Replacing valve seats is one of the basic jobs that is often necessary when rebuilding aluminum or cast iron heads with cracked, damaged or badly worn seats. But there is a lot more to replacing a valve seat than prying out the old one and driving in a new one. If the head is cast iron with integral seats, the head has to be machined to replace the seat (sometimes called installing a "false" seat). And if the head is aluminum, the seat counterbore may have to be machined to accept an oversize seat if the bore is loose, deformed or damaged. Either way, a machinist has to figure the amount of interference that is required for the new seat before cutting the head on a seat-and-guide machine. He also has to decide what type of seat to install. Replacing a seat, therefore, involves a number of decisions and steps, all of which affect the outcome of the repair job.
As you might have guessed, we encountered differing opinions about the right way and wrong way to replace valve seats while researching this article, particularly with respect to the amount of interference fit that is required to retain seats in aluminum heads. A common fear expressed by many engine rebuilders is concern over the possibility of seats falling out, particularly in aluminum heads where the difference in coefficients of thermal expansion between the head and seats can cause seats to loosen if the head overheats. Consequently, engine rebuilders expressed differing views on whether or not locking compound and/or peening or staking should be used as "insurance" when installing seats in aluminum heads.
One point everyone does seem to agree upon is that valve seats play a critical role in the longevity of the valves. The seats draw heat away from the valves and conduct it into the cylinder head. This provides most of the cooling that the valves receive and is absolutely critical with exhaust valves. Anything that interferes with the seat's ability to cool the valves (such as a loose fit or deposits between the seat and its counterbore) can lead to premature valve failure and expensive comebacks.
The seat alloy and hardness must also be matched to the application and compatible with the type of valves that are installed in the engine. Again, we found differences of opinion regarding the selection and use of various seat materials.
To better understand the issues behind the differing opinions regarding valve seat replacement, let's start with the seats themselves and why they fail.
WHY VALVE SEATS FAIL
Nonintegral valve seats can fail for a number of reasons. Most of the seats that end up being replaced are replaced because they are either cracked or too worn to be reground or remachined. Seats can crack from thermal stress (engine overheating usually), thermal shock (a sudden and rapid change in operating temperature), or mechanical stress (detonation, excessive valve lash that results in severe pounding, etc.).
A small amount of valve recession results from normal high mileage wear, but it can also occur when unleaded gasoline or a "dry" fuel such as propane or natural gas is used in an engine that is not equipped with hard seats. Recession takes place when the seats get hot and microscopic welds form between the valve face and seat. Every time the valve opens, tiny chunks of metal are torn away and blown out the exhaust. Over time, the seat is gradually eaten away and the valve slowly sinks deeper and deeper into the head. Eventually the lash in the valvetrain closes up and prevents the valve from seating. This causes the valve to overheat and burn. Compression is lost and the engine is diagnosed as having a "bad valve." The seat also has to be replaced, but it many instances it may not be recognized as the underlying cause of the valve failure.
As a rule, a seat should be replaced if the specified installed valve height cannot be achieved without excessive grinding of the valve stem tip (less than .030 in.), or if the specified installed spring height cannot be achieved using a .060 in. spring shim. This applies to integral valve seats as well as nonintegral seats. The only other alternative to replacing the seat is to install an aftermarket valve that has an oversized head (.030 in.). This type of valve rides higher on the seat to compensate for excessive seat wear or machining, and can eliminate the need to replace the seat.
A seat may also have to be replaced if it is loose or if the cylinder head is cracked and requires welding in the combustion chamber area (the seats should be removed prior to welding).
One way to check a seat for looseness is to hold your finger on one side of the seat while tapping the other side with a hammer. If you feel movement, the seat is loose and should come out (so it does not fall out later!).
The seats in an aluminum head may also loosen or fall out when the head is being cleaned in a bake oven or preheated in an oven for straightening. The same thing can happen to the guides. Whether or not this occurs depends on the amount of interference fit between the seats and head. The less the interference, the more likely the seats are to loosen and fall out when the head is baked. If you do not want the seats to fall out, turn the head upside down or stake the seats prior to baking.
VALVE SEAT REMOVAL
A variety of techniques are being used by engine rebuilders to extract nonintegral valve seats from cylinder heads:
Though not a tool for removing seats, the "Seat Ring Factory" by the K.O. Lee Company, Aberdeen, SD (800-874-9215) is a lathe for making your own seats from semi-finished nickel alloy rings ranging in size from 5/8 in. (16 mm) ID to 2-1/4 in. (57 mm) OD, and 13/32 to 1/2 in. deep.
ANYTHING BUT A BORING SUBJECT
Once a seat has been removed from a cylinder head, a determination must be made as to whether or not the counterbore needs to be machined to accept an oversized seat. If the original seat was loose, if the counterbore is flared more than .001 in. (wider at the top than the bottom), or if the difference between the counterbore inside diameter (ID) and a standard seat outside diameter (OD) is not enough to provide the desired interference fit, then machining will be necessary.
