PTS Flowbench Forum

Hey guys, new member here. I'm a student at Queen's University in Kingston Ontario. As part of the miniature formula design team (http://qfsae.ca/) I've been given the task of building a flow bench so we can figure out the valve lift on various parts. I've done a good bit of reading but I'm still a little fuzzy on some of the concepts, so I was hoping you guys could help me out and make sure my design will work before I waste any of the teams budget.

Mostly because it has a pretty simple design and a comprehensive build document I was going to build a flow bench based off of the plans posted on the Dtec website (http://dtec.net.au/Flowbench%20Design%20Guide.htm). It's a single cabinet with an internal settling chamber and orifice plate with two water manometers. Does anyone see any issues with his design? Also because there is no speed shops with flow benches in my area for office plate calibration, I was going to order a set from the PTS website. The problem is, I don't know what my max CFM range will be, which Bruce (the fine gentleman running PTS) told me I should know before I order plates. The engine we use is a 600cc 4-cylinder Honda superbike engine out of an F4i. It's a fairly small bench so I'll probably be testing at around 15" (maybe the bench could handle 28", not sure that the size makes a difference). How can I figure out what an appropriate range would be? Is it purely based on the size of parts being tested? We don't have a large budget for this so the less plates we need to order the better for us.
Also the Dtec design incorporates multiple orfices side by side that can be covered individually (so any combination can be used at once). Is this necessary? Or would it be sufficient to have a single swappable plate?

Final question (bear with me guys) to make sure I understand the basic concept of how these things work. To calibrate the bench I set up the inclined manometer to show 0% (at the top) when the bench is off/sealed, and 100% (at the bottom) when air is flowing at a specific inches of water and no test piece is attached? How does this change when I swap to a smaller plate?

Thanks guys, any input is appreciated.

First thing is probably to select your vacuum cleaner motors, that will pretty much fix your final flow capability and mains power draw. Design your bench to eventually use all of the available electrical power, even if at first you can get away with fitting fewer motors.
Just block off the unused motor holes. At some future time you will be very glad you thought that far ahead...

The only real problem with water manometers is that they only work accurately over a very limited flow range, because the scale at the low flow end becomes extremely cramped due to the square law flow scale.
This usually requires switching ranges during a test, which can cause big problems, apart from the extra effort involved.

Suppose you reach the practical limit of one range at say 50 CFM, and you do nothing but switch range, and the bench now reads 48 CFM or 52 CFM. That is going to put a kink in your readings and add some uncertainty.

The solution is to use an electronic manometer and some software that will work over the entire range with only one size of measurement orifice, if your budget will stretch that far.
Maybe not possible for you, but at least cost it all out and think about it.

The most important thing with any flow bench is repeatability, so that you can reliably detect very small changes which are very repeatable. It does not really matter if your bench reads within 1% or 5% absolute as long as it always reads the same.
If its always out by the exact same amount, it really does not matter, as long as that absolute error never changes.
If you can test the same item several times spaced over days or months, and it always reads exactly the same, you have built the perfect bench, even if it reads a bit different to someone else's.

Alright thanks, I'm actually planning on using a shop vac coupled with some large blow off valves as my air source. And yea we definitely don't have the budget for an electronic meter, unless I buy one of those cheap cfm meters off ebay that are designed for air ducts (would that actually be a viable option?).

So for accuracy, I want enough flow plates so that my scale is always measuring in the higher range? Hmmm maybe putting the plates adjacent to each other would allow me to get by with fewer plates, as I could use various combinations of plates or just one at a time. This is a side project for the team and we barley have enough money to build the car in the first place, so the fewer plates we need to buy the better.

O/k fair enough on the budget.

Your best bet then is to just build the bench and fabricate a pair of water manometers, and carefully mark out the square law percent flow scale. You will quickly see that the graduations on the flow scale get very close together down at the low flow end.

You may decide that one quarter the way up the length of the scale is as low as you really want to go, and that will be at the 50% flow mark. So that will give you usable 2:1 CFM flow range for each measurement orifice.

The next step might be to carefully make a series of orifice plates in 2:1 flow steps with diameters like X, 1.414X, 2X and 2.828X, 4X and so on.
The dimension X will depend on the chosen orifice pressure drop, and you can use the spread sheet provided here at the Forum to work it out.

