Cylinder head cooling system mods for the AMC 195.6 OHV engine

Hot-rodding the Nash/Rambler 195.6 overhead valve six

23 May 2010

The single most serious problem with this engine is a design flaw in the cylinder head, involving the cooling system. The TSM (factory Technical Service Manual) for this engine states that cylinder headbolt torque must be "checked" every 4000 miles, and retorqued every 8000 miles. For an engine designed (new cylinder head) in the late 1951's this is very very often, and a friend with more engine experience than I states that a retorquing schedule is an admission of a design error. I have to agree.

The problem

The problem is two-fold: one, while the engine is cold and the thermostat is closed no water circulates in the head; two, the thermostat is not thermally coupled to any combustion chamber. These related problems cause huge temperature excursions within the head at warm-up. When the cold engine is started, heat builds in the combustion chambers and indirectly heats the head water jacket. With the thermostat closed, temperature rises rapidly, likely with many hot-spots, until heat reaches the thermostat via conduction only through the cast iron and water -- there is no circulation due to the water pump and jacket design.

The problem is exacerbated further by the very deep wedge combustion chamber design needed to accomodate the integral intake trough and attendant long head exhaust ports. Large combustion chamber area and long exhaust ports put a lot of head into the head to begin with.

On the later AMC six (199/232/258 and eventual 4.0L), the thermostat is adjacent to #1 cylinder; the copper button of the thermostat is within an inch of the cylinder wall. On the 195.6 OHV, the thermostat is at least 4 inches from #1 cylinder; but the most critical problem is that there is no way for water to circulate until the thermostat opens. Since the thermostat is warmed only by conduction, the thermostat pod remains cold while the head casting between the cylinders is too hot to touch. I have heard "water hammering" from steam pockets in the head during warm up. You can verify this situation without instrumentation by feeling the underside of the thermostat housing vs. the head with your hand!

Once the thermostat does open, coolant temperature regulates normally; and once open, as long as the engine is producing sufficient heat for the thermostat to remain open, coolant circulates through the block, head and radiator.

How this affects head sealing -- via head bolt loosening

I conservatively estimate that during warmup, the head heats 150 - 200 degrees F above the block temperature in the time before the thermostat begins to open. Given the temperature coefficient of cast iron and the head length, with a 150 degree rise the cylinder head grows .024" longer than the block, upon every cold-engine warm-up! Once the thermostat opens and block and head temperatures more or less equalize the head contracts somewhat as the block lengthens. Therefore, the cylinder head slides back and forth on the block (transverse motion), shifting on the headgasket. This lengthwise temperature excursion causes the cylinder headbolts to back out, hence the required re-torquing schedule.

Don't believe me on this bolt-loosening business? Check out this page at the Jost Effect. There's even a video showing transverse motion backing out a bolt!

With hundreds of cold/hot/cold cycles per year, head bolt torque drops, increasing the likelihood of headgasket leaks. The gasket situation is not good by design, and there are not really enough head bolts to begin with.

The solution

The solution -- time will tell how this behaves long-term -- consisted of the following, each covered in some detail below. I believe in redundancy and overkill when it comes to reliability!

  1. Modifying the cylinder head coolant path to circulate coolant through the head and over the thermostat;
  2. The addition of two more temperature sensors to be able to see what is actually happening in there;
  3. Replacement of crappy old head bolts with ARP studs;
  4. Careful attention to block deck and head mating surfaces;
  5. Lower temperature thermostat;
  6. Computer-controlled electric cooling fans.

Fix #1 -- cylinder head coolant flow

It turns out it is very easy to introduce coolant circulation within the cylinder head and block during warm-up, drawing coolant past the thermostat in the process, thereby transmitting head heat directly to the thermostat.

The fix is simple: drill and tap the bottom of the thermostat well for a pipe fitting -- I used 3/8" NPT -- for a 90-degree street elbow and hose barb, and plumb this with hose to a tee into the water pump's heater suction line. (This water pump has two inlets; the large one that draws from the bottom of the radiator and another that draws block water through the heater core.)

With this fix, coolant is circulated through the head (and block) continuously during warm-up, when the thermostat is closed. The downside is increased engine warm-up time, but my 1963 American still warms up faster than my 232-powered Classic wagon.

Fix #2 -- temperature sensors

Clearly this doesn't "fix" anything but information is power in this circumstance!

I decided I wanted to know head coolant temperatures a little more intimately than the lone factory sensor up in the isolated thermostat pod, so I drilled two holes, between cylinders 3 and 4, and 4 and 5, and inserted aftermarket Stewart-Warner electric senders and dash gauge. I'm only using the front one at the moment, which also drives the fan computer. The hole locations were determined by probing my sliced up cylinder head (that thing has been very handy...).

Fix #3 -- head studs

The factory torquing specification is very strange, it calls for hot-torquing, there is very little gasket area around critical areas, and since head clamping is especially critical in this motor, I jettisoned the ancient head bolts and replaced them with ARP studs; please refer to that page for details. They were not cheap but what does a leaky cylinder head cost? Factory spec is for 60 ft/lbs hot; ARP demanded 75 ft/lbs cold. I went with ARPs recommendation. I also slightly chamfered both block and head holes to accomodate thread pullup (no sign of it in the stock block before rebuild).

Fix #4 -- mating surfaces

Since I was doing a full tear down/rebuild this was feasible. Nothing special was done here, I just made sure I did not get "it's just an old car" work done on the block and head. I specified 100-microinch surface grind and personally checked the block for flatness with a 24" straightedge before I took it home. Only old style metal/composite gaskets are available for this engine, not the self-sealing blue type Felpro gaskets. There is very little gasket area around head steam holes. There's no reason to not pay very close attention to this stuff!

Fix #5 -- coolant temperature

I always run my engines with 195 degree F thermostats, but not this one. I'll sacrifice a bit of combustion efficiency with the idea that, if it's not running due to cylinder headgasket failure, "efficiency" is academic. Given the other changes I could probably have run hotter, but overkill reliability is the goal here.

It is common practice to drill a small hole in the thermostat to allow purging of air while filling, and I did that -- but I made the hole 3/16". I oriented the drilled hole towards the radiator, so that heated water runs under the thermostat on it's way through.

Fix #6 -- electric cooling fans

Not directly a solution to head problems, but when was inadequate cooling a good thing? The early American chassis is very small, but I manages to fit a pair of 12" electric fans I got from Summit as pushers, between the radiator and grille. Always going for the overkill solution, I made a microcontroller-based fan controller that slowly varies the speed of the fans depending on cylinder head temperature.

It seems that the factory may have introduced this very fix in some heads, likely very late, before the motor was discontinued in 1965; see photo right. This was a junk head removed from a Classic sedan; the plug was in place presumably from the factory.