Seats are available in various oversizes. But the amount of metal that can be safely removed from most aluminum cylinder heads is minimal, so the less the amount of machining that is required the better. Cutting a seat counterbore too large or too deep may weaken the head, cut into the water jacket or cut into the adjacent seat.
The amount of interference required to lock a seat in place depends on the diameter of the seat (the larger the seat, the greater the interference that is required), the type of head (aluminum or cast iron), the application (hotter running applications typically require more interference to keep the seats from falling out), and in some cases the type of material used in the seat itself (hard seats cannot take as much interference as softer seats).
For cast iron heads, recommendations range from .003 to .006 in. for valve seats up to 2 inches in diameter. For aluminum heads, some rebuilders and seat suppliers said more interference is needed because of the difference in the coefficients of thermal expansion between the head and seats. Aluminum expands several two to three times as much as cast iron when it gets hot, so recommendations ranged from .004 to as much as .0085 in. interference for valves up to 2 inches in diameter. But others said seats in aluminum heads actually require less interference than those in cast iron. One such person was Ray English of AADCO, Solon, OH.
English says he has been rebuilding aluminum heads for 15 years and has never used more than .005 in. of interference. "You really do not need any more than that. Most of the factory specs call for .003 to .005 in. of interference. On a Jaguar, it is only .001 in. Yet many people think you need a lot of interference with aluminum to keep the seats from falling out. But that is just not the case. Aluminum provides such a good heat sink that you do not need more than .005 in. With cast iron, though, you sometimes need as much as .007 or .008 in. of interference because the seats run hotter."
What about using a locking compound (such as #640 red Loctite) as added "insurance" when installing seats, or peening or staking seats to keep them from falling out? English says neither is necessary.
"I have never staked a valve seat in my life. If you have to peen or stake a seat to keep it in place, you did a crappy job installing it. I also would not use a locking compound because it can create a thermal barrier between the seat and head."
English said a common problem he sees in aluminum heads that have been rebuilt by others is improperly machined seat counterbores. The bore should have a smooth finish so the seat will fit tightly and won't broach or shave the head metal as it is being driven in.
"We have seen bore finishes that looked as if somebody used an auger to cut the hole. The finish was awful because the machinist did not use a cutting fluid. When you cut aluminum, you have to use a lubricant to stop the metal from balling up on the end of your tool. Oil will also help your tools last a lot longer. We only have to buy about 10 tips a year for all the heads we do."
A good finish also requires sharp tools and plenty of cutting speed, says English. He recommends cutting at 600 rpm. He also cautions against using the same tools on aluminum that have been used on cast iron.
"A lot of machinists use the same cutters on cast iron and aluminum. But when you cut cast iron and then use the same tool on aluminum, it won't cut for anything. That is why we have one set of tools for aluminum heads and another set for cast iron."
When replacing a seat, English says he measures the OD and depth of the original seat and then goes .020 in. over, allowing for a .005 in. interference fit. "If you go with too large an oversize, you will end up removing too much metal and weakening the head."
English also said he has been making a lot of his own custom seats because many of the seats he needs for import heads are oddball sizes. "We plan to introduce our own line of replacement seats for aluminum heads early next year. They will be .020 oversize, and they will be easier to machine than most of the aftermarket seats that are currently available.
"We have found that most of the original equipment seats in aluminum heads are not much more than cast iron, which is very easy to cut. But most aftermarket seats 18 to 20% chromium and are a monster to cut, which is ridiculous. You do not really need that hard of a seat in an aluminum head because the seats never get very hot. So that is why we are introducing a softer material that will be easier to machine, easier on tooling and give a more precise seat."
CHOOSING THE "RIGHT" REPLACEMENT VALVE SEAT
The original equipment manufacturers use a variety of seat materials, including cast iron, iron alloys, nickel alloys, cobalt alloys (stellite) and powdered metal (which generally contain no chrome or nickel, only vanadium and iron). Most OE seats in passenger car aluminum heads are a high grade of cast iron or powder metal. The better (more expensive) materials are usually found in high output and turbocharged engines, with hard seats and stellite being used mostly in diesels and industrial engines.
When replacing a seat, you should use one that is at least as good as the original if not better. Hard seats are a must for high temperature, high load and dry fuel (propane or natural gas applications). In fact, most seat suppliers have special alloys specifically designed for dry fuel applications. But hard seats are not required for light duty passenger car applications. Even so, many aftermarket seats are made of premium grade alloys or heat treated iron to provide improved longevity and performance.
As one seat supplier put it, "considering the insignificant difference in price between a so-so seat and a good seat, would not you rather sleep at night?"
Seat and valve materials must be compatible with one another as well as suited for the application. A hard valve generally requires a hard seat and vice versa. A stellite faced valve in an industrial engine, for example, would require a stellite seat. A titanium racing valve, on the other hand (which is relatively soft), would require a soft cast iron or beryllium-copper seat.
