I really want to try this method of building a pipe.  It looks easy - but I am sure there is some learned skill that needs to be applied here.

I don't believe my Rosebud torch or regulators can flow enough heat to get this done......maybe if I borrowed a friend and his torch we could use 2 torches to get the heat applied.  The pipe must be filled with dry sand to prevent the pipe from kinking at the bend.....the sand must be dry so that the heat doesn't get dissipated into steam.


http://www.youtube.com/watch?v=v8mMQbEkr8w

http://www.youtube.com/watch?v=vyM65MKxDNI

http://www.youtube.com/watch?v=E0qiLDFWQS8

If I can find a local welding/fabrication shop that already has a big torch......maybe I can get them to help me bend the pipe.

DragBikeMike wrote on 12/01/23 at 19:09:19:


………

There are no 97’s left.  Getting more requires a minimum order of 12, which I have no plans to do unless someone wants to order at least 8.  Selling the initial batch was a long process, just not much interest.

DBM did you the get the PM I sent you ?

Guys,
I admit to loosing patience with that first video, but heating it for over, was it 4 minutes?, before bending it, for what it is wortk, it might as well be a solid rod, certainly not thin wall tubing, cos that would have been red hot long before , even with an air assisted propane torch.
The sand filled method is doable. I know because I did a VW beach buggy exhaust back in 1976 in Mombasa. Takes some technique, though. You have to start the bend from both ends of it, doing only 1/3 of it in each end, then finish it off in the middle of the bend last.
That produces a passable bend, but still not round.
The last video, well, was that really thin wall tubing, I think not.

Plumbers have a way to bend copper tubing, and that is inserting a spring into the tube. The spring must be made from square section steel, or if made from round steel, it must be ground down by 1/2 of the spring wires diameter, so that in each case it presents a smooth surface out to the tube. Combined with quality dies to support the internal of the bent (and some superiour lubrication) this method might just cut it. And perhaps this is the way to bend the double walled original header.
One final method is called mandrel bending, that is the industry standard, and the method used in the manufacture of the part I have ordered:
https://www.ebay.com/itm/385110366508

We will see soon enough if it is usable.

Cheers, Finn Hammer

After rounding the corners with the router, and opening up to the plenum gallery, I slapped some paint on the thing, just to make it look a bit nice, after all.
The point here is to look at this part of the bench as equivalent to the cast iron work on a lathe or a miling machine. Not very smooth, not very accurate, but sturdy and sufficient to support the parts that matter: The cylinder support up top, the orifice plates and the door into them, in the middle.

I am planning on these orifice plates, measuring at 8" differential
CFM      Dia.
10-20      0.72"
15-30      0.88"
25-50      1.14"
40-80      1.44"
65-130     1.84
100-200   2.28"
150-300   2.80"

This should give me a  good overlap, and some ability to check them against each other for fine calibration.

Man, I'm gonna have a flow bench! that is really exiting.

Cheers, Finn Hammer

Mike,

I intend to test at 28", just because I can, but I will definitely also test at 15" as you have requested. Perhaps even also do it on a regular basis, since keeping numbers comparable with yours seems like a good idea.

The 8" is the differential pressure across the measuring orifice, as monitored by the inclined manometer.

When it comes to calibration of the plates, the proper way would be to take at least one plate to a headwork shop, and have it calibrated on an already accurate bench. This is not possible for me here in Denmark, so I will have to make do with calculated hole dimensions, and hope they produce remotely accurate results.
When I said calibrate, I mean to use one of the plates I make, to be the local standard, and then adjust the other plates to deliver identical numbers in the overlapping area.
Let's take the 40-80 CFM plate as the standard.
It should not be operated beyond 90% of full bore, so it is really a 40-72 CFM orifice.
That is why I chose 65-130 (65-117)CFM as my next step, so that there would be 7 overlapping CFM's to compare. No, I did not read about that anywhere, just seemed to be common sense.
I will measure by "First Principle", that means using water columns as suggested in the Dtech article you recommended in your own bench build thread. There are excell solutions to make the scale, havent yet looked into it in detail, but will have to quite soon.
In the end, I will say, that the bench here is for Savage only, and really also just to record any beneficial changes I may make to the head.
If we were to modify big block chavy heads and the like, we might have to produce comparable numbers, but who knows, yours and my heads may be the only savage heads to be flow tested.

Below is a picture of the top fixture for the head, with mock cylinder protruding into the bench, o-ring seal to the gasket surface, 12mm hollow dowels to align the head, and 8mm studs to tighten the head down with.


Edit.
I would like to add that the litterature I have read about flow benches is the Dtec design guide:
http://dtec.net.au/Flowbench%20Design%20Guide.htm
I have also found some good things in the forum here:
https://flowbenchtech.com/
Finally, I skimmed David Vizards: How to port and flow test cylinder heads.
That's it.
On the use of manometer gauges: The best ones are MKS Baratrons, they are quite expensive, and the only ones I would trust to below one percent.
But starting to use gauges would expand the amount of work involved to a level where the flow bench itself becomes the project.
It would be possible to derive a PWM signal for the motors to arrive at the desired test pressure automatically, and ever so many other nice features, but:
By using water manometers I can keep the bench simple, which is important to me,.

Cheers, Finn Hammer

I hope to be able to blow through the exhaust, but I need to get the turbines hooked up to the bench first, and get a feel for how many motors I need, and how well I can handle them with the variable speed thyristor drivers. With that data in, I will build a reversible box like dtec did.
Odd, I thought those bypass valves would be good. But when I lower the rpm with the thyristors, less heat too, so perhaps just small leaks for fine tuning test pressure.

Cannot really say more before that.

Cheers, Finn Hammer

Edit, I think I misunderstood the part of the snubber valve, will check up on it. ch12305]

Wow, this took a lot longer to finish than at first antissipated, but the flowbench is now operative, although not fully (like: not at all ) calibrated.Anyway, here is a short walktrough of the device:
Here you see the powerhouse, 4 vacuum cleaner motors pulling together, speed adjusted by a SCR based motor regulator. This works very well.