Developing a bstaley Disco

[rsterne]

After having a delay due to a burst disc failing and draining my SCUBA tank, and subsequently having to refill it (a long process with a ShoeBox).... I got back to doing more testing today.... I installed a stack of four 90 Durometer O-rings, this time without the 0.055" thick backer ring.... I figured that since the O-rings deflect a lot less than the 70D, I would have sufficient adjustment range between Max. CCW and 4 turns in (the settings I used before) to see what we need to see.... Well, at max. CCW (minimum lift) I got just about what I expected, but the rest of the results leave me totally confused.... Here is what happened....



I used the same colours for all the stroke settings as on the other two graphs.... The vertical scale is the same, but it is shifted down 50 fps in velocity to include all the curves.... The graph is even the same length as the original (no O-ring) graph.... The black "max. stroke" curve starts at nearly maximum velocity, indicating that the hammer spring preload is too high.... The efficiency is 1.23 FPE/CI, in between what I got unbuffered and with the 70D buffer.... I think that if the entire curve was included, it would be a pretty usable string, and the data fits well with what I got using the 70D buffer.... Now for the bad news....

First of all, the other four curves show the OPPOSITE trend for the velocity to what you would expect in a bstaley gun.... As you turn the striker in CW, the valve lift increases, and the velocity is suppose to increase with it.... That was plainly the case with the 70D buffer.... With the 90D O-rings, however, the highest velocity occurred at only 1 turn in, and the velocity DROPPED as I turned the striker in CW, with the lowest velocity (not counting the black line) occurring with the striker in 4 turns.... Three of the curves, on the face of it, look pretty decent in length, 14-15 shots.... The problem is, that the efficiency of all four of these curves was WORSE that the unbuffered gun.... It ran from 0.93 to 1.08 FPE/CI, improving as the velocity dropped.... The velocity with the buffer ran about 30-40 fps lower than the unbuffered gun, and the efficiency was poor as well, so this setup is pretty much a bust, IMO.... The big question is WHY ???

A day or two ago, I had a request to measure the valve lift with a buffer in place.... While I had the stack of four 90D O-rings in the gun, I added the Lift Indicator Rod, set the striker to Max. CCW, readjusted the preload to just shy of cioil bind, and measured the lift at three pressures, 2000, 1500, and 1000 psi.... I then pulled the gun apart, wound the striker in four turns CW, and repeated the measurements.... Here are the results....



A word of caution here.... I was having trouble with the O-ring cocking on the rod (ie not staying perpendicular to it), so it was hard to measure.... I would give a tolerance of about 0.005" plus/minus to the above numbers, but the TREND is pretty clear.... There is VERY little lift occuring when the striker is Max CCW, but it is still about twice as much at 1000 psi as it is at 2000.... and it only takes about 1/64" of lift to produce 640 fps (13 FPE) at 2000 psi.... AMAZING !!! ....

I'll be looking forward to seeing your analysis of these rather confusing results.... So far, it would appear that with a Disco, the 70D O-rings are far superior.... So where do we go from here?....

Bob

[rsterne]

Well, I learned a valuable lesson about the O-ring buffers, today.... If the hammer isn't hitting them, they can shift out of place and cause weird results (you were right, Eric!).... I began to suspect that last night, and bstaley suggested it this morning, but I had already confirmed it.... I think the slight air blast coming out around the stem of the valve pushes them back a bit, or maybe just makes a space between them if the hammer isn't already contacting them as it would be before that blast of air occurs if the buffer was even partially engaged.... In addition, I made a new Lift Indicator Rod, even lighter and with a thinner head than the original.... This one is made from 5/32" OD aluminum hobby tubing, and only weighs 1.4 grams and only requires 1/2 turn adjustment on the preload.... The larger diameter means it doesn't wiggle around in the hole in the RVA adjusting screw, and the larger O-ring tends to stay nice and straight, allowing much more accurate and consistent measurements.... I then redid my previous lift measurements, and took a bunch of new ones, with the following results....



I have confirmed that at 4 turns in, with either O-ring stack, the hammer does NOT contact the buffer (or so lightly as to have no effect), providing the O-rings haven't shifted.... Therefore, those curves on the previous graphs should be ignored.... Anything happening there is a result of the buffer not remaining properly in place.... In addition, the previous chart I posted about the lift numbers with the 90D buffer are incorrect, the numbers above are the accurate ones....

