Pingu's Next Project - A Flowbench

[pingu393]

I make (and reject) a lot of modifications, especially engine modifications.

I test my engine modifications with a manometer before I try them on the car, then I test them on the car with a datalogger.

The manometer I use simply measures the pressure drop across the OEM part and the pressure drop across the modified part. It works fairly well, but I am aware that the inlet pressure is not consistent in the tests, even though I am using the same air pump (either a vacuum cleaner, or a leaf blower).

So, I've decided to build a proper flowbench based on the Wikipedia description ( https://en.wikipedia.org/wiki/Air_flow_bench)

The pump will be my stripped-down vacuum cleaner.

Does anyone know anything about minimum plenum sizes, or anything else that I might need to know?

I've just measured the vacuum cleaner at 318 cfm (60mm orifice, 54.6 m/sec)

[227bhp]

You're not being very specific about what you're testing, but there are generally two ways of doing it and two places to go.
One is David Vizards airflow book which sounds like it would suit you (he outlines a simple way of testing called floating pressure drop), the other is the PTS Flow bench tech forum which is for building a 'proper' bench which sounds out of your league.

https://www.musclecardiy.com/cylinder-heads/build-...

Sorry it's page three, you'll work it out.

[pingu393]

Mr Vizard was a hero of mine when I was at college in the late 80s. It's great to see that he seems to be as generous with his knowledge as I hoped when I was a yoof.

I've never heard of "floating pressure drop", so I'll give it a Google. - TAVM.

I read the link last night and the "Low Budget Flow Bench" looks like what I have already.

I am looking at the intake system (upto, but not including the plenum) - the intake tube, the air filter box, the air filter and the MAF.

I may also look at cylinder heads, but that is not very likely, other that to test only. I think that modifying cylinder heads for performance is best left to the experts (like Vizard, not your local engineering shop). There's much more to air flow in a cylinder head than volumetric air flow . Swirl and tumble needs to be properly designed into a system, and your local engineering shop just don't have the mixture of skill and knowledge.

[pingu393]

Interesting images on "floating pressure drop flow bench" on Google. ( https://www.google.co.uk/search?q=floating+pressur...)

Does anyone know the pros and cons of the order for

1. Pump - Test Plenum - Metering Plenum
2. Test Plenum - Pump - Metering Plenum (as on Wikipedia)
3. Test Plenum - Metering Plenum - Pump

I suspect that option 3 is best as that is basically what a vacuum cleaner is designed to do. My initial thought was that it could cause the motor to burn out if the flow is restricted, but if I control the Test Plenum pressure by allowing air to enter the pump AFTER the Metering Plenum - Bingo.

Any thoughts?

[pingu393]

It works - in a fashion .

It needs some fettling. I can test items with <300cfm flow rates using a 10mm orifice. I will try to add additional suction devices and will hopefully be able to get some mega-suction.

Time to find some broken cylinder vacuum cleaners with working motors.

I'll post up some video of it later.

[pingu393]

Done - https://youtu.be/QC0trNMPRns

[tapkaJohnD]

There's another book about DiY flow benching, "Practical Gas Flow" by John Dalton (Brooklands Books)
Probably out of print, but there are copies on sale on Amazon

https://www.amazon.co.uk/s?url=search-alias%3Dstri...
JOhn

[Boosted LS1]

Where is Peter Burgess :-)

[Mignon]

You're making a fair old meal of this when a properly calibrated flowbench like mine is even easier to make. It can all be done from plastic drainpipe and an old wooden bench.

Top of bench is where the cylinder head goes. Make a 60mm hole and chamfer the edges nicely to improve flow coefficient. On this sits your bore adapator and cylinder head all sealed with plasticene. Under the hole is a down tube of 110mm soilpipe about 18" long running into a 90 degree bend and then sleeving down to 65mm plastic drainpipe running out horizontally. You want about 4 feet of this for the flow to stabilise and then fit a metal orifice plate, square edged, of about 22m hole size. This should give good readings for a standard wet and dry vac. Then another few feet of pipe and connect the vacuum source.

