Over my ten years building wbx’s commercially, it took to my 45th engine before I had one develop a fire ring leak* . Fire rings, or combustion seals, are multi-layer steel (MLS) rings that seal the cylinder sleeves to the head. These normally install dry and depend on the continuous tension of the 8 surrounding head-studs to maintain the seal that keeps combustion gases in (hence “fire ring”) and keeps liquid coolant out (the green cyl top o-rings doing this job or aiding it in anyway is a fantasy, they really don’t provide any necessary function).
.
These seals do very hard duty. The burning gases can reach almost 2000ºF, pressures often exceed 800psi, and at that pressure there’s force as high as 6 tons as often as 40 times per second in each cylinder, alternating between the two sides of each head, trying to separate these clamped seals while the hot high-pressure gases try to force their way out. The seals also have to allow a bit of shear or “creep” caused by differential metal expansion rates between the cast iron cylinders and aluminum head (aluminum expands with temperature at twice the rate of ferrous metals). It’s a tough job, and naturally with so many variables it doesn’t always prove dependable (frankly, I’m amazed it ever works at all).
.
So I decided some extra measures were warranted to avoid the problem. I began treating the fire rings with Permatex Ultra-Copper spray, which is specified for composite head gaskets which incorporate MLS fire rings just like those used separately in the wbx. Over the next couple years I had yet another fire ring leak under warranty, so I looked for something better. I found out that Hylomar, which I had used a bit in the distant past (as an anti-squeal to coat the back side of disc brake pads, for which it works very well) , was specifically developed as a non-setting polymer to seal Rolls-Royce jet engine casings, which have extensive flat-machined no-gasket junctures that undergo extreme thermal expansion and contraction, at temperatures comparable to those that metals would reach in a cylinder head. Since the polymer never hardens, it’s always able to creep between the surfaces and allow differential expansion while maintaining a seal. They claim these parts can actually be disassembled and reassembled without a new sealant treatment, it stays soft and sticky basically forever. Pretty amazing stuff.
.
However, you don’t just gob this stuff on (not that you do that with any sealant, right?). To work as advertised it must be applied first to both joining surfaces, allowed to air dry for a few minutes, then the surfaces can be joined, partially torqued, allowing a little while for the polymer to flow, then brought to final torque.
.
The problem with classic Hylomar in a tube is its workability is simply awful, it uses an acetone vehicle which evaporates very quickly leaving the stiff gooey polymer behind, and it’s nearly impossible to spread out in anything like an even layer, even if you work really fast. It’s like a bright blue version of gorilla snot (aka 3M weatherstrip adhesive), very stiff, very sticky (it never loses tack).
.
Then I found out that there are some variants to original Hylomar Universal Blue in a tube, including a spray, and the AF type (Advanced Formula: anything in a tube, toothpaste, foot cream, whatever, has to have an Advanced Formula), which comes in a tube but promises better workability. It’s somewhat better, it’s still stiff bitchy stuff to deal with, but the AF doesn’t flash off as fast so it does allow a little more work time.
.
The spray would be great for speed and the ability to lay down a thin even layer. But it has to be applied to both mating surfaces to work as it’s designed. So I needed a way to spray the fire rings separately, on both sides without the first side getting stuck to anything (did I mention this stuff is sticky?), and a way to spray only the narrow fire ring sealing area around each combustion chamber, plus the narrow cylinder top sealing surface without spraying down into the bore. In a production environment there isn’t time to hand-mask all this stuff over and over again, I needed a rig that was quick to apply and usable multiple times.
.
So, pictured is what I came up with, which is quite fast and easy to use. I measured the on-center distance between the two cylinder openings on a wbx cyl head (~112.5mm) and on a piece of stiff cardboard large enough to cover the whole head bottom surface, I drew two circles just slightly over the ID of the cyl openings (~106.5mm, so do 107-108mm) on 112.5mm centers. Cut out the two circles and I had a mask for the bottom surface of the head.
.
Then I wanted to both locate that mask accurately on the head and to mask off the inside of the cyl openings where the green o-rings ride. Plastic one-quart yogurt containers from the recycling bin came in handy here (see, saving all that recyling crap actually does accomplish something!). I cut off two at a height where they would wedge into the openings and just about reach the bottom inside surface, this was about an inch below the top rim of the containers I had. When these ring masks were stuffed into the holes in the cardboard the rims met, so a short bit of each rim needed to be trimmed off so they weren’t pushing against each other and distorting and trying to pop out of the holes.