Note that the lift at 1000 psi with no buffer in place changes 0.022" between max. stroke (full CCW) and 4 turns in, or about 0.005" per turn, and that should be linear.... That means that at 3 turns in, with no buffer, the lift at 1000 psi should be about 0.086".... I made another test, with the 90D buffer at 3 turns in, and the lift was 0.084", so the buffer is either not in contact, or it is just barely touching it at 1000 psi.... It would NOT be producing any changes to the shot string, so again any difference between that string and the unbuffered one would be due to the stack moving around.... I also tested the 70D buffer with the backer ring at 3 turns in, and the lift was 0.075" at 1000 psi.... This indicates that the hammer is just making contact at that pressure, but it would not be making contact with the buffer at over 1200 psi.... so the data for that string (unfortunately, the best string I had initially) may be somewhat suspect.... Perhaps that is why I couldn't duplicate it last night.... On the other hand, if the hammer is contacting the buffer at 1000 psi and below (only), that may be the reason that first string extended down to 850 psi....

Now that I have a better indication of what range of adjustments will and will not work, I am going to go back and retest both the 90D and 70D (with backer) stacks at the appropriate settings, but this time with 1/2 turn increments.... I'm hoping for a lot more consistent results....

Bob

[AirGunEric]

I kinda figured that would occur at some point or another- but I'm not the one doing the testing!

Good luck with the next steps.

[rsterne]

I completed the second set of tests, changing the striker setting by only a half turn at a time instead of a full turn.... I also eliminated the settings where I was sure the buffer was not playing a role.... All strings were started at 2000 psi, and the Disco hammer spring was set just shy of coil bind as before.... and only shots within a 4% ES are graphed.... The four 70 Durometer O-rings had a 90D Backer ring under the stack.... and here are the results....



Basically, once the striker is screwed in 2 turns CW or more from the fully CCW position, it is ineffective in altering the shot string to any significant degree.... These results are different for the 3 turns out curve (blue line) which I found in the first test.... My only explanation is that the buffer in that test was moving out of place slightly, and interfering with the hammer in just the right way to skew the results.... That curve was a lower velocity, longer, and flatter, but since I haven't been able to duplicate it in several tries, all I can say is that it was a tease, unfortunately.... The 2 and 3 turn in shot strings are virtually identical to the unbuffered results....

When we explore the striker adjustment range from maximum CCW (minimum lift) to 2 turns in, however, we have some very interesting things happening.... As the velocity drops, the efficiency increases, and the strings get longer.... At 1.5 turn in, the average velocity is 775 fps (at 1.18 FPE/CI), and a 1 turn in, it drops to 718 fps (at 1.36 FPE/CI).... In both cases, a large portion of the first part of the shot string appears to be missing, an indication that the hammer strike is too great for the 2000 psi fill pressure.... For both of those strings, further experimentation is needed.... At 1/2 turn in, the velocity is only 617 fps, but the gun got 32 shots within 4% ES, and the efficiency was 1.65 FPE/CI, very efficient for a 12 FPE gun.... If you remember from the first tests with the 70D buffer, at maximum CCW, the velocity was about 430 fps and the gun did 65 shots at 2.11 FPE/CI between 1650 psi and 1000.... This soft, tall buffer is excellent for drastically reducing the power of the gun and raising the shot count and efficiency.... If that is what you are after, this is something you should definitely consider....

Now to look at the stiffer buffer, consisting of four 90 Durometer O-rings.... Here are the results....



With this stiffer buffer, and without the extra height of the 0.055" thick backer ring, the buffer was basically ineffective once the adjuster was past 1 turn in.... The top group of curves pretty much mirror those of the unbuffered gun at the same adjustments, although for some reason they seem to peak at a slightly higher velocity, although no higher than the unbuffered gun did at maximum.... The efficiencies are similar as well, so IMO no advantage can be found operating in that adjustment range....

