Poppet extrusion - key contributing factors?

[nervoustrigger]

Today over in the darkside gate we are discussing the practical limits of PEEK as a poppet material, initially in terms of how low an operating pressure makes sense (given the relative difficulty of getting it to seal at low pressures) and then later the necessary sealing margin to avoid extrusion failure.

On the latter topic, I ran a simple calculation as follows.  BTW you can probably just glaze over the actual numbers...they simply represent an arbitrary example for calculating the demands on the material.

To put some numbers to it, the compressive strength of PEEK is listed at around 15,000psi.  In the example above, the force acting on the poppet is about 80lbs (1500psi acting on a 0.260” dia).  The annular ring carrying this load (rim of the poppet) has an area of about 0.0115 sq in.  So the material is seeing a loading of about 7000psi.  That’s only about half of the material’s rating.  Or to say it differently, I could probably get away with 2x the pressure and use this valve in a 3000psi application.

Scott pointed out we need to also consider the violent impact the poppet experiences when it slams closed.  However it's not so obvious to me that it would be a big contributor.

I figure it has been discussed before so I wanted to do some research before reaching out to you, and I ran across this post from Lloyd:

...can we look at the closing half of the cycle because that seems easiest  to analyze. The main component of closing force on the valve, as has been pretty thoroughly discussed and agreed upon, is the pressure times the valve stem area.  So with a .156 stem dia, and 2000 psi, the closing force is 38 pounds.  If the valve weighs .08 ounce, or .00017 slugs mass, and accelerates from zero at the full open position for 0.10" to the seat, the velocity of the valve when it hits the seat is only 43 ft/sec, which is 29mph.  So even though the closing acceleration was over 6,000g's, the velocity at impact was relatively slow.....I guess.  Those plastic valve heads hold up forever, so the impact on the seat can't be too bad.

So that seems to be two heavyweights whose conclusions vary substantially.

Can you comment on the extent to which the impact of the poppet snapping closed tends to affect its propensity for extruding?  If not in specific terms, at least in relative terms (a lot, a little, or it depends).

[Motorhead]

The effect is a form of PEENING.  Like a steel rivet ... it will support a huge load in shear and by itself in compression much weight before it would squish.
But start pounding on a steel rivet it mushrooms quite easily extruding its shape.

What properties PEEK has is a high elastic recovery with minimal yield ( being within limits )  Acetal also has very high elastic recovery but a much lower yield threshold before the material breaks its bond to itself and distortion starts to happen.  Acetal slightly shape shifts due to this yield and peek pretty much resists shape shifting.

All materials have a YIELD threshold be it a wood board, a super ball or piece of steel.  Load em up too far the molecular bond that holds them together starts failing.

Just sharing simple thoughts w/o all the math and deeper conversation.

[JPSAXNC]

It depends on the seal and seat design, the area of the sealing surfaces, and the area of the metering ring, between the valve head and the Id. of the valve body. If the area of the metering ring is the same as the area of the transfer port or larger, the seal will hold up very well. If the area of the meter ring is smaller than the area of the transfer port, the valve will hammer closed with huge force.

[sb327]

The air in the plenum has mass as well. To stop it abruptly causes a force that needs to be accounted for.

It would be traveling quite fast.

Like water hammer in plumbing.

Me thinks.

Dave

[PikeP]

Can you comment on the extent to which the impact of the poppet snapping closed tends to affect its propensity for extruding?  If not in specific terms, at least in relative terms (a lot, a little, or it depends).

Dynamic question bruh, there are variables here you know that alter the answer to which you receive? Its fun to toss a bit of math at the problem like lloyd did but his formulations are not entirely correct, and I don't feel like correcting them.

The 3 primary variables would be pressure differential on either side of the poppet head, terminal velocity of poppet head as it contacts the seat, and the tensile strength of the area the poppet head overlaps the seat. Poppet deformation at some micro level is nearly always guaranteed in the closing of a valve in a pcp, but to a level that is perceptible or detrimental to the poppets structure is rather unlikely outside of extreme conditions. Certainly one can arrange conditions that will cause permanent deformation from said hammering effect.