One manometer measures from the top of the down tube under the cylinder head to atmosphere. The second measures across the orifice plate with tappings 1" either side of it or very close to it.

The manometer readings will pulsate so you need dampers in the plastic tube which can be made from scrap aluminium or brass drilled with a 1mm hole. About 40mm total of damper in each tube should do.

I can explain how the manometer readings can be converted to accurate flow later if you like.

[pingu393]

Mignon said:

Stuff +

I can explain how the manometer readings can be converted to accurate flow later if you like.

Several things about your set-up that wouldn't suit me, even though I would prefer it.

I don't have the space for a permanent set-up.
I don't have any of the materials that you used, so I would have had to buy them. I only spent £10 on the manometer backboard and some 2"x1".

I'd like to see your set-up in action.

What is the largest orifice plate that you can use? I am limited to a 12mm orifice, as my vacuum cleaner can only just get a pressure difference of 27.5" if I use a 13mm orifice.

I need to make some properly machined orifices, as the ones that I have at the moment are just drilled holes in wood .

I'd be really interested in how you calibrate the flow bench. I have an anemometer, and my plan was to use fluid dynamics calculations based on Area and Velocity. The accuracy of each measurement would be poor, so the calculated flow rate would be even worse.

[Mignon]

pingu393 said:

I don't have any of the materials that you used, so I would have had to buy them. I only spent £10 on the manometer backboard and some 2"x1".

Christ, you're tight. Is 20 or 30 quid too much to spend on an absolutely accurate flowbench?

pingu393 said:

What is the largest orifice plate that you can use?

Not a relevant question. My bench is a floating pressure drop bench so it is not relevant what pressure drop is actually pulled across the test piece. The flow at a nominal 25" of water is worked out mathematically from the actual pressure drop achieved using the square root law.

pingu393 said:

I'd be really interested in how you calibrate the flow bench.

If made correctly the bench needs no calibration. All parts of the design were done to properly follow the guidelines in British Standard 1042. The measuring orifice plate needs to be measured accurately and made properly sharp (square) edged and then it will achieve the expected flow coefficient.

I made a set of 4 reference orifice plates to test it with and achieved within 1% accuracy full range from 20 cfm up to over 250 cfm. It has also been tested over the years against several Superflow and other types of proprietary flow benches costing thousands and agreed exactly.

If you are prepared to calibrate your bench against a good known one then you don't need an accurately made orifice plate. Any sort of hole in a sheet of metal will do and you just work out its flow coefficient. You could actually make almost all of it from wood.

I see no point in a flow bench that is not accurately calibrated because you can't compare your results with known targets or anyone else's work.

[Mignon]

This document shows a diagram of the essential parts of a flowbench like mine.

ftp://ftp.ecs.csus.edu/sae/2002%20and%20before/FSAE%201999-2000/FSAE/engine/Air%20Flow%20paper.doc

Mine doesn't use the control valve as it is not meant to achieve a fixed pressure drop and it has a downpipe and 90 degree bend rather than a big plenum but either design will do fine.

Yonks ago me and my friend Dave Vizard got very deeply into home made flowbench design on a car forum and I gave detailed instructions for my design plus the necessary maths but I'm damned if I can remember the forum name and it's long since closed down. When I first built my bench back in about 1990 or so I made a dumb error by following someone else's work and put a manometer tapping in the wrong place and I got flow numbers which were all over the damn shop, nowhere near what they should have been on my test plates. It was actually Dave Vizard who worked out what my mistake must have been when we were chatting about it on the phone one night. Next day I fixed the problem and all 4 reference plates flowed exactly what the maths said they should which was handy.

We have had several flowbench discussions in here over the years including back in 2012 when I invited Dave Vizard to contribute here for a while. I suggest you search under "pumaracing" and "flowbench" or similar.

[Mignon]

If you are wanting a flowbench to test cylinder head modifications then I suggest you read this thread.

https://www.pistonheads.com/gassing/topic.asp?h=0&...