Once again, however, when we explore from maximum CCW (minimum lift) over the first part of the striker adjustment, we see more shots at increased efficiency, but at lower velocity.... The shot string at 1/2 turn in (791 fps average) started out with the first shot the fastest, so it definitely needs to be explored with less hammer spring preload.... With the striker at maximum CCW, the velocity averaged 671 fps, and it looked like it too, could benefit from a bit less hammer energy for the 2000 psi fill.... The efficiency of that 15 shot string was 1.30 FPE/CI, and I have a feeling that both can be increased a bit.... However, when you consider that the 1 turn in string with the 70D buffer was higher velocity AND more efficient, I have my doubts if the 90D buffer will show any advantages over the softer one....

There are a couple of general observations I have to make.... When the striker adjustment first starts to have a serious affect on the velocity, it appears that the shot string shifts to the left, ie it peaks at higher pressure.... In an unbuffered gun, that usually indicates too much hammer strike, and that is likely the same case here.... Only adjusting the preload to a lower value will let us find out if there is more shots and efficiency hiding there.... Once the velocity is reduced further, however, the shot string returns to the original peak pressure, and eventually, as the velocity is lowered enough, then the fill pressure must follow, just as in detuning by hammer energy alone.... While the basic idea of reducing lift directly by decreasing the striker engagement with the valve is simple in concept, and very effective.... it still needs that fine tuning of trying various combinations of hammer spring preload and/or fill (and refill) pressure to find the optimum setting....

When time permits, I plan to concentrate on those parts of the envelope where less hammer strike is needed to optimize the results.... In the meantime, there is certainly lots here to chew on and discuss....

Bob

[rsterne]

I had a chance this evening to do some testing of preload changes on the 90 Durometer stack for the striker position that was 1/2 turn out.... That tune had a shot string that started high and declined, so I wanted to see what would happen if I reduced the hammer spring preload, if the gun would react in a "normal" fashion even with the buffer.... Well, as I expected, it did.... The colours have changed on this graph for clarity, so don't use them to refer to earlier graphs....



All the solid lines are at maximum hammer spring preload, as before.... The black line is with the striker full CCW (0 turns), the solid red line is 1/2 turn in on the striker, and the blue line is 1 turn in on the striker.... The other two red lines show what happens with the preload reduced two and four turns.... As you can see, the peak velocity drops and shifts to a lower pressure, reacting just like a conventional unregulated PCP, as the preload is reduced.... In addition, the efficiency increases, and the shot count increases, again as per usual.... The interesting thing about the efficiency is that it increases from top to bottom on that graph, starting at 1.06 FPE/CI for the blue line, and progressing to 1.30 FPE/CI for the black line.... I have no reason to believe that any "magic" will happen in between, although in between tunes are certainly possible.... As an example, if you wanted to tune for the flattest curve peaking at 800 fps, you would use a striker setting roughly 3/4 of a turn in CW, and a preload likely in the neighbourhood of 2 turns out from coil bind.... and you could expect about 12 shots at an efficiency of about 1.20 FPE/CI.... Using a full size Disco tube, that would be about 24-25 shots within a 4% ES.... The problem is, that is basically no better than you can get with a stock Disco.... From what I have seen so far (and I'm running out of ideas), the 90 Durometer buffer doesn't seem to offer a significant improvement over stock.... I detuned a Disco with just an RVA (preload adjuster) and achieved 33 shots at 18.6 FPE within 4% ES from 1800 psi down to 1000.... That is 614 FPE from 452 CI of air, which works out to 1.36 FPE/CI.... That is more power AND more efficiency than anything I've seen so far with the 90 Durometer buffer.... It is possible, of course, that shortening the valve stem and increasing the hammer stroke might provide better results.... I'm still hoping that the 70 Durometer buffer will provide greater efficiencies as well....

Bob

[AirGunEric]

Well, let's look at this from a different angle- 90 Durometer is pretty rigid- so the amount of flex possible is significantly less than a 70D unit. As this is about establishing a "buffer" or what I believe we could also label a "pad" or "cushion" some amount of "give" in it would seem to be required. 90D has very little compressibility to it, relatively speaking. Perhaps a stack of some 70D units and a couple 90D units would provide better results? You gain some height on the stack to reduce the travel of the hammer, but still retain a fair bit of "cushion".

