Fabricating velocity probes

[jsa]

[86rocco1]

[slracer]

It's been a couple of weeks and it seems everything I touched has not worked or broken so I wasn't too interested in building a prototype. As John said, the proof is in the pudding so I went ahead and started one today. (As proof of the above snake bite period, I didn't get the post a pic on my last post! The block showed up here so I will try to edit that post later.) More proof! :p When I was sizing the tubes, I forgot that the adjacent sizes "nest". Therefore, a 1/8th outer tube is needed with a 1/16th inner tube. Of course, I didn't order any 1/8th but found them at the "local" hobbyshop (only 65 miles away). After I got home, I remembered that I also needed some 5/32nd size, that is why I didn't finish the prototype today. I did buy the DuBro #785 tubing bender that John mentioned and I also got the K&S spring style bender set. The DuBro works great and bent a 1/8th tube into a 180* turn without kinking with an outside dimension of 1/2 inch. As John said, take it slow.

Try to follow the pic below along with the description as I might get it complicated. I was able to hold the 1/16th tube tight in the die with a finger and get a .365 outer dimension for a 180* turn. Not quite to my desired size, but good enough for a first try. (I haven't tried the K&S bender yet.) The hole in the 1/16th tube is a #65 drill (.0345). I hand filed and sanded the tip to a hemisphere looking shape and chamfered the inner edge also. Then I moved off to the 1/8th tube. Using a 3/32nd tube as a spacer, I decided where the static holes would go and drilled 4 #68 (.0310) holes at approximately 90*. I was able to use my Dremel for these holes as I have a 3 jaw chuck attachment for the tip. I used a #54 drill and hand spun it to deburr the holes. The inner dia of this tube just fit a #41 drill so, about 1-1/2 inches from the other end of the tube, I made a #41 hole to "vent" the static air. I assembled this much using a short piece of 3/32nd to center the 1/16th in the 1/8th and soldered the 3 together . You can see part of the 3/32 sticking out of the 1/8th in the pic. Then came the first design change! I started to do the same to the cut tubes at the other end when it was apparent that the 1/8th could have extended to the end of the 1/16th and been sealed the same way. This would have provided a 1/8th OD for the flex tube to attach to. I will cut this section off for the same result, but will lose about an inch of length. When I get the 5/32nd tube I will be able to "tee" off the 1/8th for the flex attachment there, but I'm not ready for another 130 mile drive just now. So far, no problems with construction and if the "tee" arrangement works, I'll be finished in no time! Yeah, right! ??? Then, of course, it will be test time! Everybody got a spoon for the pudding?

Doug

BTW - The straight tube shows how the 3 sizes nest. I added a black band at the end of 2 so you could see where the joint is.

[slracer]

Here's the latest on the new design front! As shown in the pic below, I have finished the prototype construction. I got the 5/32 tubing and built a "T" arrangement for the static pressure tap. I made a rather "lazy" S shape of the larger size and drilled through one of the corners. Then I drilled a hole in the side of the 1/8 tube and slipped the 1/8 through the hole in the 5/32 until the diagonal lined up with the hole in the 1/8. I then soldered the two together. The 3/4 inch overlap provides the strength for the joint and the solder provides only the seal. At the open end of the 5/32, I inserted a short piece of 1/8 and soldered it into place so a 1/8 ID flex hose could be attached. I also slipped a 1/8 with 3/32 spacers over the exposed 1/16 (pitot) tube at the open end so the same hose size fits there nicely. In the previous post, I said I was going to cut it off but the added piece saved the overall length. I then blew some air through to verify the passages are open. This model is ready for test!

Also shown in the pic is the start of the second version. My original goal was to get about a 1/4 inch width. With the added width of the 1/8 tube, the overall width of the original is about 0.396 in. The limiting dimension for construction was the diameter of the mandrel used to bend the tubing. I made a new mandrel and bender and bent a new pitot tube. The 1/16th tubing is easy to bend as I suspected. This new "core" is just under 0.290 outside the bend and with the 1/8 width added, the total will be about 0.321 inches. I also reduced the length of the 3/32 spacers so the static holes could be closer to the tip and the length of the unsupported 1/16 tube could be shortened. As all the parts have been cut and fitted, this version will be finished tomorrow (Sunday). I may try to make a tighter radius (a .157 mandrel) and shoot for the 1/4 inch outside dimension but I think the latest version will be sufficient to test. I'll keep you up to date. -- Doug

[jfholm]

Doug,

It really looks nice. fine craftsmanship there.

