From (email me there please)… This is a sub-site to main site at

This web page last updated 17 March 2019


Designs for Passively, Thermally Gated Fluid Flow Switches


Abstract / Pre-Summary

            This sub-page to is meant to describe methods for designing a passively powered (unpowered by electronics) flow-switch for the control of the flow of fluids.  The designs are kept as simple as possible.  The temperature-based discriminatory function is based on the differential expansion coefficient of different metals (or other materials), AKA “coefficient of thermal expansion”.  See .

            The variations of such a design (as described here) could be applied to many different applications of a thermally controlled fluid-flow switch.  However, the primary envisioned application is for a “transpiration cooling” of a spacecraft skin during atmospheric-re-entry.  A common example of “transpiration cooling” would be the evaporation of sweat from the skin of a human or other mammal.

             As with other sub-pages of , the intent here is to “defensively publish” propulsion-related (and “misc.”) ideas, to make them available to everyone “for free”, and to prevent “patent trolling” of (mostly) simple, basic ideas.



            To acquaint oneself with the basics, please see , where we see that two dissimilar metals can be bonded together, in a strip or in a disc.  We read that “Common bimetal materials are steel/copper and steel/brass which are welded, brazed or riveted together.”  See also for reference.  Elsewhere ( ) we learn that “…iron and carbon steel will expand at about 75% of the amount copper will at the same temperature.”  It appears that common thermostats is use today depend on differential thermal expansion between copper (or more copper in the alloy) being high-expansion, while more-steel (more iron) and less copper in the mixture or alloy, is less thermally expansive.

            (Boring details now can be skipped for a little while if desired).  The target here is to have one of our two metals (in the thermal-expansion mismatched pair) be the same as stainless steel (310S alloy to be specific) for reasons to be spelled out shortly below.  (This is especially true of 1 of the 3 designs).  However, these kinds of stainless steels have a “thermal expansion coefficient” which is roughly the same as that of copper.  So copper won’t do, for one of the metals.

            For reasons also to be clarified soon, we want high melting points in both of our metals, and we also want highly mismatched “thermal expansion coefficients”.  For reference, see for “thermal expansion coefficients” for some common engineering metals, and  for the melting points of common engineering metals.

            Now please excuse me as I will, from time to time, drop the impersonal (stilted) style of writing, and write as “me”…  This is still boring.  Skip some more if you like.  I first looked for metals that will expand significantly MORE than stainless steel…  And also have reasonably high melting points.  Ideally, the melting point will reach or exceed that of 310S stainless steel at 2,400 F.  Aluminum expands more, but melts at 1,221°F…  Not good at all.  The only other vaguely possible candidate might be “Magnesium Alloy AZ31B” at a melting point

of 1,120 to 1,170°F.  Not good enough either.  But there are your best candidates, if you want to still consider thinking along these lines.

            But what I do from here is to now go the other way, and find a mismatched “thermal expansion coefficient” in the other direction, significantly lower than that of stainless steel, which still has a high melting point.  I find “Kovar”.  is “Kovar”, Super Alloy KOVAR (UNS K94610), an iron-based alloy with nickel and cobalt…  It has melting point = 2,640°F  And an optimally way-low “thermal expansion coefficient”.  I will now use this along with the “Stainless Steel 310S” (or similar) in the designs here, as has been selected by Elon Musk and his SpaceX engineers, as we’ll see in a little while.  Spec sheets for 310S are at .

            Now back to more interesting things.  The applications for the devices here could vary greatly, but the intended target application is described best here:  

From the above…

Musk has said the nose of Starship may be exposed to temperatures of about 2,700 degrees Fahrenheit. The type of steel alloy SpaceX may use on Starship's outer skin, called 310S, melts at about 2,400 degrees.”

Now if you read the article in more detail, you’ll find that the envisioned SpaceX design is summarized by the Elon Musk quote below:

"On the windward side, what I want to do is have the first-ever regenerative heat shield. A double-walled stainless shell - like a stainless-steel sandwich," Musk said. "You flow either fuel or water in between the sandwich layer, and then you have micro-perforations on the outside - very tiny perforations - and you essentially bleed water, or you could bleed fuel, through the micro-perforations on the outside. You wouldn't see them unless you got up close."

But these tiny cooling-flow-bleeding micro-holes could become clogged with (for examples) bird poop, as the rocket awaits launching from Earth, or with wind-blown dust or sand, sitting on the surface of Mars.  So the application for the passive (non-powered) temperature-sensing flow-gating switches described here is to supplement SpaceX’s design, to add higher amounts of cooling-fluid flow where needed (where temperatures go too high, from failure of the tiny weep holes, or because of any reason, really).  The gated flow switches described here will have bigger, harder-to-clog holes.

            One can also read (at ) that water and methane can both provide clogging problems as cooling fluids.  To pull a quote from there:

            Musk has said that using methane as a coolant might be better than water.

