How to get accurate/comparable cfm numbers

[bpowell]

[color=#000000]The bench is up and running giving consistent numbers, but my flow numbers are higher than they should be...I'm in need of help with my calculations...I think ???

The orifice plate has (8) positions and is 0.065" thick, the holes were cut with wire EDM, but the finish is not perfect. This may be due to the material having a galvanized finish. I calibrated each orifice by using 100, 200 and 300 CFM calibration plates from bruce, and my intake/exhaust CF

[larrycavan]

Oh....I"ve been there and I'll tell you this. Their flow calculations are off. On some holes....waaaaaaaay off.

Forget their flow values. I spoke to John Clark back in 94 and he told me himself that the magazine article contained misprints and he had worked some of the holes with a die grinder.

That disk will leak on exhaust as you're already seeing. It's not uniform either so that makes exhaust calibration next to impossible.

If you can get to the disk, rip it out and go with maybe 4 ranges insead of 8 or 9. Go with rubber stoppers instead of the rotating disk.

There's a ton of info on that to be found on this site. We've been over that very, very thoroughly. Use the search feature or google.

Get the spreadsheet from the spreadsheet section to figure your ranges and Cd values.

I'm more than willing to help you get that thing working properly but you really can find everything you need in the archives to get the design correct. Do that first, then holler back.

Good news is....you came to the right place to resolve your problem!

Forgot to add....You'll need a couple of calibration plates. You need to decide on your ranges before you can decide on what you need for calibration plates.

[bpowell]

Hello Larry, I also heard some of the information in the article was off...wish we would have known this before we built it. The exhaust use to leak quite a bit more than it does now, the orifice plate is now spring loaded which reduced exhaust leakage by 18% (still way off I know).

I have searched and read quite a bit on this site but saw conflicts on how to calculate the leakage. Could you point me to the correct information or re-state it here? This has been quite trying over the last 5 months so any help is greatly appreciated

[LEVAS VHF]

[color=#000000]In English:
Hello, I am Victor of Argentina, I made the FLOWBENCH of

[Tony]

Welcome to the Forum LEVAS.

We have other posters here for which English is not their first language. It can sometimes be a little difficult, but we can still communicate.



bpowell,

There are several rather fundamental problems with the original design that unfortunately you are just now discovering. Leakage is one, the orifice flow coefficients changes when the airflow direction is reversed, but the biggest problem by far, is that the measurement orifice will be flowing highly disturbed turbulent air from the test hole.

It is unfortunate that the measurement orifice has been placed directly below, and directly in line with the test hole. Any swirls, eddies or turbulence from the test hole will effect the orifice flow coefficient in an unpredictable way. There is no quick fix either, but as Larry says you have definitely come to the right place.

Others here have had similar problems and there are ways to improve the design, but I will leave that to people with more direct personal experience with your style of bench.

But in a nutshell, an orifice flow coefficient is only predictable if the air upstream flowing into the orifice is completely undisturbed. Very turbulent air can either increase or hinder flow, and the resulting pressure drop across the measurement orifice can become highly unpredictable and inconsistent.

The usual symptom of the problem is that the measured flow increments have no obvious consistent relationship to orifice size increments.

[MMack]

I was fortunate to have Bruce guide me to this site prior to going very far in the construction of my bench. From the work of others, I added a box to the top of the bench that is one 2X4 tall, and a deflector under the inlet hole. I have about 1 3/4" from the under side of the top plate to the top of the deflector. The deflector is 1/8" aluminum. The top pressure tap for the depression and the low side of the incline manometer is about 1" above the deflector in the middle of the inlet hole. I have changable orifice plates that I access through deck plates I bought at Boaters World. I made my own plates to start with, ultimately I will buy some from Bruce, but can't do it yet. I started the bench with a 70 CFM plate in the bench and put a 41 CFM plate on the inlet. I read 61% on the incline manometer, or 42.7 CFM. I have not been able to spend much more time on the bench this week, but I am very happy with what I have seen so far.

Thanks to everyone on this site for your help and ideas!

Mike

[Thomas Vaught]

[color=#000000]Hola Acepto Victor,

Your English will be better than my Espa

[bpowell]

[color=#000000][quote="bpowell"]Leakage Calculation A
Leakage was factored in by taking the leakage say 7.5% on the intake times the cfm rating for orifice 1 (28cfm) to get a cfm leakage of 2.52 cfm. This 2.54 cfm leakage for the intake is then subtracted from the result cfm (regardless of orifice used) then multiplied by the CF. Here's an example:
Flow at .100

[larrycavan]

What is that big bolt / stud, on the top right of the sealing ring?

[bpowell]

Larry, that's a spring loaded ball bearing to enable easy alignment of the orifice plate position over the fixed "flow hole" (under the orifice plate). Each orifice position has a corresponding dimple relative to the hole, so you can feel (through the orifice handle) when the spring loaded ball drops into the dimple, then you know the selected orifice is perfectly centered with the "flow hole". The internal spring can be adjusted by turning the set screw in the center of the bolt, you can just see the hex end.

I'm thinking calculation A is correct...thoughts? ???

[Thomas Vaught]

Leakage should be the same unless the orifice plate is not flat. Calculation A sounds right to me.

Tom V.

[bpowell]

Thanks Tom!

Quick thought, I've got lots of data, and need help interpreting it. Anybody use DOE (Design of Experiments) to evaluate variables and their influence on other variables? I did these years ago but can't find my book... Anyone have an Excel spreadsheet for a basic DOE experiment or other reference? Thanks.