Tractorsport Flowbench Forum Archive

Posted: **Tue Nov 21, 2006 7:30 pm**

Hello,

I've been looking to build a flowbench for a while now. I currently port heads for my engines but I can only do a minimal of porting as I don't want to ruin the heads. I want to go further with the porting so now the time has come to look at building a bench.
I looked at various designs and decided upon the flowbench described in the Mopar Tech articles. This is a simple bench using ordinary PVC pipework and an S-type pitot tube.
Rather than use the gate valves with the 2" orifice for going to a lower range, I want to use two pipe runs instead. Here's where I'd like some advice;

I've looked at using a 2" pipe and a 3" pipe for the ranges. If I make the pitot tube in accordance to those used by the EPA and Dwyer it seems I can flow 160cfm with the 2" pipe and 350cfm with the 3" pipe whilst pulling 4" depression across the device.
Here's a graph of the 2" pipe.

3" pipe.

Charts were calculated using the formulea on Dwyer bulletin H-12.

I was looking at using the Dwyer -4-0-4" Digital differential manometer for measuring the depression as it has 0.01" resolution and a Dwyer 36" U-tube manometer for test pressure reading.

Would this combination give good results?

For motors, I've been looking at either Ametek or YDK as these are readily available in vacuums here in the UK. For estimating the number of motors I need to use is the following correct?

Say I need to pull 350cfm at 28" vacuum (test depression). The motor flow chart shows that at 28" vacuum it produces roughly 100cfm. Allowing for some losses due to the piping runs etc. would I simply need 4-5 motors?

I look forward to any comments, suggestions, etc.

Stef

Posted: **Tue Nov 21, 2006 7:39 pm**

This is a sample from the spreadsheet I developed for the flow numbers. I included calculations for the Reynolds and Mach numbers as it seems to be important to keep the Mach # below 0.3 and the Reynolds above 20,000 for the best readings (according to various sources I've read like ISO 3699). Mach numbers never go over 0.1 for either pipe sizes.

Stef

Posted: **Tue Nov 21, 2006 8:51 pm**

Welcome to the board Stef!

The bench you are considering is more like a pitot bench. Bruce or one of the pitot bench guys will chime in here and help you out with this....just wanted to say welcome..

Larry C

Posted: **Wed Nov 22, 2006 12:57 am**

Stef,

Yes Mach 0.3 is where air will start to compress and change density giving results that are differnt to the formulae used below mach 0.3.

At the low end the air needs to be turbulent, in the reynolds number sense, to maintain results that follow square law. Most literature suggests flow is turbulent at Re>=10000.

I started off doing a Pitot bench in the hope I could do my flow range with just one pitot. I found that approaching Mach 0.3 generated a lot of pressure loss through the piping to and from the pitot. Yours being lower velocity will be much less.

Have a look back through recent posts in this forum, there is a link to a pressure loss calculator that will help you determine how much pressure your motors will need to generate. It is likely to be more than 28"+4". Be aware the stated insertion loss for a pitot is only part of the story, the other part is the pipeing losses.

Posted: **Wed Nov 22, 2006 6:18 pm**

Thank you for the comments so far.

jsa
So far the calculations I've done indicate the pipe sizes will flow within the min/max ranges for a pitot tube. I did the spreadsheet so that I could see where I would get a sensible reading from the manometer i.e 0.01" change = 1cfm change. This gives a working range for the 2" pipe of ~23cfm - 160cfm and for the 3" pipe, ~150cfm - 350cfm.
I think this should be enough to do 4v - 4 cylinder heads... Any thoughts?

I spent some time looking for that link... But no luck. Can you remember what the thread topic was or the poster?

I may have more questions to ask as I look into this in more detail.

Stef

Posted: **Thu Nov 23, 2006 8:28 am**

Posted: **Thu Nov 23, 2006 2:06 pm**

If you are having trouble finding a post in a search, be sure to click on "From the beginning" on the search page. I forget to do that once in a while, too!

