updated october 2018
you can contact me at it is as if at protonmail dot com
(click images for full size.)
this is an old engine designed before freeways were even under construction. luckily oil pressure, volume and internal distribution have proved more than adequate for even high-performance driving, but severely lacking in two crucial areas: cooling and filtration.
in 2010 i had done what i then thought was a decent if stock rebuild, but i melted all the rod big-end and main bearings in an admittedly stressful situation (W.O.T. top gear, climbing a 10% grade in Death Valley in August 2015, 115F air temperature). a then-brand-new-from-Kanters oil pressure relief spring failed (see below) dropping oil pressure from 60's to 40's psi, probably not unrelated.
when i got it home, before the teardown for the current build, stuck a temperature sensor in the far end of the main gallery and drove 10 miles up the Glendale freeway, modest grade, 5%? and at the top the oil was 230F. i can't even imagine what it was on that 3000 mile drive culminating the Death Valley grade.
the current system below has so far solved all problems, in easily 10,000 miles of often extremely hard driving, routinely exceeding that Death Valley grade.
i strongly suggest that anyone working with this engine insert a temperature sensor in the oiling system somewhere (the far back end main gallery accepts a Stewart-Warner 280F sender without blocking flow) and see for yourself what's going on.
the rest of this page is in reverse chronological order; newest system at the top with prior work below. the new work is based on the old.
choice of oil is a near-religious subject amongst car folk, but here's my choice and reasoning. new oils are incomparibly better than oils of the past, both for lubricity and especially important here, viscosity control over the operating temperature range.
the factory Technical Service Manual recommends a range of viscosities centered around the common 10W-30 for most climates. today's 10W-30 is too thin, and most oils today do not have the zinc (ZDDP) that this engine needs (the too-small mushroom-head tappets are particularly wear prone). these engines were not built to the close tolerances they are today.
i currently run, specifically, Valvoline VR1 20W-50 Racing oil. only VR1, only 20W-50. this oil has the zinc absolutely necessary here. though my engine oil is now controlled, the "thick" 20W-50 produces factory-correct pressures once it is warmed up. cold, the relief valve is open at idle!
my engine builder, Pete Fleming, built this engine to run 20W-50 VR1 and so that's what i run. my choice, other things beiing equal, would be Mobil1 15W-50, and again, only 15W-50. Mobil1 15W-50 is the only Mobil1 synthetic to contain zinc for flat-tappet engines. this part isn't opinion -- i called and talked to an application engineer at Mobil1 to discuss the right oil for this engine.
neither oil is cheap, who knows which is "better" but those at least meet/exceed documentable requirements for this engine. oil's cheaper than motors.
the current system consists of the stock pump modified for full flow oil filtration, a fairly large stacked-plate cooler with fan, and an Accusump accumulator system from Canton Racing Products. how and why i got here is described below. click the picture at the top of this page, on the right. the full oiling system is visible, other than the oil filter hiding beneath the alternator. the photo below shows the oil pump (red, protruding from the block), the line running to the alternator forward, and the line from the Accusump feeding the main gallery, top and center. lines are PTFE lined stainless steel braid, crimped 6 AN fittings.
the first mod i made was for full-flow filtration, in the mistaken belief that the partial-flow system wasn't adequate. it actually is... adequate. an AMC forum member research found that given the oil turnover rate through normal engine operation all oil passed through the filter in a surprisingly short period of time. but adequate isn't necessarily good enough, and the full-flow mod ended up enabling the rest of the system.
the oil pump must be modified for an external full-flow filter; the construction details follow below and date to 2010. that has been entirely trouble free. the mods are required for the plumbing necessary for filter and cooler.
active oil cooling is flatly necessary for anything more than casual Sunday drives. however without the substantial mod to the oil pump described below there is no way to get at the oil flow due to the bypass system.
with the mod however it's just a matter of mounting and plumbing. my installation includes a temperature sensor screwed into a bung welded into the oil pan right above the oil pickup. my (homebrew) "ECU" system controls the oil cooler fan speed.
a note on oil cooler mounting: i originally had the oil cooler mounted down in the valance, a big screened hole below the right headlight. in Death Valley (or equiv.) this was great -- however in cool/cold weather it took over 30 minutes to warm up. it's since been moved to it's current location under the hood.
here it is mounted on a bracket bent to fit the cooler onto the curved inner fender and to provide about an inch of clearance for air flow. i'm not totally happy with it circulating under-hood air, but it seems adequate. (if i 'hole' the inner fender the fan will then draw cooler air from the wheel well area, should i need it.)
