Modeling PCP Valve Lift and Dwell

[Motorhead]

Maybe this "PARTICULAR"  part of this has or has not been openly talked about ?

That being the SPEED of the impact on valve stem from a lighter hammer. * Certainly a Lock-Time speed up having rifle cycle faster.
But where I am going is based upon my findings that Fast & Light hammers in small caliber PCP's have tended to make the guns less pellet fit picky.
Rather than having pellet stuck like a cork sitting there tightly until sufficient pressure is behind it & dislodging it at some point. * This fundamentally with a slower opening valve which could be a light hammer spring, heavy hammer or some of both.

The Light Fast strike of these LW hammers appears to get the valve off the seat & at max flow height so quickly that the pressure rise behind pellet is instantaneous with the loose or tight fit being a pressure check to release from a stationary position less important.

This not saying correct pellet fits not important ... it is for consistency indeed !!

Another interesting thing I have found with LW hammers is that the EFFICIENCY swings from a light pellet to a heavy pellet in the same gun have become less.   Can only think it is a dwell issue that results from Valve open / Valve close duration being so minimal.
* Get MAX pressure behind pellet fast as you can, launch it and shut the valve quickly you simply are not wasting much air.
Pellet weight variances of 3-6 grains in a .177 caliber gun at @ 15-20 ft lbs just don't seem to matter as much as it did with a heavy hammer.

Just sharing thoughts here 
Scott
 

 

[Bill G]

Theory sounds solid too me. Kinda goes hand in hand with material compression of the poppet.  The energy delivered by a LW vs HW hammer is, in theory, the same. The speed at impact with the stem is different. To me, this would make the upswing of the dwell parabola more steep with the light wieght hammer.  The clipping affect would occur earlier in the cycle thus not need to be as long to perform the same work.  Or stated differently... The decompression of the poppet material would occur over a shorter period thus allowing the stored energy to get to work earlier.  The closing cylcle would be nearly the same excluding inertia differential of LW vs HW hammer. 

Scott, Your work with the LW hammers has really proven a lot of theory.  Couple this with all the discovery of sealing properties of less compressable poppet material and we are going to see very substantual gains in efficiency.  I would think that it is only a short time away that we will be seeing Big bores, that are traditionally hard to cock, employing much lighter hammers and much lighter springs.  Especially with the movement toward balanced valve theories.

[Bill G]

This thread is a real treasure of information. Bob you mentioned moving it to the Geek Gate. Is there a parent gate to this geek gate? gate?

[rsterne]

The "Geek Gate" is my nickname for the Engineering-Research & Development child board in the Machine Shop gate.... Scott, I'm not arguing that light hammers don't have their place, they certainly work great in moderately powered (eg. regulated) guns, no question.... Bill G, I still question how you can achieve BOTH light hammers AND lighter springs in a big bore, unless like you mention you use a balanced valve of some sort.... Perhaps that is the way of the future, particularly for regulated guns, but we have a LOT more to learn about how the balancing of the valve affects the self-regulating properties over a wide range of pressure....

The main thing I wanted to pass on was how little the pellet moves by the time the valve reaches (nearly) full flow, less than 1/4".... It means that the pressure rise is VERY fast, whether the hammer is light/fast or heavy/slow.... A lot is obviously happening during that very brief period of time, and the modelling is VERY complex, so all I'm really trying to say is that we need more data and then just work with the trends instead of worrying so much about the math.... Lloyd's model is excellent for looking at the overall picture.... but trying to separate out what happens in the first 1% of the pellet travel is IMO asking too much....

DATA, we need DATA....

Bob

[Bill G]

It is certainly asking alot for my math ability I have. If I had done a better job at learning calculus it would probably be easier.  connecting the dots and discovering how it all comes together is certainly alot of fun.

[rsterne]

^X2....

Bob

[PakProtector]

I want to turn loose GT-Power on this issue. Discarding piston anybody? Might as well go into the acoustic development as well. not sure it can be done, but their stuff looks thorough.
cheers,
Douglas

[rsterne]

No idea what that means, sorry....