"Rapid water vaporization can counter-intuitively cause it to snap freeze & block cooling channels," he said in a Tweet last month.

“But Engelund also sees challenges with the methane option. When exposed to high temperatures, carbon atoms in hydrocarbon fuel (like methane) can "coke" or stick together and turn solid. Such debris can then block fine structures like pores.”

          For these reasons, I would propose the use of liquid nitrogen instead.  It would have none of these above problems.  It can double up to be used for cold-gas mini-thrusters for deep-space maneuvers.  OK, so nitrogen would be hard to source locally on Mars…  But the space craft could have its nitrogen tanks topped off in Earth orbit (by another spacecraft) before re-entering Earth’s atmosphere.  So nitrogen (N2) will be shown in drawings here.  For reference, the boiling point of liquid nitrogen is -320 F.

            Note that by design, the cooling fluid (N2 here) inside the stainless steel sandwich MUST be considerably higher-pressured than the pressure of the atmosphere that is being re-entered, on the windward side, during re-entry.  Otherwise it simply won’t flow right!  However, the flow of the N2 into that “stainless steel sandwich” cavity could be turned OFF normally, and turned ON, only when needed (obviously, since we don’t want to lose the fluid when it’s not needed).  So, pressurize this cavity only at, say, 1,500 F and above, since we want to avoid melting at 2,400 F.  Keep these facts in mind for the designs below…

          So then, there will be FOUR major design-types described here; A simple one (a “spoon”), a very-very simple one (a “bar”), and a slightly more complex (but still fairly simple) one (a “shovel”), and finally a “volcano”, with variations within the four major sub-types described as well.


Simple Design; the “Spoon”

            Imagine a roughly spoon-shaped device.  It is located INSIDE the N2-filled layer, so that (when cooling is needed) the N2 (in liquid or gas form, either way) is pressurized, pushing the bowl part of the spoon against the outermost hull.  The outermost stainless steel hull has a hole in it.  This hole is much larger (and less easily clogged) than the normal, un-gated micro-holes.  The bowl part of the spoon is round, not oval as in a household spoon.  The bowl part is a disc thermostat-type switch (it bounces back and forth, with regards to which face is convex, and which face is concave, “popping” at the designed-in temperature set-point).  Let’s say we set the set-point (tripping point) at 1,800 F “panic level heat”, for example.

            This following element may or may not be needed, in reality, but the drawings will show it:  There will be a cylindrical “collar” around the hole in the outermost hull, protruding into the bowl of the spoon at cold temperatures.  This is probably needed to concentrate the force arising from differential pressure inside and outside of the outermost hull.  We don’t want inadvertent mechanical interference (contact) between the outermost hull, and the spoon-handle for example, or an undesired area of the Kovar bowl-face more likely, to cause excessive leaks.  So the spoon-face to “collar” limited-contact region concentrates the fluid-pressure-differential-derived contact force, which at least partially seals against undesired fluid leaks, until such time as way-high (“panic level”) temperatures enable more fluid flow.  This design will probably have some undesired leakage to begin with.  Also the “collar” will present a flat face to the spoon bowl, not a curved surface as exists at the stainless steel hull.  This will help prevent premature cooling-fluid leaks.

            The handle of the “spoon” is flat and rectangular, unlike a household spoon.  The tip of the handle is welded, screwed, bolted, riveted, or otherwise hard-fastened to the inside of the outermost hull.  This allows the length of the spoon handle to bend as the “thermostat switch” disc expands and contracts, and “pops”.  That is, it relieves mechanical stresses on the point where the tip of the spoon-handle is attached to the (inside of the) outermost hull.  If the handle were to be too short, these stresses would increase.  The internal cavity-pressure forces the spoon-bowl against the hull-hole (especially with a collar on the hull-hole), and minimizes fluid losses until the bowl “pops” at the temperature set-point.

            It might be possible (but seems highly implausible) to locate this whole assembly outside of the outermost hull, and have metallic spring-tension in the spoon-handle countervail against fluid loss.  This seems to be so highly implausible that it will not be further discussed or diagrammed.

            The entire spoon and spoon-handle should be made of stainless steel, with a thermal expansion rate to match that of the outermost hull.  This way, there should be minimal (thermally caused) location mismatch between the middle of the spoon-bowl and the hole in the hull.  The only part of the spoon assembly which is NOT made of this stainless steel will be on the face of the spoon-bowl which is touching the hull-hole.  This face will be of a low thermal expansion rate-type material, made of what we previously selected as being “Kovar”.  It will be fastened at multiple points (or continuously) to the stainless steel part of the spoon-bowl, by tiny rivets, screws, bolts, or by other fastening methods, or (probably best) by welding.