To put it simply, there is more going on when the poppet is at rest with the extreme pressure differential opposed to when in motion where pressure differential and opposing forces are not at their most extreme...

Another thought experiment. Would you want your finger underneath that poppet when its potentially at 3000+ psi on one end and ATM on the other, or would you want it under when its 3000 psi vs anything higher than ATM?

[nervoustrigger]

Perhaps a simpler way I could have asked the question was “how much sealing margin is needed to reasonably assure a plastic poppet of X material having Y properties will not extrude?”.  The question arising in the other thread for which I’m seeking guidance is the extent to which the impact forces of the valve closing play a role.  But more broadly, just how miserly can we be with the sealing margin before the material is at risk of extruding in a meaningful way?
 
Meanwhile yes it can be quite useful to throw in a little math as a sanity check.  Often times it puts things into perspective enough to recognize if a variable is worthwhile to pursue further, or if its effect lies somewhere down in the noise floor.  I’m not sure what there is to be gained by being so dismissive toward the concept of using a simple calculation to estimate something.  It need not be “entirely correct” to be useful.  However if you’d like to point out that Lloyd’s example misses the mark by an order of magnitude or fails to account for an important factor, please do so.  But to simply cast shade on it does not tell us anything about how wrong it is, nor offer an alternative way to approach the question.  It’s just inimical.
 
The finger thought experiment is a fun one.  I’d have to say I’m less concerned about my finger being struck by a poppet moving at 43fps than I am having my flesh pinched against a valve seat by 7000psi.  Are those two scenarios sufficiently analogous to the stresses arising from the impact of the valve closing and those arising from the applied force of the air pressure holding the poppet against the seat?

[Motorhead]

Let me add some more info ....

You / We have a few ways to make PCP valve seats & poppet heads to use with them ... EACH creates differing stress's on Part A or Part B pending which is presumably Metal, other a plastic of some sort.

Valves such as what BSA has used for many years has the THROAT at the bottom of an @ 45* V angle.  Poppet used is a square edge cylinder with a threaded in stem.  In application the poppet head is WEDGED into this taper.   Poppet head rather than trying to be sheared along the outside of throat sealing margins, is actually put into compression trying to crush the poppet inward and stuff it down the throat.  If the material does not yield a seal is made and no extrusion.
BSA valves have proven Very very reliable offering high flow rate potential.

Valves such as DAYSTATE uses are about 180* from this.  They use a one piece Steel poppet that has an angled sealing margin on the poppet head.  They then wedge this into a PEEK seat that has a square edge on throat.   The sealing disk / seat in these rifles is much larger than poppet head with result in no way can it radially yield to the poppet head beating against it attempting to extrude the seats I.D. larger.

 IN BOTH above examples the only wear or very subtle extrusion happening is on the softer material becoming slightly flattened on the sealing margin.
BSA on the poppet heads throat side edge, Daystate the inner edge of seat I.D. against poppet.   SOLID AND FAILURE FREE !!


Only when we use 90* Seats or slight variance and similar Poppet architecture do we have a design where we are trying to place and maintain an AKA: "oversize pin over the top of an under size hole" placing all the compressive forces on a BAND of material outside of the throat diameter.  This places the material in SHEAR and just like most any over size Pin, Peg etc .... You beat on it long enough the outside diameter when fail / yield being shed away and it will go right down the hole.

As to PEEK and these type seats above ... Larger the throat greater the sealing margin required ( generally ) with pressure valve is processing also a consideration. Lower the pressure, thinner the sealing margin etc ....

On poppets with valve throats on .250" or less on pressures under 2K  have used sealing margins / diameter differential of .030 -.040" X2 using PEEK w/o issue.
That being a .250 throat having a poppet head of .310" to .320"

Beyond that I'm not sure MATH has an anwser and one must either start large & be happy or slowly reduce diameter differential until you note extrusion happening.  There are IMO far too many forces at work here .... As in a SSG equipped hammer will create MUCH MORE poppet pounding than a hammer that runs with spring preload against poppet.

Deep deep rabbit hole on this one !!!


Scott S

[rsterne]

When I talk about poppet extrusion, I am invariably talking about the head of the poppet flowing into the valve throat, like this.... This is a Delrin MRod poppet used in a 0.281" ID throat at 3000 psi, which is too much load....



