Orifice Wheel Bench, Leakage on Exhaust Flow - Minimizing leakage past the wheel.

[Terry_Zakis]

I am looking at building two benches. One that uses all motors in parallel, which will give higher CFM flow for testing exhausts, intake manifolds, and carbs. But I've also thought of building a second bench that uses two banks of motors in series, say 8 motors feeding 8 motors. This should significantly increase the available test pressure.

As you may have read in some of my earlier posts, one bench will have the measurement orifices outside of the bench, by means of ASME code orifice meters in several pipe runs. No chance of leakage this way. But the second bench will be the MSD / Superflow style with the orifice wheel.

The delema that I'm wrestling with at present is how do you minimize leakage on a wheel type orifice bench, when the flow is running in reverse, when doing exhaust work?

For flowing intakes, I see that the differential pressure across the orifice wheel will help to seat that wheel and increase the sealing effectiveness. But my concern is when flow is reversed, and there would be a tendency to partially bow or lift the orifice wheel, and allow flow to exit through the non-active orifices.

The idea that Tony is working with, using the ridgid melamine to mount the orifices would probably reduce this effect, but I just received my 1/8" and 1/4" plates of 24" x 24" 6061 aluminum, and hate to think of cutting them up.

What have people done out there to minimize this kind of leakage when flowing exhaust?

Could this be why the newer Superflow benches look like they use fixed orifices that are accessed through a window on the front of the bench?

Thoughts would be appreciated. Again Tony, you're ideas are great!

Terry

[bruce]

Spring load the shaft the orifice disk rides on so it is always pushing on the intake direction of the orifice plate? All you would need is a spring that has enough force to over come the pressure. Just an idea . . .

[84-1074663779]

Yes, give it a bit of spring preload, but operating air pressure drop across the orifice is going to seal it very tight, especially if the centre bearing is of the self aligning type.

It might be best to always keep the air flowing in the same direction through the measurement orifice disc, and move the test piece between a "blow" port, and a "suction" port in the bench top.

If you calibrate your orifice in one direction, how do you know it will flow the same in the reverse direction ? The flow dynamics of the whole bench are going to change if you reverse the flow.

The only disadvantage that I can see of doing this, will be that the measurement orifice always sees ambient air temperature, but the test piece sees slightly heated air when blow testing. It might still be the lesser of two evils.

If this worries you, you can always compare the delta P across two identical orifice plates, one in the turret, and the other over the "blow" port in the bench top.

[Shawn]

I know that our Superflow does change the rating of the orifice depending on blowing or sucking. Off the top of my head range 3 on the intake is 300cfm and on the exhaust it is 321 cfm.So your assumptions are correct.
Shawn

[Terry_Zakis]

Thank You Bruce, Tony, and Shawn,

The spring loading the disc is a great idea, which will help seat the disc when you reverse flow across the orifice wheel.

For an MSD/Superflow style bench where you only use the aluminum or steel orifice plate, the plate is thin, and even though it's spring loaded, the sealing between the partition that the wheels mounted to, and the orifice that you want to use can distort and cause air to flow between the partition and the orifice wheel, and then exit from other orifices.

I guess this is what Shawn is getting at on his response of the Superflow bench, that flows more in reverse than in standard (suction) direction. So I guess Superflow's approach is to "calibrate" this leakage out?

Tony, I agree with you, that changing flow direction will change the flow dynamic. So I imagine this type would absolutely require calibrations in both directions. Which in turn only adds more uncertainty to measurement.

This is probably why your setup only flows in one direction? And you have discharge and suction flanges on the benchtop, and just move the test piece between the two ports depending on the testing that you'll be doing? Brilliant idea! With that approach, the airflow would work to seat the disc and provide a superior seal as well.

And I think the benches that change flow direction, need more valving too. Whereas on Tony's style bench, with air flowing in one direction only, you can control flow by selecting the number of motors you're running, and vary speed of a few with a variac.

Tony, on the temperature aspect, I think that either style bench would need compensation for heating of the airflow by the motors, with a second temperature measurement. An MSD/Superflow style bench that reverses flow, would be sending heated air across the orifice, and that same heated air across the test piece. There would be little difference in air temperature from the orifice to the test, so as long as the temperature at the orifice plate is being used in the calculation there should be little error introduced.

For your style bench Tony, that flows air in only one direction, I would think it to be even more important to correct for air temperature when flowing exhausts. Because the inlet air to the orifice would be at atmospheric temp, but then the air is heated through the motors, and so the airflow through the test piece is at a higher temperature, and is therefore flowing at a greater volumetric rate. So the two temperatures would need to be read and a correction applied. What are the thoughts on this?

I'm not sure if I follow on the self-aligning bearing Tony. Are you implying that a compression spring is used below the orifice wheel (or melamine carrier), and then snugged up in order to pull the wheel/carrier to the partition. And to rotate the wheel from the opposite side, you use a flex-joint or some king of universal joint, so only rotational torque is transmitted and you don't run the risk of creating leaks?

Thanks for the information and discussion!

Terry

[84-1074663779]

Hi Terry,

I only wish I had a camera to show you my setup, but I will attempt to describe how I mounted my turret wheel.

There is a large flat melamine bulkhead built permanently into the main bench structure. In the exact centre is a 12mm diameter projecting steel stud, secured with two nuts and two large flat washers. One nut is deeply recessed (flush) into the melamine sheet.

The turret is approximately two feet in diameter, and was cut out with a router from a melamine sheet. In the centre is a low profile pressed steel self aligning NSK ball race (UBPF-201) which is bolted and epoxied to the turret. This slips over the stud, and the bearing can self align so both melamine sheets are in perfect flat rotating contact. A spring is then slipped on the stud with a washer and a self locking nut. The nut can be screwed down to give appropriate spring preload.

