Page 7 of 9

Re: Modified EZ-Flow bench problem

Posted: Sat Dec 19, 2015 7:48 am
by SWR
I have now hooked the 2 chambers in my big-chambered bench up as a settling chamber before the orifice. Test pressure taken from the chamber. Flow orifice low and high pressures taken from the corner taps of the orifice.

Using a 60.0mm / 2,3622" test orifice in a 3 feet long 4.92" pipe fitted with a bellmouth, dumping into the combined volume of the settling chambers - no center orifice, that's 14 cubic feet of resting volume...- and using the correct temp and viscosity index and all that, it should - due to the pipe wall effect - flow 237.36 cfm. The manometer still says 272 cfm using the 305 cfm test orifice. 5 cfm leak at 70".

That's 1.5 cfm better than the "just pipes" bench. And I tried swapping the test pressure and the manometer low side. No difference. Unless 0.1% is a difference.. :(

I am sure better men have gone insane over things easier than this...

Re: Modified EZ-Flow bench problem

Posted: Sat Dec 19, 2015 10:18 am
by larrycavan
SWR wrote:I have now hooked the 2 chambers in my big-chambered bench up as a settling chamber before the orifice. Test pressure taken from the chamber. Flow orifice low and high pressures taken from the corner taps of the orifice.

Using a 60.0mm / 2,3622" test orifice in a 3 feet long 4.92" pipe fitted with a bellmouth, dumping into the combined volume of the settling chambers - no center orifice, that's 14 cubic feet of resting volume...- and using the correct temp and viscosity index and all that, it should - due to the pipe wall effect - flow 237.36 cfm. The manometer still says 272 cfm using the 305 cfm test orifice. 5 cfm leak at 70".

That's 1.5 cfm better than the "just pipes" bench. And I tried swapping the test pressure and the manometer low side. No difference. Unless 0.1% is a difference.. :(

I am sure better men have gone insane over things easier than this...
Can't accurately visualize what you've described as the setup.
I don't know what formulas you are using.

All I have to offer at this point that might help you is to revisit your formula for calculating CFM. IF it uses a coefficient of discharge value, then adjust that value until you get the correct reading.

The bench you have built with the pipes should follow HVAC measuring rules. At least I feel it should. Here's a video that might help.

https://www.youtube.com/watch?v=YmKhE1d1Wws

Also this http://www.efunda.com/formulae/fluids/c ... wmeter.cfm

Re: Modified EZ-Flow bench problem

Posted: Sat Dec 19, 2015 6:39 pm
by SWR
They say a picture says more than a thousand words... so Paint to the rescue.
Testbench layout small.jpg
This is a schematic of how it is now. The pipe lengths are a bit longer before and after the orifices, but it is the general layout.

Re: Modified EZ-Flow bench problem

Posted: Sat Dec 19, 2015 7:14 pm
by SWR
larrycavan wrote:I don't know what formulas you are using.

All I have to offer at this point that might help you is to revisit your formula for calculating CFM. IF it uses a coefficient of discharge value, then adjust that value until you get the correct reading.


Larry, I do not use a formula. I use a high-end HVAC airflow calculation program that is very similar to, but more extensive, than the efunda calc you linked below.
The bench you have built with the pipes should follow HVAC measuring rules. At least I feel it should. Here's a video that might help.

https://www.youtube.com/watch?v=YmKhE1d1Wws

Also this http://www.efunda.com/formulae/fluids/c ... wmeter.cfm


The online calculator says 239 cfm for my 60mm / 2.362" test orifice vs my programs value of 237.39 (due to viscosity values changing with temp). The coefficient is 0.619 for this orifice, pressure loss of 7.552".

Re: Modified EZ-Flow bench problem

Posted: Mon Dec 21, 2015 10:53 am
by larrycavan
I guess my point went misunderstood so I'l explain it.

With an orifice you have a known fixed value, that being the diameter. THAT is THE ONLY KNOWN & PHYSICALLY MEASURABLE VALUE.

Coefficient of Discharge [Cd] is a theoretical value. It is the only adjustable value assigned to the orifice itself. The only other adjustable value is the Dp across the orifice.

Cd is not a steady state value. It can and does change with changes in DP. That is why you calibrate at a specific pressure against an orifice with a reliably rated value at that specific pressure.

That all aside, what I suggested was to adjust your Cd value until you get the correct CFM Reading. It doesn't matter what the Cd is. What matters is that your formula arrives at the correct value for CFM.

So, let's say that you had to adjust your Cd to .55 to get the correct CFM reading because at .61 you are reading high. There is absolutely nothing wrong with that approach. Who cares what the value is? What you care is that you get the correct CFM value.

Sounds a little off the beaten path perhaps? Then again, you chose to follow a path that is different than a traditional orifice bench. The map coordinates may well vary to arrive at the same destination. Who cares if you went over the hills and through the valley of death instead of jumping on the Interstate and getting there fast and easy??? Getting there is what matters and unless you can account for the peculiar behavior of your Orifice In A Pipe design system by some other explanation [which so far has defied explanation], give that approach a try and see what happens.

Use that approach with more than one orifice. See if a trend exists. I wish you Good Luck :)

Re: Modified EZ-Flow bench problem

Posted: Tue Dec 22, 2015 5:57 am
by SWR
The "problem" is that I cannot adjust any Cd value. It is calculated by the program to be a correct value in regards to the Reynolds number for that size orifice and pipe/chamber dimensions. So I did a little test, seen below. Cd for the chamber bench is 0.597, pressure loss of 9.758".

My bench may be strolling through the valley of death, but if you keep walking, you get out at the other end. You guys know normal orifice benches, so I did a "normal orifice bench test" against any layout I could imagine with my stuff. Except flipping the flow orifice in the chamber upside down to get even more still air..
Test modes small.jpg
These are the results.

