Login    Register
User Information
Username:
Password:
We are a free and open
community, all are welcome.
Click here to Register
Sponsored
Who is online

In total there are 73 users online :: 4 registered, 0 hidden and 69 guests


Most users ever online was 155 on Mon Aug 15, 2016 1:40 am

Registered users: Bing [Bot], Google [Bot], Google Adsense [Bot], Yahoo [Bot] based on users active over the past 5 minutes

The Team
Administrators
Global Moderators
global_moderators.png CS

GGDT and in tank Quick Dump Valve Performance.

Post questions and info about pneumatic (compressed gas) powered cannons here. This includes discussion about valves, pipe types, compressors, alternate gas setups, and anything else relevant.
Sponsored 
  • Author
    Message

GGDT and in tank Quick Dump Valve Performance.

Unread postAuthor: Technician1002 » Thu Apr 30, 2009 12:39 pm

I was trying to model the performance of my valve in GGDT. I started listing it as a barrel sealer, but it was rejected as the seat and piston diameter were equal. I went to a Generic and even with very optimistic valve coefficients, it didn't match the results even close. I notice that any valve opening time between 0.7 and 5 ms didn't make any change to the results that mattered. I couldn't reproduce the results I was getting even by telling GGDT I was using 100 PSI instead of the 60 used in the test shots. I think GGDT is assuming the generic valve has a path of fittings between the valve and air supply.

With that in mind I told GGDT it was a burst disk.

I plugged in my values and got a match. To get a match I had to increase the valve coefficient to the point I got a warning it was optimistic. Is it really possible a piston valve can match the performance of a burst disk with light projectiles?

I used a projectile mass of 60 gm and valve coef of 75%.

:idea: . I just thought I would share my results and get your comments.

The modeled shot is of the 4 inch Poof foam ball in the 4 inch barrel. The measured valve diameter is 2.08 inches. I guessed on projectile weight. I used 60 Grams in GGDT. I don't have a postal scale.

Chamber size is 8 inch diameter X 15 inch. Inner diameter for the coaxial design is 2.25 inches for a volume of 694 cu in.

Graph of the actual in barrel shot is shown below.

To read the graph, the ball has a small magnet from an ear bud headphone embedded in the ball. As it passed through coils of wire on the barrel 1 foot apart, the voltage swung from negative to positive as it crossed the center. The scope measuring cursors are on the last 2 crossings for a travel time of 2.2 ms for the 1 foot with an average speed of 454 FPS.

If anybody has some scales and a 4 inch Poof ball, please weigh it.

Comments?? Am I missing something? Does this really compare favorably with a burst disk? :D
  • 0

Attachments
spudfilesfoamball.jpg
In barrel launch of the 4 inch foam ball at 60 PSI.
GGDT.jpg
GGDT
4 inch test.JPG
4 inch barrel test setup. Foam ball, barrel with pick up coils, coil pre-amp, scope, printer.
Last edited by Technician1002 on Thu Apr 30, 2009 5:42 pm, edited 1 time in total.
User avatar
Technician1002
Chief of Staff
Chief of Staff
 
Posts: 5190
Joined: Sat Apr 04, 2009 11:10 am
Reputation: 14

Unread postAuthor: jeepkahn » Thu Apr 30, 2009 1:41 pm

Yes, your QDV will prolly flow the same as a burst disc, you basically have a manually activated toolie valve, and with my toolie valve, ggdt models showed the same thing you're seeing when comparing ggdt #'s to real world performance...
  • 0


jeepkahn
Colonel
Colonel
 
Posts: 747
Joined: Fri Jan 09, 2009 1:11 pm
Location: Triad, NC, USA
Reputation: 0

Unread postAuthor: Carlman » Thu Apr 30, 2009 1:53 pm

It wont flow the same as a burst disk nor will the opening time be close to burst disk specs. Barrel sealer is still the closest modeled valve
  • 0

Image
Aussie spudders unite!!
User avatar
Carlman
Major General
Major General
 
Posts: 1618
Joined: Sat Nov 24, 2007 1:18 am
Location: Western Australia
Reputation: 0

Unread postAuthor: Technician1002 » Thu Apr 30, 2009 2:13 pm

Carlman wrote:It wont flow the same as a burst disk nor will the opening time be close to burst disk specs. Barrel sealer is still the closest modeled valve


GGDT only gives an error message. Piston Diameter greater than or equal to the valve seat. :( .

When modeling a generic valve, any opening time under 5 ms made no real difference, but nothing I put in even came close to my measured results. I even plugged in 100 PSI and a valve efficiency of 90% and still GGDT modeled it way slow.