Now the absolute accuracy is not hugely important, as long as the readings are repeatable, and the ends of the scales match up. Just to consider some completely fictional numbers, you might decide to make measurement orifice plates that read 8CFM, 16 CFM, 32 CFM, 64 CFM, and 128 CFM all at the 100% flow mark on your sloping manometer, for example. Or it could just as easily be 10, 20, 40, 80, 160CFM if you wish.

So you build it like that and use it like that, and it will give you some great results.

But then later on, you decide to buy a 100 CFM calibration plate from Bruce. What should happen is the reading should be 78.125% flow on the 128 CFM range (100 true CFM) but there will inevitably be some error.

What you can do to fix it is to adjust the slope of your manometer very slightly to get it to read 78.125% as it should.

All your ranges should now be spot on.

Its possible to build a perfectly usable and repeatable bench without ever having calibrated it against an external standard. But eventually you will wish to know for sure where you really are, and any error is very easy to correct as long as you design your sloping manometer so that the slope is adjustable.

Philip_Robinson wrote: And yea we definitely don't have the budget for an electronic meter, unless I buy one of those cheap cfm meters off ebay that are designed for air ducts (would that actually be a viable option?).

Yes, you can use one of those handheld digital manometers you just have to convert the numbers manually. Setup a spreadsheet to do the math for you.

I'm actually planning on doing a thread on this after the New Year on my PTS build as an alternative gauging setup for the DIY'er

That sounds pretty appealing as it would make the process allot simpler,

Ok guys there's alot more theory going into this than I thought, the more I look into this, the more confused I am.

So please tell me what I'm not understanding.

The end result of what I'm trying to get is the flow coefficients for the parts I am testing. Do I get this by using the equation
Cv = (Gas flow rate/962) x [(specific gravity of water x Abs T)/ (P1² - P2²) ] ^0.5 ???

So the inclined manometer/anemometer gives me the gas flow rate and the vertical manometer gives me delta P. How do I relate delta P with (P1² - P2²)^0.5?

Or am I looking at this all wrong. A flow coefficient is equal to the actual flow/theoretical flow rate right? So I get the actual from the inclined manometer/anemometer and I calculate the theoretical using flow= (delta P)^0.5 * combined coefficients*orifice area
Does this mean I wouldn't be able to get the flow coefficients using an anemometer as the theoretical flow is based around orifice parameters?

Also does any think using a digital anemometer would impact the accuracy of the bench? People seem to have alot of mixed results with them. How would I even implement one? Would I just have to make sure that all of the air flow travels through the impeller?

Have a look at my spreadsheet, you can find it: HERE, on one of those sheets, I give a detailed description of the formulae used.

You just might need to compensate that anemometer for changes in temperature and humidity, and every extra required measurement you must make for correct compensation adds another potential source of introduced error.

What makes a ratiometric airflow bench so powerful is that it requires no corrections whatsoever.

Whatever air flows through the test piece flows through the measurement orifice. Its exactly the same air, with the same temperature and humidity.
So altitude, barometric pressure, winter or summer, day or night, your tests will always repeat exactly.
You can read straight off the flow manometer and know it is right.

The RELATIVE differential pressure drops across your home made measurement orifice and test piece will remain the same.
And if you always manually correct the test pressure back to some constant (28 inches ?) the pressure drop across the test piece then ALWAYS accurately reflects flow figure directly.
Ain't mother nature neat !

Anything else involving velocity measured in a tube with anemometers or pitot tubes introduces the spectre of having to add several potential (and very evil) corrections, not to mention rounding errors, to every individual reading you take.

Ok so flow coefficient = actual flow / theoretical flow

where actual is measured by the bench, and theoretical is calculated using the excel file posted by 86roco.

Does this mean that using an anemometer I wouldn't be able to get the theoretical flow, because the equations for theoretical flow are based around orifice plate parameters?

OR are the orifice parameters in the excel file based on the test piece itself, and not the office plate used for testing.

OR am I just over complicating this and the flow coefficient is simply equal to the percentage reading on the inclined manometer?