1/3 Scale Ford 289 Hi-Po

[Sparky_SC]

Always amazed at your work. To call you a engine builder would be a huge disservice, you are more of a engine scientist.

 

[Sparky_SC]

I discovered I that can rev the engine up and keep it running by opening the throttle a max of a couple degrees from idle with the needle valve opened about 2-1/2 turns. (The air bleed hole is currently blocked off.) This particular needle opening is more than twice what I've used in my other (but somewhat smaller) engines. The puzzling thing is that rpm is so sensitive to throttle position. From idle to what may be 3 krpm or so the throttle range is maybe only 2-3 degrees. - Terry

Been pondering those symptoms. If this was a test question on a exam I was taking somewhere my best guess would be that there is air being drawn into the intake somewhere other than through the carb. The far open needle valve and nearly closed throttle plate are compensating for this.

I am not making suggestions to the master ! I just wonder if my reasoning is sound. I learn so much from your posts, thank you !

 

[Arild]

While waiting for the spark plugs to arrive, a pair of Tygon fuel lines covered in a protective wrap were installed to connect the fuel tank to the carburetor. The bundle was secured to the bell housing with a clamp that keeps it out of the port-side exhaust. The .018" restrictor used in the earlier fuel pump tests of the fuel pump was installed in the pump's output hose.

An issue with the elongated Cobra style air cleaners was that they extended over the distributor with very little clearance and as a result required right angle boots on the cap. Dorman 47408 soft vacuum elbows were repurposed for use as right angle high voltage boots. The 289's cap was designed with conical towers to closely resemble the original Ford cap, and this created problems fitting them to the towers. The boots had to be chilled in dry ice and manually opened up with a tapered end mill for acceptable fits.

Another issue with the cap involved the high voltage wire contacts. Each tower electrode is a pressed-in brass insert center-drilled to accept a particular mystery connector pin that I've used many times before. After finishing the cap I realized I didn't have any more of these pins, and the ones I did have had difficulties with the 90 degree bends inside the boots. Instead, I simply tin-soldered together the stranded ends of the high voltage wires to make a crude pin. Even though the eighth inch diameter high voltage wire is a snug fit inside the boot, the wires were doubly secured to the boots with heat shrink tubing. The actual electrical contact is simply a tinned wire supported inside half inch deep hole in the center of the brass electrode.

The spark plug ends of the wires were terminated with straight boots as I've done many times before. A half inch of insulation was removed from the end of the wire which was balled up into a 'spring' contact inside the boot. Shortened Dorman 47400 straight fittings happen to fit snugly over Viper Z1 plugs and were used as straight plug boots. Heat-shrink tubing around the ends of the boots doubly secure the wires to the boots. The plug wires were routed below the valve covers so the covers can be easily removed for valve lash adjustments without disturbing the wires. An assortment of miniature wire looms machined from plastic stock tidied up the plug wiring and kept them off the exhaust manifolds.

A pair of coolant hoses were fabricated from 1/4" silicone tubing. Both the upper and lower hoses need springs inside them to prevent the tubing from collapsing at the sharp bends needed between the engine and radiator. The radiator will be re-installed later.

With everything finally in place for a first start I decided to perform a brief check before finally installing the radiator and adding coolant. The good news was that the starter had no problem with extended cranking times and never got warm during all my testing. The engine wanted to run and eventually did but only for several seconds at a time. So, there's a fuel problem somewhere. The fuel level inside the carb remained where it should while running, the needle valve has an effect, and there's plenty of suction at the carb's air inlet at all throttle angles. Sustained running was only possible though at nearly closed throttle. When it did run it sounded great, and eight warm exhaust runners told me all cylinders were firing.

A large amount of oil being spit out of the exhausts indicated the initial 270 ml oil in the sump was probably too much. I reduced the oil in the sump in steps and eventually arrived at 80 ml which seems low but for now is giving an acceptable light smoky exhaust while the engine is running.

A unexpected problem with the belts coming off their pulleys also showed up during the many starts and stops. In the process of removing the crank pulley to see if its grooves could be deepened, I discovered I'd never installed the bolt that was supposed to secure the keyed damper to the crankshaft (yikes!) The pulleys may or may not need addressing before the radiator is installed.

Feels like a carburetor problem . . . - Terry

View attachment 162267View attachment 162268View attachment 162269View attachment 162270View attachment 162271View attachment 162272View attachment 162273View attachment 162274View attachment 162275

Looking fantastic!!