Note that three things have occurred.... the seat has pounded into the poppet head, flaring the material out behind that, and in addition the Delrin has extruded into the throat.... This does not occur with this poppet if the throat is reduced to 0.266", or if the pressure is reduced to below 2700 psi, or some combination that reduces the load in a similar way.... I have never pushed a PEEK poppet far enough to see a similar failure, even in my .457 Hayabusa at 3600 psi (.500" poppet OD, 0.422" throat ID)....

The situation that Scott is describing, where the seat flattens or widens if it starts as a knife edge is a different form of distortion, caused when the loads exceed the compressive strength of the material.... If you start with a true knife edge seal (inside taper, outside taper, or a concave poppet on a flat body), then even small loads will flatten out that knife edge until the surface area of the seal is wide enough to support the loads.... The higher the compressive strength of the material, the narrower the required seat would end up, for a given force on the poppet when closed.... Hence, PEEK can use a narrower seat than Delrin, which would be in turn narrower than Teflon.... If you are really interested in how narrow a seat you can use with PEEK, in a given gun, simply make the poppet head slightly concave, running on a flat valve seat, fire a few rounds, disassemble and measure the seat margin.... Presto, you have your answer.... 

Regarding the 'hammering" of the poppet on closing, that would be a function of the closing velocity, the poppet mass, and the modulas of elasticity of the poppet/seat material.... It becomes a matter of how much mass you are stopping in how short a distance.... Compared to metal on metal (hammer on nail), however, even PEEK compresses a significant amount when the poppet hits the seat, slowing the poppet gradually to a stop.... gradually in this case being a few thou of compression.... The softer the material (lower modulas of elasticity), the further it compresses, and the LOWER the peak force during the brief time it is decelerating.... I have measured the compression on a Delrin poppet in a Disco at about 0.010" at 2000 psi, for example.... You can calculate the peak deceleration, and hence the peak force, during this part of the closing cycle (when the poppet is in contact with the seat and decelerating).... I have not done so, but I believe it to be less than the force on the poppet when closed.... If this were not the case, I think we would see most valve stems coming out of the poppet they are (usually) just pressed into, and (generally) we don't....

So, I don't have an exact answer for you, but in general terms, if you keep the compressive load on the poppet/seat material less than the compressive strength (related to but can be greater than the tensile strength), I sincerely doubt you will have any problems.... My Hayabusa .457, doing that calculation, works out to 12,500 psi, and that has not been an issue with PEEK.... Note, that if you use a tapered poppet or seat, or a concave one, if the load ends up too great, the seat will just widen slightly until a balance is reached....

Bob

[PikeP]

Another thought....you're simply asking if its more likely that poppet extrudes due to compressive failure or impact failure.

Do you think a valve is easier to open when at rest with maximum pressure differential, or when the poppet is in motion with far less pressure differential..... (hint...hammer bounce...)

IMO no need to overthink these things. Just my 2c again

[PikeP]

To jostle your mind further, you also have 2 stages of closing, or near infinitely more if you get deep enough, but the two of concern are impact closure when the seat and poppet make contact and no more air is passing, and then you also have the compression stage where the pressure differential peaks post former stage...

To touch on impact failure...failure due to impact will likely be of the fracturing nature due to micro fracture build up imo...

[rsterne]

IMO, Lloyd neglected the hammer mass (many times the poppet mass) when calculating the velocity that the poppet approaches the seat at on closing.... In a conventional valve, I don't think "impact" forces are enough to cause an issue.... That may not be the case in a valve that "blows closed", like a Cothran valve if you remove the "metering rod" from the valve stem.... In fact, that is the only circumstance where I have seen a poppet "self-destruct", and then only because it is a very soft material, and operating in a mode it was not designed for....

Bob

[mackeral5]

This has been a great read.  I do not have much to add, as most of my experience  is based on leveraging information previously  written by Bob and Scott.  Bob's writings hooked me on peek as a poppet material.  Scotts's writings gave me the understanding of how to design and tune my own Cobra valves as well as eliminating excess sealing margins and keeping poppet OD to the bare minimum.  While these are not all inclusive design parameters, these general guidelines have served me well in my tinkering.