I moved to Burst disk to get in the ball park at 60 PSI.

Try it. Use a opening time of anything under 7 ms. My tank, barrel, projectile, and try to model anything close in any valve style.

To even get close to a burst disk, I had to up the valve efficiency. I presume this is because burst disk has a ragged edge that adds turbulence that I don't have. A burst disk has a length of pipe before the disk (coupler half) that I don't have.

Until I get an actual weight of my projectile, I'm holding off on concluding the QDV is exceeding the performance of the burst disk. It has to be slower.
  • 0

User avatar
Technician1002
Chief of Staff
Chief of Staff
 
Posts: 5190
Joined: Sat Apr 04, 2009 11:10 am
Reputation: 14

Unread postAuthor: jeepkahn » Thu Apr 30, 2009 3:07 pm

have you tried using barrel sealer with a cv rating instead of effeciency rating in the config... somewhere d-hall had a rough formula for calculating actual cv rating..the wiki maybe...

http://www.spudfiles.com/spud_wiki/index.php/GGDT

I'm guessing here, but for calculating Cv you would use 3 for burst disc because this formula is for calculating fully open flow and not for calculating opening times etc... where you're valve, like a toolie, WHEN OPEN is going to flow like a burst disc because there is NO obstruction between the hi-pressure area of air and the barrel breech....
  • 0


jeepkahn
Colonel
Colonel
 
Posts: 747
Joined: Fri Jan 09, 2009 1:11 pm
Location: Triad, NC, USA
Reputation: 0

Unread postAuthor: Technician1002 » Thu Apr 30, 2009 3:14 pm

jeepkahn wrote:have you tried using barrel sealer with a cv rating instead of effeciency rating in the config... somewhere d-hall had a rough formula for calculating actual cv rating..the wiki maybe...


No, this is my first try at using GGDT. I use Ubuntu, so I had to borrow a machine to run it. After I get the ball weighed, I'll play again with better values.

The valve opening time is very quick without a high pressure pilot to slow it.

I gotta find the scale. The 2 inch piston weighs about the same as an apple... :idea: Apple.. Grocery store.. Scale.. Guys, I'm off to the market. :D
  • 0

Last edited by Technician1002 on Thu Apr 30, 2009 3:23 pm, edited 2 times in total.
User avatar
Technician1002
Chief of Staff
Chief of Staff
 
Posts: 5190
Joined: Sat Apr 04, 2009 11:10 am
Reputation: 14

Unread postAuthor: jeepkahn » Thu Apr 30, 2009 3:15 pm

check my edit above you're last post

also 60gms is pretty heavy for a foam ball that's nearly 4oz... a nerf footbal weighs less than that...

and does your barrel actually protrude into the chamber??? if not put 0 as the inside diameter on your chamber

I'm guessing 31gms for the foam ball...
  • 0


jeepkahn
Colonel
Colonel
 
Posts: 747
Joined: Fri Jan 09, 2009 1:11 pm
Location: Triad, NC, USA
Reputation: 0

Sponsored

Sponsor
 


Unread postAuthor: Ragnarok » Thu Apr 30, 2009 3:49 pm

jeepkahn wrote:also 60gms is pretty heavy for a foam ball that's nearly 4oz... a nerf footbal weighs less than that...

I don't know where you got that figure from, but an avoirdupois ounce is 28.35 grams (as opposed to a troy ounce of 31.1 grams).

For simplicity, it's usually easiest to reckon an ounce is ~30 grams. So it's about 2 ounces.

It'd just be so much simpler if everyone worked in metric... (Okay, I use Imperial on occasion, but mostly for rough guessing - "It's roughly 10 feet")
  • 0

Does that thing kinda look like a big cat to you?
User avatar
Ragnarok
Chief of Staff
Chief of Staff
 
Posts: 5339
Joined: Tue Dec 19, 2006 8:23 am
Location: The UK
Reputation: 8

Unread postAuthor: dewey-1 » Thu Apr 30, 2009 3:52 pm

The Nerf Pocket Vortex footballs weigh in at 30 grams.
  • 0

User avatar
dewey-1
Major General
Major General
 
Posts: 1298
Joined: Wed Mar 28, 2007 10:24 am
Location: NE Wisconsin USA
Reputation: 0

Unread postAuthor: Technician1002 » Thu Apr 30, 2009 4:17 pm

jeepkahn wrote:check my edit above you're last post

also 60gms is pretty heavy for a foam ball that's nearly 4oz... a nerf footbal weighs less than that...

and does your barrel actually protrude into the chamber??? if not put 0 as the inside diameter on your chamber

I'm guessing 31gms for the foam ball...