 

[mayhugh1]

Sparky's diagnosis makes just too much sense. If it hadn't been for the strong suction at the carb air inlet even at wide open throttle, I might have come to a similar conclusion. But it's well worth investigating.

When old-school troubleshooting an intake leak on a full-size engine, I'd typically spray some carb cleaner around the intake manifold and carb base while the engine was idling. If the idle speed was affected the carb cleaner, I'd be in the general area of a leak. With a model engine though it's hard to localize the spray, and keeping the engine idling hands-off is like balancing a razor blade on its edge.

Newer tools have been developed for troubleshooting similar problems in modern cars whose 'money lights' will turn because of leaks so small they would be unnoticed in this particular case. One of these tools I've used to find intake and evap leaks in my own newer cars is a 'smoke machine'. Basically, it creates a slightly pressurized stream of smoke (heated baby oil). For an intake leak test the stream would be injected into the throttle body through a sealed bladder. If there is any any leak at all, the tell-tale smoke will pinpoint its exact location.

For my test it was most convenient to plug the exhausts and inject the smoke into the Venturi opening through a conical rubber fitting stuffed in the throttle barrel. There was no smoke visible anywhere indicating a vacuum leak is probably not my problem.

I'm in new and territory here and will follow up on any suggestions others might have. Although my needle valve design is nearly identical to what I've used successfully in my last two engines, I'm beginning to wonder if I'm just not getting enough fuel to the Venturi when the throttle barrel is opened more than a few degrees.- Terry

 

[Sparky_SC]

Dang ! A smoke leak detector, you have all the nice toys !

I have also heard of using small propane/butane torch, unlit, probing for leaks. Same effect as the carb cleaner method but cleaner. It might be worth a try just for good measure next time you are in running configuration. It wouldn't pinpoint a leak but might show a reaction. I am just wondering if there is any possibility of the smoke tester giving a false negative????? The tester is intended for engines that flow hundreds of CFM in operation, maybe it won't spot a tiny leak that could still cause problems on something this tiny? Beats me, just thinking out loud.

One other thought I had. What about fuel flow? I think of a engine that runs with the choke nearly closed but not when opened, common with a dirty/plugged carb. The nearly closed throttle plate and wide open needle could possibly point to possibly needing excessive vacuum to pick up fuel reliably. Just another of my hair brained ideas LOL Perhaps the same unlit torch near the intake when running to "richen" the mixture would yield some hints?

I have zero doubts you will find the problem soon. I learn much from your builds, the resolution to this problem will also be educational.

 

[mayhugh1]

Dang ! A smoke leak detector, you have all the nice toys !

I have also heard of using small propane/butane torch, unlit, probing for leaks. Same effect as the carb cleaner method but cleaner. It might be worth a try just for good measure next time you are in running configuration. It wouldn't pinpoint a leak but might show a reaction. I am just wondering if there is any possibility of the smoke tester giving a false negative????? The tester is intended for engines that flow hundreds of CFM in operation, maybe it won't spot a tiny leak that could still cause problems on something this tiny? Beats me, just thinking out loud.

One other thought I had. What about fuel flow? I think of a engine that runs with the choke nearly closed but not when opened. The nearly closed throttle plate and wide open needle could possibly point to possibly needed excessive vacuum to pick up fuel reliably. Just another of my hair brained ideas LOL

I have zero doubts you will find the problem soon. I learn much from your builds, the resolution to this problem will also be educational.

Sparky,
The most popular use for a smoke machine is to find a leak an evap leak in a car. These leaks can be small pin hole leaks in the charcoal canister, vent valve, purge valve, or any of the interconnecting lines.
I've also used an mapp gas torch (not lit of course) with a rubber hose on the end, but I'm not sure with a model engine I could stay far enough away from the carb to avoid false positives.

By the way, if you ever have to pay a garage to find one of these leaks for you, you'll find you'll be able to buy a bi-directional scanner, smoke machine, and the replacement parts to fix it for what a dealer will charge you to 'maybe' fix it. - Terry

 

[Sparky_SC]

Sparky,
The most popular use for a smoke machine is to find a leak an evap leak in a car. These leaks can be small pin hole leaks in the charcoal canister, vent valve, purge valve, or any of the interconnecting lines.
I've also used an mapp gas torch (not lit of course) with a rubber hose on the end, but I'm not sure with a model engine I could stay far enough away from the carb to avoid false positives.