yet another serious oiling system problem appears only when the car it's in is driven hard and fast on mountain/canyon roads like we have here in California. i do planned/routed "tours" with a vintage sports car crowd and the roadster is now fast enough to cause severe loss of oil pressure in turns due to sloshing in the pan. i neglected to take the time to baffle the pan when the engine was built and plan to do so this winter (2018/2019). the sump only holds four quarts, though i often drive with five, but then i get oil misting issues and increased oil consumption. (a carefully designed passive crankcase de-mister and rejiggered PCV system solved most of the oil consumption; that's documented below.)
someone on the 2018 California Melee suggested i look into an Accusump; after research and email discussion with Jeff at Canton Racing Products i bought and installed one.
briefly, the Accusump accumulator is a cylinder with a sliding piston inside, oil on one side and pressuried air on the other. the oil side is connected to the engine's main gallery through two separate and distinct valves: a small, low flow orfice with checkvalve and a larger, electrically controlled dump valve.
the small orfice and check valve allows high-pressure engine oil to fill the accumulator relatively slowly, the piston compressing the air side until equilibrium. (here, the accumulator has two quarts of oil at something over 70 psi, which is the cold-engine-oil startup pressure). the slow-fill assures that an empty accumulator won't starve the engine while it fills.
the second valve on the Accusump is large and electrically controlled. its purpose is to allow the pressurized oil in the accumulator to feed the engine when the oil pressure form the main pump temporarily plummets; in my typical case, in severe turns at speed in mountain roads (or panic-type stops) when enough oil has sloshed to one side or the other, allowing the oil pump to suck air. it typcically takes 2 to 10 seconds to recover from this during which oil pressure is zero. plain bearings don't like this especially under load.
Canton has different manual and electric solenoid valves, and different range pressure switches to control the electric solenoid valve. the switch closes (applying power to the solenoid, opening the valve) at a fixed pressure, and opens (powers off) at a somewhat higher fixed pressure. the trick is to pick a pressure switch that remains closed during typical idle (lowest speed/lowest normal oil pressure).
my engine idles at 600 rpm where oil pressure is typically 35 - 40 psi and i selected a switch that closes (turns on) at 35 psi and off at 40 psi. this would mean that when i throttled back the accumulator would dump oil until it equalized to idle pressure (eg. 35 psi). in my case i added a computer output that is "off" below 800 rpm and drive the switch from that. any aftermarket ignitions (eg. MegaJolt Lite Jr, Megasquirt, etc) have outputs for this. or simply set the idle up, or put up with it as-is; in my case at speed the thing would be full anyway under load/speed when i needed it.
the second thing that sold me on the Accusump is that it provides full-pressure engine pre-oiling. at ignition-on the switch is closed (0 psi), solenoid open and the accumulator pushes oil into the engine until it reaches the "off" pressure (40 psi here). takes about two seconds.
the devil is in the details.
the critical hoses that appear to stick up exposed are carefully placed to fit into recesses in the hood and aren't as exposed to harm as the appear to be here. it's tight in there, it's a small car.
there's a lot of hose in there, and it's "only" 6AN. bends are all large radius, two 45-degree fittings and one 90, necessary for clearance out of the pump, but it was bored and smoothed out. also see the smoothing and debugging done to the pump.
but this isn't a big modern V8. the pump drove the main gallery through a 5/16" hole. with the admittedly wimpy test of blowing through the hose, there's little restriction. objectively, the main gallery clips at 60 psi, the relief valve pressure, by 1600 rpm when hot. there are pressure and flow losses through all lines but experience here shows this to be more than adequate.
with cold 20W-50 oil pressure at idle has the relief valve open; even 2000 rpm pushes cold (60F) pressure past 75 psi.
the Accusump makes oil changes slightly more complicated. before
draining the oil the accumulator must be emptied; simply power the solenoid
on until the air pressure gauge reads minimum. after draining and
refilling, pull a wire off the switch/solenoid to ensure that the big
valve does not open and starve the engine. post-oil-change
is the one time when the accumulator is empty and there is no oil
in the gallery. ask me how i know! lol
the stock oil pump is an external but typical gear pump. the pump inserts into the lower side of the block left, pulls oil from the pan and pushes directly into the main gallery. top end (rocker shaft, etc) is lubricated by an external line that carries oil from a tap on the block, up to the head casting where it flows upward through a rocker shaft support, into the hollow shaft and from there to each rocker. on earlier engines the top-end source is the main gallery, eg. full engine oil supply. on later engines the top end is fed by an intermittent source generated by a flat on the camshaft's front journal that pulses the main gallery feed, to limit oil flow to the top end.
crankshaft main journals (4) are fed directly from the main gallery as is each cam bearing. connecting rod bearings receive oil via drilled crank. the connecting rod big end has a squirt hole that lubricates the cam journals. the cam is also splash lubricated. some years have piston squirt lubrication via conn rod squirt hole. the specifics of the oiling system make it very easy to modify.
i modified the pump for full-flow filtration by fabricating a new pump top cover with pressurized oil outlet and blocking the outlet on the bottom of the pump body with a steel gasket that blocked the factory oil outlet hole that fed the main gallery. with this mod oil is pumped out the cover, through a PTFE 6AN line to the rather large oil cooler mounted in the front valance, from there into the oil filter, the filter outlet feeding the main gallery.
