Choose the best design...Your opinion needed

[smooth blend]

[color=#000000]Which design do you think its best from an accurate flow standpoint. I

[larrycavan]

This is a rather long post. I apologize for that but getting concepts across takes a bit of explanation.

I believe any of the 3 design concepts will work. Design "C" is pretty close to what I currently have.

As it has been stated by other forum members, [I agree completely] trying to achieve identical pressure drops across the calibration plate and the flowdisk, proves to be the difficult part. What that would achieve is verification of the formula for calculated flow 'IF' the actual Cd for the orifice in question were known to begin with.

I don't feel it plays as much into a successcull design as one might think it does. That's not to dismiss the idea in any way but rather to point out that you can make the MSD design work with some design changes that are easy to make.

Here's my supporting data on that statement.

During calibratin of my bench I found a calibration plate that I had forgotten I even had. It was tucked away inside a folder where I had kept all my flow sheets from several years ago.

The orifice is a 1.24" and was punched from the same material my flow disk was made from. It was punched with the same punch. It should be pretty close to the actual hole in my flow disk in terms of CFM capicity and Cd.

I had flowd it on a SF110 in both the 105 and 185 ranges at 6" of test pressure. I've noticed that flowing on this SF110 you'll get a slightly larger number if you use a higher range. They don't correlate 1 to 1, exactly.

So the flow for the SF105 range worked out to be 54.6CFM @6".

Using the formulas from the forum, the Cd for the plate proved to be lower than it's twin on the flowdisk. No surprise really because the conditions for flow in the top plenium are really not the same as the conditions seen by the top side of the calibration plate.

Once I had settled on what the ranges actually were vs what the formulas said they could be, I grabbed a couple of 34mm carbs, a KZ head and a ZX900 head off the shelf and flow tested them after I had calibrated my bench. I compared my flow tests with the flow sheets I had from those same items after flowing them on the SF110 and everything lined up quite nicely. Each item flowed to within 3-5 CFM maximum difference.

To me, once you're in that ballpark, it's not important that your numbers match identical to another flowbench.

The key thing I think that needs to be relized here is that the formulas found in this forum are absolutely crucial to calibrating. With them, you can figure out everything you need to calibrate. Without them, you're right back at ground zero. Having a couple of items that you have already flowed on a SF makes the job that much easier.

No matter which design you happen to decide on, I feel you can end up with a reliable flowbench. Keeping the direct blast of air off the flowdisk, sealing the ranges from leaking one to another and accurately scaling your manometers will give you the foundation. The formulas will tie it all together.

One thing that I have not seen any talk on is the location of the lower plenum pressure tube for the manometer. I havent' done any experimenting with this because it's such a pain to get at the thing.

I feel it may be worth investigating the possibility of having one for Intake and a second for exhaust. There is no good location that keeps it out of the way for both flow directions in the MSD design. Possibly shielding it would be something to consider. Putting a piece of large core foam over the end of the tube would work as well. Mine is located off to the left corner of the lower plenum near the exhaust flow control valve. I've often wondered if it's being affected by velocity in that location when performing exhaust flow tests.


Larry

[Mouse]

From what I have learned working with Pitot tubes, I would favor design B, if you have the space. One thing you have to think about is will there be an effect if a head is flowed then turned 180 degrees. Will the numbers be the same. Watch the angles of flow through the system for symmetricy.


"Which design do you think its best from an accurate flow standpoint"

Sorry for the change in subject matter, but you asked the million dollar question and I would like to respond. Unless you have unlimited time and money to tinker with settling chambers, flow disks and orifices, diffusers, pressure tap locations, etc... for the home builder, there is no subsitute for the Pitot tube. With a Pitot tube design, you can flow test calibration orifices all day without worrying about the effects of having different environments on your calibration orifice, and your measuring orifice. Pitot tubes are simply not effected by this condition that can cause uncertainty. What that means is that you can flow test calibration orifices on your Pitot tube design, and if the numbers are correct, what more could you want?

There are some basics that must be followed using Pitot tube designs, and you must have precision and tested calibration orifices (sharp edge orifices, not a .125" plate with a hole in it. A precision machined and tested orifice) to pull it off, but quite doable. Also, when it comes down to the actual Pitot tube, some variations are superior to others, especially when it comes to pressure drop, pressure differential, intrusion and stability of readings or turbulence tolerance and lead-in/out length requirements. And one last point, since Pitot tubes are much less intrusive to air flow, they leave more flow for your test piece and typically require much less air power.

I believe there is no perfect design, I have yet to see a design that doesn't deal with some compromises. Every design has it's problems. But flow measuring is all about confidence. If you can prove your design, you can flow test with confidence. I think Pitot tube designs are easier to proof test, making them easier to build and tune.



John

[84-1074663779]

Agree with what has been said. My bench most closely resembles design "B" turned on its side, simply because it is easiest to make, but the other designs should also work well.

I simply used full sized six foot by four foot sheets of one inch MDF bolted to a heavy welded steel angle frame. It is large and heavy, but size is no real disadvantage, provided you have the space.