By the way, if you ever have to pay a garage to find one of these leaks for you, you'll find you'll be able to buy a bi-directional scanner, smoke machine, and the replacement parts to fix it for what a dealer will charge you to 'maybe' fix it. - Terry

You have one of those little pencil butane torches in your toybox?

 

[mayhugh1]

You have one of those little pencil butane torches in your toybox?

No, just your typical 'thaw the water pipes' torch.

 

[mayhugh1]

For me the most interesting part of a build can be the steps taken to get a newly finished engine running. Too often it's done behind closed doors, and readers who've been following the build miss out on an opportunity to learn about what is sometimes the hardest part of the build. Some of life's things have come up recently and are limiting the time I can work on the engine. But I'll sneak into the shop when I can and detail progress as it's made.

I discovered the problem with the fan belts (repurposed 1/8" o-rings) was that they were sitting in pulley grooves machined for 3/32" o-rings. I deepened the grooves as much as possible but not as much as needed because the crank and water pump pulleys didn't have the extra needed material. The belts were still problematic after re-machining, and the final solution was to order the correct o-rings.

The carburetor was removed from the engine and carefully inspected with the focus being on the needle valve assembly. I was hoping to find debris or a machining issue, but everything looked and measured as expected. While the carb was off the engine, a vacuum port was added to the rear of the high rise spacer below the carb so manifold vacuum could be measured later.

With an air leak ruled out, I next considered the possibility that the needle valve design was so far out of whack that the mixture was much too lean at any throttle setting. Although there's some interaction between the jet and the Venturi, I decided to work just one at a time.

The 289's Venturi size was derived using an equation that I've used before (see the posts on the 289's carb). This equation shows Venturi diameter being proportional to the square root of cylinder displacement. My Inline Six engine with its .382 cubic inch cylinder displacement has a .138" diameter Venturi and has been working reasonably well.

With a 1.076 cubic inch cylinder displacement, the 289's theoretical Venturi worked out to a diameter of .232". However, I elected to take only half of the increase and instead used .180". Right or wrong, I had my reasons. Unlike the large intake valves on the full-size engine my intake valves are identical to the smaller exhaust valves, and if I'm being honest their lift also wound up a bit shy of what's needed to create a theoretical maximum flow. The camshaft was also designed for low speed vacuum rather than a screaming top end because I'd never been comfortable with the sharp unblended runners that resulted from machining the intake manifold from a single piece billet. So, for now I decided to leave the Venturi alone.

The jet size however was another matter. The 289's needle valve assembly was machined identical to that on the Inline Six. The .0225" jet and .026" needle along with the 80 tpi needle gives the Inline Six's carb needle about 2-1/2 turns between full ON and full OFF. The Inline Six ran fine with only 3/4 turn, and I had assumed at least another full turn would be available to the 289.

In actuality, the 289 tried to idle with the needle fully open and the throttle within a flea's eyelash of being fully closed. Rpm was extremely sensitive to the throttle, and an additional degree or two of throttle caused the engine to surge before dying.

I decided to scale the new jet diameter up by the square root of the piston displacement ratio. The 289's new jet diameter became 1.678 x .0225" = .038", and a search through my wife's sewing basket turned up a .042" diameter needle. In a perfect world I'd only have to machine a new needle holder and reuse the existing needle valve body. However, the spray nozzle was too small to be re-drilled for the larger needle, and so I'd have to machine a whole new needle valve assembly.

The 8-80 and 10-56 threads on the micro-machined valve body make it the definition of a PITA part. After many, many hours and scrapped parts (same problems as with the original part) I finally had a new valve body. In the meantime the new fan belts had arrived, and testing showed they stayed in place. The radiator was finally installed and the system filled with 32 oz (1 qt) of 50/50 antifreeze.

I immediately happened across an operating point at about 1/4 throttle and a needle opening of 3-1/4 t (don't really know how many turns is full open with this new needle) where the engine would run at a few thousand rpm. With the engine now capable of sustained running, I didn't try any tuning but instead figured I'd better determine a proper running oil level.