Regarding Cothran closing forces, with the rod removed....  pushed hard enough with an excessive stem vent fighting a heavy hammer, a peek seal will literally bounce of the seat like a machine gun, with the muzzle blast scattering the chronograph  and lights across the shop.  And literally shattering  the peek sealing washer in the process.  This occurred during >250fpe testing of a .357.  Switch to delrin seal and no more machine-gunning.... no more destroyed seals, and YES you can reliably run a Cothran Powerhouse valve in a Simplified Balanced Valve configuration  by removing the rod and converting to a delrin seal,  but that is a different subject for a different thread.

The above mentioned shattered peek washer is the only time I have successfully destroyed a peek poppet.....

[JPSAXNC]

Hi Bob, When the black delrin poppet was in the prosses of flattening and extruding did you notice any positive or negative changes to the shot string? Thanks James

[Hack21]

As has been discussed, the seat geometry will affect the force that is applied to the poppet and since stress is force divided by area, it will affect the stress proportionally. Please forgive my rough sketch.  In the case of a chamfered seat, the resulting force is the axial force divided by sine of half the seat angle. The table in the lower right shows the multiplier for different seat angles.  For example, a square seat would have an angle of 180° and would have a multiplier of 1X.  For a 60° seat, the multiplier would be 2X and the force would double. I hope this is helpful.

Thanks.

[Motorhead]

YES .. this picture above is HOW the BSA valves are configured.

[Hack21]

YES .. this picture above is HOW the BSA valves are configured.

Scott,

I believe this is how Tim McMurray has been setting them up for a few decades as well.

Thanks.

[Motorhead]

Only thing I can think of too add further in this conversation based up the above 3 posts talking about a TAPERED sealing surface would be this:

I terms of Un-DISRUPTED flow of air passing around the poppet head once opened & Especially when first opening with minimal radial gap occurring is what happen to flow if or when seat RECESSION had damaged either part A or B of the sealing components.

Lets just say we're speaking of a BSA type valve ... Seat being Brass & poppet head being Acetal.   because the softer part is the poppet head the only recession is that of the sealing margin getting flattened creating wider sealing surface.  Once poppet is up off its seat the flow still travels past 2 opposed smooth surfaces.
** IF .... The Seat was the softer material and poppet head being the harder,  RECESSION would be the seat being compressed / dented where the poppet is in contact with it while closed.   This now creates a radial crater / pocket etc type shape that very much Disrupts the air flow as well tends to damp flow early in the opening cycle until sufficient lift is present where this is less an issue. Tho issue again shows up upon closing where the valves dwell will tend to become shorter due to again to flow disruption upon closing.  ( A sealing surface damaged in this way cuts total flow potential & perhaps efficiency ? )

Just some FYI ....

Scott

[rsterne]

James, as the poppet gradually extruded into the valve throat, the velocity dropped a bit.... probably due to reduced area for flow.... but it could have been causing the poppet to stick a bit and be harder for the hammer to open.... In any case, the performance did decrease a bit....

The diagram for an angled seat, and the explanation, explain why a bevelled seat can wedge and be harder to open.... Taken to the extreme (a very shallow angle, ie a deep, long taper).... it would be almost impossible to open.... but it would seal extremely well.... 

Bob

[Hack21]

Is creep a factor or is it strictly yielding from the impact load only when a poppet distorts over time and cycles?  Creep is the phenomenon where some materials may yield slowly over time from loads that are well below the measured yield strength of the material under prescribed standard tests. For example, wood creeps and this is why old houses like mine are crooked.  The floor loads in the house haven't changed signicantly from the day it was constructed, yet every year they become ever so slightly more lense shaped.  I read posts about Beeman QB Chief poppets that are found to be failed after long periods with minimal use and I wonder.

[Motorhead]

Is creep a factor or is it strictly yielding from the impact load only when a poppet distorts over time and cycles?  Creep is the phenomenon where some materials may yield slowly over time from loads that are well below the measured yield strength of the material under prescribed standard tests.

GLASS does the same ... over time.  Tho said has no attributes even remotely relatable to the materials we're discussing here.