            When the assembly is cold, the Kovar is large and the stainless steel is small (compared to the Kovar), causing a convex bowl-face to be mated to the hull-hole.  Now we heat up.  Approaching set-point time, and at set-point time, the spoon-bowl may briefly be flat…  That is shown in the drawing as well.  But at higher temperatures, the stainless steel has expanded a lot more than the Kovar.  So the Kovar is smaller now, and the stainless steel is large (compared to the Kovar), causing a concave bowl-face to be mated to the hull-hole.

            In the middle of the stainless steel bowl-face (that touches the hull-hole), there will be a circular array of indentations scoured into the outer face of the Kovar (one design variation not diagramed here), or, better yet, for simplicity and less mass, let’s (as shown here) just drill that circular array of holes clear through the entire spoon-bowl.  At cold temperatures, these voids or holes in the spoon-bowl do nothing.  At hot temperature (in excess of the trip-point) they vent cooling fluid (N2 for example) to the outside of our rocket ship.

            For a simple design, the length of the spoon-handle should run parallel to the body of the (assumed cylindrical) spaceship.  We do NOT want to add excess complexity in having to deal with the curved shape of the outer hull, when designing the flow-switch.  Also note that this document is size-agnostic as far as the designs of these flow-gates are concerned.  They can be as large or as small as is practical, functional, and affordable.

            Drawings below (for consistency) will show fluid flow from left to right.  This follows conventions in reading a page of text, left to right, or in electronics schematics, where inputs to a page are usually on the left, and outputs are on the right.




Figure #1


            Figure #1 (above) should be fairly self-explanatory.  The same color schemes will be kept (for different materials) below, for consistency and ease of understanding.  The mounting of the handle-tip goes off-screen as being uninteresting (for now at least), below.



            Figure #2


            Figure #2 (above) gets us acquainted with the basics of this design.  The spoon-bowls bend and change which face is convex, and which is concave, just as in a disc thermostat electrical switch.  Here, however, our purpose is not switching electrical current, but rather, gating the flow of a cooling fluid (nitrogen for example).  The holes through the spoons gate the fluid flow, as the holes change their tilt.  Free flow is permitted only at way-hot temperatures.

However, it contains a “cheat”…  Notice that the spoon-handles are always straight, and their separation distance from the outer hull-wall change in the 3 different drawings for the 3 different temperatures.  A good analogy here would be to photography:  If you’re taking a photo of the moon, or even a mountain 2 miles away, you set your focus to be set to “infinity”, even though the moon and the mountain aren’t “infinity” away; they are just nominally “far” away.  So the above drawing might be semi-accurate for a nominally “long” spoon handle.  But we ALWAYS want to reduce mass in a spacecraft!

So the repeated drawing below is basically the same, stripping away the clutter of extra (not repeated) labels.  But shorter spoon handles are shown, and the “cheating” is eliminated.  An outer-hull-wall “collar” (just like the collar around the fluid-flow hole) is added to the bolt that fastens the handle-end of the spoon.  Both collars are made of the same type of stainless steel (or at the very least, they have matching rates of thermal expansion).  This keeps the pressure of the spoon-face (Kovar side) against the fluid exit hole more neutral, during temperature changes; the only changes in the bending of the spoon handle will now be the relative changes in the Kovar’s thermal expansion, and the convex-v/s-concave spoon-bowl changes.  It may be wise to keep the spoon-handle material as thin as possible (while still retaining mechanical integrity), to allow internal fluid pressure to freely bend it as much as possible, in order to NOT allow fluid flow (at mid-high temperature ranges), allowing fluid flow only at VERY hot (“panic”) temperatures.

Please recall that a bolt is shown here, but a rivet, screw, or welding (or other fastening method) could be used at the end of the spoon handle.  The important thing is that there should be a spacer, washer, “collar”, or other method to “jack up” the spoon-handle away from the outer hull wall, to roughly match the dimensions of the “collar” around the fluid-flow exit hole. 


            Figure #3


            Figure #3 (above) allows a shorter spoon handle, and does not “cheat” (it diagrams the design more honestly).  Note that on the (cold) left, we have a leak (a gap) at the bottom of the fluid exit hole.  We do NOT care, since the cavity has not yet been filled with cooling fluids.  At the middle drawing (hot, not yet panic-level-hot, and cooling fluid present), the spoon-handle is free to bend slightly (due to pressure differentials), and the Kovar face of the spoon-bowl is parallel to the exit hole lip (minimizing fluid loss).  At the far right (way hot, panic-level hot), we have a fluid-exit gap at the top of the exit hole, but this is desirable anyway.  Thus, the round array of holes through both layers of the spoon-bowl may not really be needed…  They are shown in the drawings anyway.

            This concludes the description of the first design.


