Okay, thought I'd stop beating my head against this and actually ask for advice. I have a homebrew peripheral port 13B; based off an S5 NA engine if it makes any difference. Sleeves sunk in for the intake are 2" ID 6061 tubes bored out to, off the top of my head, 46mm ID. The holes aren't cut all the way through the housing; end of the tubes is square to what's cut out. In the water jackets is Devcon Plastic Steel liquid.

The problem is the thing keeps leaking water around the sleeves. Nothing when it's first put in the car, then fire it up once, twice, a couple times maybe... and let it cool down and now water's seeping around the sleeve. Redid the filling at least once; ground away a lot of the old stuff looking for cracks or air voids, used a plate bolted onto the side of the housing to put the devcon under air pressure and force it through any cracks, gave it a couple day cure, pressure tested it with that same plate, dye penetrant tested it, bolted it together and pressure tested the coolant jacket on the complete engine... and then fire the thing up, let it cool, and voila, the bastard's leaking again.

What's going on here? I looked up the datasheet for the aluminum filled epoxy, but it looks like the coefficient of thermal expansion is even greater than the steel stuff... should I try some other epoxy? Do something different with surface prep? Any advice would be appreciated.

 

You should have precision machined for a press fit

 

Did. Had to heat the housings up and chill the sleeves to get them in. That was before the epoxy.

 

well, i guess whenever i get mine fired up, i should expect the same then? i wish i had something constructive to advise, but sadly, i don't at this time. however, i had long since made the decision to cut the bore all the way through on my next semi-peripheral (or full peripheral) project and this only serves to cement that decision.

 

I had water leaks but they were always external, not internal.

My tubes were steel (STEEL!) and they didn't even meet the inner aluminum surface of the rotor housing and were nowhere near a press fit, except for the 1/2-13 threaded rod I had in the unused tension bolt hole that I welded them to, and the short Allen bolt that I threaded into what was left of the old intake stud's hole (worked like a set screw). That and a whole lot of JB Weld...

My leakage problems were because of an air pocket in one rotor housing that allowed coolant to seep out around the #2 tube, and the coolant got there because I didn't put the silicone at the iron/aluminum interface at the right height (think sealing a J-bridge).

BTW - You should have made the tubes go all the way through, if only to provide support for the rotor housing. My housings got wasp-waisted at the intake port area, just barely enough to cause the chrome surface to warp a little.

 

Damn! I'm at a loss then. I wish I could help.

 

Yeah, may just chuck them on the mill, bore 'em out, get the devcon out (somehow) and do it again. Bleh. Still, in general, no problems with Devcon plastic steel liquid?

 

forgive the ignorance, but why would you want a lower coefficient?

 

The linear coefficient of thermal expansion for aluminum is around 13 microinches/inch per degree F. Devcon publishes the coefficient of thermal expansion for their epoxies; the steel filled liquid stuff is 38 and the aluminum filled stuff is 50. So either one is going to expand more, but the steel filled stuff is closer to that of aluminum. My thinking is that what might happen if the coefficients are too far off is that when heated, the epoxy would deform under the stress; the added volume has to go somewhere, so it would go the only way not constrained by aluminum, call it outwards... but then when cooled, it shrinks back in all directions, possibly peeling away from the aluminum.

Obviously since neither is exactly the same as aluminum they'd both do it somewhat, but my thinking is that the steel would be better since it's at least closer.

 

My rear port seeped coolant externally for about the first hour of run time. Hasn't leaked since. My sleeves were slip fit 6061, with aluminum Devcon.

 

makes total sense. thanks for the explanation. i went with the aluminum epoxy for mine, but maybe i should have taken the time to compare and contrast everything on the data sheet. i have heard of epoxy peeling away, but i don't know which ones the people that reported it used.

 

Why not try a RTV type of sealant, it could be flexible and it can compress and expand without a problem.

I am just adding a idea, I have the pipe and some rotor housing and a Metal Lath/end mill so i am going to do this sometime in the furture

 

Stupid question, but did you check to make sure the housing is not warped.

 

I was pretty successful with my setup. I used the copper gasket maker silicone at the points where the sleeve touches the housing and made sure I sealed all the way around. Then I filled the area wit JB Weld industro with steel fileings from my buddy's brake rotor turning mixed in. I left a space of about 2mm at the top to make sure it did not have any epoxy touching the wall and then I finished that last 2mm with the same copper gasket maker. So far I have 600 miles of hard driving including daily use with no problems at all.

My stoff was cut using a hacksaw and a holesaw. So if you are milling yours then the results should be much better. Just my 2 Euro Cents. Good luck with your build.

 

Got any pics?

 

Hmm I'm liking the cooper gasket thing your doing. Me, I've been experimenting with some junk housings and actually tig welding a steel sleeve to the chrome inner housing surface. I got the idea from that Rx8 PP air plane engine a few years back. I've been just using 500 degree gasket silicon on the outside. I'll then assemble a bare short block and water pressure test it. It works but I need to build a running engine to see how it does during heat cycling. Oh yea, holesaw works for me too. LOL!

 

I'm in the process of installing some sleeves right now and called up Devcon and asked them what they recommended. They have a special sealant that you put on one side of a part, in this case the outer sleeve and then press fit the sleeve into the housing and the sealant will seal it. Then you would use Devcon alluminum putty to fill the rest of the ports. So it is a two stage process. I'll let you know the outcome. You want the sealant to expand the same rate as the housing. If you use a sealant that expands at different rates, then the sealant will just crack and fail, thats why most guys use the Devcon alluminum putty to fill the waterjackets and not the steel putty or other.

 

Sorry. No pics. I was so excited to see it running I completely blew it. lol. But I will def open around christmas so Ill take some pisc then.

 

Where's the video?

 

I'll work on getting one up this weekend.

 

On all the semi p-ports I have put together so far, our machinist does a 3 step o.d. sleeve. Then hits it with liquid nitrogen and presses it into the rotor housing. I then use araldite epoxy, wait to dry 2 days and finish it up with Jb weld. Its VERY important to remember aluminum will oxidize very fast, so you have to use a sanding bit to totally sand the surfaces clean and very quickly put on the epoxy. Then re-sand for the JB weld. This technique has had 0% leaks internally or externally for years. If that helps anyone trying out p-ports

 

Thanks for the input Gto, I just freshened up the motor and still had no leaks after years and about 10k miles of abuse. I know I missed it on the video but here is a link now.

https://distilleryimage8.s3.amazonaw...a819f6_101.mp4

https://www.facebook.com/photo.php?v=10201486719835797

changed turbo from track video to a GT3576 and thats how it sounds on the street now.