[pingu393]

Mignon said:

Christ, you're tight.

Tighter than a penguin's ars....

Mignon said:

My bench is a floating pressure drop bench so it is not relevant what pressure drop is actually pulled across the test piece. The flow at a nominal 25" of water is worked out mathematically from the actual pressure drop achieved using the square root law.

I was under the impression that you had to actually pull 28" in order to get a realistic air velocity. As air is compressible, the faster it flows, the more compressed it is. I need to do some more reading.

Mignon said:

If made correctly the bench needs no calibration. All parts of the design were done to properly follow the guidelines in British Standard 1042.

More reading to do.

Mignon said:

The measuring orifice plate needs to be measured accurately and made properly sharp (square) edged and then it will achieve the expected flow coefficient.

I should be able to do this quite easily on my lathe.

Mignon said:

If you are prepared to calibrate your bench against a good known one then you don't need an accurately made orifice plate. Any sort of hole in a sheet of metal will do and you just work out its flow coefficient. You could actually make almost all of it from wood.

This was my fallback option

Mignon said:

I see no point in a flow bench that is not accurately calibrated because you can't compare your results with known targets or anyone else's work.

I agree with you, but this is the only reason to calibrate a flow bench. I was only planning on comparing my own test pieces, so a calibrated flow bench is only a "nice to have" at the moment. It may become a "need to have" in the future.


Many thanks for your time

[Mignon]

pingu393 said:

I was under the impression that you had to actually pull 28" in order to get a realistic air velocity. As air is compressible, the faster it flows, the more compressed it is. I need to do some more reading.

Definitely not. There just needs to be enough flow so that it's turbulent not laminar and that happens very quickly. A couple of inches of pressure drop is enough which is about what a single vac motor gets to with a big valve at high lift.

[GreenV8S]

My book learning says that the boundary layer behaves differently in accelerating and decelerating flows compared to steady state so the flow patterns could vary during the induction cycle. Do bench flow testers ever try to model that sort of thing?

[227bhp]

You can't replicate exactly what goes on in a port if that's what you mean. Not when the flow stops and starts constantly due to the valve opening and closing.

[pingu393]

Mignon said:

pingu393 said:

I was under the impression that you had to actually pull 28" in order to get a realistic air velocity. As air is compressible, the faster it flows, the more compressed it is. I need to do some more reading.

Definitely not. There just needs to be enough flow so that it's turbulent not laminar and that happens very quickly. A couple of inches of pressure drop is enough which is about what a single vac motor gets to with a big valve at high lift.

PH really would benefit from a "Like" button.



What you say makes sense and agrees with my observations with a smoke generator.

[pingu393]

Mignon said:

If you are wanting a flowbench to test cylinder head modifications then I suggest you read this thread.

https://www.pistonheads.com/gassing/topic.asp?h=0&...

Excellent read. Made even better by David Vizard's contributions, he is to engines what King Kenny is to football

[pingu393]

Also posted in https://www.pistonheads.com/gassing/topic.asp?h=0&...

Hopefully, one of the earlier posters will read this...

According to Helgesen, the orifice sizes for the following volumetric air flows are...

5CFM --> 0.210"
10CFM --> 0.296"
20CFM --> 0.419"
40CFM --> 0.594"
80CFM --> 0.840"
160CFM --> 1.185"

From the equation, Diameter = SQRT ( ( 4 * Q ) / ( SQRT ( 2 * ( Change in Pressure ) / Density ) ) ) we get...

5CFM --> 0.209"
10CFM --> 0.295"
20CFM --> 0.418"
40CFM --> 0.591"
80CFM --> 0.836"
160CFM --> 1.182"


Can anybody explain why some of the calculated orifice diameters are outside the tolerance values of the Helgesen diameters?

Is it a case of "close is good enough" and just ignore the +/-0.001" tolerance?

Edited by pingu393 on Monday 3rd September 23:41