® More on Cylinder Heads

It really is a fascinating thing, the four stroke motorcycle cylinder head. In a space the size of a twenty-five cent coin, you have the engine distilled down to its essence, the most important part of the air pumping character of the powerplant: the valve seat.

So very much depends on that steel sliver, that keeper of passageways, together with its mate the valve itself. The absolute seal so elusive, and degrees approaching perfection dearly sought, each offering yet another height of performance satisfaction, the added pound of pressure so earnestly sought. Its physical beauty like Saturn's rings, like the diamond's facets, its charm ever more potent due to its essential utility.

Valves and valve seats
Valve job work is focused on valves and valve seats. The two actually wear very differently. Valves show their wear mostly as recession, that is, a groove that is worn into the valve's face as the valve gradually recedes into the cylinder head. The softer intakes recede more than the exhausts, but regardless, recession results in two bad things. One, it gradually ruins the valve's seal, and two, it moves the valve toward its lifter, which reduces the valve's clearance. Both lower cylinder compression. If left unchecked, valve recession results in burned valves. Before the early 1970s, it was common for receded valves to be corrected by machining their faces flat again. But this has long been inadvisable due to a plasma coating found on major powersports manufacturers' valves. The valves are replaced instead. Valves can also wear on their tips, with certain models more prone to this than others. The valve seat, for its part, is the part of the cylinder head the valve rests on. Valve seats on vintage Japanese bikes wear by widening, getting out-of-round, and by pitting or cratering. Interestingly, intake seats exhibit mostly widening and rounded edges, whereas exhaust seats are heavy on the pitting. Valve guides also wear, of course, more on rocker arm engines, and usually the exhaust more than the intake.

These then are the reasons for a valve job on a Japanese cylinder head: receded valves, and worn and pitted seats. And frankly, lapping valves -- the term used to describe putting abrasive paste on a valve and spinning it against its seat -- fixes neither. Although some OEMS legitimize it, valve lapping is not best practice. A valve-to-seat interface that needs to be lapped is one whose parts are faulty. The valve job itself is inferior from the outset. I haven't owned a valve lapping stick since Carter was president.

Valve seat angles
A cylinder head's valve seat has three basic angles. The sealing surface, where the valve will reside, is usually a 45 degree angle. This is the heart of the valve seat, the holy grail. Valve seats also have two additional angles, a 60 degree transitioning from port to seat, and a 30 degree transitioning from seat to combustion chamber. These angles are sort of overgrown chamfers that ease airflow across the valve seat. So critical are these transition angles that the Superflow corporation, makers of the most widely-used cylinder head airflow measuring equipment in the world, holds that these three angles constitute the most important place in the port, flow-wise, and thus the most impactive on performance. Do a careful valve job and you will reap at least 80 percent of what you might hope to gain by "porting". I've proven it.

Manipulating the angles
A proper valve job looks like this. First the head is inspected for damage that would make spending time and resources on the valves and seats inadvisable. Once the head is deemed usable, it is ultrasonically cleaned and painted. The next step is to machine the seat's 45 degree to square it up relative to the valve guide and dress away pitting and unevenness. Then Prussian Blue is carefully smeared on the seat and the valve inserted, to "register" where on the valve's face the seat is contacting it. It's important that the contact be centered on the valve's face. If it is off center, then the 30 and 60 degree angles are machined as necessary to move the contact where it belongs. If for example the valve-to-seat contact is off center toward the combustion chamber, then the 30 degree angle is machined slightly to move the contact the other way. If on the other hand the contact is off center on the opposite side, toward the valve guide, then the 60 degree angle is dressed a bit to move the contact appropriately. Once the seat is smooth and centered respective to the valve, you're still not done. Now that 45 contact zone needs to be either widened or narrowed, as needed and more or less to factory spec, for good sealing and long life. Narrow seats seal the best, but wider ones last the longest, so a compromise is sought between these two goals. The 45 degree is narrowed by machining both 30 and 60 an equal amount. It is widened by machining the 45 alone.

If the cylinder head us from a shim valve train type engine, the work involved in a valve job is potentially doubled. Many extra steps are needed due to the constraints of available shim sizes. The tech must be sure that the end result does put the installed shims near the small end of their range. Even more work is required to prepare the head for aftermarket cams.

First, lapping. Why do it? If you have done a good valve job, there is no need to lap. In fact, you will be degrading, undoing, much of your hard work. Lapping at its very best is not part of a valve job but a very poor substitute for it. If you understand the what and why of valve recession, why then would you do a procedure that hastens it, and actually produces it where there was none to begin with? Makes no sense to me. Now though I avoid lapping, my friend and colleague Nigel Patrick (five-time consecutive ProStar champion) very lightly and faintly laps to check contact and concentricity. Just a few strokes. And that makes sense, in a way. He doesn't do it to improve the seal, only to confirm it. But even there I have to ask -- why? Why substitute the traditional Prussian Blue to do this job, with lapping? With him I think it's just habit born of many many years. No harm. Third answer: 70s Japanese valves are plated. Why defeat or harm that plating by abrading it? I can't say this strongly enough: though OEMs countenance it, lapping is a hack procedure best avoided and with virtually nothing to commend it.