I'll make some weight adjustments. It was a guess.

The barrel.. Since the external barrel detaches, no, but the valve is inside the tank as a coaxial and extends all the way through the tank, so the tank volume is correct, about 700 cu in. The volume in the tank does have a 2 inch pipe all the way through it. The tank with rounded edges were "Adjusted" in GGDT to closely match the real tank volume. It was water volume measured when I finally pressure tested the tank.

Unlike a traditional coax design, the valve is at the other end of the tank so it is very close to the screw on barrels for a dead space of only a few cubic inches. This limits the length of the plumbing between the valve ports and the projectile when using barrels larger than 2 inch. (the 2 inch barrel permits loading projectiles against the piston.)

I'm back from the store. The valve core is just about 1/3 LB.

Edit, Update..

I did some plugging into GGDT using a barrel sealer valve. To get the thing to calculate, I lied and said the piston diameter is 2.1 inches so it was happy. I still had to run the valve coef high to match my measured results. The graph shows a very sluggish valve start as it assumes only air is used to open it, but in reality, the moving rod hits it so it starts opening while already in motion. With those parameters it shows a valve opening time of about 4 ms.

Plugging in a 2.5 inch piston with the same size seat (and weight) shows it fully open in under 2 ms. In actual measurements with the impact start, reality is just under 1 ms.

Again, in GGDT, I had to run the valve coef way up to fit reality. 75% is a good match.

Yikes I just plugged in the stuff for the 80 PSI 1 Liter Gateraid launch in the 3 inch barrel. :shock: No wonder I got a bruise. Just short of 2,000 Ft*Lb :D
  • 0

Attachments
PICT0324a.JPG
View into the open2 inch QDV valve.
PICT0302a.JPG
Edit, Added view of early construction photo with valve body in the tank ready for more sanding. A foam dust plug is in the far end to keep the dust down.
Last edited by Technician1002 on Thu Apr 30, 2009 6:04 pm, edited 5 times in total.
User avatar
Technician1002
Chief of Staff
Chief of Staff
 
Posts: 5190
Joined: Sat Apr 04, 2009 11:10 am
Reputation: 14

Unread postAuthor: jeepkahn » Thu Apr 30, 2009 4:27 pm

Ragnarok wrote:
jeepkahn wrote:also 60gms is pretty heavy for a foam ball that's nearly 4oz... a nerf footbal weighs less than that...

I don't know where you got that figure from, but an avoirdupois ounce is 28.35 grams (as opposed to a troy ounce of 31.1 grams).

For simplicity, it's usually easiest to reckon an ounce is ~30 grams. So it's about 2 ounces.

It'd just be so much simpler if everyone worked in metric... (Okay, I use Imperial on occasion, but mostly for rough guessing - "It's roughly 10 feet")


I was smoking something I think.... And yes, even though I'm Amerkin, I'm a big fan of the metric system...
  • 0


jeepkahn
Colonel
Colonel
 
Posts: 747
Joined: Fri Jan 09, 2009 1:11 pm
Location: Triad, NC, USA
Reputation: 0

Unread postAuthor: btrettel » Fri May 01, 2009 10:35 am

Technician, I'm a fan of your magnetic projectile and coil method. I plan on using the same method to produce a projectile position vs. time plot and from there I'll use a modified version of the simulation I wrote, BAGS, to figure out the valve flow coefficient as a function of pressure.

I haven't had the opportunity to collect any empirical data due to a lack of equipment and time, but what you've posted seems usable. I count 6 times the projectile crossed a coil on the oscilloscope output. Is the blue tape around the beginning of the barrel obscuring a coil? If so, that'd make sense and the coil at the beginning of the barrel would allow for this data to be used for the procedure I've detailed.

You'd probably find the results of the analysis interesting or useful for your own simulation in GGDT, so if there isn't a coil at the beginning of the barrel then you might want to retest with evenly spaced coils from the beginning to the end of the barrel.

Also, GGDT can run in Linux. I run Linux and getting GGDT to run took a little cajoling but it does work. If I recall correctly I followed this guide.
  • 0

All spud gun related projects are currently on hold.

btrettel
Major
Major
 
Posts: 380
Joined: Sun Feb 03, 2008 4:40 pm
Location: Maryland
Reputation: 0

Unread postAuthor: Technician1002 » Fri May 01, 2009 11:16 am

btrettel wrote:Technician, I'm a fan of your magnetic projectile and coil method. I plan on using the same method to produce a projectile position vs. time plot and from there I'll use a modified version of the simulation I wrote, BAGS, to figure out the valve flow coefficient as a function of pressure.

