How many orifices do I need

[Old Grey]

I was wondering how many internal orifices I really need.
I have 8 #115923's - like a SF600 -, and a FP-1 with a 40" inclined manometer and the vertical is 100".
The SF600 has 6 orifices, but after reading, they use that many because of the water gauge resolution is poor under 40% incline.

I read that people tried 1 - 600cfm orifice, but it seams to not be perfectly linear.
Was the single orifice repeatable? - slight errors are ok so long as it's repeatable (all SF benches are the same) -.

One of the problem I have with the PTS bench is that it has to be internally controlled - it's too slow for the 100 flow tests to develop a single head -, so I am going to make a wheel with single direction orifices, but with 3 working in opposite directions to the other 3. Because sharp edge orifices are more linear compared to SF600 square edge orifice I am going to make one 3mm Perspex wheel with 6 orifices, but 3 will be intake orifices (sharp edge on top and bevelled on the bottom), and 3 exh (bevelled on top sharp edge on bottom).

Most of my tests are from 300cfm@28" to 40 cfm@28" - 2V are 40@0.100", and 4V are 40@0.050" -, so that's my main accuracy point, but I will tack on more capacity for big stuff but that doesn't need to be super accurate.

Do you think 3 each way is good enough? - I'm might be able to squeeze 4 int and 3 exh on the disk -

[jfholm]

If you are using a digital manometer many of us have found all you need is one internal orifice. I did some testing on my bench and with the digital 1 orifice inside will work. It is only with using the water manometerss that I have found the need to change internal orifices.

John

[Old Grey]

If you are using a digital manometer many of us have found all you need is one internal orifice. I did some testing on my bench and with the digital 1 orifice inside will work. It is only with using the water manometerss that I have found the need to change internal orifices.

John

Are you flipping the orifice, or are you using a thin square edge plate.

Based on this post
http://www.flowbenchtech.com/forum/view ... p?f=5&t=79
I'm a bit worried about the 7% error at 100cfm

I want fast external operation and could make a 4 orifice disk - a 300cfm, 600cfm int, and 250cfm, 500cfm exh

[Brucepts]

The FP1 uses a 40" DP sensor where as the PTS DM uses a 16" there in lays the difference as to why the PTS DM can get away with one or two internal orifice plates. With a 40" sensor using a 16" DP you are only using 40% of the sensors range on a 16" sensor at 16" DP you are using 100% of the sensors range.

What does FP recommend for their DM for a range of plates? I would venture to say they have no clue

On the PTS DM, I go with a range of 0-450 (which would go from 0-450 or 150-600) for most accurate or single plate at 25-600 which will be repeatable but high-low on top and bottom by a few CFM. With tweaking it can be dialed in but it takes alot of playing.

In regards to the shared link, I am no longer supplying a 3.490 plate which would be 660@16
but rather a 625@16" plate is the top range, the DM is now all USB and the one back in 2010 was a serial model or early USB not sure what I used for that test.

What head are you working on that would require a range outside of 0-450? I don't know anyone who is testing at 25 to 600cfm on a single head? I would go back to FP and ask them what they recommend for ranges on the FP1 for accuracy with the 40" sensor.

I've never tested it with the PTS Design sorry to say, they were not willing to let me have an orifice style DM to test out. I could gladly send back my pitot one I have for an upgrade but at the time John wanted to charge me for this upgrade. I was not willing to pay him to advertise his product for free here sorry to say. Since I have one of the first FP1's and helped with the R&D during design we kinda parted ways after that and the PTS DM evolved

[Old Grey]

What head are you working on that would require a range outside of 0-450? I don't know anyone who is testing at 25 to 600cfm on a single head? I would go back to FP and ask them what they recommend for ranges on the FP1 for accuracy with the 40" sensor.

Not that I do them any more, but DCRE and Fueler stuff. But now that I'm retired injured, I don't think I will be going over 300cfm@28" - 350 tops -.

I've never tested it with the PTS Design sorry to say, they were not willing to let me have an orifice style DM to test out. I could gladly send back my pitot one I have for an upgrade but at the time John wanted to charge me for this upgrade. I was not willing to pay him to advertise his product for free here sorry to say. Since I have one of the first FP1's and helped with the R&D during design we kinda parted ways after that and the PTS DM evolved

When I bought the FP-1 back in 2007 it was the bee's knees - Audie cost double and the Perf Trends was released soon after -, so I'm stuck with the FP-1. Mine is switch-able between pitot and orifice, but that just does the math internally so it spits out the final number for either.
It seams the resolution of the 40" sensor through it's full range is the problem, barring getting info from FP or opening it up for the part number to get the specs, would testing the sensor against a home made U-tube be accurate enough.

[1960FL]

Grey,

The FP1 is a 10 bit device so the 40” sensor is only as sensitive as .039” so in its crude for NO Averaging mode it would measure .039, .078. .117 on a 3.0” internal plate that is like 24CFM steps. Now averaging fixes this but it is not the end all solution. So my guess is you will need three ranges with specific CFM limit on each range to control any errors. I/E a 3.0” ID plate can flow 386CFM at 10” DP and will flow 24CFM at .039” DP A 2.4” plate will flow 247CFM at 10” and 15.4 at .039” and a 1.75” plate will flow 131CFM at 10” and 8.2 at .039” these are just examples.

What does all this mean well the averaging is huge here if I have a 100 sample of data where each step is +/- 8.2cfm it is a lot easer to get a valid 100CFM number than it is trying to average 100 samples of +/- 24CFM numbers.

