Intake Design and Building: How to?

ProDarwin said:

I meant these 2 parts:

 

Also, send me a CAD file of some sort and I'll be happy to take a detailed look

Oh those.  Just an external corner that looks kind of weird, it didn't fillet cleanly.  What email and what file type you want?

This is looking awesome. You can seriously cast something like that DIY? I'd love to see your setup. 

Mr_Asa said:

ProDarwin said:

I meant these 2 parts:

 

Also, send me a CAD file of some sort and I'll be happy to take a detailed look

Oh those.  Just an external corner that looks kind of weird, it didn't fillet cleanly.  What email and what file type you want?

my screen name at the google mail

STP/STEP 214 preferred

maschinenbau said:

This is looking awesome. You can seriously cast something like that DIY? I'd love to see your setup. 

Maybe?  The stuff I've personally prepped and cast before is usually something that can fit in one hand, but I've helped setup and done the pouring for larger pieces approximately this size.  I do most of my casting at USF with the guy that runs the ceramics studio

In reply to ProDarwin :

Email on the way.

I'm not used to sand casting, but assuming the principles of die casting/forging apply, I have some questions/concerns.

From an air intake design standpoint I am also not an expert, but there seem to be a few unnecessary transitions on the inside.

From a cad standpoint there are some other issues, but some of that is just me nit-picking stuff :)  

If you want to discuss over zoom, let me know.  If you want me to leave all my comments here and turn this into a design review, that works too

I was planning on doing a shell casting, not sand.

Lets just do a review here, more people that look at it and comment the better off the design will be

Sounds good.  I'll dive into it in some detail in a couple hours.  Have real work to do in the meantime :)

Ok here goes.  Apologies in advance for sloppy mouse annotations.

Apologies if any of this comes off as aggressive.  This is all intended to be constructive criticism and I hope it helps.

Item 1:  This is a section through the approximate center of the runners (it was hard to find exact center).  Note these tow corner pockets that are not needed.

On the left side, would it make more sense to have the "end" wall of the manifold shared with the runner (shown in red).

On the right side I feel like it would make sense to do the same thing.  Even if it doesn't help the flow stay attached, it would be less material (weight), and eliminate stress concentrations from the thin sections, and eliminate the oddly large thick chunk of material right in the middle.  (shown in blue)

Item 2:  This is a zoomed up section on just the runner entrance,  Its hard to tell at first glance, but these walls are kind of all over the place.  Notice the messy chunk on the left side.  I'm assuming you want a consistent draft on them?  Would it also make sense to do some G2 blends as opposed to a simple radius here (same with the blue line above too)?  Its been a while since I have taken fluids, but maybe someone who ports heads can shed some light on this.  See curvature combs below.  Note that they aren't symmetric, but they should be.

 ITEM 3:  This is a view looking in from the TB into runner #1.  It kind of messy.  Note the big lump on the LH side, also mentioned in item #1.  Also lots of little imperfections right at the TB inlet.

Another angle:

Another angle:

Item 4:  This is aview looking down a runner into the head.  Note where the injector port comes through it isn't very clean.  There is a bump of material that can disturb flow.

Here is a different runner:

This one has a chunk on the opposite side as well:

Also, are all the ports supposed to be leaning slightly in one direction?

A lot of what I've run into with this are the limitations with the existing sketches and Solidworks not wanting to play nicely.  I had to redo the lofted extrusion and lofted cuts for the runners four or five times to get them to all look somewhat like the same.  (Fig 1)  That being said, #1 is definitely not the same as the others, which led to other issues later on.

The first pic: on the left I played around with it a bit and it didn't give me a clean fillet when I went to make that rear-most wall flush with the runner.  Similar issues with the "lump" on the front, minimum wall thickness when I first ran the loft from the throttle to the base (See Fig. 2) was less than a sixteenth so the lump was a work around.  I can think of a few ways to fix that one.  Making the wall flush with the runner might be possible.  I'll try it out.

Wall at the base of the loft is 3/8, so it should give a solid thickness, but it needs some love.

Second Pic: Blending the edges of the runners into the body of the intake was also a pain due to the problems with the previous sketches and Solidworks being Solidworks

The radius for the fillets are the same, but they didn't want to select all on the same operation, so I needed to do some mild edits in between and then go back and do the next fillets (check Fig 3 & 4)

As for the curvature combs, the line section itself is different there which is why the curve is just slightly different.  See the comments above on the lofted cuts.  I haven't messed with many other fillet types other than a simple radius.

