Disk Jet DP Reloaded

[George Schmermund]

Well, the last attempt to present a coherent discussion about a DP design went south pretty quickly. I suppose that I'm partly responsible for it's race to the bottom. I shouldn't encourage group lunacy by responding to and amplifying absurdities. But after a few beers it's hard not to have some fun with the keyboard!

Anyhow, one of the main ideas that didn't come across clearly was because of my use of the term 'momentum transfer'. I'll concede that my semantics were flawed and that I should have thought more about how it would be interrupted. A better engineering description would have been ENTRAINMENT. It should have been presented something like this:

The gas molecules entering the intake of the pump would DIFFUSE into the curtain of supersonic oil vapors and become ENTRAINED and transported to the cold pump walls, whereupon the gas molecules would be ABSORBED into the rapidly condensing oil vapor before they could DIFFUSE back out of the now liquid oil.

I claim that this is the principle mechanism of transport for this pump. This also explains why a vertical mercury pump could not work successfully with horizontal jets. I hope this correction provides some clarity. Also, with a properly designed 'jet assembly' there is no need for a 'tree' of any kind!

[Jerry Biehler]

Didn't read the other thread but I see no reason why you would expect gases would be absorbed by the working fluid and then released again.

[Jon Rosenstiel]

ENTRAIN: (Courtesy Merriam-Webster)

To draw in and transport (as solid particles or gas) by the flow of a fluid.

Jon Rosenstiel

[Doug Browning]

As you have a test apparatus available, could you put in two adjacent horizontal disks, with a porous spacer between them, to see if it still pumps. (so nulling out any momentum transfer to see if "entrainment" is still operating) Your design may require specific oils that would absorb gas well if that proves to work. I suspect such an arrangement could only operate for the intake stage though; that momentum transfer will be necessary once the pressure approaches the viscous flow region. But it could make for a novel high conductance upper stage if it works, just drill lots of small radial holes in the central vapor column. Likely to also have an issue with oil vapor backstreaming if the input stage is not directed downward.

While condensed (cooled) oil vapor on the walls might absorb the gas in the fluid, it seems there is somewhat of a problem at the other end in the boiler. Boiling the oil would certainly free the gas again, but it would just travel back up the "tree" assembly to be expelled in the vapor jets. Seems like you will need a two stage boiler, one to free the gas just below the oil boiling point, then further heating in an isolated chamber to actually boil the oil again for re-use. This might be implemented as just a temperature profiled diffractionating boiler design, with a liquid only connection between the pre-heat and final boil sections.

The Edwards Diffstak series, Edwards EOx pumps, and the Veeco EP-3 do have exactly this type of diffractionating boiler setup, where the outermost (and coolest) concentric ring of the boiler is used exclusively for powering the ejector jet. Oil then flows thru a submerged bottom notch to reach the inner boiler section for the tree topside jets. Varian uses the outermost boiler ring for powering both the ejector and the bottom 2 tree jets.

[Dennis P Brown]

Have you tested the unit yet? If so, could you please post any results (if available!)?

By the way, I really like the overall design and frankly, don't really care what principle makes it work. Rather, the fact that your finished DP is such a simple and straight forward design (which looks much, much simpler than the complex cone jets use in conventional units.) For me, that feature alone really is a step up in DP pump design regardless of the principle of action! I'm with Mr. Hull - but I would like to add that for the average, low skilled machinist (me) your design is an order of magnitude better than any commercial unit since the basic design can be so easily machined (compared to the classic design DP - I don't mean that no high skills weren't used in making your's but rather the basic design is easier in principle to make.)