John

[86rocco1]

Definitely fine craftsmanship. I'm still not convinced the the basic design concept sound but it looks like you're pretty much ready to test it. What I think will happen is that at areas in the cross section of port where there's are large variations in velocity, you will be able to generate different velocity reading without moving the tip simply by changing the location of the static port. That said, I very much hope I'm proven wrong because this probe is pretty, nice work!

[slracer]

John and Ed, Thanks for the comments! I hope looking good helps it work!

Ed, jsa, (and others), I've been thinking about your concerns and trying to relate them to a Pitot bench flow element. In my bench, the element is in a 1.57" ID tube (which is a small size for most benches). The static pickup is located 1/2" in front of the pitot pickup and flush with the internal wall as suggested here on the forum and this location is also in line with the Dwyer installation recommendations for static pickups. That locates it an "enormous" (relative) distance from the pitot pickup holes in one of Bruce's averaging pitot tubes (which I use). In trying to reconcile how this can work (and the probe not work), I believe the difference is that the purpose of the flow element is to find the AVERAGE velocity in the tube while the the probe is looking for the LOCAL velocities (which seems to be YOUR concern when I reread the posts). With that in mind, I hope the small distance from my pitot to static pickup points minimizes any errors (and suggests that I use only 1 static pickup located on the inside of the bend, nearest the pitot). However, it also says that a "standard" setup has a built-in error due to the diameter of the probe and the location of the static pickups in an area which has been disturbed by the perturbation of the streamline being measured. (A 1/8" probe in a 1" port like mine would change the velocity 1-1/2% or 5 fps at 300 fps tip speed). It also points out how critical the positioning of the tip might be (for either design) as some of the static pickup holes will be in an area of separated flow if the angle relative to the streamline is too great. It would be interesting to see what those results would look like. (Yeah Bruce, I want more data! :p )

All of this theory is interesting, but when I finish the second probe (should be today), I hope to get a volunteer to test it relative to a standard model and report the results. Since I have no experience doing a standard test, nor do I have a standard probe, I would like someone else to do the testing. It should include rotating a standard probe holding the tip fixed and changing the angle of the probe relative to the flow to see how much the results there change as well as "wringing out" the new design. Anyone interested?

While we are discussing it, is there a "primer" on using a velocity probe? A short (or not so short) list of Do's & Don'ts would really be useful for all of us new guys!

Another rambling dissertation from Doug complete! :p

[86rocco1]

I certainly won't disagree with you concerning the short comings of the conventional Pitot tube, I think this design may eliminate one set of problems but create a whole new set. I'm looking forward to seeing some test results. BTW, do you have a conventional Pitot tube you can use for comparison?

[slracer]

Hi Ed, No, I do not have a conventional pitot which is one reason I would really like someone else to test. I have also never done a velocity probe and could mess up the test so that any results would not be worthwhile. I hope someone is interested enough to give it a go! -- Doug

[Dave W]

[86rocco1]

Normal solder either the plumbing type or the electrical type should work GREAT on brass just make sure the brass is clean and use a little flux.

[jfholm]

I see the argument for getting errors with either design on the pitot tubes. When I built mine it was stated that the static holes should be 300% of the tube diameter from the tip. So mine are 3/8" from the tip.

After seeing the velocity changes in the port in such small distances which one is going to be better? Conventional or Doug's. Either design places the static holes in an area where the velocity could be much different than the pressure tip.

And on the solder I used some solid core that was very small in diameter I got from Radio Shack and then as Ed mentioned use some flux. Heat the brass tube first then put a little flux on it and then heat again and add the solder and it should suck right in.

John

[slracer]

Dave, As Ed says, just use a little flux and be sure the brass is clean. I lightly sanded the surface with 400 grit wet & dry to clean it. Then I used a paste flux from Home Depot (tinning flux I think it was called - I'll check tomorrow) and a Radio Shack electrical solder. -- Doug

It was called Tinning Flux and came from the Plumbing Dept as it is used on copper and brass pipes.

[slracer]

I did complete the 2nd version of the probe last Sunday, but it has taken me this long to get a pic and get it ready for here. As you can see by the comparison to the penny, the tip is really small. The other end also got a minor mod in that the hoses will end in different places. Because these are not barbed fittings, clamps of some kind may be required. The offset allows the clamps to also be offset so there is less chance of interference. Another "plan ahead" feature. :p

Still looking for someone to test this one. -- Doug

[jfholm]

[color=#000000]Doug,
I would be happy to test for you.

Also try this. Put your probe in the port with the pressure tip at a location. It would be easier to do it on the floor, wall or roof where you can hold it steady. Hold the probe so the static ports are in the 9 o'clock position and take a reading. Then rotate the probe 90