On Sunday we conducted a series of tests. I had been planing on trying out a new type of catalyst media (molecular sieve) I sourced, however I discovered when preparing it that the permanganate doesn't actually get absorbed into the sieve, only the water in the solution (possibly). The result is that you have a material that looks purple but doesn't actually contain any permanganate. Its funny that the sieve does its job exactly, it absorbs what it was designed to (water) and not other stuff.

 As it turns out desicants can work through  absorption or adsorption, which are two different things entirely, although they both have the same result. The former is actually a chemical process (the desicant changes its chemicial structure) where the latter is purely a physical process (like a sponge) relying on surface area alone. For example when sodium hydroxide "absorbs" carbon dioxide from the air, and its chemical formula changes (not sure what to) to include an extra CO2.


The sieve actually works by absorption (chemcial one), which explains why there are different types of sieve that absorb different sized molecules. In addition, when in a solution potassium permanganate is ionic, which also significantly complicates things. I am not sure I really understand extelly why the zeolite wouldent absorb the solution because its not based on the size of the molecules, but I do know that one of the ways in which zeolite absorb is via ion exchange, so I believe it might have absorbed the potassium or the manganese depending on if it is - or +. They seem to be usually anionic, I think because they are usually used to dry hydrocarbon gasses, which means it would be absorbing the potassium, and leaving manganese. I did notice some discolouration in the media after impregnation, which could posibly be manganese, although I am not sure if you can just get manganese from a manganese ion, you would have to get something with it (manganese dioxide for example). See the bellow photos for a comparison between the clay (Above or possibly left) which is a deep purple and the bellow molecular sieve pellets (bellow or right) which are almost brown.

So I had been planing on testing some 3 and 10 angstroms sieve, but couldn't do that because I knew that the permanganate, if indeed there was any on it would just have gotten washed away resulting in a nice fountain of uncatalised peroxide.

The other thing we were testing was the new engine and that data acquisition system. It is a significant improvement over the old bi-prop engine we were using, both in the time it took to change catlyist and the smoothness of the exhaust due to the post decomposition chamber. I dident time it but I think that between the first and second tests we changed catlyist in about 15 minutes and allot of that time was taking photos.

We got some nice data from the acquisition system, although there were some issues with reliability. It is unfortunate that on the third test (the only one to be successful) the data cut out a few seconds in. I have yet to identify what the cause is but I have noticed that sometimes the readings will go to 5V and stay there. I think this may be because of communication issues between the Arduino and computer. With the lab view firmware on it, communication is at 115200 baud. I used to run it at 9600, so I am not surprised that there are communication issues. The problem is I think it needs to talk this fast in order for lab view to get all the information it needs. I really need to just use a NI DAQ. I have one I can use (its Jamie's) so I might give that a try and see if that fixes my issues. If i doesn't then the problem must be in the usb server or wifi link.

I could convert the voltage reading from the pressure transducer into pressure just from the manufactures data, but as I changed the gain on the load cell amplifier I can't convert the load cell voltage into force without putting some weight on it and taking measurements, which I haven't had time to do. Also I am not sure what units the time axis are in. They appear to be counts, but I Havant been able to find out what the sample rate of the Arduino NI firmware is.

Here is a video of the three tests:

Test 1:

Instead I retested the bentonite clay (kitty litter) I had previously done some early  experiements on but never tested in an engine. Earlier I concluded that the clay was a bad idea because it became too soft, as it adsorbed resulting in sludge, however I only ever tested it with %50. From this you would think it a good idea to test with more concentrated peroxide, but I dident want to waste the concntrated peroxide I had so I tested with %50 just so see what would happen.

The result was steam for the first few seconds then a peroxide fountain. This can be seen in test 1 of the video. I dont think the %80 would have been much better, but it would be good to find out for certain and I should have tested with the %80.

Test 2: 

For the second test, I really didn't have anything new to try in the way of catlyist so I decided to go with the old zeolite catlyist (we know it is good for at least 5 seconds) to give the chamber a workout. Unfortunatelly there was a problem with the engine and peroxide could be seen spraying from the top end of the engine, so I stopped the test a few seconds in. Upon reviewing the video (while the system was still charged) we concluded that the stream was coming from between the injector and engine and that there must be a problem with the o-ring. Indeed upon taking it apart we found that the oring was broken. On a side note I should point out that I didnt actually machine a o-ring grove in the chamber, because I was pressed for time and concluded it wasn't important. Groves are important for radial and dynamic seals, but for a static face seal with a high clamping force (much more than was needed) an o-ring isnt going to extrude the way it should so I figured that it would be aright (also I have gotten away with it in the past). Indeed the problem wasn't pressure but that the grove dident keep the oring in the correct position during tightening.   I think that it must have gotten too far to one side and when we clamped the faces together so got stretched into the sharp bolt threads which damaged it.

A friend, Ashley pointed out that for static face seals a gasket is usually used anyway which is completely true and something I cant believe I had never thought of. We were using an o-ring as a not very good gasket. For the next test I would like to make up a gasket out of a thin sheet of nitrile.

I don't really understand how the thrust which is in volts could go above 5, as this is the maximum the micro controller can read. Also the pressure scale is wrong in this plot and I forgot to label the axies.

Test 3:

This was just a repeat of test 2, but paying careful attention to the assembly to make sure the o-ring was in the correct spot. I also dident tighten the chamber in the hope that the o-ring wouldn't deform as much.

The result was much the same as the previous tests in terms of catylist lifetime, but the exhaust was much smoother and diffident seem to waggle at all. I was glad that we got a decent run in and tested the chamber. 

Unfortunatelly on test three the pressure data cut out a few seconds in (for about 300 counts) (went to 5V) after count 300, but the load data continued