Saturday 31 March 2012

A change of direction

Today Scott and I did some tests on pulsing a solenoid to vary flow rate which  gave encouraging results. Eventually I would like to work out a relationship for flow rate vs. duty cycle, but today all we did were some basic tests to determine if we could get the flow rate to change with duty cycle and by how much.

We used a frequency of 30Hz which I read worked for Armidillo, and tested 4 duty cycles between 25 and 100 percent with an initial tank pressure of 500PSI. For each duty cycle we filled the tank with 2L of water and ran three tests measuring the flow rate over 10 seconds.

Initially we got some quite inconsistent results which can be seen in this table:


Frequency (Hz) 30
Duty Cycle Water Output (g)
1st 10 s 2nd 10 s 3rd 10 s
0.25 303 224 147
0.5 299 226 276
0.75 312 369 290
1 294 146 337


When testing we could see air come out with the water which we thought would have affected the results. We have encountered this problem before but it hasn't really affected anything until now. My theory is that because the fill cart is still connected to the setup and contains air initially at the same pressure as the tank, as the pressure in the tank drops air is injected into the main 4 way cross as the pressure in the tank drops. We added a ball valve at the cross to isolate the fill cart which both eliminated the air coming out with the water and made the results much more consistent.

Results after adding in the isolation valve can be seen here:


With valve addition
Frequency (Hz) 30
Duty Cycle Water Output (g)
1st 10 s 2nd 10 s 3rd 10 s
0.25 180 174 1045
0.5 210 329 1276
0.75 186 485 454
1 413 535 488




There is still some inconsistency in the first test, but I think this can be explained by the fact that for the first few seconds after a new fill only air comes out as there is air in the lines.

We found that for the low duty cycles every now and again we got these huge flow rates that was even greater than the flow rate when the solenoid was fully on. We noticed that these coincided with a pocket of air coming out. Scott and I had a few ideas which could explain this. Firstly, for the low duty cycles we could actually hear the solenoid opening and closing as opposed to the later tests where no could could be heard from the solenoid. One theory is that the mass spring system which is oscillating inside the solenoid has its natural frequency close to our test frequency of 30Hz and that the air perturbs the system enough to make it unstable which results in the poppet travelling much more in the open direction than is possible due to the electromagnetic force alone. If this is the case we might want to change the duty cycle as a sudden increase in flow rate wont be good.

It is interesting that for the low duty cycles the solenoid was actually oscillating. I think we will want to make it so that it doesn't if we can. This will probably mean increasing the frequency.

The results look slightly  erratic but that the 2nd and third for both runs are good, and would be better if it wasn't for the crazy high flow rate of the low duty cycles.


On another note it is clear now that using a consumable catalyst isn't going to provide the lifetime I need for the hovering vehicle.

As I have previously said I would greatly prefer to buy peroxide rather than make it myself, but the problem is suppliers that will sell to individuals are few and far between. The lack of success in the consumable catlyist department has motivated me to push a little harder and I have now found a supplier and last week I placed an order for 20L of %90 rocket grade peroxide, which should arrive next week.

With this I swill be able to travel the well worn path of silver mesh catlyist. I am slightly disappointed, as I really liked the idea of a granular catlyist but we learnt a huge amount and other than the time lost, I hadn't invested a large amount of money if equipment specific to the granular catlyist.

No comments:

Post a Comment