Simplest Design; the Contracting Bar

            The simplest design doesn’t have a “snap point” like the first design does.  It’s continuously variable instead.  It consists of a thin (easily bent) bar of Kovar, somehow mounted (ideally spot-welded) on the inside of the outer hull.  The Kovar expands more slowly, and the outer hull (stainless steel) expands more rapidly, as the hull heats up.  So the Kovar “contracts” (relatively speaking at least) and uncovers holes in the outer hull wall, as the temperature goes up.  The bars should be aligned parallel to the body of the spaceship, to minimize problems caused by the curvature (assumed cylindrical) body of the hull.  One or two fluid-exit holes could be used, with the bolt (or spot-weld or other fastener) at one end of the bar (one hole) or in the middle of the bar (two holes).  See the drawing below…



Figure #4

            Figure #4 (above) shows (cold) one or two (collared) holes in the outer stainless steel hull-wall.  The collars are needed, as before, to concentrate pressure-induced forces that militate against undesired leakage.  They also convert the curved hull-wall surface to be flat, instead, before meeting the bar surface.  Then the single-hole version is shown “hot” and “panic level hot”.  There’s hardly anything fancy going on here…

            Parenthetically, note that we could even go to 3, 4, or 5 (starfish style) thermal expansion (Kovar) “legs” pointing out to 3, 4, or 5 hull-wall holes, radially arranged around the central bolt (or spot weld, etc.).  Curvature of the hull-wall makes this design implausible or too expensive (one could add a flat, patterned “N”-armed “landing pad” or mounting surface to the inside of the inner hull wall, to mate to our many-armed starfish).  This design is implausible, so will not be diagrammed, or mentioned again.

            Now before moving on to a simpler (“shovel”) design, please note that the style of this document has been to move from the simpler ideas (with simplifying shortcuts or “cheats” in place) to more honest or complicated representations of engineering reality.  The above figures 2-3-4 (and associated comments) refer to a “collar” or tube around the cooling-fluids exit-holes.  No mention is made (above) of how these “collars” are formed, or how they will behave (in more detail) thermally.  For more details about that, please see figures 8-9-10 below, and the associated text.  Please note that thermal-expansion details (below) for these “collars” will actually work in our favor, especially for the “spoon” design.


Simple Design; the “Shovel”

            A probably-superior (less expensive) design variation on the above is to convert the simple bar to a “shovel”.  Below, two spot-welded joints are shown…  One at the tip of the shovel handle (to serve the same purposes as have already been described, with respect to the “collar” at the mounting hole in the previous design), and one where the “handle” of the shovel meets the oval-shaped “blade” of the shovel.  The blade of the shovel has curvature to meet and mate to the curvature of the hull-wall.  Collars are eliminated.  One or two holes (with one or two shovel blades) could be used, the same as with the simple bar design.  Note that the shovel blade is curved, but the handle is not.  This facilitates bending of the shovel blade, in order for internal cooling-fluid pressure to (mostly) seal against excessive leakage, below panic levels of heat.

            The below drawing shows the “shovel” design configured at high heat levels, but short of “panic” heat levels.  That is, in the below drawing, the lowest tip of the shovel blade (tip of blade furthest away from anchor point at handle end) is just barely on the edge of opening up a flow-hole into the hull-hole).  If the hull differentially expands any more with respect to the shovel and blade, cooling fluid will flow.


Figure #5

            Figure #5 (above) shows the basic “shovel” design.  Note that the entire handle is made of Kovar, which expands (with increasing heat) a lot less than stainless steel does.  So as heat increases, relatively at least, we could say that Kovar “shrinks” compared to the stainless steel of the hull wall.

            Note that UNLIKE the “spoon” design, we WANT the handle of the shovel to expand as differentially as possible, with respect to the hull wall.  This brings us to the next refinement of the “shovel” design, which is to increase the differential (in the handle) even more.  The below shows the straight handle as being straight and LONG, to provide maximum total differential length changes over temperature changes.  Then it shows the handle as being flat serpentine shaped, to get medium total thermal differentiation per length (and so the total handle length can be shorter).  Then it shows the handle as being designed to be helical, in which case the thermal differentiation per length is maximized (and so the total handle length can be shortest of all).

            10 March 2019 update:  Note I didn’t catch this on the first-pass publication on 6 March, but think now about the following:  In the first case (on the left in the drawing below), the straight-handled shovel should be made out of Kovar, and it will expand less than the mated-to stainless steel hull.  The hole in the hull (as is already shown in the figure above) should be at the end of the shovel-blade, which is opposite the end of the handle that is spot-welded (or otherwise fastened) to the hull.  The hole at the extreme outer tip of the shovel-blade will be uncovered as the entire shovel expands less than the hull, at high temperatures.  The shovel expands LESS than the hull, differentially speaking, that is.

            However, when we move to a serpentine-shaped or helically-shaped shovel handle, the entire shovel should now no longer be made of Kovar.  It should now be made of stainless steel (or other material with a higher thermal expansion rate).  AND, the ABOVE drawing is no longer accurate, in these latter two cases, in that the hull-hole should now be located at the end of the shovel-blade which is closest to the anchor-point of the handle-tip (where it is spot-welded or otherwise fastened to the hull).  In the latter two cases (serpentine or helical handle), the hole should be uncovered by the stainless steel shovel-blade expanding-moving AWAY from both the hole and anchored handle-end, while in the former case (straight handle), the hole needs to be uncovered by the Kovar blade differentially (relatively) contracting and moving TOWARDS the anchored handle-end, at higher temperatures.