Brian Bender of S.B. International (J-LOY), said his company generally follows the OE material specifications on replacement seats. "If they use iron, nickel or cobalt in a given engine, we do the same."
Material compatibility is very important, said Bender, especially with fuels such as propane and natural gas. "We have our special Star Series inserts that provides extra hot hardness. It is a nonmagnetic nickel based alloy and contains some cobalt. This type of seat can really take the heat, especially if the air/fuel ratio or timing is off."
Bender said that most of the seat failures he sees are the result of abnormal engine operation (overheating, detonation, wrong air/fuel mixture, etc.) or because someone chose the wrong type of replacement seat.
"Some people are still trying to use plain cast iron seats in unleaded engines. It just will not work because it is too soft. They are a thing of the past so that is why we have discontinued plain cast iron seats in our line."
Tom Tucker of the Tucker Company, said 440 stainless steel seats or Silicone XB (an iron seat with 18% chrome) are probably the most popular aftermarket seat materials today. But he also stressed the importance of choosing a seat that is designed for a specific application.
"We have a tool steel tungsten carbide material for natural gas applications that holds up especially well. We also have an E-series material that provides superior hot hardness but is not as hard or abrasive as #3 stellite."
Tucker said one thing you have to watch out for is that some seat alloys can develop soft spots as they age. "I do not recommend reusing worn seats because the hardness of the seat can vary from one spot to another."
A HARD SUBJECT
There are two schools of thought on the subject of hardness. One equates hardness with quality. The other recognizes the importance of hardness but realizes that other factors are just as important.
The hard schoolers say things like you need a seat with a hardness of Rockwell C 37 to 45 for unleaded fuel, and 40 to 50 Rc for propane, natural gas and high compression/turbo applications. But Joe Keon Jr. of Martin-Wells, which is one of two companies in the U.S. that manufacturers valve seats (L.E. Jones being the other), toughness and durability are better measures of quality. Keon emphasizes the metallurgical aspects of selecting a seat material.
Keon said the first hard material that was used for industrial valve seats was stellite #3 (cobalt with 30% chrome, 12% tungsten and 2.5% carbon). The seats were made by welding stellite onto a tool steel base material.
"In those days they did not think about heat transfer. All they wanted was a seat platform that was as hard as possible. The next step in the evolution of seat materials was to synthesize hardness. By heat treating tool steel, seat hardness could be increased to Rc 43. But if a heat treated seat is subjected to overheating, you get grain inversion and the molecules return to their original position. You lose that hardness.
"Forty years ago we came up with our Well-Tite formula that achieves the same wearability of a 52 Rockwell C stellite type of product but with a hardness of only 35 to 37. And it does a far better job of dissipating heat. The Well-Tite alloy contains 42% nickel, which sucks the heat away from the valve. It has 10 to 12% chrome for oxidation, and 7% moly for toughness. We also found that our formula produces an oxide layer that works up to the surface through a chemical action and acts like a lubricant to prolong valve life."
Keon said Well-Tite seats are not heat treated, but are machined and sold "as cast." He says this allows them to handle high temperatures without danger of grain inversion. He also said the unique Well-Tite ally has an excellent memory characteristic that allows a seat to return to its original size after heating without distortion, which means the seats will not loosen or fall out if an engine overheats.
Keon said rebuilders should be especially careful about the quality of the seats they buy. He said many offshore suppliers are less than thorough about their quality control measures, yet charge as much for cast iron seats as ones that contain superior alloys.
VALVE SEAT INSTALLATION
Once the counterbore in the head has been machined for the desired interference fit and a replacement has been selected, the next step is to install the seat.
As mentioned previously, the hole must be clean and have a smooth surface finish. The seat should be placed with the radius or chamfer side down and lubricated (ATF works fine) prior to being pressed or driven in with a piloted driver (recommended to prevent cocking).
If the replacement seat has a sharp edge, it should be chamfered or rounded so it won't scrape any metal off the head as it is being driven into position. If metal gets under the seat, it will create a gap that forms a heat barrier. This, in turn, will interfere with the seat's ability to cool the valve and premature valve failure will likely result.
Preheating the head and/or chilling the seats with dry ice or carbon dioxide (do not use Freon because it damages the ozone!) will make installation easier and lessen the danger of broaching the counterbore as the seat is being installed.
If you choose to peen or stake the seats after they have been installed as added insurance to prevent them from falling out (which should not be necessary if the seats have the correct interference and were properly installed), several engine rebuilders we interviewed recommended rolling or peening rather than staking. Their reason? Staking creates stress points and potential hot spots.
That wraps up our overview of valve seat replacement. As we said at the beginning of this article, seat replacement involves a number of important decisions. So measure your counterbores carefully, chose the "right" seat for the application, double check your machine work and make sure the seats are installed correctly.