data capture, and I have absolutely no idea what the response times of pressure sensors are (yet). :)

BJJB

[AGD]

Yes the dynamic sensors we use have faster response times. We actually use both depending on the situation and the time resolution required. It's also possible our scale is off because we have to pulse 100 psi air into the sensor to calibrate it. It is by no means an exact method. I would just scale the curve to get the right speed.

AGD

[Redkey]

stuff

Pressure transducers, at least mine anyhow, have a typical full scale response time of about 0.1 ms or about 1/10,000 of a second.

I'm not sure about the whole dynamic pressure transducer conversation. There are two main types of pressure transducers... absolute and differential. My understanding is that a differential pressure transducer measures the difference between two pressures, ie they have two inputs. An absolute pressure transducer is basically the same only it has one input... with the second pressure being open to air.

DACs run at many speeds, My card, for example, will collect about one channel at 1 MHz, 1,000,000 samples per second. Splitting that into multiple channels will reduce the sampling rate ie three channels will each run at about 333 KHz depending on how your card is configured.

bjjb99 if you're going to build your own system I can help you out in private since it sounds like the AGD folks don't want the details discussed in public. It's not exactly a cheap project and it's definately not simple unless you live in the world of testing.

How can you tell exactly when the ball starts moving by looking at the plot?


Later
Jack

oh... and...
68 cui @ 3,000 psi =
68 in*in*in X 3,000 lbs/in*in = 204,000 in-lbs
once the units are cancelled.

[AGD]

Redkey,

Yes we do use an AD card but I don't know what the max rate. It's multi channel so I think it depends on how many you scan.

The peak of the graph determines when the ball starts moving because it expands the chamber volume. We used to put a sensor in the breach but they correlated so well we just look for the peak now.

AGD

[bjjb99]

Redkey wrote:
>
> bjjb99 if you're going to build your
> own system I can help you out in
> private since it sounds like the AGD
> folks don't want the details discussed
> in public.

I think I've got a pretty good idea of what needs to be done. For the time being, this is just a set of thoughts rolling around in my head; I currently lack the funds and the workspace to set up a decent testbed.

> It's not exactly a cheap project and
> it's definately not simple unless you
> live in the world of testing.

Well, at least I've got something going for me. I have been responsible for experimental design, specification, setup, execution, and post-collect data processing for more than a couple of years during my various work-related incarnations... nothing dealing with high speed pressure measurement, though. I agree that this won't be one of these "rubber band and chewing gum" budget experiments. The combination of fast response time, small packaging, high precision, and low cost is extremely rare when it comes to instrumentation/sensors.

Time to add this project to my ever-growing list of things to do once I get a workshop set up. :)

BJJB

[Redkey]

ball movement

AGD...

from the plot it looks like the arrow is a bit below the peak of the curve... I was wondering if there was another measurement taking place.

Does your system also measure volume of air used? Or do you somehow calculate the volume based on the duration of the event?

thanks
Jack

[AGD]

We are not concerned with volume right now. The arrows just generally point to the peaks there are no other measurements going on.

AGD

[hitmanng]

I have been following this post with interest and thought it was time to jump it.
Remember that there are a lot of factors determining the amount of pressure needed. Most importantly how much air is wasted and not converted into kinetic energy and the friction on the ball.
Remeber Tom's graph measures the pressure behind the ball. As the gas expands it moves the ball forward decreasing the preasure behind the ball. When the ball (cork) exits the barrel (or porting is hit) the presure drops to zero.
CHK6 did the math showing that an average pressure of 28 psi is needed over a 10 inch distance. AGD has verified that with wastage in escaped air and friction, 35 psi is closer to the real number.
I have now seen 3 numbers for the amount of travel time the ball has in the barrel. In CHK6's example 0.003 sec. In the graph of the Angel above 0.004 seconds and in the Mag from barrel tip #1. 0.006 seconds. This actually makes sense. An increase in initial presure will cause an increase in the accelleration of the ball. You would of course have to stop the excelleration earlier than 10 inches if you don't want to go over 300ft/sec.
In the Angel example above the pressure goes from 100 psi to 0 psi over 10 inches. The Angel runs out of pressure before it runs out of barrel. This is the most efficient process.
You could easily use a lot more presure and have the barrel run out before the pressure. This would be very inefficient of course. Theoretically as long as all of the force if the air pressure is used before it exits the barrel all guns should have the same efficieny ie it takes the same amount of force to move the same object the same distance.
The only factors that we have control over are the volume and the initial pressure of the gas. But guns do not all have the exact same efficiency. Why not?? The bolt.
Where does the bolt come it? The bolt has mass. Is moved by air and drastically changes the volume of the chamber.
Everything I have seen so far indicates that you are shooting a blowgun not a paintball gun.
Any thoughts??
Hitmanng
Killer of sacred cows

[MikeCouves]

Whooooooooooosh!! There it goes, all that stuff right over my head . How old are you guys that do this? Is this like university grade math? Sounds pretty tough.

[AGD]

Having the barrel run out before the pressure has the interesting effect of giving you tighter velocity groups which means better accuracy. Wasting gas in this case can be a benefit especially to a back player who needs the range and can go in tighter to 300 fps. The worst thing to do is have too much barrel and the ball starts pulling a vacume before it leaves. This actually does happen on a poplular low pressure gun I won't name. With too low a chamber pressure the air runs out faster than you would like.

AGD

[steveg]

Tom just out of curiosity. Most low pressure markers have
a reputation for very good consistancy , so is the fact
of running out of pressure before barrel of real practical
concern, or academic interest

RedKey thanks for the in-lbs thing, I thought it might be
that but the bit if math I had ran away as fast as it could
and hid as soon as I finished school. Use it or lose it