[rsterne]

Yeah, if you look back at the second post in the thread, there is a little chart showing the difference in compressibility between 70D and 90D.... basically the 90D is about twice as stiff.... While it's possible that they might be better in a gun with a more powerful hammer strike, in the Disco my gut feel is that the 70D is looking better.... What I find quite amazing is that there is still a difference in lift between high and low pressure even with the buffer engaged to the point of limiting the velocity to non-PAL numbers.... with lifts between 0.022"-0.028".... Yes, you could use a mix of 70D and 90D to achieve a "spring rate" in between, but I'm thinking that adjusting the striker "in between" will give similar results....

In some ways, the buffer is working like a very stiff valve spring, in that it is limiting the lift and making the difference between the lift at high and low pressure smaller.... The major difference is that it is easily adjustable without tearing the valve out of the gun.... The more I think about it, the more I like that analogy.... The 90D O-rings work like a stiffer spring with less preload....

Bob

[rsterne]

I decided that for the Preload testing on the 70 Durometer buffer I would try removing the 0.055" thick Backer Ring.... Here is the test at Maximum CCW on the striker, and maximum preload (just shy of coil bind).... shown on the graph as the black line with square markers....



You can see that my intuition was pretty good about using the Backer Ring to check out the lower end of the velocity range possible using a bstaley buffer.... Without it, and with the striker set fully CCW (minimum lift) the average velocity of the string was 750 fps, about halfway between the 1 turn in and the 1.5 turn in strings.... However, this is perfect to explore reducing the preload, as most of the first part of the shot string is misssing.... The plan is to try a few reduced preload settings with this striker adjustment, and then wind it in 1/2 turn and test again.... Incidently, once again on this graph, the efficiency increases from top to bottom, from 1.07 FPE/CI at 2 turns in to 1.65 FPE/CI at 1/2 turn in, with the "Max. No Backer" string coming in at 1.20 FPE/CI....

Bob

[rsterne]

I completed the preload adjustment trials this afternoon, working on the strings that were "left shifted" ie the first shots were the fastest when the hammer spring preload was at maximum.... Here are the results for the stack of four 90 Durometer O-rings....



The string at 1/2 turn in on the striker (solid red line) was shifted left.... I shot additional strings at 2 turns and 4 turns out on the preload, shown in the dotted and dashed red lines.... The string at 4 turns out was the longest and the most efficient, returning 14 shots averaging 738 fps with an efficiency of 1.25 FPE/CI.... The highest efficiency overall was still the minimum lift string (black line - striker fully CCW), which was 15 shots at 670 fps. and 1.30 FPE/CI.... That is decent efficiency, but not as good as I shot with a stock Disco with the power just dialed down to the same level with an RVA.... On to the stack of four 70 Durometer O-rings (without the backer ring)....



The first thing to note is that at 1 turn in on the striker (the blue line), the shot string is almost identical to the same curve using 90 D Orings (top), and indeed virtually identical to the unbuffered gun at the same setting.... In other words, by that time the buffer is pretty much ineffective.... The set of red lines (1/2 turn in on the striker) are the same as on the first graph, solid is maximum preload, dotted is 2 turns out, and dashed is 4 turns out.... The most efficient of those was the latter, 16 shots at 758 fps with an efficiency of 1.24 FPE/CI.... The black lines are at fully CCW on the striker, using the same line types as above.... Once again, the curve at 4 turns out on the preload was the most efficient, giving 19 shots at 713 fps (16 FPE) and 1.31 FPE/CI.... Once again the efficiencies were beaten by a stock Disco detuned with a simple RVA, but in this case there were more shots in the string within a 4% ES than I have been able to achieve before at these power levels.... The bottom string on that graph would be 38-40 shots using a full size Disco tube....

I can now offer the following conclusions for my testing of the four #113 O-ring bstaley buffers in a stock Disco fitted with a Challenger hammer and RVA....

1. The 90 Durometer O-rings provided a wider range of velocities for the same range of striker adjustment.... eg. 1 turn was ~175 fps compared to ~125 fps....

2. The 70 Durometer O-rings required an additional spacer to get under 700 fps, but when used with a 0.055" thick backer ring could be adjusted down to about 440 fps (6 FPE).... I preferred the 70D....

