Flow bench development on 16v Cosworth head

Competition engines and ancillaries - general discussion
Post Reply
jmcpower
Posts: 19
Joined: June 23rd, 2006, 8:59 pm
Location: Cyprus
Contact:

Flow bench development on 16v Cosworth head

Post by jmcpower »

Hi Guy and everybody,

I have currently purchased a flowbench from Superflow - the SF120E along with Flow Comp with motor control, flow bench swirl meter, 4 valve universal opener with dial gauge, exhaust velocity pitot tube, inlet velocity pitot tube, head adaptor for cylinder simulation!

In the past I have been porting cylinder heads (still do), for racing engines based on the information From Guy's Fiat Lancia Twin Cam book, (The Bible as I refer to it!), and I had very good results, but as Guy would agree there is a point that you need the right equipment for the development of such a serious and high standard job!

Its the first time that I have been using a flow bench and I must admit that Guy's tips were a great help to get going!!

My 1st project Job is a Cosworth head (currently standard) and after sand blasting and cleaning thoroughly, I was ready to do some flow testing with the bare head at 10" Test depression.

To my surprise the max bare head flow from a standard Cosworth head is only 117.5 for the inlet ports and 89.6 cfm for the exhaust - 76% compared to the inlet flow. Has anybody any flow tests for the bare standard head for comparison ?
I have also measured on all tests the velocity in the inlet port at half way into the port about an inch before the valve guide and I will try and see the difference during the porting steps! The velocity test was performed with a pitot tube at nine different points on the cross sectional area on the port one inch before the guide. I will try and attach the files with the 4 different tests.
First was a bare standard head test. Then the first step (Test Bare head 1) was porting (with 50 grit) the inlet ports 1 mm oversize up to the valve guide. After this modification the results were 119.4cfm. The second step was another 1mm of porting up to the valve guide and the results were again surprising ¢‚¬Å“ only 119.6 (test 2) and indicating that there was a restriction on the other half of the port towards the valve seat and short radius.

Conclusion from the 1st two test is that Cosworth had done the math correctly...the first half of the inlet port flow is well balanced compared to the other half of the port towards the short radius and valve seats. Opening up just the first part of the ports will have no effect on the flow since there will be a restriction on the other half.

Third step was 1mm porting on the other half of the port with a minor ‹Å“sorting out¢ž¢ on the short radius - the result was a big increase - 136.5cfm !!

My next step will be the splitter vane and then with the valves in, so I can see what is happening. I will try and measure the velocity on the short and long radius and compare.
I know there is still a lot to get from this head but I would rather do it step by step slowly so I can learn from it than just get carried away and rush things just to get the big numbers.

Guy, I would like to ask what do you use for the initial quick opening of the port diameter....I used the Ata bands 50 grid but it takes time !! Any recommendations?

I find difficulty to attach the test files ....they are converted to ASCII files from the original Port Flow analyzer Pro software.

Best regards,

Mike

JMC Power Automotive, Cyprus
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Post by Guy Croft »

hullo Mike,

Last standard Sierra Cosworth 16v head I did flowed 114.7cfm at 10" test depression with guides removed. One might expect some variations from head to head, this is reasonable. More important than the absolute figure is the difference ie: improvement. So I say trust your measurements and look for gains. And look out for losses!

That head had about 23mm diameter thru the splitter, enlarging that to 23.7mm, knife-edging the splitter vane at entry and thinning it a bit plus all over smoothing at 80 grit gave me 128.1 cfm, ie: plus 12% bare port flow. This was a low budget job so I left it at that. I only smoothed the short side radius, I was not engaged to do the valve seats, but I know they will be done properly at another firm (friends of mine). I think that the flow with valve in will be similarly better.

When you're doing a multivalve head remember these are the main flow control points.

1. Choke point no 1 - head outer face. The entry section. Is it big enough? Making a radiused thin steel or alloy template of inlet gasket size will soon tell you.
2. Choke point no 2 - thru the splitter. Again, test an orifice with radiused entry and of splitter diameter and see what it flows.
3. Choke point no3 - valve throat. Usually plenty big enough but again, test as (2).

Essentially the measurements will tell you that the entry section could flow, say just for argument's sake it's 180 cfm.
Now massflow (kg/sec) = air density x cross sectional area x velocity, and massflow per second going in must equal that coming out of the valve throat, you don't lose mass, and that much air will not exit the valve throat because of the reduction in velocity due to reduction of cross sectional area and viscous (drag) losses and turbulent flow losses deeper in the head. And how big you can go inside is more often than not governed by casting thickness. The velocity is not a constant across the whole port, anywhere; there are regions of high velocity (low regional pressure and good flow) and low velocity (poor flow). The massflow is just the sum of these.

In between the choke points there are losses due to imperfect surface contour (bumps and changes of section) and surface roughness. I have not found any more flow by going from oil-wetted 80 grit to 120 grit (or finer) although 80 grit can look a bit rough to the untrained eye. Around the bend at the short side radius the loss is quite high (though less than a sidedraft head with a sharp turn) and you cannot really get rid of it, only reduce it, but that region must be developed with valve-in tests after every mod or your low-medium flow could end up hopeless. Yes, you can and should - treat the region outboard of the valve throat (and short side radius) as one entity and modify it in isolation. Then move onto the short side rad and valve throat and see if you can optimise the flow behaviour with the valve fitted and make use of your outer region flow gains. You're sure to lose some because of the discharge loss around the valve.

On the exhaust side you will realise that the E/I flow ratio changes as you improve the inlet - and you've got to be very modest how much metal you take out of the ex ports on that head or you'll break through. It's a clever casting, but there are water galleries everywhere.

The flow test of a full spec rcae head shown below might help you, note how the final E/I, depending on inlet cam lift - can be as low as 63%. High pumping loss at that low level, in normally aspirated trim anyhow.

GC
Attachments
Sierra Cosworth full spec flows.JPG
Sierra Cosworth full spec flows.JPG (102.16 KiB) Viewed 6388 times
Post Reply

Who is online

Users browsing this forum: No registered users and 29 guests