In the digital monometer world it is always better to use a smaller orifice when possible I/E if you only flow 250CFM heads then don’t use a 400CFM internal plate.


Rick

[Old Grey]

Grey,

The FP1 is a 10 bit device so the 40” sensor is only as sensitive as .039” so in its crude for NO Averaging mode it would measure .039, .078. .117 on a 3.0” internal plate that is like 24CFM steps. Now averaging fixes this but it is not the end all solution. So my guess is you will need three ranges with specific CFM limit on each range to control any errors. I/E a 3.0” ID plate can flow 386CFM at 10” DP and will flow 24CFM at .039” DP A 2.4” plate will flow 247CFM at 10” and 15.4 at .039” and a 1.75” plate will flow 131CFM at 10” and 8.2 at .039” these are just examples.

What does all this mean well the averaging is huge here if I have a 100 sample of data where each step is +/- 8.2cfm it is a lot easer to get a valid 100CFM number than it is trying to average 100 samples of +/- 24CFM numbers.

In the digital monometer world it is always better to use a smaller orifice when possible I/E if you only flow 250CFM heads then don’t use a 400CFM internal plate.


Rick

Now I'm depressed.
It seams better to scrap the whole thing, like I did in 2007, than have +/- 24CFM error.

[Brucepts]

Have you tried the flowbench with individual plates and saw calibration results?

If not I would say lets see what you find out and go from there with getting plate ranges that work with what you have?

I'm all for working with you on this . . . I really don't think it's going to be as bad as it sounds like it will

[1960FL]

Grey,

I am not saying you will have a +/- 24cfm error and sorry if my writings had you misguided. Remember I do not know all the programming in the FP1 but I know how these things work. What i am saying is how the DM works to give you the data you see on the screen. Remember the DM is going to be working electronically at a rate of say 250hz or 250 samples per second. So lets say that the DM reads 250 samples, stores them, then averages them to give you the output (crude average) in that 250 samples it reads say 100 are +24, 100 are 0 and 50 are +48 the system output would be (2400 + 0 + 2400)/250 = 19.2 so you can see the DM can extrapolate based on what it is reading and the averaging solves the problem of getting the right data out. So lets do this for the smaller plate 1.75” say it reads 100 at +8.2, 100 at 0 and 50 at +16.4 or (820 + 0 + 820)/250 = 6.5. The concept here is that the greater the number of averaged samples the more accurate our reading will be statistically, in my last case what would happen if the DM made a mistake and made one reading wrong that would be equal to +/- 8.2/250 or +/-.032 of a CFM….

I wrote this as I want you and others to understand in this CRUDE depiction of the digital world that DM does not read some infinite scale, in crude math it is the sensor length / bit sample. Example 40/1024 = .039 or 16/1024 = .015 in fact the PTS dm uses technology that provides 12 bit resolution so you are looking at 16/2048 = .0078 increments it can read. The smaller the sample the higher the resolution…

The whole reason I am going into this diatribe is that you have to be careful not to over Orifice (to big) the internal plate for what you are flowing, again if you are using an FP1 you need to set a mental limit of flow range for a plate say a max of 10” Delta P. what this means is if I put a 3.0 in internal orifice in the machine though it can flow 386 CFM at 10” it can also flow 772cfm at 40” thus if you are testing and you push the internal plate past your theoretical limit (386) you will now be testing in an (uncharted) un calibrated area and your errors will increase. I say this as in the PTS DM we use the whole 16” sensor thus if someone try’s to push past, flow stops reading and eventually they damage the sensor so they have a hard stopping point.

I think your plan to use an orifice disk is good but the execution and accuracy of the machine work will be important in getting the best calibration.

So in this I would look back at your original question and tell you 5 hole square edge plate, then figure out just how much of the sensor you want to use 10” 16” ???? no less than 10 and nor more than 20, less than 10 will hurt accuracy and more than 20 uses to much motor power to drive the flow readings.

Enough for now.

Rick

[Old Grey]

Sorry, this project gets me down because the person that was supposed to help me left me hanging with a big financial bill, hence why it got canned in 2007.

Even though this new project bears no resemblance to a SF600, I wanted to address all the problems of the SF600, like it's non linearity at higher flows. I know a guy that has a SF110, a SF600, and a SF1020(not currently running) all in one room, and directly comparing a head on the SF600 and SF110, the SF600' always reads considerably lower at high flow rates. Now this could be a velocity problem on a convoluted port - even though it also does it on fairly straight ports as well - where the air can turn corners easier at the slower air speed at 10", or more likely, the unsealed square edge SF600 disk is doing it.

I want to take the unsealed square edge disk out of the equation with an internal clamp, O-ring, and sharp edge bevelled orifices, so I would be disappointed going back to a square edge disk. I was mentally toying with a bidirectional bevelled orifice, by putting 2 square edged bevelled orifices back to back and hoping that the Vena contracta would negate the Coandă effect of the lower edge by clearing it. It would solve the square edge problem and negate the need to flip bevelled orifices, but that's just a theory.

I currently have a 450mm 3mm perspex disk and would like to fit all the single direction inlet and exhaust orifices on it, but I think I will be limited to 6 orifices in total - I will have to set them out -. Also the FP-1 only has adjustment of discharge coefficients for 3 inlet and 3 exh orifices, but that's pretty much a mute point because sharp edged orifices seam to have a pretty consistent DC's regardless of size.