Bob

[Bill G]

Douglas, I spun out at turn one and you left me behind......

[Bill G]

Googled it!  Software for engines... ie minus the piston.  Got a copy?  My youngest brother worked for a company out of germany called called ANSYS.  Flow dynamic stuff.  They do flow through Turbines, turbos, ventilation ect.

[rsterne]

Computational Flow Dynamics (CFD) is an interesting field, but wayyyyyyyyyyy beyond my comprehension.... At some point in time somebody with the modern (desktop) equivalent of a Cray Supercomputer will undoubtedly figure out how to build the world's best PCP, but I'll be long dead and gone by then.... At least I'll have the comfort of knowing that about 90% of the modelling he did and decisions he made based on that didn't pan out in the real world.... at least the first time around....

Bob

[Bill G]

no doubt.  Also, when they look back and ask, "how we all got so close", once they get it figured out. 

[rsterne]

Kind of reminds me of a story about a Ferrari F1 car back in the 60's.... They spent weeks working on a tuned exhaust system, countless hours on the Dyno, and gained "X" horsepower with it.... Then at the first race, they found NO improvement in performance, if anything they were fractionally slower.... During the race, one of the cars had a minor shunt and folded over the end of the exhaust pipe, so during a pit stop they cut the end off with a hacksaw to remove the damaged portion.... and promptly gained a bunch of power, with the lap times dropping drastically.... They quickly pitted the other team car, cut off the same amount of pipe, and voila, both cars were now running faster than ever before....

Bob

[lloyd-ss]

Yes, an outwardly simple dynamic process that is in reality extremely complex and probably almost impossible to model and more successfully optimized by experimentation.

Bob has already done this, but I worked some more on calculating the THEORETICAL travel of the valve stem of a Discovery valve with the spring pre-loaded to near coil bind.  I ended up with a graph that looks good but is really of little use.

All theoretical:  The 60 gram hammer hits the 4 gram valve at about 19.7 fps.  Using conservation of momentum and (for ease of calculation) a fully elastic collision, the valve  accelerates to 37.7fps off the seat and the hammer slows to 17.9fps.  The graph shows the theoretical valve stem rise to .08" in a little over 4 milliseconds, and then the mirror image drop back to the closed position in a total of 8 milliseconds.  In the calculation, only the 2000psi action on the valve stem diameter of .156 ( 38 pounds of closing force) is considered as acting on the valve to close it.

Hmmmm.... basically junk to be picked apart.  The lift height of .08" looks reasonable, but in reality, that 8 millisec is enough time to mostly empty the gun's reservoir.  Going on empirical data, if the valve opened and closed instantaneously, it would only need to remain open for about one millisecond, and the pellet should be out of the end of the barrel in about 2 and a half millisecs.
So where are the holes in the theory? Please add your thoughts!

One is that it is not an elastic collision.  In reality, I think that fact makes the situation worse.   There has already been mention of valve stem deflection during the hammer strike, which I think is like a tennis racquet deflecting from the tennis ball and then the ball accelerating to a much higher velocity: preloading a spring and releasing it.  But it seems like that would make the valve open faster and higher. But several people have reported better performance with a less flexible seat and stem.

So I am wondering why the valve closes so much faster than theory suggests?  Or even if the valve doesn't fully close, the flow basically stops.
Maybe the flow past the valve has more effect than has been thought.

Theoretical valve open time based only on hammer strike
is about 8 millisec.  Empirical data suggests open time is
closer to one millisec.

[rsterne]

Interesting data, Lloyd.... Steve in NC did a similar calculation and got about 1 mSec at 2000 psi and 2 mSec at 1000 psi.... BUT I don't know what he used for a closing force.... IIRC, he also got about 0.06" lift at 2000 psi and 0.12" at 1000, which was right about what I have measured.... I've quit going to the Green Forum and have no idea where the post could be found....