[Thomas Vaught]

All three of the designs will work fine especially on a "suck" only
type bench.

You do not need the lower baffle with any of the "suck" only benches
as the bench will be more efficient (per SF) if you have a "Wall" of
air being drawn to motors vs a second baffle where the air must
go around the baffle.

I am basing this on your statements about the bench you are working with.

If it was a "suck and blow" type bench the baffle below the orifice
plate would be a good thing.

JMO

Tom V.

ps A friend built a bench like the "A" bench but with the bench on its side and the right side up. He did not have to use a defuser as
the air made a 180 degree turn from the floor into the orifice

[Terry_Zakis]

One thing that I don't recall reading on the forum is the acceptable range of flow that one would expect a typical orifice to work within, where the calculations would produce valid numbers which would agree with a calibration point. What I'm getting at here is that it's generally accepted that orifices have a turndown ratio of 4:1, and from the work that I've done with Lambda Square on my flow prover orifices, is that a calibration would be done at 75% of maximum flow. And with the 4:1 turndown ratio, that calibration should produce good numbers from 100% down to 25% of the flow range. Some industrial flow computers that are made for measuring orifice flow, even make allowances for a dual-calibration range on the same orifice. This allows you to enter both a low range and high range calibration point for the same orifice. Point is that for each given orifice size, there will inherently be more error from 25% down on the flow range.

So it would only stand to reason that you'd get different readings from two flow ranges, especially is you're near the 25% flow range on one of those orifices.

I'll also clraify that the above information is based on gas flow in closed pipes, where the accurate flow range is where the flow is "turbulent" (Re above 10,000). However, from what I've read, Tony's work is based on slowing the flow down (bigger box is better), so that the flow approaching the orifice is not turbulent (Re below 10,000), or possibly laminar (Re below 2,000).

Does anyone have more input on this topic?

Tony's background in calibration lab work, and extensive experience with his bench may have some useful observations here. Perhaps the orifices used in the larger plenums, which rely of laminar flow conditions have a larger turndown ratio (better accuracy at lower flow conditions) ??

Thank You Gentlemen,

Terry Terezakis
Western Massachusetts, USA

[larrycavan]

[Terry_Zakis]

Hello Larrycavan,

Apoligies up front. This is long.

I quickly found some infromation on turndown at:



Flowmeters and Turndown Ratio
An introduction to turndown ratio for flow measurement devices as orifices, venturi meter etc.
Turndown ratio is often used to compare the span - the range - of flow measurement devices.

Turndown Ratio
Turndown ratio can be expressed as:

TR = qmax / qmin (1)
where
TR = Turndown Ratio
qmax = maximum flow
qmin = minimum flow
Maximum and minimum flow is stated within a specified accuracy and repeatability for the device.

Example - Turndown Ratio for an Orifice Meter
The turndown ratio - TR - for an orifice meter with maximum flow of 12 kg/s and a minimum flow of 3 kg/s can be calculated as:

TR = (12 kg/s) / (3 kg/s) = 4 - normally expressed as turndown ratio of 4:1
This is a typical turndown ratio for a orifice plate. In general a orifice plates has turndown ratio between 3:1 and 5:1.

Turndown Ratio and Measured Signal

In a flow meter based on the orifice or venturi principle, the differential pressure upstream and downstream of an obstruction in the flow is used to indicate the flow. According the Bernoulli Equation the differential pressure increases with the square of flow velocity. A large turndown ratio will cramp the measurement signal at low flow rate.

Here is another source for some good reading on measuring flow with differential pressure meters, at Omega Instruments:



The orifice meters that I purchased for my "flow provers" have an integral orifice and orifice carrier, machined into one assembly, for being inserted between piping flanges. The orifices were set up based on the ISO Code 5167-1998, which parallels the ASME MFC-3M-1989, for flow within closed conduit (pipes).

Correction, please note that my orifices were calibrated at 70% of the maximum flow range.

My plan has been to use one wall of my shop, 26-feet long, to run a set of flow provers, which I could run from from my bench through, and use the provers as a check on the accuracy of the benches that I build/develop. For my provers I have four orifice meters, one 3-inch turbine meter, a 4-inch LFE, and two 2-inch LFE's.

For the orifice meter provers I have 4 of these setup's : one 6-inch, 0.725 Beta, a 4-inch at 0.725 Beta, a second 4-inch at 0.600 Beta, and finally a 3-inch at 0.600 Beta ratio. All were set up with a dual range calibration, that is, there's a calibration point for 0-20" of water colum, and a calibration point for 0-10" of water column differential across the orifices.

According to the ISO 5167 code, the calibration point is at 70% of the maximum flow. Maximum flow would be at the respective 20, or 10 inches of w.c. The 4:1 turndown ratio was used to determine the mimimum flow range for each of the orifice meters. Based on this information, each of the orifices should be able to measure flow within an uncertainty band of +/- 1%, or less. I made sure to observe all of the rules for lenghts of straight pipe runs, upstream and downstream of the orifice meters, and on the direction coming from the bench, the 6-inch line has a flow straightener that I fabricated out of 1" and 1/2" copper tubing, about 2-foot long, inserted into the line. This straightener will get rid of any rotational flow caused by the two "out of plane" elbows upstream.