            When we add the serpentine or helical handle, we get MORE, not LESS, thermal expansion, so Kovar (less thermal expansion) no longer makes sense, for the design.  So in these latter 2 designs, we go to stainless steel, not Kovar.  So the below drawing is now corrected, for the latter 2 cases.  An additional corrective / supplemental drawing is added below that (to reduce confusion, for hole location) in the above case as well.  Here as elsewhere, please advise me (at ) if any additional drawings are needed.



Figure #6

            Figure #6 (above) shows three different handle types.  Which is best?  My best guess is that flat serpentine or helical is better than straight-flat.  A wilder guess is that serpentine is the very best, because the helical version is probably significantly more expensive to manufacture.  Note that these different handle-type choices apply to the bar and to the shovel designs, but not to the spoon design.  In the spoon design (unlike the others), we want the linear thermal expansion rate to match (or nearly match) that of the hull.


Figure #7

            Figure #7 (above, newly added 16 March 2019) makes it clear where the hull-hole should be added with respect to the shovel blade, if the shovel design has a serpentine or helical handle design.  The hull-hole location has moved from one end of the shovel blade to the other end of the shovel blade, that is, with respect to what was shown in Figure #5.  Figure #7 above shows “panic level” heat, or, cooling fluids flowing.

            The following comments probably deserve no drawings (but email me at if you really need a drawing to make sense of the following).  The handle-tip ends (in all of the designs above) are firmly fastened to the hull.  The lengths (of the bar, or spoon or shovel handles), though, are free to “slosh around” under shock and vibration stresses.  Especially if the assembly is nominally “long”, this could cause metal fatigue, and eventually, cracks in the metal, and even breakage. 

            The “fix” for that, here, would be to place along the lengths, here, 1 or more, per length, of what is variously called a conduit clamp, tube strap, one-hole strap, or two-hole strap.  These could be off-the-shelf or, more likely, custom-designed for the application here.  Such a strap could even be placed across the middle of the “shovel blade”, for example.  A sample image of a one-hole strap (clamp) is at .  A sample image of a two-hole strap (clamp) is at .


The “Volcano” Design

            The “volcano” design’s name will become justified in a little while, read on…  But first, let’s cover some basics.

            Solid metals will expand as they heat up, at rates that differ from metal to metal.  If there is a hole drilled in a large flat sheet of metal, the hole will NOT expand significantly as the metal is heated, if the surrounding metal is “constrained”.  This (“constrained”) is the case for our spacecraft hull.  In this case, the heated metal can NOT expand towards the hole or void, so the hole-size will NOT shrink much, as the temperature goes up.  One can think of molecules or crystals (grains) of metal, all trying to shoulder aside their neighbors, trying to expand towards the hole.  They can’t do it to a significant degree, because they are all “fighting each other” in all trying to do this at the same time.  If the surrounding metal is NOT constrained, then think of a ring.  The limits to the surrounding metal disc around the hole are not much bigger than the hole, so we have a ring (AKA torus or doughnut).  When heated, the entire ring (including the hole) will expand outwards, since there are fewer close-by molecules (or grains or crystals of metal) that are all cramming in together, all around.  That is, the entire ring has room to freely expand.

            If we go back here in this document and look at the spoon design (figures 2 and 3) and the bar design (figure 4), and we look at what we called the “collar” around the cooling-fluid-exit hole in the outer hull, we can think of the “collar” as a “nut” (to a bolt), or a short piece of pipe, or a sleeve…  Whatever you want to call it.  It needs to be joined to the hull.  We didn’t really talk about that, further above.  If we weld it, it will be a focal point of stress, there, under thermal cycling, where the hole in the sheet metal is “constrained”, but a short distance away, in the “collar” (sleeve) the metal is NOT constrained.  The sheet metal hull-hole hole wants to get larger (as we heat up, but the hole is NOT allowed to get bigger, because it is “constrained”), and the void in the sleeve wants to get bigger, and IS allowed to get bigger (is not constrained).  Stress point!!!  A weld there is clearly a pretty bad idea…  And we need a seal against the flow of cooling fluids.

            See for a discussion of associated matters, but we want to compress this metal at this stress point.  The below drawing can be associated with the “collars” (sleeves) as shown in figures 2-3-4, and with continued discussions of the “volcano” design below.  But what we want to do, is to drill a small hole (at normal temperatures; there’s no need for heat, here) in the hull-wall, and force a “mandrel” through the hole, using a “hydraulic puller”.  This compresses the metal around the hole, so that the metal is pre-stressed and strengthened.  In this particular application, we would want the being-compressed hull-metal to be squeezed between the mandrel and a funnel-shaped piece of tool steel or even glassy metals (AKA “sheet metal tooling”), to create a pre-stressed funnel shape in the stainless steel hull.  We can then weld a custom-shaped extension to it to create the “collar” (sleeve) in its final form.