Posted: **Fri Nov 24, 2006 9:15 pm**

Posted: **Sun Nov 26, 2006 7:27 pm**

jsa thanks for the links. I think I may add another smaller pipe size, say 1", to cover very low flows. I will juggle the numbers around for the other pipe sizes to get a set of pipe sizes that will cover 0-350cfm maybe going higher if possible.
I won't be swapping pipes in and out of the bench but rather have all the pipes installed and use valves to switch between pipes... Well that's the plan anyway!

I do a sketch and post it up so you guys can see what I mean.

Stef

Posted: **Mon Nov 27, 2006 3:20 pm**

Here's a quick sketch of what I had in mind.

One thing I noticed whilst doing the calculations is that the useful or accurate range is quite dependent on the resolution of the measuring device and it's range. With a 4" range and 0.01" resolution of the Dwyer digital manometer, nearly half the pipes range on the larger pipes is "wasted". With a better resolution, say 0.001", this range can be extended quite a bit and the number of pipes reduced. I guess that the Flowquik, E-Z Flow, etc. must have a high resolution to get good readings across the ranges they operate at.

When it comes to measuring devices and the primary element (pitot) I have some more questions.

1. Has anyone tried using or making an averaging pitot tube with square cross-sectional area? i.e Square brass tubes, drilled as per the averaging pitot tube on one side and then joined together back-to-back. I've read that a square cross-sectional pitot is more accurate and can read down to lower flows despite being more "restrictive".

2. Has anyone thought of using model RC aircraft electronic airspeed sensors or readouts? Some of their sensors seem quite accurate.

Stef

Posted: **Mon Nov 27, 2006 10:08 pm**

OK....I see what you're up to......that should afford quick range changing..

Posted: **Tue Nov 28, 2006 3:05 pm**

Been doing some more number crunching today. I've been comparing the pitot tube with the orifice plate for flow vs resolution.

I also found a nice differential digital manometer - a Testo 506-3.

This has two ranges - 0 to 4" with 0.004" H2O resolution and 4 to 80" with 0.04" H20 resolution (if I've got the conversion from mbars to inches H2O correct ??? )

Even with this I still need a couple of pipe sizes to avoid too much velocity with the pitot tube.

But, using an orifice plate similar to the E-Z Flow i.e. 4" pipe with 2.5" orifice, It seems I can go from as low as 4cfm to nearly 400cfm with the resolution offered by the Testo manometer. It would be simple to build too.

So what route should I now take? Pitot or orifice - the eternal question :p

From what I've read on here it seems best to keep it simple at first... I'll do some more sketching and maybe play around with some pipe/board...

Stef

Posted: **Wed Nov 29, 2006 3:19 am**

Posted: **Sat Dec 02, 2006 2:28 pm**

I got the conversion right then!

The LFE option I looked at and got the brochures for, but I won't go there. They seem too expensive.

An orifice does indeed change Cd once the pipe Re number goes below 10,000 and so does the uncertaincy percent value of the Cd value. The orifice design I've been looking at, made to ISO 5167 requirements, has a fairly steady Cd value of 0.6 up to Re 10k, this raises to Cd 0.7 at Re 5k. Uncertaincy or accuracy values are normally 0.5% up to Re 10k and 1.25% under Re 10k.
There are other factors that change the Cd value but if the setup is kept within the confines of the ISO or the ASME standards accuracy can be kept to within 0.5%.

Stef

Posted: **Sat Dec 02, 2006 3:11 pm**

Might want to do a little reading on the post related to the plates that were passed around last year and tested on various benches.

Might save yourself a whole lot of engineering (my head is spinning just reading this post, can't imagine machining the parts required and what that would cost), you will find that some of the DIY flowbench designs were just as accurate as the Pro built systems. The goal is repeatability in your design with a level of accuracy you are comfortable with.

Tractorsport Flowbench Forum Archive

New member, comments on my bench.