Test of PEEK as a high pressure valve material

[rsterne]

Yep. Lloyd, you're absolutely correct, if you make the stem(s) large enough you can make the forces, both to crack and to close, pretty much anything you want, including stay open (C larger than A) and dump whatever air is available in a fixed volume chamber.... If the chamber is about 1/2 the bore volume, I understand that can be pretty efficient, though I've never done the math myself.... IIRC, that 2:1 bore:valve ratio is a Perotti design idea....

Bob

[lloyd-ss]

Here is a variation on the design that uses a fairly conventional valve package.  A little easier to build, I believe.  There is a pressurized biasing chamber behind the middle part of the valve stem. It should work if the bleed passage is large enough.
Again, it greatly reduces the force required to crack open the valve. The light blue circles are o-rings. Forces are in pounds.
Lloyd-ss

[csdilligaf]

   I cant help myself but I think this valve should be called  "the Big Bang Theory".  You guys are very impressive with calcs and graphs.  Sure beats the trial and error method.

[rsterne]

Lloyd, I guess the question is how important is the opening force (energy to "crack" the valve) compared to the total energy required to provide the necessary lift and dwell once the airflow starts.... The measurements I did on a stock Delrin Disco poppet only showed 0.009" compression at 2000 psi, which calculates out to an energy required to crack the valve of 0.052 FPE, or about 21% of the total hammer energy available, leaving the rest to do the work required producing useful lift and dwell.... Yes, increasing the pressure and the diameter will increase that energy requirement, but changing to PEEK will reduce it significantly.... I did some testing on my Disco Double before and after replacing the Disco poppet with a PEEK one (same hammer strike), as follows:



The important column for this discussion is the first one, with a pellet in the chamber and the ports wide open to maximize airflow.... At 2800 psi, the PEEK poppet gave about 35% more lift, which means there was 35% more RESIDUAL energy after the valve cracked.... In a previous thread (Reply #25.... http://www.gatewaytoairguns.org/GTA/index.php?topic=57538.20 ) I showed my conclusion that the energy to crack the valve is on the order of 10-25% of the total.... While this is nothing to be sneezed at, it certainly doesn't mean you could cut the hammer energy in half or less.... even if you eliminated all of the excess opening force, and only left the closing force acting on the stem diameter.... To REALLY reduce the total hammer energy requirement, IMO, you have to reduce the "back" stem diameter (or counterbalance it)....

The design in your last post removes one of the potential drawbacks I saw to your previous drawing, and that is with the balancing stem out in the front of the valve you introduce more twists and turns to the airflow entering the valve from the reservoir.... While this isn't an issue in a 20 FPE gun, I think it would be a major concern in a 200 (or 500) FPE gun where the ideal plenum is going to be 200-500 cc....

Bob

[lloyd-ss]

Bob,
For the specific application I have, I tend to disagree that changing to a peek valve will improve the situation enough for my liking.
The valve now has a .51 diameter throat with a .187 stem and a valve head made of glass filled Delrin.  The Peek may compress less than the glass filled Delrin, but I would not expect a change as great as changing from standard Delrin. The closed valve has 613 founds of force on it at 3000 psi.
The hammer and spring set up I am using now is 167 grams with a 3" stroke, delivering almost 7 foot pounds of energy. And that still isn't enough to get the valve working properly at 3000 psi.   I have tried many combinations of hammer and spring and none are to my liking.
Here is what I think is a large part of the problem.  Lets compare a .255 throat discovery valve and the .51 throat big bore valve both operating at 3000 psi. 

                              small    big   
throat  dia              .255    .510
area   sqin              .051    .204
force  lbs                  153     613
margin load              192     382   lbs/linear inch
margin pressure       12.7    25.5  ksi assuming .015 wide margin

The last 2 lines are the important ones.  We have doubled the throat diameter but quadrupled its area and force.  The length of the valve margin (circumference) has only doubled, but must handle 4 times the load, so the margin is seeing twice the force per sqin of contact.  I think the big valve is compressing much more than expected and thus spreading the load over a larger area.  Otherwise, I think the valve would break. So now that .015 wide margin is maybe .030 wide,  and has compressed maybe .020 instead of .010. That doubles the opening energy required. 
All of the high powered big bores take significant effort to cock.  I think reducing the load on the closed valve head will greatly reduce the hammer force required, and thus reduce the cocking effort.
I will find out.
Lloyd

[rsterne]

Good point about the margin load being doubled IF the seat is the same width.... That would certainly double the energy required to crack the valve if the compression distance doubled.... I have no doubt that cancelling out the opening force on the valve (just leaving the closing force) will reduce the hammer strike necessary to cycle it.... I'm just concerned that the RESIDUAL energy available to cycle the valve throughout the lift and dwell required to produce the FPE you need won't increase as much as you hope, allowing you to reduce the hammer strike by a large amount....

