Static test of "one pounder" PVC rocket engine

*I had mistakenly called this a Bates grain in my original post.  Thanks to Seth Leigh and Al Bradley for correcting me, and to Jerry Irvine, Russel McMahon, and Serge Pipko for providing details on how a Bates grain is constructed.
Also thanks to Troy Prideaux for his juicy bit of gossip on the elusive rocket genius, Dr. Bates.  :)

New acknowledgement:  Looking back at Richard Nakka's website, I realized that I have re-invented a wheel here, in that the engine I have documented on this page is very similar to the I-class PVC engine built by Chuck Knight and described in detail at:

There are some differences in grain size and nozzle design - Mr. Knight very cleverly imbeds steel washers in a concrete nozzle to reduce erosion.  The largest difference is that I am using KN/Sucrose, Mr. Knight uses KN/Sorbitol.

For perspective, below is a photo of a small, re-loadable PVC motor I have fired 4 times with an uninhibited grain of about 40 grams of rcandy.  This is moderately large by my standards, having made mostly fireworks-type rockets burning 8 to 10 grams of this fuel.

This small engine is part of a unit I am working on for developing an ejection system.  I want it to propel a test-rocket to an altitude of about 200 feet.  That way I can fly it at home, watch its entire flight, get it back, and stay on good terms with neighbors and local authorities.  Four reloads seems to be about the maximum lifespan of this pvc engine, but it will be cheap and easy to replace.  No, I didn't drop the grain in the dirt - it looks that way because it was rolled in black powder which serves as priming.  And the thing about black powder is...well, it's black.  It is also very good priming for this fuel.  I like black powder.  I use it often.  But I am NOT a sulfur addict.  I can quit anytime I want.

But today I am working on something else.  Inspired by Foy,  I plan build an engine that burns about 1 pound of recrystallized KN/sucrose.  I have never made an engine this size before - my largest previous engine burned 100 grams of fuel.  Foy reports working on an engine with a 2 inch diameter with an uninhibited grain 8.8 inches long, which would make it fast-burning and high thrust.  Since I am taking a big step by making an engine this size, I will be conservative and use an outside-inhibited grain for a longer, less-intense burn.

The Fuel:

I have been making recrystallized KN/sucrose in 150 gram batches as per the process described at:

But the current project requires a lot more fuel, so it is time to upscale my operation.  Here I am making the recipe x 4 to yield about 600 grams of propellant.  It will be evaporated in two rectangular Pyrex baking pans, scraped up and smoothed-out in a food-processor.

I inadvertently overcooked this batch a bit.  In the photos below, notice the darker color of the candy on the left - it is what was left over after making the engine described here and it is too dark.  A 1/4 inch strand does not burn continuously in open air.  I kept relighting the test-strand with a propane torch and determined that it would take about 30 seconds for it to burn one inch, if indeed it would burn.

The batch on the right is about the right color - it burns continuously in open air at the rate of about 11 seconds per inch.  This is why I am reluctant to call this "caramel" since I wish to avoid caramelizing the sugar.  Soon I will try Ray Calkins' idea of reducing the temperature towards the end of the bake.  But I believe that the darker fuel will work OK when well-primed and burned under pressure, so will use it anyway.

The Grain:

I am not keen on re-heating this amount of rocket fuel in the house.  Fortunately, my brave little toaster-oven can go to the picnic table.  Unfortunately, it is a bit windy and sprinkling occasional rain.  So a candy cooker cover is contrived.  Good thing I am a pack rat.  I never keep anything unless I can imagine a use for it.

The fuel is heated at 200 degrees for about an hour.  It takes a long time for big blocks to get heated all the way through. When fully heated, it is the texture of soft putty, and easily formed or molded with slight pressure.  Re-heating at 200 degrees does not seem to degrade the candy, even when done for several hours, but it will eventually dry out and lose its fine texture.

So now I have everything together and can load the grain tube, salvaged from a roll of Handi-Wrap.  I am a bit ashamed to not be using a tube I have made, but this one fits perfectly, and it is available.  The candy-lumps come out of the oven looking for all in the world like flan de leche, and smelling a bit like it too.  I try not to salivate on this hygroscopic fuel.

Small hot chunks are cut off with a knife, quickly rolled into snakes, slugs or balls according to my whim, and pressed into the tube with the PVC rammer.  Sorry, didn't get a picture of the rammer in action, use of imagination is recommended.  The bottom end of the grain is pressed on a bit of black powder for priming, adding a mild sulfurous stench to the sweet candy...mmm.  The tube is loaded to within an inch or so of the end, since forming the core will displace some fuel forward.

The core is formed by ramming a 1/2 inch hardwood dowel down the length of the still-hot candy.  The dowel is pointed and has been sanded but not otherwise finished.  Hopefully I will be able to get it out - this stuff is pretty sticky.

Well it did come out. Twisting helps.  In fact, I took advantage of this opportunity to do some internal priming by sprinkling  homebrew black powder down the core, shaking it back and forth a few times, and running the coring-tool back through to press it into the candy.  I think that is why the grain looks so dark in the last shot above.

The Case:

I cut a piece of 1-1/2 inch white sch 40 pvc pipe 14 inches long.  One end-cap is drilled with a 1 inch hole and glued onto the tube.  This end is packed with bentonite clay, a heaping tablespoon at a time, packing each addition with an old hardwood chair-leg as a rammer and a dozen whacks with a 2-pound ball pein hammer.  The finished plug is 1.75 inches long, making the whole nozzle 2 inches, including the thickness of the PVC cap.