Who says orifice in a pipe would always differ a lot from a big chamber bench? Not if you have the right flow calculation program.. but it shows too much. Consistently. I still need to think, any pointers welcome.

Or maybe just change the inclined to show what it really should at 10"..

Re: Modified EZ-Flow bench problem

Posted: Tue Dec 22, 2015 8:02 am
by Brucepts
SWR wrote:Who says orifice in a pipe would always differ a lot from a big chamber bench? Not if you have the right flow calculation program.. but it shows too much. Consistently. I still need to think, any pointers welcome.

Or maybe just change the inclined to show what it really should at 10"..
Recommended procedure for calibrating a DIY incline manometer is change the rise to match your plate cfm, so it could rise 10.125 or 10.25 or what ever rise it works out to be. You are never going to get exact numbers in my opinion, to many variables in build quality compared to the "ideal" of the math.
Consistently. I still need to think, any pointers welcome.
I'm biased . . . build a two chamber flowbench with the same size chamber before and after the orifice plate, get away from anything that has to do with an orifice in a pipe so you don't have to go through what you are with the math and still come out with different numbers.

But, it all comes down to what you want to do. If you like the challenge of trying to figure out "orifice in a pipe" keep moving through the Valley of Death until you come out the other side. Personally, I look for the short cuts out of the valley :D

Re: Modified EZ-Flow bench problem

Posted: Tue Dec 22, 2015 11:08 am
by SWR
Brucepts wrote:Recommended procedure for calibrating a DIY incline manometer is change the rise to match your plate cfm, so it could rise 10.125 or 10.25 or what ever rise it works out to be. You are never going to get exact numbers in my opinion, to many variables in build quality compared to the "ideal" of the math.
12.7", then..
I'm biased . . . build a two chamber flowbench with the same size chamber before and after the orifice plate, get away from anything that has to do with an orifice in a pipe so you don't have to go through what you are with the math and still come out with different numbers.

But, it all comes down to what you want to do. If you like the challenge of trying to figure out "orifice in a pipe" keep moving through the Valley of Death until you come out the other side. Personally, I look for the short cuts out of the valley :D
The math is super easy, I'm not struggling with that. The pc does all that, likewise for any chamber bench. I'll make a 2 chambered box with identical size chambers, orifice in center, baffle, etc.. and then see how much it differs. I guess it will show the same numbers.. we'll see. :D

Re: Modified EZ-Flow bench problem

Posted: Thu Dec 24, 2015 8:55 am
by 1960FL
SWR,

I have some other info on pressure loss across the orifice https://en.wikipedia.org/wiki/Orifice_plate look at section on pressure loss, I am researching, I might recommend a hybrid you are using an application recommended for HVAC there margin of error is big, not to discredit your application I am just not sure it is best for what you are doing. I suggest in the testing phase to put the standard Orifice in a pipe formula in Excel I can help if you need it. This will allow you to focus on Compressible Gas calculations and as Larry suggested have the ability to calibrate by moving CD up or down. Remember you application most likely uses formula's based on Harris/Gallagher and thus incorporate Reynalds number in which this too is a guess. There is no actual given CD every engineering document I have seen will provide a suggest starting point chart for edge types (square, round sharp, angled). For true Sharp edge I use .620.

By using the calculation for Compressible Gas we bring in the aspect of Ratiometic to the bench which helps simplify unless you truly are taking all the needed information into account on both sides of the orifice including a wet bulb reading on the test room and time facility. All this is way over kill for what we do.

SWR what we need is solid calculation to get you calibrated. I/E 25 cfm to 275 cfm on a 305cfm internal orifice.

Rick

Re: Modified EZ-Flow bench problem

Posted: Thu Dec 24, 2015 9:29 am
by SWR
1960FL wrote:SWR,

I have some other info on pressure loss across the orifice https://en.wikipedia.org/wiki/Orifice_plate look at section on pressure loss, I am researching, I might recommend a hybrid you are using an application recommended for HVAC there margin of error is big, not to discredit your application I am just not sure it is best for what you are doing. I suggest in the testing phase to put the standard Orifice in a pipe formula in Excel I can help if you need it. This will allow you to focus on Compressible Gas calculations and as Larry suggested have the ability to calibrate by moving CD up or down. Remember you application most likely uses formula's based on Harris/Gallagher and thus incorporate Reynalds number in which this too is a guess. There is no actual given CD every engineering document I have seen will provide a suggest starting point chart for edge types (square, round sharp, angled). For true Sharp edge I use .620.

By using the calculation for Compressible Gas we bring in the aspect of Ratiometic to the bench which helps simplify unless you truly are taking all the needed information into account on both sides of the orifice including a wet bulb reading on the test room and time facility. All this is way over kill for what we do.

SWR what we need is solid calculation to get you calibrated. I/E 25 cfm to 275 cfm on a 305cfm internal orifice.

Rick
Rick,

Thanks for that. I do however doubt the maths used have much margin for error, if it was not fairly accurate, why describe the answer to the third decimal? Maybe I'm biased because it is a $1.000+ program, though. I did use the Excel spreadsheet and it did say more or less identical numbers to my program given that I made the Cd in Excel match what the program said it should be. It is an ISO 5167 certified calculation.

For ease of finding answers, I'll calculate using the Excel program so we arrive at the same values. I did ask in a different thread how big a baffeled chamber should be to get good correlation to, say, a PTS bench? Does a 27" x 27" x 27" box either side of the orifice do the job, if I have a big baffle over the orifice and no direct airflow? I'll make one to see how far off the results will be from what I see now. :)