One could use a rotary table, rotary cross slide table, or a rotary head machine. All really show you how nice CNC machines are.

Love the rear CW!

Pure rotary hotness! do they have vids?

youtube has some vids of their 4 rotors running on that test stand.

The straight edges are machined, but I used a vertical beltsander for the round edge. Doesn't need to be precision because it doesn't touch anything. Got it pretty good though.

They are nice, I want one bad.


Update

Finished the front counterweight

https://lh3.googleusercontent.com/-c...2/CIMG1550.JPG


Finished the rear counterweight. I mounted the 4-rotor shaft in the lathe, and made sure the runout was under 0.0003". Then I mounted the counterweight and machined the oil seal surface and the flywheel mating surfaces.

https://lh6.googleusercontent.com/-e...2/CIMG1540.JPG


Also modified my OS giken flywheel a bit, just removed some weight.

https://lh6.googleusercontent.com/-6...2/CIMG1548.JPG

So I've been trying to design my radiator. Already got a aluminium radiator core, some 1050A sheet metal, and 2 12" cooling fans. Now I need to figure out how I'm going to design and fabricate the end tanks.

Here is the first design I made. I'm going with the endtanks on the side, because else the coolant hoses will interfere with the hood vent.

https://lh3.googleusercontent.com/-l...werp%25201.jpg


Also did some simple flow analysis

https://lh6.googleusercontent.com/-h...281%2529_1.jpg

The colours represent coolant velocity, this is assuming the waterpump can pump about 125 l/min. I don't know if it pumps that much, but won't matter because this analysis is just to compare different designs. It's clearly visible that the water has to flow through a pretty sharp bend, which doesn't do good things for flow. This causes a few parts of the radiator core to be pretty ineffective. Also the pressure differential over the radiator was 4.5 psi. So I looked at some serious racing radiators, and almost all of them have something like this (look at the lower part, the upper part is the oilcooler)

http://www.vprace.com/carsales/lolaf..._radiator3.jpg

Instead of a sharp bend they put in a piece that smooths out the flow.
So I designed a new radiator:

https://lh6.googleusercontent.com/-8...werp%25202.jpg

And checked the coolant flow

https://lh4.googleusercontent.com/-Y...281%2529_2.jpg

Problem is still there, it's better than before but I was hoping it would be more effective. Pressure difference is reduced from 4.5 to 3 psi and flow over the radiator core looks a bit better, but I'm going to fiddle with it a bit more.

Now that I read upto this point after 2 hours my eyes are burning, Mad skillz, R&D and countless hours of hard work are going to pay off soon enough! Keep it rotating ! Josh

What happens if you put the ports on the edge of the tank instead of the face?
Alternatively, leave the bungs where they are but angle them.

I have no grip on the math but I wonder:
Given that there's a hard defined envelope for the radiator, you seem to have opted for the largest possible core, leaving minimal room for the tanks/ports.
The end tank size and placement results in penalties in flow and pressure.

If the end tanks were increased in size to allow less disruption, the core would have to be reduced.
There must be a crossover point in overall performance, where the increase in one factor is balanced by the decrease in the other.

In other words, does improving the intake/outflow by enlarging the end tanks and decreasing the core work better than the reverse?

along those lines the BMW radiator in the pic has the radiator elbows straight out the sides, that probably won't fit, but maybe as it gets closer to straight flow gets better?

since there is more engine, it might actually matter if the radiator has some pressure drop, but really who knows?

Can't you make it that you have a corner (or just cornered, so the flow is better. Example 45° and then another 45°.

Did I miss the results you got back from getting the rotors themselves balanced? Was curious about that.

As many have said holy hell this is amazing! Very VERY exceptional work.

Look forward to watching it.

Don't think I'll have room for angled bungs. The end tanks really aren't that small, their 45mm's wide, which is wider than with most radiators. I did want the core a bit smaller though, but the size I wanted wasn't standard, so it had to be custom made which was expensive. This one is 20mm's wider than I wanted, but it was half the money and will do. I can probably modify the chassis to give me a bit more room at the top end tank if needed.



Yes, I also think it would be better, but there just isn't any room. There is always a pressure drop over the radiator, or else there wouldn't be any flow through it. But I think reducing that pressure drop while keeping coolant flow is good.