3. With both buffers, as the buffer first becomes engaged, the peak of the velocity curve shifts to higher pressures, and must be compensated for with reduced hammer spring preload.... This is over a range of about 1 turn for the 90D buffer and about 1.5 turns for the 70D buffer....

4. I was unable to achieve increased efficiency during this first stage of buffer engagement, compared to what could be achieved at similar velocities with a simple RVA.... At the lower power levels, however, longer than usual shot counts were obtained with the 70D buffer....

5. As the buffer becomes fully engaged, the velocity drops rapidly, and the peak of the curve shifts back to the original pressure range.... Eventually, as the velocity is decreased further, the peak shifts to lower pressures, and the fill pressure must be reduced accordingly....

6. During this second portion of the buffer use, with the 70D buffer with backer ring, the efficiency climbed to higher levels than I have acheived in a Disco with a simple RVA, peaking at 2.21 FPE/CI, an extremely high value, at 6 FPE.... These high efficiencies were obtained with very large shot counts as well....

The bstaley buffer, when used in a stock Disco, provides the ability to change the velocity over a wide range, in a very quick and simple manner, without having to adjust the fill pressure.... I didn't find any gains in efficiency unless the power was reduced below 15 FPE using the 70D buffer.... although some lengthening of the shot strings occurred when settings were optimized for any given velocity.... I personally preferred the 70D buffer over the harder 90D O-rings.... although I didn't test the 90D below 670 fps.... When the Disco is detuned to below 600 fps using the 70D buffer, the efficiency skyrockets, and the shot strings soar....

All in all, an interesting experiment.... I learned a lot....

Bob

[rsterne]

Today I got a chance to try a buffer made from three 70 Durometer O-rings instead of four.... That required replacing the flanged striked in the Challenger hammer with one made from a 1/4-28 SHSS with the point ground flat.... The hammer ended up slighty lighter, but with the striker set flush with the end of the hammer, the stroke increased from 0.50" (with the flanged one fully retracted CCW) back to the stock travel of 0.58".... This increased the velocity of the shot strings, making the first shot pretty much the fastest, and moving the energy of the gun up to 24 FPE with the striker extended a turn, which made the buffer ineffective.... I left the hammer spring preload just shy of coil bind and shot a string with the striker flush, extended CW 1/2 turn, CW 1 turn, and also with the striker below flush (CCW) by 1/2 turn.... Those results are show in the graph below as solid lines.... The efficiencies are all below 1.0 FPE/CI, as we would expect with the excessive hammer strike eliminating most of the first half of the shot strings....



The buffer is ineffective with the striker 1 turn CW, and having little effect at 1/2 turn CW, but becomes effective when flush, and the velocity continues to drop as the striker is recessed in the face of the hammer.... I didn't try it further CCW than 1/2 turn, however, as I was interested in velocities not too far below stock.... I then optimized the hammer spring preload for three striker settings, with the buffer just disengaged, starting to engage, and well engaged, and those are the dotted lines.... The colours are the same as the strings taken at maximum preload.... The dotted red line is fairly close to stock Disco power and shot count, although down a bit on efficiency at 1.07 FPE/CI.... The dotted green line is with the striker flush with the end of the Challenger hammer, with the preload set 5 turns out from coil bind.... That turned out to be the optimum setting for a 2000 psi fill, and the string is 17 shots at 18 FPE for an efficiency of 1.29 FPE/CI.... The dotted black line was with the striker 1/2 turn recessed in the hammer face and the preload out 6 turns, and returned 20 shots at 16 FPE with an efficiency of 1.34 FPE/CI.... Once again, these efficiencies, while pretty decent, are no better than what I can get with a simple RVA in a stock Disco.... However, you can still use a 2000 psi fill, and because of that, the strings are longer....

In reality, there is no practical difference between today's testing and what I did last week with the stock Challenger striker and one more O-ring in the buffer.... Either setup can be adjusted for about the same velocities and efficiencies.... although today's strings are 1 shot longer than with the taller buffer.... When you consider that with a full size Disco tube, the shot count will double from the above.... 40 shots within a 4% ES at 16 FPE is a pretty decent string starting from 2000 psi and shooting down to 1150....

Bob