.
Next, round masks that could be placed in the combustion chambers so only the narrow perimeter where the fire rings sit would be left exposed. The bottoms of the quart containers were the perfect diameter, I cut them so there was just a very short rim so they stay stiff.
.
Spray cans work best held upright, plus Hylomar’s expensive so I want to use up the whole can, so the head surface couldn’t be laying flat, but these heads will present at a nice angle when just laid down as they are. Head set on edge of spray table, cardboard mask laid over, plastic rings wedged into the cyl openings (trimmed part of top edges facing together), and finally the round masks put in place concentric with the ring masks. To accomplish that I made two wire hooks penetrating the centers of the round masks so they could be hung from the ring masks and centered up by eye. Then its a quick spray around each circle, one light even coat, and the masks come off and are put on the next head.
.
For the fire rings I traced four 100mm circles on a cardboard sheet, punched short brads thru, three evenly spaced around each circle, flip the cardboard and I have a little pygmy forest of four sets of three pointy pins to set the fire rings on. Spray one side, flip the rings gripping only by the edges, spray the other side. After the brief wait period, these can be dropped in the pre-sprayed head openings, use a pointed tool to poke them down in a few places, and they’ll stick in place, ready for the head install.
.
I tried using food container bottoms to mask out the inside of the cyl sleeves, and wrapped masking tape around the outside, but the process of installing the piston/cylinder assemblies involves a lot of handling and the pre-sprayed top surface shouldn’t be touched. It was hard to avoid marring the sealant coat with all the other details you have to watch. In the end I found it easier to just install them dry and run a bead of Hylomar AF around each top after they’re in place when the head’s ready to go on.
.
So that’s the way I prepped the parts. Then there’s an extra step in the final head assembly, because Hylomar should be partly clamped, given 20 minutes or so to flow under moderate pressure, then the fasteners can be brought to final torque.
.
Most engine assembly specs call for torquing heads in at least two or more stages. I do that on wbx’s, of course, and if I’m doing both sides I fit up both heads at 10ft.lb, final check fitment of everything, then, alternating sides, I go to 20, 30, then 37 and a quick go-over or two. Alternating sides is fairly standard practice on boxers to balance stresses across the case, the cases do distort a fair amount under these applied stresses. For the Hylomar treatment I torque the head nuts first to 10ft.lb., then to 20, then after a 20-minute wait take them to 30, and on to the final 37ft.lb. and the go-over.
.
And by the way, this is one place where even clamping matters a lot. The used capnuts usually have uneven faces, too, so they get resurfaced by hand on a belt sander to ensure a flat sealing face. I always chase all head stud and capnut threads, blow off clean, and just before assembly put one drop of clean 30w oil on each stud and run a nut up and down it to evenly lube the threads. Clean off excess oil and wipe the capnut faces with Brakleen.
.
I also thought Hylomar might be a good sealant to use under the capnuts. Since it’s non-setting it should tolerate retorquing without the need to remove the capnut and apply new sealant. I would complete the full torque-up, then remove the capnuts one at a time, apply a bead of Hylomar to the nut face, run it down to lightly touch the head to leave a layer of sealant there, back off for air to get in, wait a few minutes, then final torque and move to the next nut. Very time-consuming, and despite all the care taken the results were less than perfect. I decided it wasn’t as good as Reinzosil there, which I had been using for years without problems (I abandoned the mystery yellow VW-spec’d goo years before due to less than stellar performance). It makes sense since Hylomar is intended for flat mating surfaces clamped together which have a few thousandths of an inch creep, whereas torquing a nut means wiping the stuff a long distance under high pressure. I think the film probably breaks in places under that effort, and I was also concerned about its effect on the measured torque since it creates so much drag under the nut face. Siloxane sealants like Reinzosil have such nice workability that they aren’t overly shear stressed and don’t create undue drag to affect torque measurements.
.
So, a bit of extra work, but only a few minutes with my mask and spray rigs, and I never had another fire ring leak complaint.
.
.
*I have two related articles here, one on how to diagnose this type of leak:
Bad coolant pressure cap, or bad head gasket?
the other some observations about the corrosion process undermining the water jacket seal:
Thoughts on Vanagon WBX Water Jacket Seal