I haven't had the opportunity to collect any empirical data due to a lack of equipment and time, but what you've posted seems usable. I count 6 times the projectile crossed a coil on the oscilloscope output. Is the blue tape around the beginning of the barrel obscuring a coil? If so, that'd make sense and the coil at the beginning of the barrel would allow for this data to be used for the procedure I've detailed.

You'd probably find the results of the analysis interesting or useful for your own simulation in GGDT, so if there isn't a coil at the beginning of the barrel then you might want to retest with evenly spaced coils from the beginning to the end of the barrel.

Also, GGDT can run in Linux. I run Linux and getting GGDT to run took a little cajoling but it does work. If I recall correctly I followed this guide.


Good observation. The barrel is actually about 5.5 feet long. It worked well for a length for the canceled foam ball distance competition, but we did cut it a little short for 100 PSI. We were not using GGDT to design this.

Starting at the muzzle the coils are 1 foot apart with the first coil slightly beyond the projectile at rest position for a total of 6 coils. The first coil crossing is at about 2~4 inches of ball travel. If the barrel was fully 6 feet long there would have been a 7th coil right on the coupling.

Thanks for the tip on Linux. I'll look into it.

If you do use the method on the larger barrels, be aware in metropolitan areas, the coils make excellent AM radio antennas. You will need to filter the RF out. I used a 0.01 uF mylar shunt cap. Each coil is 4 turns.

Edit, all the coils are simply connected in series in a daisy chain fashion.

For the valve coef, It is looking like it is pretty good. To get the acceleration curve to match my data, I had to be optimistic in th value. I think the lack of a plumbing space between the "burst disk" and the chamber and no ragged edge of a disk is why.

If you have trouble getting the radio noise out of it, PM me and I'll share more info on how to.

As for a lack of a coil at the beginning of the barrel, I used the printed graph in GGDT and worked back from the muzzle. This is part of the way I was confirming my valve opening time. A slow valve has a slow rise in launch position. Working backwards is showing an opening time of just under 1 ms at higher pressures. (>50)
  • 0

User avatar
Technician1002
Chief of Staff
Chief of Staff
 
Posts: 5190
Joined: Sat Apr 04, 2009 11:10 am
Reputation: 14

Unread postAuthor: clide » Fri May 01, 2009 8:03 pm

I think part of the problem may be your setup. With a barrel significantly larger than the valve and a really light projectile you have a setup that is hypersensitive to valve flow rate. While it is not a bad idea to have the setup sensitive to the variable you are looking to find, I think you may have it to a point where it is magnifying errors in GGDT that don't show up in a more typical gun configuration.

Do you have something like a 5' tennis ball barrel that you can use to get test data and model the gun with?
  • 0

<a href="http://gbcannon.com" target="_blank"><img src="http://gbcannon.com/pics/misc/pixel.png" border="0"></a>latest update - debut of the cardapult

clide
Donating Member
Donating Member
 
Posts: 785
Joined: Sun Mar 06, 2005 3:06 am
Location: Oklahoma, USA
Reputation: 0

Unread postAuthor: Technician1002 » Sun May 03, 2009 6:49 pm

clide wrote:I think part of the problem may be your setup. With a barrel significantly larger than the valve and a really light projectile you have a setup that is hypersensitive to valve flow rate. While it is not a bad idea to have the setup sensitive to the variable you are looking to find, I think you may have it to a point where it is magnifying errors in GGDT that don't show up in a more typical gun configuration.

Do you have something like a 5' tennis ball barrel that you can use to get test data and model the gun with?


We only tested t shirts and foam balls for the competition, then a few apples, oranges, water bottles and an occasional soda. Much of the non-competition items were not magnet tested to get an acceleration curve.

We have a small selection of tubes. The red and white candy stripe is TB size. It was a 10 foot that is now a 3 foot and 7 foot section.

I haven't taken the time to put a magnet in a tennis ball yet. My yard is too small for testing them. Maybe later.

We did get some launch curves of t shirts in 2.5 and 3 inch barrels. We did not clock anything out the 2 inch or golf ball barrel. We did get a spud almost 350 yards.
  • 0

Attachments
Shootoff tubes.jpg
Launcher tubes and tanks
User avatar
Technician1002
Chief of Staff
Chief of Staff
 
Posts: 5190
Joined: Sat Apr 04, 2009 11:10 am
Reputation: 14

Next

Return to Pneumatic Cannon Discussion

Who is online

Registered users: Bing [Bot], Google [Bot], Google Adsense [Bot], Yahoo [Bot]

Reputation System ©'