It's surprising how even a little bit of sump oil gets whipped up by the crank windage and the pressure pulses inside the crankcase. For a model engine doing no significant work, I like to run with the highest oil level possible without an unacceptable amount of oil smoke in the exhaust. If everything in a V-8 wet sump engine is working as it should, this smoke will show up first in the starboard exhaust as the oil level is increased. This happens because the webs of the clockwise rotating crank tend to douse the undersides of the pistons in the starboard-side bank.

The sump was drained, and only 50 ml of the original 80 ml showed up in the pan. Hopefully there's 30 ml distributed throughout the engine. The good news is that even without an oil pump, oil has been collecting under the valve covers hinting that my PCV top-end oiling scheme may actually be working. The sump was refilled with 100 ml of 0-W20 and the engine re-started. The oil level was then increased 50 ml at a time until an unacceptable amount of exhaust smoke showed up at the current throttle and needle settings. The starboard side exhaust showed no smoke with 250 ml but 300 ml fogged up the shop. From now on, FULL is 270 ml.

The upper radiator hose is clear silicone and coolant can be seen flowing when the engine is started. The water pump seems to be working well, and the coolant temperature in the radiator after the oil tests was about 50F above ambient.

A quick check found a needle setting that will allow the engine to rev up at wide open throttle, but now idling is a problem. The air bleed doesn't seem to help, and so when I get the time my next focus will be there. - Terry

 

[GreenTwin]

Lots of lessons in this latest post.
My observations:

1. Building model engines requires a lot of perseverance.
Building a complex V8 model engine requires exponentially more perseverance.

2. One has to learn how to get everything to fit correctly, without binding.
I guess the term is "fitter".

3. If everything fits correctly, and is timed correctly, then the fun begins.
It is like writing a complex computer program; you write it; then run it; then you get to debug it.

4. You can see how Mayhugh is very systematic about checking, verifying, and solving issues.
I think a more random approach would not work very well with such a complex engine.
He solves one issue at a time using a pragmatic approach.

5. It is probably not a good idea to make a V8 for your first engine.
It would be easier to build upon your previous less complex successes.
Previous successful engines gives one confidence that the success can be repeated.

Not trying to put words in Mayhugh's mouth, but this is what I am seeing as far as how the master builders make these fantastic builds possible and functional.
Thanks for posting the detailed behind-the-scene explanation, since this helps a great deal in understanding what one is up against before they start a similar build.
.

 

[mayhugh1]

A Holley carburetor is an impressive assemblage of 'kludges' (only from a fuel injection perspective, of course) that supplies fuel to an engine over a wide range of speed and load conditions. The choke provides a rich mixture during a cold start, and a fast idle cam keeps the engine running during its warm-up. Idle circuitry maintains an idle, and off-idle circuitry provides a transition to the main jets. An accelerator pump covers up the stumble that would otherwise occur when the gas pedal is pushed, and a power valve meters in extra fuel to help with detonation when the engine is under heavy load.

Model engines aren't typically loaded but they are expected to start, idle, and rev over a reasonable range including short pulls into w.o.t. Although the choke valve is currently my index finger, the new needle valve is providing good running settings for either idle, mid-range, or wide open throttle up to 4500 rpm. It can also provide a compromise setting for idle through midrange or for midrange through top-end. However, there's no setting that includes a 1k idle which was one of my goals for the engine.

A major limitation of a simple one needle carburetor stems from its Venturi being on the throttle. The Venturi discussions in the earlier carb design posts pertained to w.o.t. In mid-range though where the Venturi's geometry changes as the throttle is rotated toward idle its effective size becomes progressively smaller. Higher and higher velocity air draws more and more fuel from the spray bar, the floor drops out of the A/F ratio, and the engine dies.

An air bleed added to this simple style carburetor can extend its idle. An air bleed is an atmospherically vented hole that's cross-drilled through the throttle bore to intersect the Venturi as the throttle is rotated toward idle. When this hole is uncovered by the Venturi the additional air flowing through it tends to lean the mixture and extend the idle to a lower rpm than would otherwise be possible. The diameter of the hole is determined experimentally, but mine usually end up around 1/3 the diameter of the Venturi. The 289's vent hole has a sliding cover to fine tune the amount of air entering the bleed passage. Depending upon exactly how the hole intersects the Venturi, the additional air may also draw some fuel from the spray bar.

One of the photos is a cross-sectional rendering of the 289's carburetor showing the original air bleed being uncovered as the throttle is rotated closed. Tests with the vent hole fully covered and the needle valve adjusted for a midrange setting of 2500 rpm showed the engine stalling at about 1600 rpm as the throttle was decreased. With the vent hole fully open, the carb was able to reach a minimum idle of some 1450 rpm. Again the engine stalled when the vent hole was re-covered.

