Paintball Marker: Velocity too low, poor efficiency

[berninator]

Hey Folks,

I've been working on a piloted QEV style paintball marker and am running into results that I do not understand:
- Paintball velocity is ~75% of predicted velocity
- Efficiency (shots per tank) of set-up is ~20% of predicted value

I'm hoping someone here might be able to help me understand my results.

Description of Set-up:

Before building an actual paintball marker, I've build a proof-of-concept test cannon. It uses all of the parts that the paintball marker will use, but mounts them in an easily accessible and exchangeable configuration.
- HPA tank to a low pressure regulator (Palmers). Then to a 3-way, 2-position pilot valve (McMaster). Then to a QEV (the chinese black box QE-04). The exhaust of a QEV leads to a 7 inch barrel.
- I've built a sear-based trigger mechanism to operate the pilot valve to remove the human element from triggering. A pull string releases a sear allowing a spring to quickly un-press the pin on the pilot valve.

Description of Operation:
- trigger mechanism is set, pressing the pin on the pilot valve which pressurizes an air chamber (0.56" ID x 12" long)
- paintball is muzzle-loaded into the barrel
- pull string is pulled, trigger mechanism releases pin on pilot valve, QEV opens and air chamber empties through barrel

Results
For my test set up, I get the following results:

Code: Select all

Velocity:
PSI     Pred.     Actual FPS
80      256       187
100     296       228
120     326       257
140     354       276
160     381       429
180     402       318
Note (10 shots were taken at each pressure, std dev are between 5 and 10 FPS)

Efficiency:
3000 PSI, 48 in^3 tank should yield ~550 shots. I am getting about 110 from a tank.

Thoughts on the results:

Let's start with efficiency. When I turn the pressure to 140 or above, there is an audible hiss as air leaks from the system. I would expect this to result in some loss of efficiency, but it seems like that wouldn't explain the huge difference between my prediction and my results. I am really at a loss for what to make of this. I'll rebuild the system with better thread sealant to see if this makes a difference.
- Can anyone comment on how much a leaky system affects efficiency?

The velocity measurements are ~75% of my predictions, which is not bad, but I'm frustrated that they are not closer. I suspect the cause is one or more of the following (please let me know if any of these match your experience):
- Pressure loss as air must take a 90 degree turn to get from QEV to barrel
- slow to pilot QEV caused by low Cv (0.11) of pilot valve
- slow to pilot QEV caused by too small diameter of tubing from pilot valve to QEV.
- air leaking around paintball (most paintballs needed a tiny push to go in)
- friction between paintball and barrel (seems likely)

Notes on predictions:
- I built a spreadsheet to simulate the operation of the paintball marker. It is based on the model described in the link below. Of particular note:
- The model takes into account choked flow through valves.
- I have further modified the model to account for the temperature decrease as the gas expands into the barrel.
- I have verified my model against the test results in the paper, so I am confident that the model is valid at those scales.
- My spreadsheet agrees with the predictions produced by GGDT (Gas Gun Design Tool).


Model used for results predictions: http://www.iontrap.wabash.edu/publicati ... cannon.pdf.

Your thoughts and suggestions would be appreciated.

[jackssmirkingrevenge]

Welcome to the forum!

Interesting project, though your image showing the setup is not showing, it would really help to assess your work.

I'm assuming that you took all the dead volume into account when it came to predicting efficiency? All the losses caused by the air expanding from the main bottle through the various regulators until it reaches the actual firing chamber are going to drop the amount of available energy

A leaky system is definitely not going to help, you should eliminate this variable post haste for a meaningful test, otherwise it's like trying to determine your mileage with a hole in your gas tank!

With regards to velocity, I would say a +/- 25% match is reasonable within the parameters involved, barrel friction for example is a very significant factor especially since you have a relatively short barrel, and paintballs aren't exactly the most uniform projectiles in the world. Also silly question, have you weighed the paintballs for consistency?

[berninator]

Thanks for your response! For some reason, I can't embed images in the post using the img tag (google drive or imgur). I've included two grainy images as attachments. I'll take a better picture and post this evening.

For my efficiency calculation (shots per tank = Ptank*Vtank/(Pacc*Vacc)) I have not taken into account dead volume. I'm unsure how much energy will have been wasted between the bottle and the accumulator (iron pipe next to aluminum barrel, attached to the QEV), but my instincts tell me that it is low. Does your experience suggest otherwise?