As the closing force increases, the lift and dwell decrease.... You are neglecting the spring force (~7 lbs on the seat, more when open) and as you mention any additional drag / pressure differential across the head of the poppet.... You are using the straight ratio of masses in your conservation of momentum calculation (which makes sense), but there is an energy loss (conversion to heat in the seal?) as the seat material decompresses over 0.006-0.009" of distance in a stock Disco.... That energy can be found by using the closing force (pressure x area + spring) in lbs. times the compression distance in the material (in feet), over 2.... Since the hammer now has less (residual) energy, it's velocity is less, so it has also less momentum.... The reduced residual energy reduces the lift, and the reduced momentum reduces the dwell.... Here are some measurements on a Disco poppet, showing the amount of energy to "crack" the valve (ie decompress the seal material)....



If you read over my post on "Sonic Choking" you will find that I have concluded that for most common PCPs, once the valve begins to close past the point where there is a sufficient reduction in area to push the velocity past Mach 1 (at the ambient pressure) then I feel that the FLOW may drop off to near zero long before the valve actually hits the seat.... This will tend to "clip" off the last part of the flow....

I have some limited data on how much affect the drag on the head of the poppet is.... and Sean also has some evidence from his work on full sized valves on board ship.... where if the valve is near the seat the forces are huge, but when fully open you can literally change the lift by hand....  Here is some data taken with my Disco Double, by closing off the barrel port, and also shooting with no pellet, showing what happens to the valve lift....



The preload was set at maximum, and lift measurements were taken at two pressures, and with two different poppets.... The first column is with a 50 gr. .30 cal pellet, the second column with no pellet, and the third column with the barrel port choked off to virtually no flow (maybe 1%).... My idea is that compared to the standard case (with pellet), with no pellet the flow past the head of the poppet is increased, so the drag / pressure differential across the poppet should increase, and the lift should decrease (and it did, by 0.005").... In the case where the flow was nearly completely blocked, there should be almost no drag / pressure differential across the head of the poppet, so the lift should increase (and it did, by 0.006-0.007").... The PEEK valve showed the same trend at 2800 psi, but less change in lift (I think because it spends more time open past the "clipping limit").... The 2000 psi data with the PEEK poppet is useless as the hammer was hitting the back of the valve, limiting the lift to 0.162" in all cases....

This data, of course, shows only the result of the change in drag with flow rate (different measured lift), it does not quantify it in a meaningful way.... other than to give some idea of its relative importance to total lift and dwell.... In addition, the pressure differential across the head of the poppet is nowhere near linear.... when the valve is just clear of the seat, there is virtually no flow, so the cracking force still applies.... This then drops off extremely rapidly, and by the time the lift reaches 1/4 of the throat area is likely so small that it can be ignored, and all the closing force is being supplied by the area of the stem and the spring, which is what you used in your model.... Then as the poppet starts approaching the seat, before the closing force grows to any great degree, I think sonic choking takes over and the flow essentially stops before the force returns to the cracking force as the valve reseals....

Bob

[lloyd-ss]

Bob and all,
I pretty much agree in principal with all that you have stated and am working on a spreadsheet that will model the valve lift based on spring, hammer, throat, valve  mass, pressure, seat deflection, etc.  A lot of variables and a lot of energy and momentum being transferred around. 

Scott’s comments a couple of posts ago about light weight hammers got the wheels turning, too.  What happens when a high vel LW hammer hits the valve?
And I looked at Steve in NC’s equations for lift and dwell (in your opening post), and although I agree with them in concept, instead of using hammer energy and momentum to determine the lift and dwell, I am thinking that using the residual energy and momentum that is imparted to the VALVE   is more appropriate.

Using a slightly more robust version of the spreadsheet model that I used to create the single graph a couple of posts ago, here are some numbers.  The platform is a Discovery, basically stock, but with adjustable spring preload, i.e., RVA.

The following chart is interesting, but some work still needs to be done to make it more accurate.  I feel that the calculations are valid, but that some of the forces that affect the valve are not defined in the model.

