Just was wondering how a flow bench works, and can one be made to used in someones shed, using parts just laying around? Any thoughts? Jmac's opinions are welcomed!!!!

Just google DIY flow bench, there is stacks of info about it!

Try here for a starter:

http://www.dtec.net.au/Flowbench%20Design%20Guide.htm

you can definitely diy this - you'll be looking at a few hundred bucks for enough fan motors (vacuum cleaner motors on a larger scale essentially) and some relays - to pull enough vacuum/flow to do cylinder heads (you'd only need to get to about 400cfm for most stuff, 600cfm would see you right for even decent big v8 race gear,

You can power it from 240v. Believe it or not, there's more than a few head porters out there who use flowbenches they made themselves. There's nothing particularly hard to duplicate/equal about some of the commercially available flowbenches either, so there's no reason not to diy one if you want one. Even the best flowbenches all vary a few cfm from one to another - really no big deal - all you need is one that is consistent to itself - so even if it reads 10cfm high or low it doesn't matter as long as it's always that way, you can then track the improvements (or occasionally find you've gone backwards with a new attempt!)

ssskiwi's link is a good one, it pretty much spells it out. If you want I've got some jpg files I can email you which were scans of a magazine article (essentially it's the same idea, as are all flowbenches generally) of how to diy one.

As I said, it'll be a good few hundred for the fan motors and relays. Making the box/enclosure is pretty straightforward basic carpentry really - no huge costs or dramas there.

There's also a forum:

http://www.tractorsport.com/cgi-bin/forum/cgi-bin/ikonboard.cgi

which moved to :

http://www.flowbenchtech.com/forum/

I've never actually built one, but it is one of those things I want to do. I could make most of it off the top of my head, without any thought at all (once you get your head wrapped around how it works, it's pretty simple to build) About the only thing I couldn't do off the top of my head, is the comparative (the angled) manometer scale (which is non linear) and calibrating it, but thankfully that information is freely available if you check out the links.

Have a look at the stuff posted. Here's a pic (I forget which website I grabbed it from, if it's a copyright issue, let me know I'll take it down immediately)




That's a pretty good all in one pic. basically you set the head up with a valve open etc, start up the vacuum motor. Using a controlled bleed you adjust the suction, until the first manometer (vacuum guage) shows 28inches of water (or whatever pressure drop you choose). Then the air flows down through a cylinder attachment (to simulate hte bore of the engine, which obstructs flow a little bit - beware of porters who flow test using something like a 4 inch common V8 sized sleeve, it will inflate the readings big time for small engine head development). Ok then it is flowed through a flat plate (around 5mm thick or so) with a specific sized hole drilled in it - extremely accurately and with sharp 90 degree edges, no radiused or smooth edges. There's known data as to what a given hole will flow with those edges, so you have a few different sized holes you can shift into that spot for the flow to go through. Then you run a manometer with one 'end' of the tube each side of that hole (the fixed orifice) , and by knowing the flow of that orifice, well the difference in vacuum from one side of it to the other will alter according to how much the head flows at that valve lift, by the time you adjust it to that pressure drop, so you then read off the manoeter that is on the angle, and it'll have figures that are percentages. So if it flows as much as the hole it'll read 100% (if it flows more, it'll read 100% too, sort of, so if you had that occur in real life, you'd re-test it with the next size hole up from that). So you look at the cfm the fixed orifice can flow, then multiply it by the % on the manometer and you have the port flow at that lift. You then switch it off, open the valve more (using a rocker and a block to adjust the valve lift, with a dial indicator to measure lift etc) to the next lift (most people test every .05 or 0.1 inches, preferably the former) and re-test, and then you have basically built up a chart of what the head flows at each valve lift. Simple, just takes a little time to do it. But can be a lot of fun.

You can also reverse the flow to test the exhaust in the direction it normally flows (that's another thing that surprises people - ports don't flow the same amount in different directions! so it makes sense to test them the way they flow).

To get 'real world' applicable figures - as mentioned you need a cylinder sleeve attachement under the head that is the same bore size as the intended engine it is going on, and about 1.5-2 times the stroke in length (I dunno all the reasons why, but in practice that works out very well). additionally, you can do stuff like testing the intake with the manifold attached (you often get less flow with the manifold on there, but developing them as a pair yields better results than separately even if the peak numbers are lower for the combined test). Ximilarly - to get real world applicable figures, people fit a length of exhaust tubing, the same diameter as the extractor primary tube to the head for testing. Just make a single flange and length of pipe - somewhere around 12 inches should be enough to properly measure flow in a meaningful way fr the exhaust.

If you had unlimited funds, modern tweaks on flowbenches include attachments that allow liquid droplets (simulating fuel) into the airflow that are also luminescent, and with uv lighting and high speed film, one can watch how the thing will flow,with the fuel, and help design port shapes that get the best mixture suspension (on stuff like nascar this would be a huge deal I would imagine). Obviously the bench needs special attachments to add the liquid then filter it back out without shorting out the motors or losing the liquid.

Another option is velocity and (I forget the word, but basically a measure of turbulence vs straight/uniform flow) and other probes that can be fed into the port to find the high and low flow areas within the port (on something like a 4v cleveland head, you can fill a lot of the port floor, getting rid of practically zero flow dead areas, it doesn't show upp as a big flow increase on the flowbenchm which only shows total flow, but the improvement in velocity and thrttle response, and how low the engine comes on the cam (with no loss of top end) is practcally night and day.

The other thing they are doing is trying higher and higher pressure drops, to try and find even the minutest gains here and there. And lastly I've heard of (but never laid eyes on, nor have I seen any results from) people experimenting with rotary valves within the airflow so they could open and close the vacuum source/fan suction quickly on and off on and off to simulate the stop start nature of real intake flow. That'd be a nightmare to try and diy at home of course, and it's the sort of thing probably only used at formula one budget levels.

This post came courtesy of my insomnia. My keyboard has made two attempts to get a divorce, citing slave labour laws :)