Anyway, with great expense and time, I'm looking at just under +/-1% uncertainty, after you take into consideration the orifices themselves, the differential pressure transmitters, static pressure transmitters, (Honeywell Smart Transmitters), the thermocouples and flow computers. But this accuracy would only apply between my Max and Min flow ranges, within the 4:1 turndown ratio. Yes, I could measure flow below the Min level, but with the understanding that my uncertainty would increase, and therefore accuracy would decrease.

This is one of the reaons that I sprung for the new 4-inch LFE, because they are very accurate and have a much better turndown capability.

I didn't follow your statement that you perform your calibrations at 4 different flow levels. I wasn't sure that was possible, unless your orifice was calibrated at multiple flow levels to some kind of primary calbration standard?

Well, I hope that some of this has made sense. Hopefully without offending anyone on the excellent forum, I've contended for a while that I just don't belive the +/- 0.5% accuracy numbers that are so frequently stated capable of internal orifice benches. Yes, I agree that absolute accuracy is not paramount, but that repeatability is. Still, this is the long route which I've been taking.

I still think that Tony is quite brilliant on this stuff, as I've always been very impressed with reading his work. So I really do belive that the internal orifice bench, with large plenum, and laminar flow approach, could produce outstanding accuracy numbers. For my own learning, I'd just want to make some comparisons between a few differnt bench designs.

Best Regards,

Terry Terezakis
North Hatfield, MA USA

[larrycavan]

Terry,

Thank you for replying to my request. I appreciate the information very much. My book of knowledge has expanded another few pages in facts and formulas plus an entire volume in thought and concept.

I'll try to explain this and be brief. That's not going to be easy and still make it clear.

As I said, I ran into the situation on a SF110 whereby flowing a calibration plate at the same pressure in two different ranges, produced slightly higher results in the higher range.
It seems logical to me that it happens because of the turn down ratio effect you described.

It's either that or you can't rely on Superflow numbers which would be a real shame because I know I [and I believe most everyone that has ventured down this path] am/are trying to do that very thing. I don't care if it's 100% but I do care that it's reasonably close.

Thank you again,

Larry

[Terry_Zakis]

Hello Again Larry,

There's actually a few possibilities. I don't know if Superflow mentions turndown ratio's in respect to their different ranges? So it's hard to tell if you are approaching the Min flow level on the larger orifice size.

Another possibility can be correction for gas temperature. Don't you have to do a correction for that on the SF110? Just an idea.

And lastly, you should consider the magnitude of the difference you're noting. Keep in mind that each orifice will flow +/- some percentage. So even if Superflow states this bench to be able to measure +/- 0.5%, that would be for each orifice. But since you're comparing two orifice ranges, then the maximum expected difference between the two should be (+/- 0.5%) + (+/- 0.5%), or +/- 1%.

So in comparisons between two ranges, you can expect a potential of twice the stated error for the bench. Then there's the consideration of what accuracy level these benches are actually supposed to provide, which can altogether increase the expected difference between the two ranges.

The realization that there can be differences between ranges is all the more reason why its important to chose the right range to start with for a given project, and then make sure that all of the successive tests on that head are done on the same range. In theory, the bias error which I've been talking about will cancel out, and you'll get very good indications of what the "changes" in airflow are. Make sense?

Best Regards,

Terry Terezakis
North Hatfield, MA USA

[larrycavan]

Terry,

That makes good sense. The differences aren't large, I dont' have the plate in front of me at the moment. It seems to me it was about 1.5CFM at 6" between the 105 and 105 Ranges.

I calibrated using this method. Here are 3 of the ranges. Using the SF calibration plate [153CFM@10"] I converted the flow to different pressures with the standard formula, then tested each range at the same 6 pressure drops. I started at 10" and used that % flow figure to determine the preliminary range value. I then continued the test at the other pressures. When I was finished testing, I compared the results and decided on a final Range Value for each range. I added up the results,took an average for each range and compared the averages.

Range % CFM Pressure
161.00 50.0% 80.5 3.00
161.00 58.5% 94.2 4.00
161.00 66.0% 106.3 5.00
161.00 72.5% 116.7 6.00
161.00 84.7% 136.4 8.00
161.00 95.0% 153.0 10.00


Range % CFM Pressure
242.00 34.0% 82.28 3.00
242.00 39.5% 95.59 4.00
242.00 44.0% 106.48 5.00
242.00 48.0% 116.16 6.00
242.00 56.0% 135.52 8.00
242.00 63.2% 152.94 10.00

386.00 21.0% 81.06 3.00
386.00 24.3% 93.80 4.00
386.00 27.5% 106.15 5.00
386.00 30.2% 116.57 6.00
386.00 35.0% 135.10 8.00
386.00 39.5% 152.47 10.00

Curiously, in comparing the results, they do vary but vary in both directions [higher - lower].

I hope that method makes sense to you.

Best Regards,
Larry