The first set of data is for a 60 gram hammer and a 4 gram valve and a stock spring adjusted to about 3 turns less than coil bind. Pressure is 2000 psi, stock valve. I did not take an energy deduction for unseating the valve because it is less than 10% of the hammer’s energy and would not affect the TRENDS that I am trying to show.

So again, in set 1, we have a stock hammer hit with the resulting lift and dwell shown at the far left.

In Set 2, the hammer weight is cut in half, (following Scott's experimentation) but everything else is left the same. Even though the hammer energy did not change (that is governed by the spring, not the hammer weight) the valve lift and dwell increased significantly.  If you look at the energy and the momentum of the hammer and valve before and after the hit, you can confirm that the laws of conservation of energy and conservation of momentum has been observed.  (Note that if the valve and hammer were the same mass and it was a completely elastic hit, the valve would have taken 100% of the velocity and energy of the hammer, and the hammer would have stopped dead.)

In Set 3, the hammer spring preload has been reduced so that the valve lift is now the same as in Set 1.  I find it interesting that the spring preload had to be greatly reduced with the light hammer to bring the lift and dwell back down.

In set 4, the VALVE weight was reduced by 50% and the spring preload again reduced to achieve the same lift and dwell as in Set 1.  This time I find it very interesting that the spring preload required very little reduction.  I think the fact that the hammer is so much heavier than the valve makes that occur.   

In set 5, the valve weight is again reduced by 50%, and again, very little spring adjustment is required to bring the lift a dwell back to the original values.

(click on the chart to enlarge it)


So what does this show? One thing is that hammer weight seems to have a large affect whereas valve weight seems to have little. I think it also shows that it is difficult to change the ratio of the valve lift to dwell very easily. In my calculations, I see that the valve dwell values are much larger than what is really observed in the gun.  I am trying to figure out why that is.  I am reasonably sure my calcs are correct, but maybe they are not.  Also affecting the dwell are the other various outside forces that have previously been mentioned: curtain affect, sonic choking, boundary layer, etc.
Still much to learn.
Lloyd-ss

[rsterne]

Wow, what disturbing results.... I hate it when calculations don't agree with reality or theory.... My understanding is that lift is proportional to (residual) energy and dwell is proportional to (residual) momentum, and all practical experiments I have done show that lightening the hammer reduces the dwell, or at least reduces the power of the shot, with no other changes.... When you lighten the hammer, you have to ADD preload to get back to where you were, not reduce it....

With the very high forces on the valve, compared to it's mass (38 lbs. vs. 4 grams), is not the collision completely INELASTIC?.... I would think from the moment of contact with the valve, the hammer and valve move as a unit, being constantly decelerated by the closing force (plus spring force and any pressure differential/drag across the poppet head).... Is it proper to use the mass of the valve by itself rather than the two masses as one?.... Something is definitely wrong, IMO, as this doesn't fit with observed facts....

In addition, how can the dwell numbers you give be in milliseconds?.... Your first graph gave a dwell of over 8 mSec, not 0.0076 mSec.... Here is a graphic representation of what I got for the lift/dwell curve after the valve was cracked....



I was not using actual weights or forces, just relative ones.... I only used whatever the residual energy and momentum was, assuming that the energy loss to crack the valve as the same for all cases.... For the mass, I used the combined mass of the hammer and valve, moving as one, so where I show the mass of the hammer as 50%, in fact it would be slightly less than that.... Using Lloyd's example, a 60 g. hammer plus a 4 g. valve = 64 g., so I would have used 32 gr. which means the hammer would only be 28 g. not 30 g.... For the top graph, I used a mass of 20 units or 10 units, and a closing force of 20 units, giving a deceleration of 1 unit for the heavy hammer, and 2 units for the light hammer.... In order to start with the same (residual) energy, the velocity for the heavy hammer was 10 units, and for the light hammer it was 14.14 units....

