Small Garage Flowbench Build

[200cfm]

Stef, your bench really looks nice.

Try this. On my sensor pickups I use a common straight cooper tube like yours, very small diameter. However, to guard against any air turbulence that might try and ram into the hole at the end of the tube, I place a "rubber hat tube" over the sensor pickup. Seal off one end of the "hat tube". That way the pickup sensor can "see" only static air and never any "direct moving air."

Also, I notice you are using two separate pickup points on the top side of the orifice. Try using just one pickup sensor and use a "T" connection to separate the one common sensor port for the two manometers. This will give better balance to the two manometers that are sharing that one top side orfice chamber looking for the static pressure in that chamber.

My bench works very nicely with a common reference for that top chamber to the manometers. Good luck.

[Flash]

[stef-1]

Thanks for the further suggestions guys, looks like I've got plenty to try in a weeks time!

John, I tried Ed's spreadsheet and got the same result as you. One thing I will say is that I've had a few British Standards on orifice design (same as your ASME standards) and there is one I've been reading today - BS ISO TR 15377:1998 "Measurement of fluid flow by means of pressure differential devices - Guidelines for specification of nozzles and orifice plates beyond the scope of ISO 5167-1". There is a section in there on measuring flow between two large spaces or a pipe with very low Beta ratio. This recommends a Cd of 0.596 for this situation. Searches on here this morning shows that the PTS bench has Cd's in the same ballpark i.e. under 0.6. With that number in Ed's spreadsheet it seems to work out spot on. Funny thing is when my spreadsheet does its iteration to home in on the "true" Cd value this calcs out at 0.596 for the plate! Maybe I can have that cigar after all

[bruce]

The PTS bench has a Cd of around .585 when using the PTS DM, reason for this is the CFM sensor is actually reading a 16.31" scale and not 16" so you have to change the Cd to match using a calibration plate. Each bench build could possibly be different.

That is the interesting thing about orifice calibration you never know what numbers are actually correct ???

You need to make sure your scale is linear not just correct at one reading. It could be on at 100% but off at 40% this is why I recommend a calibration plate for 80% of the range. We have debated flowbench calibration at some lengths on the forum over the years now and there is really no set way of doing it.

At some point you have to say "Ok, I'm happy and confident with my numbers"

[jfholm]

[blaktopr]

Hey Stef, I looked over some pages to try to make sure if this was covered, but still not sure.

Your manometer bounce issue. I have that in my first bench using your means of depression control. Not a throttle body, but a 3" pvc cap. My cap has a threaded rod and screws into the bench with the radius portion into the hole. Well I have to sometimes stabilize it to contol the meter.

Now your method is more secure, but I feel there is turbulence. The way I see it (or understand yours) that it is bleeding inwards and outwards, depending on direction you are flowing, am I right?

So possibly if you were to help the pressure recovery like we try in our heads, that it may help. How about a filter on it? A belmouth on both sides?

[stef-1]

I agree that the Cd and even the flow calculation is something that you have to be happy with. The fancy calculations they have in the standards have an "uncertaincy factor" for both the Cd calculation and for the flowrate calculation!

Blaktopr - The entry from the bleed chamber to the throttle is already radiused. I could add some plastiscine to the throttle entry to form a bellmouth to test that theory. I have a filter to try as well.

I'll have to make a list of things to test when I get back so I can go through them methodically.

Well, see you all in a week

[stef-1]

Well, after getting the manometer bounce issues addressed I've gotten on with a few other odd bits to do on the bench.
At the moment I'm saving for some electrical parts to complete the motor speed control. But at the same time I have people wanting me to test parts. So in order to allow this and to get the finer control over the test pressure that I need I installed a 22mm gate valve in the spare port in the bleed chamber. This was really cheap from the hardware store and worked really well allowing really great control of test pressure.



I also wanted to have a go at making my own pitot probes, one for intake and one for exhaust. I got together some brass tube from the local hardware store, a small tube bender off ebay and some brass carburetor t-pieces. It took a couple of goes to get the j-shaped probe right but I think they worked out quite well. They're around 15-17" long.











I then added a couple more through-hull fittings to the back board which I connected to the Dwyer gauge. I did a little test and the setup seemed to work quite well.





Just need to make a test adaptor now, a valve opening fixture and finish off the spreadsheet to go with it all

[stef-1]

Did some more testing today on the orifice plates I currently have.
I'd like to post these up and see what you guys think...
Orifice being tested is 37.94mm (1.494")
Orifice in bench is 57.16mm (2.25")
Both are sharp edge type.

10" 90.22 0.585 91.8
15" 111.25 0.589 112.4
20" 128.99 0.592 129.8
25" 145.16 0.596 145.1
30" 159.85 0.599 159.0
35" 173.76 0.603 171.7

1st column is test pressure, 2nd is cfm, 3rd is the Cd of the plate and 4th is cfm calculated from Roccos spreadsheet using a Cd of 0.6. I think mine are slightly higher as my spreadsheet takes compressibility of the air into account, though I'm unsure it'll make that much of a difference ???

Going back down in test pressure I got;

35" 173.76
30" 160.75
25" 146.35
20" 129.88
15" 112.49
10" 91.17

Seems to be 1 cfm off on the way back down. I repeated the test several times and it always had that 1 cfm difference. Is this normal or should it be the same?
Guys, is this all inline with what you'd expect?

[jfholm]

Stef,
First of all, how do you determine what the Cd of the orifice is. I have looked everywhere and cannot find a way to calculate the Cd of an orifice. It always seems to be an educated guess.

On the second question you had I too have noticed that I get a slight difference when going backwards also. With this in mind I always just test going up and don't worry about it. We are doing a comparison more than anything else. If my bench reads within a couple of cfm of another persons I am happy.

I think you have done great with your bench.

John

[bruce]

Average your Cd across that range and use it for your calculations and get testing!

[stef-1]

[jfholm]

Stef,
I notice that in this calculation the weight of air is at .0751 lb. per cubic foot. That is a sea level reading. I have a friend here in Salt Lake City and he and I have been discussing if our altitude affects this as the weight of air at our altitude, 4300 ft and barometric pressures usually around 25.7 on a good day, the weight is around .069 lb if I remember correctly. On a great day in SLC with "mine shaft" conditions our honest barometric pressure may hit 25.9.

So that would change things would it not?

John

[49-1183904562]

Stef,

Since I am the one doing the FlowSoft Development I would like to look at your math if you would not mind and at minimum add a utility to FlowSoft that would perform these calculations to derive true CD. I have been working on code to allow for multiple configurations and this might just add more value and accuracy.

Thanks,

Rick

If you would not mind sending me the sheet just mail it to rick at fx-cycles dot com

[stef-1]

John and Rick, I've sent you both an email.

John, I can input ambient air conditions for the testing including barometric pressure. It was my intention to include a worksheet with my results corrected to "standard" conditions to allow better comparisons with other flow benches.

I will post screenshots as I develop the spreadsheet.