Any reason the seat can't be wider?.... I realize that increases the OD, and hence the force, but the margin area will grow faster than the force, so there has to be an optimum width relative to the diameter, doesn't there?.... Either that, or go to carbon-filled PEEK.... *LOL*

Have you tried to measure the seat compression?.... That would allow you to calculate the FPE required to crack the valve.... (613/24 x distance) with the distance in inches.... I get 0.255 FPE for each 0.010".... If we assume the seat is compressing 0.020", that's only 1/2 FPE to crack the valve, dropping your hammer energy to 6.5 FPE.... It seems to me that by far the majority of your energy is going into producing lift and dwell, not cracking the valve.... Either that, or Steve's math is flawed, and I wouldn't want to bet on that....

If you're going to test this, I guess time will tell....

Bob

[lloyd-ss]

Bob,
All very good points and questions and I will try and be diligent in documenting cause and effect as I go along.  That is not always a strong point with me, so I will follow your lead.  You are definitely the benchmark in that department!
Your statement about the majority of the energy going to produce lift and dwell rings true, but I am at a loss as why that is the case.  Maybe when the valve head is so large, plus being so disproportionately large in comparison to the valve stem, the presure drop across the valve head tries harder to shut the valve.   I don't know.

I will measure the seat compression soon.  I haven't had this unit apart in a few months and I don't remember some of the details.  I will get a good "as-built" before I start altering.

Also, after I get the as-built documented, I will make a peek valve head before I go after the balanced valve design.  I just got a length of it from McMaster this past week.  I didn't think I could hold off any longer with Peek being what all the Cool Kids were playing with, LOL.  And yes, I looked at the glass filled Peek... and then looked at my billfold. LOL again!
Thanks, Lloyd

[lloyd-ss]

I checked the movement of the valve stem with a dial test indicator as I slowly filled and slowly released the pressure.  Not much hysteresis, but the valve was slightly slower to recover as the pressure was bled.

Valve head material glass filled Delrin, valve throat approx .500, but I will have to do a tear-down to verify that. Not sure about the margin, but it is small.  I will also measure that.

Here are the readings at teh tip of the valve stem as the tank was bled:
3000 psi .009"
2500   .008
2000  .007
1500  .0057
1000  .0045
500    .003
100    .002 (valve started to leak)
0        .000


I did a bleed from about 2000, then a fill, and then a final bleed and the readings were pretty repeatable.
There was less seat compression than I was expecting, especially if you discount the first .002 as just taking up slack.
Lloyd.
 

[csdilligaf]

  Just be glad you dont have to by Surgically implantable Peek.  I did a job last week for knee joint implant.  I normally include material in the job but for 3/4" x 3" x 12" long it was $800.  I had them drop ship it to me.  I got a little left over,  anyone got any idea's for it?

[rsterne]

Save it until you REALLY need it, Chip.... *LOL*....

Lloyd, that is about the amount of deflection I would expect, the problem is, that indicates only about 1/4 FPE to crack the valve off the seat.... I suspect you are correct that the initial force from the pressure differential across the head is much greater than what we think it is.... I was thinking that the force on the stem was virtually zero with the valve closed, but in fact the way we need to think of the system is like this....



The area of the stem "A" is always seeing pressure, which in your case means ~90 lbs (including the spring?).... We should be thinking of the area of the seat "B" as being composed of two parts, "A" plus the rest (B-A), so when closed the force is ~90 plus (613-90) = 613 lbs.... The force on that remainder (B-A) falls as the pressure is equalized between the Valve "Pr" and the Exhaust Port "Pe".... It never reaches zero, but the force due to the pressure differential once the flow is maxed, which you can also look at as the parasitic drag on the head, is very small, on the order of 3-5% in this experiment....



The first column is the "normal" case, wide port transfer port and a pellet loaded.... The second column is the same but no pellet (no backpressure, so maximum possible flow).... The third column is with the transfer port virtually completely blocked, so much so that a pellet was spit out at visible velocity ~200 fps, and therefore basically no flow through the valve, and hence no parasitic drag.... The data for the 2000 psi PEEK case was useless as the hammer was hitting the back of the valve body (max. lift 0.162"), but the data for 2800 psi is very good.... Compare the lift at 2800 psi in column 1 to column 3, and by completely blocking the flow, the lift only increased 6% in the case of (badly pounded) Delrin and 2% in the case of PEEK.... At 2000 psi with the Delrin, the increase in lift was only 4%.... Since the lift is proportional to the inverse of the closing force, we then have evidence that the parasitic drag (force during the major flow event) is only about 5% of the total closing force.... ie on the order of the spring force or so.... 