            If this protrusion protrudes OUTWARD from the outermost hull wall, the shape (and emitted cooling fluids) is vaguely suggestive of a volcano.  So there you have it (the name justification for calling it a “volcano”).

            Such a “volcano” could go on either side of the outer hull wall…  Protruding inwards or outwards.  Protruding outwards would be more permissive of inspection and possible maintenance, so that’s what is shown here.

            Putting the volcanos on the outside would also increase turbulent flow as opposed to laminar flow, though, according to my brief research. seems to tell us that laminar flow sheds heat better than turbulent flow does, so what I am showing here is probably wrong…  We need inward-pointing volcanos, not outward-pointing ones.  But let’s just move on and show them pointing outwards, anyway.  In reality we’ll probably want a “volcano” shape, but with inverted flow!

            Do we actually want to induce turbulent flow early on, during re-entry?  Is it better to stretch out the heat-shedding process over time, to reduce peak heat levels?  I have been told that more-turbulent airflow over a rougher surface reduces air resistance (think of the dimples on a gold ball, for example), so I am thinking that turbulent airflow might get us through the uppermost reaches of the atmosphere faster?  And then thicker air will absorb the heat faster, sooner?  I’m clearly out of my areas of expertise now, so I assume it will be OK for now to just show the “volcano” in drawings, and not an “inverse volcano”…  (The latter has been added further below at figure #12).



Figure #8

            Figure #8 (above) shows, in the upper left, the mandrel-punched, “tapered” hole, and a welded-on cylinder or sleeve-extender, as has already been discussed.  Also note (again) that this most excellent idea could be applied to figures 2-3-4 (spoon and bar designs) for very good reasons, as has already been discussed.  Especially with respect to the cooling-fluid-exit hole, that is, NOT necessarily so much with respect to the hole where the “spoon” or bar is joined to the hull-wall (where a simple washer-type assembly could be bolted in).

            Then we go further, though, into the specifics of the “volcano design”.  A “coin” or disk of Kovar alloy is spot-welded into the tip of the “volcano”.  At cold or very-cold temperatures, the coin is going to be squeezed (gripped) in there very tightly (because the stainless steel contracts in the cold more than the Kovar does).  At hot temperatures (say at 1,500 F, where we might first start to pressurize the cooling-fluid layer of the “stainless steel sandwich” of the hull), everything is pretty neutral (free-floating if you will, stresses-wise, expansion-wise).  At “panic level” hot temperatures (maybe around 1,800 F as we approach the melting point of 2,400 of 310S stainless steel), the disk of Kovar will have expanded so much less than the surrounding tube (sleeve) of stainless steel, that a cooling-fluid-releasing gap will open up, releasing the pressurized cooling fluids, so as to cool the being-heated hull.

            The below drawings shows the sleeve, minus and then plus the spot-welded-in disk of Kovar, at four different temperatures, to further clarify matters.


Figure #9

            Figure #9 (above) shows “A” (upper left) as “very cold”, in the condition that the mandrel-punched hole, plus the welded-on sleeve addition, “wants to be in” before we spot-weld in the “volcano tip” with a Kovar plug-disk.  Note that the sleeve has contracted in the cold (where it can contract, after the weld).

            Then “B” shows the configuration at room temperature.  This is the temperature at which we presumably will (later on, diagrammed below) force-fit the Kovar plug in there, and spot-weld it in.  Note that the sleeve is still of a slightly smaller diameter than the mandrel-punched hole.  The sleeve is LESS shrunken down than in “A”, but still shrunken down with respect to the mandrel-punched hole.  This also implies that the Kovar plug will NOT “want to” fit in there, at this temperature!  That means that we either have to fit the Kovar plug in there (and then spot-weld it) at higher temperatures (above room temperature), or, perhaps, we can (as will be shown below) turn the Kovar “disk” into a “bullet” shape instead, to facilitate force-fitting it in there at room temperatures.

            “C” shows the configuration at “hot” temperatures, where we will presumably want to pressurize the hull-cavity with cooling fluids.  This will, for example, be in the range of 1,500 F (hull-pressurization point) but below, for example, 1,800 F “panic heat” level (which is well below 2,400 F as being the melting point of 310S stainless steel).  “C” shows the welded-on extension sleeve as being in a “natural” state, or relaxed, with respect to the mandrel-punched hole, and with respect to the Kovar plug, which we will add next.  This means that at “hot” temperatures (but below “panic level” temperatures), the Kovar plug will fit reasonably tightly in the sleeve, without leaking too much cooling fluid.