I have received a bottle of loc-tite 545, I'll rebuild the system and post results. (I opted for teflon tape at first, but the threads on these chinese QEVs do not seem to seal well).

With regard to lower-than-predicted velocity, maybe I should count myself lucky that it is not worse! I did not weigh each paintball, but these are high quality paintballs and my FPS consistency (Std Dev of 5 to 10 FPS) suggests that paintball inconsistency is not the primary driver of lower-than-predicted velocity.

I'll post again when I have new results. Any other suggestions would be appreciated!

[jackssmirkingrevenge]

Ah, there's your problem.

Your dead volume is comparatively enormous, it most certainly has to be taken into account when calculating efficiency. I estimate that the volume between bottle and chamber is 50-75% of your firing chamber volume, which even with no heat losses would make your efficiency calculations way off if you hadn't taken it into account.

Even within the bottle itself, pressure and therefore efficiency is lost by simply going from the 3000 psi main fill to the 850 psi regulator exit.

edit: just to put some numbers on what I'm talking about, you have a remote line between the bottle and the regulator which has an internal volume of approximately 0.75 cubic inches at 850 psi. For an ideal gas that amount of air would be around 5 cubic inches at 120 psi. This means that (given that your chamber is around 3 cubic inches) that the air wasted filling the remote coil alone could have given you at least 1 extra shot at 120 psi per fill, doubling your efficiency.

I have received a bottle of loc-tite 545, I'll rebuild the system and post results. (I opted for teflon tape at first, but the threads on these chinese QEVs do not seem to seal well).

Are you sure it's the right thread? In my experience Chinese fittings tend to be BSP or European standard while most US fittings are in NPT, and 1/4" NPT is not equivalent to 1/4" BSP even if you might be able to force them together.

With regards to muzzle energy, consider this little exercise I had done with a virtual 0.177" BB launcher - in each case, the amount of air is the same, but at each step I half the volume and double the pressure.

20 inch chamber at 50 psi - 406 feet per second

10 inch chamber at 100 psi - 537 feet per second

5 inch chamber at 200 psi - 684 feet per second

2.5 inch chamber at 400 psi - 830 feet per second

1.25 inch chamber at 800 psi - 958 feet per second

0.625 inch chamber at 1600 psi - 1006 feet per second

While this is just a model you can see that higher pressure leads to more acceleration, which means that you can get more performance for the same air use.

I would say that if you aim for an operating pressure of around 200-300 psi and a much smaller chamber you will up your efficiency considerably, also consider using a tank which already has a low output pressure, for example Ninja tanks which even in their basic 48/3000 aluminum form can be set to a 450 psi output.

[berninator]

As promised, I have attached a higher resolution image of my test setup.

I've thought at length about your comments about dead volume and efficiency but am still confused.

Some definitions (correct me if I am wrong):
- My definition of dead volume is the volume of gas that does not contribute to propulsion and is wasted by operation of the system.
- Efficiency in this case is number of actual "charges" per bottle vs. theoretical number of charges per bottle.

To your point, I understand that the volume between the pilot valve and the QEV is dead volume that does not contribute to propulsion (and is vented to atmosphere). But the volume between the bottle and the pilot valve never empties to atmosphere and is used to fill the chamber for the next shot. That air is not wasted as your example would suggest.

You make another point about efficiency losses going from 3000 psi to 850 through the bottle regulator, and then again from 850 to 150 through the LP regulator. Pressure has indeed decreased, but with a corresponding increase in volume, so the total energy in the gas should remain the same. (Assuming that heat loss is negligible. Even if it is substantial, the gas expands as it moves through the regulators, which would cool it. Any heat transfer should be into the gas, which should increase the efficiency of the system).

In short, ignoring leaks and and friction losses, I do not understand how dead volume could contribute to such a huge loss in efficiency.

Are you sure it's the right thread? In my experience Chinese fittings tend to be BSP or European standard while most US fittings are in NPT, and 1/4" NPT is not equivalent to 1/4" BSP even if you might be able to force them together.

The ebay page suggested NPT threads, but that is far from a guarantee. If I cannot get the threads to seal with loctite, I will replace the black-box QEV with one I purchased from mcmaster.

I would say that if you aim for an operating pressure of around 200-300 psi and a much smaller chamber you will up your efficiency considerably,

I appreciate the illustration of the relationship between pressure and speed. I had not thought to run a simulation experiment to test the speed/pressure relationship at constant energy. Having seen it, it is intuitive. I agree that this marker would be more efficient at 200 to 300 PSI, but one of the design goals of this project was to have a paintball marker that operated at less than 150PSI (the typical safety rating for easily obtainable valves).