In 1964 AMC did add a full-flow filtration oil pump to this engine, but it won't fit into the pre-1964 small (01) chassis, the filter doesn't clear the suspension. I got a rusted pump from a friend, it was too far gone to use but it served as a model for cogitating on a solution. The fundamental limitation in the pre-1964 American chassis is clearance. The oil pump is external to the block, and sits very close to the lower A-arm pivots. I believe that if I had a decent Classic pump I could have modified it for my own ends, but I couldn't find one, and the non-filter pump is common enough so I based my hack on that.
the pump is Melling M61. buy a new not remanufactured one. the old ones always "look good". there is subtlety in this crude lump; bodies wear, pump volume drops, new gears don't fix cavity wear. you were warned.
below are photos of the original 2010 installation. the rubber hose/nipple system shown here was swapped out for proper stainless steel braided, PTFE lined, crimped 6AM fittings a few months after these photos were taken; i just didn't trust them. these photos however do show component location well. the oil filter mount is from a "remote oil filter kit", i tossed the screw-on adapter. it's mounted via a simple bent bracket bolted to the timing chain cover bolts.
The pump, of the common gear type, (green assembly with three cover bolts, lower left in the first picture) sucks oil from the pan and in this modified system, pushes it out the pump cover through the brass fitting and lower-most hose, into the filter inlet. Oil flows through the filter and out the top-most hose (arches up and over) and into the center of the main gallery where the AMC factory conveniently put a 1/4" NPT tapped hole, directly above the original pump feed location.
The trick was to interrupt this circuit to insert the filter. This turned out to be very easy. My first version involved drilling and tapping the pump body outlet for a plug but my final version was vastly simpler -- leave the pump untouched and replace the paper gasket with a 20-gauge steel plate "gasket" the exact size and shape as the pump-to-block gasket that didn't have the outlet hole cut. I used Permatex Right Stuff as sealant. There's a lot of leeway in pump to block insertion depth (gear mesh depth) and during normal operation there's no or little pressure difference on each side, so it can be thin.
sadly i don't have a single photo of the steel shim "gasket" that blocks the pump-to-gallery outlet, but it was easy to fabricate -- i simply traced out the oil pump base gasket onto .025" sheet steel and drilled all of the holes except pump outlet.
I then fabricated a new cover for the pump with an outlet directly opposite the original outlet in the pump body; oil under pressure now exits through the cover. The cover is fabricated from two pieces of 1/4" steel stock, the small piece stiffens and builds up height for sufficient threads in the fitting, and allowed the driven-gear lubrication well to be a simple through hole in the larger plate. The small milled groove feeds pump inlet oil (not outlet pressure) to the top of the driven gear and exactly matches the factory configuration.
The location of the new pump outlet hole was fairly touchy; note that it is not centered in the gear output cavity, but slightly to one side. This is due to interference with the top pump bolt. socket-head bolts are required. The hole seems large but the effective diameter is actually the ID of the fitting, about 3/8".
After welding, the plate warps; I milled it more or less flat then ground it flat flat with 80-grit wet-or-dry on a ground cast iron plate. Flatness matters here, this is the mating/sealing surface for the pump gears as well as the pump body gasket surface. The gasket is dimensionally thin, hard, and subject to full pump pressure, and this is a core mission-critical part. It's worth the extra effort to get this perfect. Note also that the gasket is trimmed around the new outlet hole.
To the cover I added a 90-degree 3/8" pipe to 1/2" flare tubing adapter. I used a stainless steel part instead of plumbing store brass. i needed to shave about 1/16" off one side of the flare adapter to clear the hex socket bolt head. I assembled the adapter and plate on the bench and was able to get it very tight. It should be left pointing towards the front of the car, up 45 degrees or so from horizontal; this gives maximum clearance under the car and allows for easy wrench access.
(a hardware store type brass 90 degree hose nipple is shown here; when i switched out the rubber hoses i installed much larger and higher quality stainless steel AN flare elbow.)
Note also that internally, the fitting must be flush or below cover plate flush; I pre-assembled the cover plate and fitting, then filed the fitting (not while installed in the cover!) such that it was 10 or so thousandths below flush when assembled.
the hex head bolts shown in this early photo interefered with the line fitting; I replaced all the bolts with socket head bolts. The top bolt also needs to be 3" long rather than the stock 2.5" given the additional thickness.