The last time I presented this argument, however, I was thinking that the force on the stem was zero when the valve was closed, but as in my first paragraph above, that is NOT the case, it is ALWAYS present, but the force on the head of the poppet shouldn't include it.... Therefore, we have a situation that as the valve opens we go from full (opening) force (ie 613 lbs) through a rapidly declining force until the pressure in the exhaust port rises to near the valve (reservoir) pressure, at which point we have the closing force on the stem (~90 lbs. including spring) plus the parasitic drag, which is pretty small.... This means that instead of breaking the force into two major parts (613 before cracking and ~90 after) we need to include a rapidly decreasing transition between the two....

That additional closing force, due to the pressure differential across the head of the poppet (not including the stem area) is very short-lived, and my gut feel is that it decreases exponentially with lift, to near zero at the curtain (flow limiting) lift, which is 1/4 the throat diameter.... However, it may well be that the force doesn't decrease that quickly, and perhaps we should look at it as a linear function over the initial portion of the valve opening cycle, but virtually gone by the time the full pressure reaches the pellet (1/10th of the dwell time or less).... If that is indeed the case, then balancing the valve to remove the additional closing force on the "non-stem" portion of the head may produce a larger change than I originally thought....

Boy, that was a long-winded way of saying you were right, HUH?....

Bob

[SeanMP]

Geez I leave you guys alone for a day a look what you've gotten into

Lloyd
I think your undervaluing the effect of various material selection can have on the outcome. I'll refer back to the valve I made  to test the material theory. It had a .500" throat and I am certain if I had used conventional materials I simply would not have been able to open it. Given that the test platform is essentially a QB ....with a 66gram hammer.

You've measured directly a deflection of .009". I think that's .008" too much given that your trying to move it with an effective mass that is 0.05%. It seems clear that as long as the valve head and seat are bonded some of the energy is being absorbed by all the masses, valve body, gun, shooter, earth etc etc. This would probably explain your missing energy.

I think the way forward is to reduce the elastic compression to near 0 so the instant there is any movement you have an effective separation and declining forces and energy being retained in the collision rather than spreading into the surrounding mass. BTW if it matters when I measured the seat margin with a comparator it was just a fraction over .001"
Glass filled delrin is just not strong enough to span a .500" hole without flexing....in fact PEEK isn't strong enough to bridge that hole either without deflection. I'm thinking the material really needs to be flipped upside down and optimum would be something well over 100ksi to bridge that hole with 600lbs force on it.

Beyond that your safe to reduce that stem area to .156". I haven't failed the stem at that OD yet......I'm thinking about trying .125. That is in O1 tool steel of course hardened and tempered. If you think about the forces applied to a 1/8" drift punch it seems doable

[rsterne]

Good point about the "missing energy", Sean.... When you consider that the poppet is, as you say, "bonded" to the rest of the gun with a force of over 600 lbs., it is reasonable to expect that when the hammer hits the stem a portion of that 7 FPE moves the entire rifle and shooter, although I might argue he is more likely to sway than to accelerate the Earth..... *LOL*.... Also, an excellent point that we are trying to overcome a resistance (inertia?) of over 600 lbs. (nearly 10,000 oz.) with a 6 oz. hammer travelling at about 35 fps.... That kind of puts it in perspective....

Bob

[lloyd-ss]

.................... Also, an excellent point that we are trying to overcome a resistance (inertia?) of over 600 lbs. (nearly 10,000 oz.) with a 6 oz. hammer travelling at about 35 fps.... That kind of puts it in perspective....
Bob

Like trying to move a school bus by crashing into it with a bicycle, LOL.  Yes, I certainly agree with the reasoning.
The test bed I have right now is a .510 cal that will eventually be a hunting bullpup. That is, after I get the power plant to where I think it needs to be. I plan on starting a new thread about this project and the first thing covered will be cleaning up my baseline test bed to remove those energy absorbing things you mentioned, Sean. 

What I am wondering, and I'd like opinions, for a .510 cal, the throat should be at least .500 for max power, right?  And needing a valve head that big,  I should probably use a one-piece steel valve stem and head (or titanium, Chip?), with a peek or metal seat. I know I am jumping ahead in the process but I don't want to go down too many dead ends.  The setup right now has a 43" barrel which I plan on shortening to about 36". Currently I am getting 710 fpe at 2600 psi with an air efficiency of 1.00-1.12.  That is with 455 gn bullets going 840fps. (Efficiencies have been as high as 1.45 at good velocities, even with lighter bullets.) My calculator shows that dropping back to a 36" barrel will still yield 650fpe at 805 fps, but I think I can tighten the efficiency up a bit and recover the velocity.  Douglas is casting the bullets for me and he also cast some nice 250 and 320 gn bullets to work with.  I want this to be an efficient, flat shooting gun with plenty of good shots.  I have always found the  phrase, "2 good shots, with a lower velocity back-up," to be an excuse more than anything else. [ Sorry for the soap box stand.]
Anyway, that's the direction I am headed. All opinions will be appreciated. [BTW machining is already happening on some of the bullpup's  components]
Thanks,
Lloyd

[rsterne]

I have always found the  phrase, "2 good shots, with a lower velocity back-up," to be an excuse more than anything else. [ Sorry for the soap box stand.]