            “D” shows what happens at “panic level” (say 1,800 F) heat, and higher.  This will be larger than the Kovar plug, and hence, release cooling fluids.

            An IMPORTANT NOTE here is that, as we have made our drawings (and discussions) more detailed and honest, we can now go back and re-examine the “spoon” design, with the bowl of the spoon covering the cooling-fluid exit hole (lip of the “collar” or sleeve).  See figures 2 and 3, at the rightmost or hottest (panic level) drawings, and compare that to “D” above, in figure 9.  Now we can see that the thermal expansion (of the sleeve-lip) will work in our favor, in the “spoon” design, to allow more cooling fluid to flow through the holes in the spoon-bowl, and-or at the gap between the spoon-bowl and the sleeve-lip.  This is simply because the spoon-bowl will thermally expand LESS than the stainless steel sleeve, due to the Kovar layer of the spoon-bowl.

            Now let’s move on and add the Kovar plugs, which we now make bullet-shaped…



Figure #10

            Figure #10 (above) shows the bullet-plugs added, at the 4 different temperatures, and what distortions happen in the “collar” or sleeve, as the stainless steel material of the sleeve thermally expands more rapidly than the Kovar plugs do.  “D” drawing finally shows how the cooling fluids will only “erupt” at high temperatures, thereby further justifying calling this the “volcano” design!

            Just HOW MUCH are there differentials in constrained v/s not constrained conditions, in the tube (collar) v/s the circular-spot-welded-to-the-mandrel-punched-hole end, v/s the spot-welded-to-the-Kovar-bullet end?  I don’t know, and have not modelled it.  The proportions in the drawings are wild guesses.  I can say that the funnel-shaped mandrel-punched hole is a step in the right direction (for stress relief), and that I suspect that the circular weld there (although already stress-relieved to some degree by the funnel-shaped mandrel-punched hole) will still remain a stress point, or weak spot, in this design.  This is why I would like to describe yet another alternate to the “volcano” design, but haven’t been able to devise a plausible one yet.

            One might imagine that in deep space use, repeated thermal cycling through high and low temperatures could cause metal fatigue, and eventual cracking.  The susceptibility to such cracking might be greatest at the lip of the “collar” or sleeve, where the Kovar bullet-plug is inserted and spot-welded.  One possible solution would be to find or design a material (other than Kovar) where the thermal expansion is less highly mismatched to stainless steel than Kovar is.  Such a step would reduce the risk of thermally induced cracking, but would reduce the efficiency of the design.  That is, the “D” sub-drawing would allow less cooling fluid to flow.

            Another choice (perhaps scarcely worth mentioning, due to extra complexity) would be to drill a hole through the middle of an outer “bullet plug”, with the outer bullet plug made of a material with a thermal expansion rate half-way between that of Kovar, and that of stainless steel.  Then spot-weld a Kovar bullet-plug in the middle of the hole in the outer bullet.  A “compound” design, shall we say, with (at “panic” temperatures) an inner and an outer exit-void.  Such a design would better distribute the thermal stresses.  This design variation is here judged as being too complicated to be worth illustrating.

            A design variation here that IS judged to be worth illustrating is, instead, to partially or completely score or cut 1, 2, or 3 or so cuts or scores into the very outermost tip of the collar or sleeve.  A score would be a partial step towards a crack, where the score-design feature “encourages” nature to take its course, and crack the design at the score-point…  Whereas a simple cut might be easier to implement, and just flat-out places the “crack” where you want it.  Then, right before the point where the “bullet” starts tapering down, one drills a hole at the end of the score or cut.  This hole distributes the stresses of the crack-end, stopping the crack from propagating further.  This optional idea is diagrammed below.  Note that the cracks or scores are located away from the spot-weld, and are evenly spaced away from each other, and away from the spot-weld.



Figure #11

            Figure #11 (above) shows a highly plausible method for “stress relief” from stress caused by highly localized mismatches in thermal expansion rates in the “volcano” design.  Those of us familiar with drilling a hole (maintenance wise) at the end of a crack in the skin of an aircraft, will be very familiar with this technique, of drilling a hole (at crack-end) to stop the propagation of a crack.  This technique has been used for decades.

            In three dimensions, especially as “N” (in the drawing above) gets larger, we could be well advised to think of a flower.  At very cold temperatures, the outermost tips of the “petals” of the flower (the sections between the cuts or scores) will separate (splay outwards) more widely.  OK then, a flower opens up in the heat and sunshine, while our version here opens up the widest only in the bitterest cold…  Call it an anti-flower then!  You can call me an anti-flower if you want to!  (Ask me for more drawings if need be, at ).  Those of you (among my billions of readers) with mechanical intuition will follow what I’ve written.  The rest of you (in your mere millions) may ask for more drawings!  And your wishes will be honored!