On the subject of the different cutters/stones, I break it down like this:
Two categories of tools

1. Dead vs. live pilot

2. Multi vs. single point

1. Dead vs. live pilot
a. Dead pilot = pilot stays stationary, fixed into guide, while cutter or stone rotates

  • A dead pilot is never and cannot be concentric w/seat, thus seat machining of seat is not concentric
  • All pilots are either tapered, expanding, or loose fit
  • Tapered and expanding pilots definitely cause non-concentricity -- no way around it
  • Guides, whether worn or not, will not hold tapered or expanding pilots concentrically
  • Where found: Black and Decker, Sioux, Kwikway, Neway
b. Live pilot = pilot rotates as one with cutter or stone (either as part of handle or bolted to it)
  • A live pilot self-centers, as accurate as valve itself, making seat machining more concentric
  • Where found: OEM hand-turned cutter or stone sets, Serdi, Mira, Hunger, Newen

2. Multi vs, single point
Sub-categories within the two categories above
a. Multi point = cutter touches seat 360 degrees

  • Cutter or stone, by design, can only follow original seat's level, contour
  • If seat or guide have shifted, cutter or stone can't correct that, only follow, conform
  • Seat that has started not perfectly perpendicular to valve ends the same way after cutting
  • This will produce poorly-sealing seat
  • Where found: OEM hand-turned cutter or stone sets, all powered stone type, Neway
b. Single point = cutter touches seat only at single point
  • Cutter doesn't follow seat level, doesn't conform
  • If seat or guide have shifted, cutter does not have to follow, it re-levels the seat
  • Its registration is to the guide, not to the seat as in multi-point systems
  • Result is corrected level, accurate, beautiful seat, best possible seal
  • Heads that always have shifted seats thus do poorly w/multi-point: Honda CBX1000, early DOHC fours
  • Where found: Serdi, Mira, Hunger, Newen

The bottom line in the two types and the two sub-types is where is the tool "registered", that is, what is its datum point, its reference? A tool whose cutter registers to the guide is intrinsically more accurate than one that registers to the valve seat. Single-point systems register to the guide, multi-point to the seat. Simple, eh? I wish this were all there was to it, but there are complexities, compromises.

You will conclude that the best tool is live pilot, single point. But this is true only if you have $30,000-40,000 to spend on the tool. From a practical standpoint, it's easier to reach target quality spending reasonable money with a multipoint tool centered on a live pilot than it is with a single-point tool live pilot system (live pilot is assumed in any case, there just is no good reason to use a dead pilot system).

In short, if money is no object go Newen, the ultimate computer-controlled, rigid, accurate single point live pilot system. If however as is usually the case money is a consideration, hand-operated live pilot multi-contact (stones) is best. But in any case, note that live pilot is preferred. Always.

Also note that although economy hand-held versions of the Serdi and some of its clones (three at last count) are available, reports are that the quality doesn't approach the conventional floor models of the Serdi, and I can attest to that from my own experience. The promise in this theoretically best combination, the hand-held Serdi Micro or the Mira, just doesn't play out, mostly due to the lack of rigidity in these economy models that bolt to the cylinder head's gasket surface, and turn by a wobbly hand crank. Thus my recommendation of multi-point live pilot, which interestingly enough is exactly what the Big Four made available back in the 60s and 70s.

A dead pilot can't hope to center in the guide, if it is taperered or locking. Thus the cutter, which will attempt to center on the seat anyway, is nonetheless influenced by the non-eccentrically positioned pilot, and that can never be good as it simply compounds error. That is, the multi-point cutter already is thrown off by its registering on the seat, then thrown off even more by the influence of a crooked pilot. This is exactly what is wrong with the Neway system. It's doubly inaccurate. On top of this and of little consideration in this discussion, the finish left by Neway is very poor. It looks from your pix like you're doing very good work, especially considering you're using Neways. But I wonder if you have ever vacuum tested Neway cut seats. It will be disillusioning.

Remember the pilot does not register the cutter in a multipoint cutter system. The cutter itself does, which is the multipoint system' only, and huge, drawback. The pilot only comes close to registering in a single-point system. And while that is certainly true, there is a rat in the basement. Even this is not completely reliable. In a single-point system the pilot doing the registration is only theoretical. Registration in single-point is somewhat dependent on the floating mechanism, whether in the cutter drive (small Serdi models) or in the floating table (large Serdi models). But to be fair, single-point does rely heavily on guide registration.

Email me
© 1996-2020 Mike Nixon