Third-Fifth pics: Pure brain fart.  If you go back through my posts you'll probably see that I was working on these pretty damned late, inherent night owl tendencies coupled with unemployment leads to me being up way past when I should be.  I did the lofts seen in Fig 1 before I found a model of a throttle body, so I just went with the same diameter as the base body instead of measuring the throttle body diameter and going from there.  It being pointed out, I see where I can fix that, and it'll be pretty easy.

Sixth-Eighth: Similar to the throttle body.  Issues I saw and forgot to return to, as well as me knowing that I'd have to do some clean up with a grinder after casting so I wasn't 100% concerned with.

Ninth: Yup, they are supposed to lean over. Chevy was either on some high grade E36 M3 or the flow data was exceptionally good that day.  The height difference from the left side of one port to the right is part of the reason that those curvature combs look weird.

Fig 1 - Note: Pay attention to the lines showing where the cuts were.  This is what was in there now.  I'll go into it again and see if I can get them to match up, but...

Fig 2

Fig 3 

Fig 4

Fig. Solidworks User

I haven't used SW in a long long time.  NX can do all these things well, but its definitely not 'easy'.  I will say, if you can make it do these things right, it elevates you above all your peers in the engineering world :)

Why is the #4 runner here different than the other 5?  You loft the extrusion once, then array it, then unite it, correct?  That should make all of them identical.

Remember how I cut the head up?  Means that there is a different gap between cyl 2&3 and 4&5, so any sort of linear pattern doesn't work.  As easy to just do each one individually than to figure the exactly correct spacing and then do three sets of linear patterns.

Anything worth doing is worth it being a pain in the ass.  That's the phrase, right?

Three sets of linear patterns is just 2 patterns.  1 for the "pitch" between heads and 1 for the "pitch" between runners :)

I always hate doing big models like this after not doing big ones for so long.  So many dependencies that get berkeleyed up when you go to do big changes.

The red and blue areas above that were suggested to be "cut off" I would actually make larger (maybe double as big as they are now)

Paul_VR6 (Forum Supporter) said:

The red and blue areas above that were suggested to be "cut off" I would actually make larger (maybe double as big as they are now)

Interesting.  What is the reason for that?

I've always wondered how you cast something hollow like an intake manifold. Do you use the slush casting process with your shell molds? Like where you pour in the metal, let it solidify for a bit, then pour out the rest? How do you time it to get an accurate thickness?

In reply to maschinenbau :

You can do a traditional cope and drag with a sand core (great example here, the one thing it looks like he does that I don't recommend is letting the weight of the top half of the mold be the only thing holding it down), you can do an investment casting with a large enough plaster mold, or you can do what I'm planning and do a shell casting.  Each of those has two or three methods that will work. 
The way I'm planning on doing it: dip in a ceramic slurry, let it dry 12-24 hours, repeat two-three times, dip in ceramic slurry, shake fine grain sand over everything, let dry 12-24 hours, repeat 5-6 times, dip in slurry, shake coarse sand over everything, let dry, repeat ~10 times.  Once the final coat has dried, you put it in a furnace to burn out the plastic and cure the ceramic, then while its still hot you pour the metal.

Slush casting probably wouldn't be a good choice as you would have very rough interior surfaces.  

Casting Thread for more questions so we don't derail this too bad?

ProDarwin said:

Paul_VR6 (Forum Supporter) said:

The red and blue areas above that were suggested to be "cut off" I would actually make larger (maybe double as big as they are now)

Interesting.  What is the reason for that?

You can choke or over flow a runner by having a wall too close. You would have to so some complex modeling to really figure it out but for the most even air distribution having pretty slow/even bulk flow is good. That way each port can  flow from the plenum evenly. 

I was thinking that boundary layers would be responsible for it.

Not sure about you guys, but there's always a very subtle disconnect in my head between sizes on CAD and sizes in real world.  Maybe because my CAD screen is too small for it to be full size most of the time?  I know "that hole is 1/2" and that runner is 3" long"  but then I hold the part and I think "berkeley, this thing is bigger than I thought."

Do the injector bungs sit at an angle? How does the rail fit?

Yup, they're angled to spray at the valve.  Fuel rail bolts to those bosses just behind the injectors.

Gluing session.  Fumes are fun.
I'm going to need to do some post-processing after this, the printer is definitely not liking the strain I'm putting it through.

Planning on machining the rail with the bosses at an angle too? Not sure the LS bothers with that.