A fork made from 25mm x 3mm flat steel strap also bolts onto the turret to clear the bearing and spring. Onto the centre of this is welded a universal joint from a rack and pinion steering shaft. This allows the turret disc to find its own flat plane of rotation independent of the drive shaft and stud.

The long turret drive shaft protrudes through the outside of the flow bench at one side through another self aligning ball race. An ordinary car steering wheel fits onto the tapered and splined end of the shaft. The whole turret rotation mechanism is made from salvaged automotive steering components.

It all feels very solid and smooth with no backlash. The outside edge of the melamine turret is made perfectly round by using a router in a jig, located off the central ball race after it is finally epoxied in position. Several coats of clear estapol seal the crumbly outer edge of the melamine, and a further touch with the router will give a very hard smooth and precise outside finished diameter.

A brass roller fitted to a spring loaded aluminium arm provide positive indexing to the turret wheel. The indents can be made with a router bit the same diameter as the roller in a suitable jig. The roller clicks into the detents very smoothly, and with a bit of effort it can be made to have a really nice feel. If the detents end up too deep, you can trim a bit off the disc diameter to get it right.

Some large white rub on Numbers (1-8) around the steering wheel central shroud can be viewed through a piece of flat sheetmetal with a 25mm round hole. So it is easy to look down on the steering wheel and see which number appears behind the hole. It took a while to do, but I am extremely happy with the final result.

The temperature rise produced by the blower is going to depend greatly on the required blow test pressure and the efficiency of the blower. This may be an issue for some people or it may not. I suppose the actual flow can be checked by replacing the test piece with an orifice identical to the measurement orifice and noting the difference.

I have not bothered to do this yet. Another problem that might be significant is that suction testing is going to hopefully use undisturbed air into the test piece. Blow testing is going to see far more turbulent air coming from the blower, unless steps are taken to stabilize the flow. But an exhaust valve is going to see some pretty hot angry air anyway. Is testing with calm laminar flow into the combustion chamber actually valid ?

[bruce]

Hey Tony, (please take this in a joking manner) here in the states they sell disposable cameras that you can take to the store and have pictures put on disk or CD's so you can email them or post on a website. If enough of us get together and send you some money do you have a store in the "outback" that might do the same? LOL Better yet send me a plane ticket and I'll come down there with my digital camera and take some pics . . .

[84-1074663779]

Ha-ha, point taken Bruce ! ! !

O/k I will look into this. Next time I sell some more shares and get some income, I will look into getting a camera.

[Terry_Zakis]

Hello Tony,

Thank you so kindly for sharing the thoughts and application details about your design of the orifice wheel. Again, great information.

Yes, I know what you are saying about the differences in approach air to the intake test piece, vs. blower air approaching the exhaust test piece being quite different. For the orifice plate used in piping, which is what most of my research has been on, the turndown (ability to measure flow accurately from the expected maximum down to some lower volume of flow, e.g. 5:1 turndown indicates measurement is good from 20% to 100% of range) of the orifice is based upon the ability of the flow approaching the orifice to maintain turbulent flow (Reynolds number greater than 10,000 at the low flow).

But the measurement of flow through an orifice, from ambient conditions, into a plenum, is quite different. I know there's one ISO code on this, that the vacuum motor manufacturers use to test and compare their motors. But I haven't looked at it in a few years, and it only applies to airflow being pulled across the orifice and into the plenum.

Using flow conditioning devices inside pipes are a relatively easy affair as compared to conditioning flow within a plenum. I'm not sure how one would do that.

Many of the industrial orifices may be made out of 0.125" stock, yet the actual orifice hole itself is machined with an angle, to produce a sharp edge. From what I have read, this is done to reduce the amount of non-recoverable pressure drop in process flows (cost savings in pumping horsepower).

The square edge orifice plate, is what's used when flow is reversed across the orifice. You are right about changing direction of flow in the bench, that the flow uniformity would be different when approaching from one side or the other. I belive you (Tony) have advocated using a larger plenum to give the air a chance to clean-up before the orifice approach, so best case may be to try and have equal plenum volume above and below the orifice plate.

I think performing the calibration in both directions has merit, because you should be able to identify the additional airflow due to leakage. But I think the temperatures would still need to be corrected for because the plenum air temp would now be higher when blowing (addition of motor heating), and if the temperature correction is not applied, then you would be underestimating the total flow. In addition, operating in the blowing mode for various periods of time, would change the plenum air temperature in unpredictable ways, which would not have been quantified in the calibration.

Great conversation Guys.

Thanks,

Terry

[84-1074663779]

Some really great thoughts there Terry.

I really don't know what the answer to all this is, but it is very good to share some ideas around. Another way to approach the whole problem of flowing exhaust ports is to suck the air out of the port and have atmospheric pressure on the combustion chamber side.

There is nothing new in this idea, but it does overcome the temperature rise problem, and the turbulence problem at the blower discharge.

Having a blow port in the top of the bench is quite handy for testing whole exhaust systems, mufflers, and intercoolers. For these types of measurements, extreme accuracy is not usually required, usually just an A/B comparison.

The air is already moving and possibly turbulent as it enters whatever you are testing, and this probably approximates fairly closely normal exhaust flow the test piece will see on the car.

I have always had my doubts when trying to suck air through an exhaust muffler for example. Getting ambient air into a typical muffler will require an exponential entry flare, or something like that, and then you would never be quite sure how the nature of the entry effects the total pressure drop.

There probably is not one best solution for building a flowbench, but I believe there is a need to both blow and suck, and keeping unidirectional flow through the measurement orifice is probably best if you can arrange it.