I've resealed the threads tonight. Crossing my fingers. I'll post new results when I have them.

[jackssmirkingrevenge]

Anywhere the gas can expand, it loses pressure and therefore energy. Let's look at a simplified example without regulation:

Say I have a 1000cc main chamber at 500 psi, 50cc of tubing and a 100cc firing chamber.

[1000cc at 500 psi] --> valve --> [50cc tubing at 0 psi] --> valve --> [100cc firing chamber at 0 psi]

I want to charge the firing chamber for the first shot, so I open both valves. Pressure equalizes for all the vessels, so I get the following:

[1000cc at 435 psi] --> valve --> [50cc tubing at 435 psi] --> valve --> [100cc firing chamber at 435 psi]

I now close all valves and fire, resulting in the following:

[1000cc at 435 psi] --> valve --> [50cc tubing at 435 psi] --> valve --> [100cc firing chamber at 0 psi]

Now I charge the system for the second shot and open all valves:

[1000cc at 397 psi] --> valve --> [50cc tubing at 397 psi] --> valve --> [100cc firing chamber at 397 psi]

Valves closed, second shot:

[1000cc at 397 psi] --> valve --> [50cc tubing at 397 psi] --> valve --> [100cc firing chamber at 0 psi]

Valves open to charge for third shot:

[1000cc at 362 psi] --> valve --> [50cc tubing at 362 psi] --> valve --> [100cc firing chamber at 362 psi]

So starting with 500 psi, I got three shots, first at 435 psi, then at 397 psi and finally 362 psi.

-------------------------------------------

We will now repeat the experiment with the tubing twice as long (ie 100cc instead of 50cc):

[1000cc at 500 psi] --> valve --> [100cc tubing at 0 psi] --> valve --> [100cc firing chamber at 0 psi]

charge for first shot:

[1000cc at 417 psi] --> valve --> [100cc tubing at 417 psi] --> valve --> [100cc firing chamber at 417 psi]

first shot:

[1000cc at 417 psi] --> valve --> [100cc tubing at 417 psi] --> valve --> [100cc firing chamber at 0 psi]

charge for second shot:

[1000cc at 382 psi] --> valve --> [100cc tubing at 382 psi] --> valve --> [100cc firing chamber at 382 psi]

second shot:

[1000cc at 382 psi] --> valve --> [100cc tubing at 382 psi] --> valve --> [100cc firing chamber at 0 psi]

charge for third shot:

[1000cc at 350 psi] --> valve --> [100cc tubing at 350 psi] --> valve --> [100cc firing chamber at 350 psi]

So starting with 500 psi, I got three shots, first at 417 psi, then at 382 psi and finally 350 psi.

Ignoring heat losses, doubling the length of the tubing has had a noticeable effect on the final pressure in the firing chamber.

Using Boyle's law where Pressure x Volume is constant, I'm sure you can write a simple spreadsheet for your system and see how yourself how varying the size of the dead volume is affecting your efficiency.

[berninator]

I created a google spreadsheet to explore efficiency further. The link below leads to a public copy, feel free to mess around with it.

https://docs.google.com/spreadsheets/d/ ... sp=sharing

This seems to support my suspicion that dead volume between the bottle and the valve has little impact on efficiency

My system is constructed like this:
bottle -> reg1 -> remote line -> reg2 -> air line 1 -> valve -> air line 2 -> QEV -> chamber

Using Boyle's law and the volume of the chamber+air line 2 (vented to atmosphere with each shot), the theoretical maximum number of shots per 48/3000 bottle is 355. Taking into account the dead volume between the bottle and the valve. This decreases to 354. Increasing the dead volume by multiples yields the following:

Code: Select all

multiple     # shots
1x              354
2x              354
5x              353
10x             353
50x             348
100x            341

The volume between the bottle and the valve is essentially charged once when the system is initially pressurised. After each vent-repressurize cycle, that volume is in the same state. Because the gas in that volume doesn't leave the system (it passes to another part of the system) it doesn't seem to have any bearing on efficiency except for the initial pressurization.

I retested my system after using thread sealant. The little gauges on the bottles are difficult to read accurately, but I got 80 shots from what looks like a drop of 900 PSI. This suggests I'll get about 250 shots per tank. Amazingly, the threads on the QEV were still leaking air, so I've replaced the black-box QEV with one purchased from mcmaster. I'll post results with the new set-up when I have them.