            Well OK then, back to sobriety, what I mean is, the Kovar “heart of the flower” is hard-plugged, location-wise, in 3-D, by the spot weld.  It’s not going anywhere.  The anti-flower feature permits stress relief at very cold temperatures.  If the holes at the ends of the “N” slots or scores are optimally positioned with respect to the optimized roll-off “shoulder” of the “bullet” plug, where the bullet sides go from straight to tapered, then all should be well.  Stress relief and undesired-cooling-fluid-flow can both be balanced and obtained within engineering reason.  The precise locations, though, of these holes, will be critical!  As with much else that is left vague here, modeling and experimentation (and scale or size) will be essential to getting some version of these designs to work optimally.

            If turbulent flow is desired, the “volcanoes” may be located on the outside of the outermost hull, as diagrammed above.  If laminar flow is more desired, the “inverted volcanoes” may be located on the inside.  OR, the non-inverted volcano can be located in a depression (indentation) or hollow (a “recessed volcano” design), to make the “crater rim” flush with the normal surface of the outer hull.  These ideas may be obvious enough, but a drawing is easy to provide, so here it is:



Figure #12

            The above drawing should require no further comments.


Concluding Notes

            The “bar” design is likely to be inferior, in my humble opinion (IMHO as they say in chat rooms etc.).  The “Volcano” may be the very best.  It’s certainly very simple.  Other than that, I really don’t know…  I just want these ideas to be thrown “out there”, and looked at…  Are you listening, Elon Musk and SpaceX engineers?

            Wild, wild, wild (and major-league-different) variations of the above-described designs MIGHT possibly want to use high-temperature solid-state lubricants, fibers, or even sponges.  Or even materials with a NEGATIVE thermal expansion rate!  My discarded ideas here are so utterly wild, that I’m not even going to bother to describe them to you.  However, if you want to go off on wild (wild “IMHO”) tangents along various lines, please be advised that chalk may be a good powdered solid-state lubricant (and tolerant of temperature extremes as well), and that a high-temperature fiber is described here:

Search for “Alternatives to asbestos”…   From there:

Ceramic fibres are another commonly used asbestos alternative, and one that we make great use of at Textile Technologies for their heat resistant properties. Ceramic fibre is often reinforced with glass, although this does bring down the maximum operating temperature for the added bonus of structural stability. Without glass, ceramic fibre is often operable at temperatures as high as 1000 C or more, while with glass it still works at 500C, which is enough for many purposes.”


            Also please be advised that ceramic sponges are still expensive, and not in common use.  However…  says…

“The sponges retain their resilience even when heated to a temperature of 800 ºC (1,472 ºF).”


“High-temperature ceramic sponge” may be a good search string.  says…


So the researchers used a method called solution blow-spinning, which had been developed previously by Wu in his lab at Tsinghua. The process uses air pressure to drive a liquid solution containing ceramic material through a tiny syringe aperture.”

So this type of technology may get cheaper here…


Consider also this:  From 

 Perhaps one of the most studied materials to exhibit negative thermal expansion is zirconium tungstate (ZrW2O8). This compound contracts continuously over a temperature range of 0.3 to 1050 K (at higher temperatures the material decomposes).”


Note that 1050 K is 1,430F; and is NOT reasonably suitable for our application.  It is too low!  (IMHO).


Now finally let’s briefly speculate on the following:  WHERE on the spacecraft skin should such fluid gates (switches as described here) be located?  They may want to be concentrated towards the mid-line, diminishing in densities (of their locations) as one travels away from the airflow-splitting center-line.  This is on the theory that the entire airstream will be collectively cooled (by the gated cooling fluids from several gates at any overheated areas) as the hot gasses split at the midline.  Upstream fluid outlets will cool all areas that are downstream (but roughly aligned in the airstream, with those outlets), that is, I think, so the cooling effects of several upstream outlets will diminish the need for highly dense outlets downstream.  Let’s arbitrarily show the flow-gates as red ovals, with the lengths of the ovals aligned with the length of the conical body of the spacecraft.  Like this:



Figure #13

            The red ovals are lengthwise-aligned with the length of the spacecraft, which is accurate for what I think is desired for the spoon, bar, and shovel designs.  The volcano design is round, though, and has one other, special consideration:  Which way should the spot-weld (for holding the bullet-plug in place) be aligned, as opposed to the opposite side of this design, where the cooling fluids exit the “volcano”?  I don’t know, but I would speculate that the “volcanoes” should be oriented opposite one another in orientation, symmetrically split down the centerline of the spacecraft (where the airflow splits in two directions).  Orienting the spot-weld upstream and cooling-flow exit downstream, I think, would be best for more laminar flow, and the opposite orientation would perhaps be better for turbulent flow.


I have no special expertise or any more plausible ideas concerning any associated matters here, so I will sign off at this time.  This concludes my ideas as of this time.  Once again, comments or questions (or idea contributions) are welcomed at


Stay tuned…  Talk to me!


Back to main site at