Nope, the rotors haven't come back yet, I'm hoping somewhere later this week.




Did some more testing with different radiator end-tank designs. After a few attempts I came up with this:

https://lh5.googleusercontent.com/-S...werp%25203.jpg

The coolant entry pipe is at a slight angle, and the bottom side of the end tank has an angle in it. This should redirect the coolant flow and eliminate most of the blind spots.

This is how the coolant flow looks like:

https://lh3.googleusercontent.com/-G...281%2529_3.jpg


Coolant flow velocity in different core ports

https://lh3.googleusercontent.com/-q...281%2529_6.jpg


Don't think it's any better though, Dammit, I really though this would work well. I wonder if this simulation really corresponds with reality. There are a lot of aluminium radiators out there that are designed a lot more poorly than this one, and they seem to be working ok. Ah well, I still got a few things I want to try, so I'll do another attempt.

yes. the stock radiator makes a much worse bend, and its um stock!

its still born, but my friend copied an endtank like one we saw in racecar engineering, its like that BMW one, but it gradually curves. kind of like this one, http://m92.photobucket.com/image/int...g.html?src=www

granted its an IC and not a radiator...

How about a radius on the back of the radiator?

Attachment 717496

went thru this whole thing and was hoping it was done the closer i got to the end, but i cant wait to see it done. sick build bro

how PPRE are doing it. (as I said before) duel, huge rad's single pass.

Have you thought about adding internal vanes in the endtank to help direct the coolant to flow where you want it? Something like this:

mmh good idea YellowT2, but don't add to much of you will have more vortexen then you want. So John, more data to render :lol:

what is that going in to?

I saw that on facebook. Very cool, but like comparing sortoff apples and oranges. That "engine bay" is the rear part of a lamborghini. So basically a chassis that can use pretty much all the power all the time, and little incoming air to cool it, so yeah they need to go big with cooling. Needed cooling isn't linked to the amount of rotors but to the hp your making, so fc's with a decent turbo conversion are probably better comparing material.


Good idea, but hurts flow though. I might do something like that if I can't find a decent design.



Did some more testing today. At first I followed tegheim's roundshaped idea, and came up with this:

https://lh4.googleusercontent.com/-I...werp%25204.jpg

https://lh5.googleusercontent.com/-d...281%2529_7.jpg

Still the same problem, thank god!, that was going to be a nightmare to fabricate! The basic problem is that there is a lot of water coming into the radiator. I don't know exactly how many l/min the S4 waterpump churns out, but electric pump's are known to pump about 115l/min without resistance, and about 80 - 90l/min when installed. I think an S4 oem pump can pump more when the rpm is spinning at normal - higher rpm's, so I used 125 l/min. 125 l/min is 2.08 l/sec, with a 35mm ID coolant entry pipe flow velocity is (2.08/1000) / (PI/4*0.035^2) = 2.16 m/sec, just like the flow analysis numbers say, That's pretty quick! So basically the speed and inertia of the coolant makes it not want to change direction which explains the blue swirling and all my attempts failing. So I had to think at a different angle, which I litterally did

https://lh4.googleusercontent.com/-z...werp%25205.jpg

And the flow patterns:

https://lh3.googleusercontent.com/-E...281%2529_9.jpg

Still some swirling, I think it's because the end tank has gotten too big. The coolant is divided better over the entire core though, and pressure drop is down to 0.18bar. It's also reasonably doable to fabricate. I might mess around with making the endtank smaller but we might have a winner :nod:

Goddamn that is awesome...

Why not increase the entry pipe?

Who would have thought that building the radiator was such a trial and error process? This is why im going to RIT in the fall for Mechanical Engineering Technology cause I love this technical stuff!!

You dont need to do inner radius on my example. Just do an ordinary, and round up the back of it.
Well, though you have a little intake, and it will get bigger very fast, the swirl will be there.
I would go for a thicker, but less complicated design. The flow-problem will always be there, but maybe its enough to cool 1000hp anyway?
Hot-spots in the radiator aren't that bad. Worse if its in the engine :)

This is an amzing build. such amazing Fab Skills. i know everyones said that already, but DAMN!!!! Can't wait to see this build unfold!

Stay Sideways my friend!