I agree 100%.... I'll be back with more comments after shopping....

Bob

[Big Bore Bart]

   Lloyd;  Go steel to steel, two different grades/ hardness', to prevent galling.   I've decided to use my early education in valving, and just go with tool steel and 4100 series for the entire valve.   The stem/head is made from an Datsun valve stem, the body/seat is the 4140.   It seals at shop air pressure with no spring.   I am designing for a MSWP of 6500psi.   I do not think any synthetic will handle that.

[lloyd-ss]

B³
6,500 in the tank? That's certainly something to check out from a distance, LOL! Good luck wiht that project.

Yes, the steel on steel sounds good, and starting with a small off the shelf valve sounds even better !!  What about a motorcycle valve? I bet there is one out there that might be the perfect size.  I might need to make a trip to the motorcycle parts dept!  Yes, definitely!  Thanks for a great idea!
Lloyd

[Big Bore Bart]

   The Tubing has a MTR yield of 115800psi so an actual MSWP of 7400Psi.    End plug and valve body of 4140, secured with six  6-32 HSCS, heads in shear in .221 counterbores.   The math works out.  The max pressure will be approached  slooooly.   Actual working pressure will be determined by the valve operation.

[rsterne]

Lloyd, you're getting good efficiency now with the 43" barrel, but shortening the barrel will drop that.... I still firmly believe that the majority of the efficiency comes from where you close the valve, and the closer the bullet is to the muzzle the lower it is.... Yes, if you can get a quicker pressure rise, and square off the pulse, you can increase the efficiency, and running higher pressures and higher sectional densities allow you to get more air into the barrel (and therefore more power out) before the bullet reaches halfway (which IMO is where the efficiency takes a dump).... but there are limits to both....

It would appear that reversing our current setup, as Sean has done, by using a steel valve and a PEEK seat, is a big step, and even moreso with a large valve throat.... You measured 0.009" movement on the valve stem as you depressurized your valve, but how much in addition to that does the head flex before the seal is broken?.... The larger the diameter and the more flexible the head of the poppet, the further the valve stem has to move before the sealing surface of the valve even starts that 0.009" decompression, right?.... Changing to a steel valve largely eliminates that flex, just leaving the seat compression to deal with.... I looked at CF filled PEEK, and yes, it's about 3X the price of unfilled, but so what?.... A foot of it would make a LOT of valve seats....

To me, the biggest problem in all of this is getting a steel poppet valve.... I don't have the skills to make one, so that leaves me with trying to find a commercial one.... Maybe these sources will be the answer.... 

http://saito-engines.info/specifications.html .... Saito 4-strokes, with cylinder bores from 20mm up to 38mm, some of the valves have to be in the usuable size range for us....

http://downloads.hobbico.com/evpl/osm/ev35530aa.pdf .... Typical OS 4-stroke, similar sizes, you can purchase the intake and exhaust valves as assemlies (valve, spring, retainer and collets)....

The problem is going to be finding out the sizes of the valves in the various models, that information is not easy to come by....

Bob

[Motorhead]

You guys should really start a NEW thread on Balanced Valves IMO 
This thread was about PEEK, now it is off in entirely a different direction 

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Now that said ..... My Peek poppet continues to leak down  ... lapping it in slowed it down a whole lot !!
But rifle sitting for a week, ten days and tanks empty once more 

Reinstalled a Delrin poppet and it 99.9% sealing once more 

So under VERY HIGH 2000# ++ pressure you may be good ... at 1700# it just did not have good enough sealing properties for a gun on the ready and stored charged up.

All I got 

[rsterne]

Well it's only HALF about balanced valves, for me, it's mostly about what Sean is doing a steel poppet on a PEEK seat.... Here is an OS Valve and spring (size unknown)



and here is a cylinder head to give an idea of the proportions relative to the bore size....



It looks to me like the valves are roughly 1/3 of the bore, and since the bores run from 3/4" to 1.5", we may be able to get valves from 1/4" - 1/2" or slightly larger?.... If the length is 4 times the head, they could be 1-2" long.... I've asked for dimensions from an OS Rep., don't know if we'll get anything useful....

Bob