This system puts full un-bypassed pump pressure into the inlet of the filter; the pressure relief bypass is downstream of the filter. This is fine as long as the upstream flow restrictions are low, which they are. I chose a filter (initially Wix 1374, recently switched to Wix 51088) that has both anti-drainback and a 10psi internal bypass; when the filter inlet pressure exceeds filter outlet pressure by that much, it lifts off it's seat and bypasses the filter. The stock oil pressure relief valve remains in the stock block location and does it's job from there.
since the filter is mounted upside down, the anti-drainback feature isn't needed.
This oil pump is very old technology. Huge clearances, rough castings, heavy, cheap to make, and reliable. Note the rough casting in the pump outlet! There's a lot of room for improvement here...
My first oil pump failed. Live and learn.
Stock gear-end clearances on the stock pump run about .008 - .009". Hoping to improve oiling, I carefully ground the pump body down so that total gear-to-cover clearance was about .002". Oil volume and pressure went way way up -- 40+ psi at idle, 60 - 80 psi above 1500 rpm. In fact I had problems with the bypass valve not able to dump enough oil back into the pan to keep cold-engine pressures under control. 2000 miles later, the driven gear welded itself to the cover, shearing teeth off the drive gear. (I shut the engine down immediately, it seems no further harm done.)
I now have a dead-stock oil pump in place, with the blocking plate and custom cover and full flow filtration. Pressures are a more normal 20+ psi at idle, 40 - 60 psi hot 1500 rpm and up. Somewhere between these two extremes, .002" excessively tight, .009" factory loose, is probably a happy compromise. Without a specific reason I'm reluctant to do the experiments. Probably dropping clearances to .005 - .006" would make for a healthy increase without any reliabity threat.
note the pump packed with vaseline for initial startup -- this is flatly required. it is not possible to prime the pump externally. it is geared directly to the camshaft, not via the distributor drive gear. it's a slight pain to pack but keep the gasket surface grease-free so that sealer will seal, but this is a critical feature.
when the engine was again overhauled in 2016/2017 this modded pump had been in use for six years and some 50,000 miles. there is some minor scuffing of the cover by the gears. i neglected to photograph the steel blocking gasket.
i did not re-use this pump body and gears; it was used when i started, and by sheer luck a found a brand-new Melling M61 pump to replace it. it received the same modifications and fabricated top cover and is now in service. gear to body clearance is tighter too.
probably the major contributor to the 2010 engine's failure in 2016 was the oil pump pressure relief valve spring, which i had purchased new from Kanter in 2010. some time during the 2016 LeMons Hell on Wheels Rally it collapsed, and fairly suddenly. new replacements are scarce; i bought one at a dear cost from Blaser's AMC. it's dimensions are below, which should be sufficient to source a replacement from industrial sources. (the shorter spring in the pic is the dead one.) in my experimentation before this new spring arrived, i experimented with shimming the old spring. this actually worked well, and since it can be done from outside the engine, is easy to do. shimming would be acceptable to get a new generic replacement to the right output pressure. it's likely not OK for a used spring that's lost it's temper, except in an emergency.
|AMC oil pump pressure relief valve spring, part 3112400|
note that i have found used springs with different number of turns. i did not measure overall length beause they were all 50 years old and i didn't trust that dimension.
Below are some pics of the 1965 full-flow filtration pump I got from Joe. Though the casting was too pitted to be used, I did make mods to it that would have solved the problem.
This casting is based upon the venerable old pump, but has a complicated cover that incorporates the overpressure bypass that dumps oil back to the pump inlet; therefore the filter will never get unregulated pressure. That's a required choice for an OEM environment, but not a much of a worry in mine.
Worse, this pump does not fit the earlier blocks; the block casting is wider at the pump mounting face, because there is a passageway in the pump outlet that requires the block face to seal it. The old blocks have air where the new block has cast iron.
However, I needed to block that outlet anyways, so I fabricated a steel button that would clamp under the pump body and block the main gallery passageway. Additionally, the pump outlet would be drilled and tapped as is the other pump. This made the button dimensions critical (note the paint marks I used to verify alignment and contact area) and in the end I abandoned this path; the other pump is far easier to mod, far more common and is in fact lower-profile than the 1965 pump.
the valve cover design is pretty good but engine oil flows along the rocker shaft and pours steadily right onto the spot where the cover gasket meets the head, and often develop a seep there, even with a new gasket. any tendency to leak is made worse by the oil dripping off the rocker shaft, onto the back edge of the seal.
a simple twist of steel baling wire around the far end of the rocker shaft provides a path for oil to return to the cavity in the head casting. there is now no oil leak or mess even when running with the valve cover off. the same wire twist has been in place for six years. it is tight enough to have a shape, but loose enough that it could never wedge itself between the rocker and washer. Even if it wears into two pieces they'll lay harmlessly on top of the head. here's a brief movie (AVI format) of it in operation.