The clay plug is now drilled with a 1/2 inch bit using my funky little drill-press gizmo.  It actually works pretty well and is easy to carry, but a real drill press is on my short list of things-to-buy.  After the drill, a sharp knife is used to flare the nozzle.  Not very precise, but it does help.  I test this occasionally with my small fireworks rocket tubes by blowing through a cylindrical nozzle and a flared one - the flared one offers noticeably less resistance, thus I assume lower case pressure for the same exhaust velocity.


The nozzle is sealed by pressing two layers of very heavy aluminum foil up against it with the rammer.  A little black powder is added to the case for priming, and the grain inserted.  Then an idea:  my black powder is in large grains, waiting to be broken up.  What if I put some of these inside the core of the engine?  Being homemade, it burns slowly enough that it should not cause cato or detonation, but should ensure full ignition and a quick rise in case pressure, right?  We shall see.  I was tempted to fill it up, but resisted the urge and only put in 8 or 10 chunks.

The other end is sealed with a cap, and the engine allowed to sit for about 45 minutes so the glue will set and the grain will cool.  It is barely warm when fired.  Here I am showing it off one last time before firing.  No, there is no gravitational anomaly here, I just have the camera tilted.

Static Test:

My "test stand" is a pair of concrete blocks, a few bricks and odd scraps of wood.  The earth is my thrust sensor - quite reliable, but hard to analyze the data.  The engine is wedged into the blocks so that it can't move, the foil nozzle-seal is pierced, a fuse is inserted and lighted.

Click on photo below for .mpg movie of burn
(600k download, 3 seconds of video)


As you may have observed the ignition was quick, there was a strong thrusty sound for about 1.5 seconds, and the case did not explode.  The case is just a bit too hot to hold without gloves but not melting.  You will notice that OSHA, the Goddess of Safety has whispered something to me about safety glasses.  What I really need is a respirator, never having developed a taste for the smell of burning plastic.  And I vaguely recall that the smoke from burning PVC contains strong carcinogens - is this correct?

In reviewing the video frame-by-frame, I observe a short spurt of smoke lasting about 0.4 second, having little directional energy.  I assume this to be the black powder priming burn.  This is followed immediately by 1.64 seconds of strong directional smoke.


Opening the case, I find that the nozzle has eroded just a bit, the throat now being about 9/16ths in diameter.  There are some horizontal cracks, apparently where the layers of clay were packed, and some chips too, but since these are clean I assume them to have been made when the case was cut open.  The nozzle is coated over the whole shoulder, most of the throat, and down part of the diverging section with some kind of slag, apparently a product of fuel combustion.  There is a sizeable glob of this slag at the head-end as well.  A distinct "notch" in the pvc is observed just above the nozzle, suggesting that exhaust gas erosion was most intense there.  Perhaps a converging nozzle would reduce this, and could probably be produced by making a 30 degree point on my clay-rammer. The inhibitor tube is charred and some of it burned-away at the nozzle end, but it is largely intact.


Case:  14 inches schedule 40 white PVC pipe, "1-1/2" inch nominal inside diameter (actually slightly larger,) overall length 14.5 inches since each end cap adds 1/4 inch.

Nozzle:  1-3/4 inches dry-packed Bentonite clay, drilled to create 1/2 inch throat, last  inch or so flared approximately 15 degrees

Nozzle seal:  2 layers of very heavy aluminum foil, cut from bottom of disposable roasting pan

Fuel:  490 grams of recrystallized KN/sucrose, packed in  paper tube, 12 inches long by 1-1/2 inch inside diameter, and thus an outside-inhibited grain. 1/2 inch cylindrical core.  I figure the initial burning surface area to be about 22 square inches, and at the end, a burning surface of about 52 square inches.  Thus initial Kn is 112, final Kn 264.  Rather progressive.

Priming:  Homemade corned black powder dusted and pressed on ends of grain as well as on inside of core.  Additional large particles of this black powder inserted in core.  Total priming about 1 gram.

Loaded weight:  968 grams.

Ignition:  Six inches of thin Visco fuse.  My apologies to the purists, but fuse is so easy... I promise to develop an electrical ignition system sometime soon.

Test:  Nozzle-up static test, wedged into cement blocks to secure the engine, offer some protection in case of CATO, or provide some shrapnel in case of detonation (observers cowered safely behind the crest of a hill, but the camera was exposed.)

Instrumentation:  Four eyes, four ears, two noses, other assorted sensory organs, and a digital camera.

Burn:  0.4 seconds pre-burn, assumed to be priming, 1.64 seconds strong thrust, followed by an odd yellow wafting smoke thought to be remnants of the inhibitor tube, and a neighbor 1/4 mile away asking "What was that noise?"

I did not notice any increase in the sound or apparent thrust as the burn progressed, as would be suggested by the difference in initial burn area and final burn area.  This leads me to wonder if there is a self-mediating effect of which I am unaware, or upon closer inspection of the remains, that the paper tube leaked and thus the grain was not outside-inhibited over its full length.  The paper inhibitor-tube is burned away more at the nozzle-end, but this could be from the increased exhaust flow in that area.

Autopsy:  The PVC case is intact, the clay nozzle slighty eroded but not significantly, and coated in slag which may have preserved the clay by serving as an ablative coating.  There are deep etchings in the PVC just ahead of the nozzle and at the head-end plug, which might be of concern if the burn were longer or more intense.


Without instrumentation, I have little hard data.  But the fact that this engine burned over a pound of KN/sucrose in 1.6 seconds suggests that it has potential to propel a modest-sized rocket to some altitude.  Without a substantial math background, I do not know how to calculate what this altitude would be.  Would anyone on the Arocket list be willing to hazard a guess or show me to the formulae?

I hope to add some instrumentation to my testing in the future, and one day be able to launch some things upwards!

Questions, comments, and criticism of this work are encouraged.

Respectfully Submitted,
Jimmy Yawn
rev. 02-22-02