The second graph shows an increase in velocity from 14.14 units to 16 units, to achieve the same area under the curve, which I feel represents the amount of air that needs to be released to provide the same shot.... The third graph shows the velocity increased further to 17 units, to provide the same area but including the effects of clipping due to curtain area on the lighter hammer.... These two graphs are only approximate in their velocity values to give a visual representation of what needs to happen to provide a similar shot....

Bob

[lloyd-ss]

Bob,
Ha, ha, disturbing results?  I think we'll all live, so it can't be that disturbing. 

I think what I have shown is that a fully elastic collision is not what is happening between the hammer and valve.In fact, it is so far off as to pretty much eliminate elastic collisions and anything close to elastic.  But fully in-elastic where the hammer and valve marry up and track each other?  I doubt that too.  Probably somewhere in between.  Also, I am thinking that the amount of energy that is supposedly lost to opening/cracking the valve is underestimated.  You know how if the gun is a little over-pressurized that it seems to take a very heavy hammer hit to open the valve.  But once you get he valve to open, it only takes a little more force to open the valve enough for actual operation. 
I will play around with my spreadsheet model some more and see if I can get something that tracks with empirical data. Actually, I like the results that Steve's two simplified formulas give.  But without his derivations, I am having a tough time getting to that point.  I have to say that Steve usually has a very straight-forward approach, and is usually right. 
Lloyd

[Motorhead]

On down the rabbit hole we go .... further we go thinking it's just a hole, less it seems to be.

Thanks Lloyd for at least trying and fooling around with "Lighter & Faster" hammers.
If anything it is being proven many valve systems as they come from the factory are far from optimum seemingly based on tube size where hammer gos gets X diameter & length ... to counter X size and its weight we install X size poppet and a poppet spring to control all that hammer weight ... to get all that hammer moving to create sufficient momentum to open valve we install X spring value.

Just find sometimes reverse engineering and using a practical approach of trial and error gets results outside of Text Book accepted "way it should be"

Always interesting, seldom dull is the tinkerers life 

Regards,
Scott

[rsterne]

The problem with modelling the "cracking" energy is that it doesn't stop instantly that the valve comes off the seat.... As Big Bore Bart mentioned at the end of the previous page, the valve can come a few thou off the seat before any flow occurs.... With no flow, that means you still have full pressure differential across the head of the poppet (presumably at the seal diameter), which would mean that instead of the 0.006"-0.009" I measured for the compression of the poppet in a Disco we might be dealing with twice that.... That would not so good up twice the energy before the force even begins to drop off.... Then we have an intermediate phase where the force is dropping off quickly (linearly? exponentially? whatever?), and depending on the lift and the flow rate, it may come down to (nearly) the closing force, which is the stem area plus the valve spring.... That bit of data that I took with open, normal, and blocked ports shows that there is still some residual drag or pressure differential across the head of the poppet, even in the case where it is opening past 1/4 of the diameter.... Then, once the valve begins to close, at some point the flow through the valve likely reaches Mach 1, and sonic choking begins to further reduce the flow, which again raises the pressure across the head of the poppet, increasing the closing force....

Sean's example of the force to change the lift of a (ship's) valve being hugely greater when it's near the seat than when it's fully open is an example.... He says that when the valve is just cracked if might take a crowbar to lift it against the flow, but when fully open, you can change the lift easily by hand.... I hope you manage to get a model that works, and tracks our experimental results.... I'd love to be able to better understand what is happening....

I'm curious as to why you think the collision between hammer and valve stem isn't ACTING as if it's inelastic?.... Isn't the valve stem acting like it's in a powerful gravitation field (the pressure and spring) which then makes it act like it has more weight.... It's really hard to wrap my brain around a 4 gram valve stem having a weight of 139 lbs., but that is the case in a Disco at 2000 psi.... Surely the stem can't react the same to the hammer impact with and without the force of the air and spring acting on it.... I believe Steve in NC thinks the collision should be treated as inelastic throughout the entire cycle until the poppet is completely reseated, at which point the hammer moves away from it in the infamous "hammer bounce"....

Bob