An identical second bleed hole intersecting the original atmospheric vent was drilled parallel and just above the original bleed hole. The total area of the two bleed holes is equal to that of the atmospheric vent. Repeating the test showed the minimum idle had moved down to about 1300 rpm while the vent hole was totally uncovered, and again engine died when the vent was covered.

Since testing showed the lowest idle always occurred with the vent hole completely uncovered, it's possible that more bleed would allow an even lower idle. Enlarging the bleeds and/or vent was too risky at this point though.

Drilling the bleed passage(s) to intersect the Venturi above the throttle's axis as shown in the two photos provides a slow transition to idle since the extra air also draws some extra fuel from the spray bar. Drilling a bleed hole from the opposite side of the throttle bore so it intersects the Venturi below the throttle's axis provides a more abrupt transition since less fuel will be drawn by its extra air. This third bleed is on my short list of modifications to try next.

A better flywheel (or in my case any flywheel) may allow a lower idle, and it should reduce some of the engine's annoying vibration. The starter clutch isn't a conventional Bendix-type mechanism that disconnects the starter from the flywheel while the engine is running. Instead, the starter is permanently engaged with the flywheel with a sprag clutch disengages the flywheel from the crankshaft while the engine is running. The result is that external to the engine there's no significant rotating mass on the crankshaft. Attending to this is the second thing on my to-do list.

In the meantime though ...

I've run nearly a quart of gasoline through the engine while playing and getting to know it. The top-end looks like it will be 4500 rpm which is 1500 rpm below the original HiPo's redline. This is pretty much what I had expected from the model's small intake valves and their less than optimum lift. I don't yet know if the CDI ignition is a limitation.

The engine has been burning an average of about half an ounce of gasoline per minute. The 3 oz fuel tank allows about six minutes of run time. I've done two six minute runs which included mostly mid-range and some w.o.t. run ups, but I wasn't comfortable with the resulting 175F exterior head temperatures. In both runs the coolant temperatures in the radiator matched the head temperatures. Most of the fuel was consumed with more comfortable three minute runs where the head and coolant temperatures reached 145F.

The PCV draft system that was designed to lubricate the top-end is working well. With the valve covers gasket'd to the heads, the valve train components accumulate an oil film, and oil collects on the valve covers under the vented filler caps.

The exhausts run cleanly with no smoke, and I recently removed and inspected the plugs, and their colors are very acceptable.

The engine seems to like 20 degrees BTDC at all rpms. In the process of playing with the timing I discovered the rotor tip electrode had been rubbing against the cap electrodes. By the time I'd discovered the fine brass debris inside the cap, the rotor had already lapped itself in. However I removed two thousandths from the tip for additional margin, and I'll revisit the timing later.

Other than its higher than desired idle, my biggest disappointment with the carburetion is that there's no single needle setting that includes both a cold start and reliable mid-to-high speed running. The engine takes nearly a minute to warm up to the where fuel atomization stabilizes, and during this time the needle must be fiddled with. My other (and smaller) engines heat up much quicker, and so the fiddling time is much less.

The plan is to next tackle the flywheel issue and to add an opposite-side bleed hole. - Terry

 

[GreenTwin]

When I flew RC airplanes, I had a lot of trouble with getting a consistent low idle speed, and the engine would often quit running on approach.
I ended up pressurizing the fuel tank, and that solved my idle problems completely, but I can't explain why.
.

 

[dsage]

Terry. You should be suspicious of the ignition system keeping up. I believe you said you were using one of Roy's CDI modules. He states the maximum RPM for the module is 20k for a SINGLE cylinder engine. If you divide that by 8 you get about 2500 rpm. I do get better than that with a CDI on my V8 but it does appear limited somewhere above that. You might try the Sage/Gedde ignition module and a COP coil temporarily if you are suspicious. You won't have any limitations then. I believe George uses that arrangement and claims 7800 rpm for his 302.
I believe the saying goes "most carburetor issues are ignition" (or maybe it's the other way around )

 

[petertha]

When I flew RC airplanes, I had a lot of trouble with getting a consistent low idle speed, and the engine would often quit running on approach.
I ended up pressurizing the fuel tank, and that solved my idle problems completely, but I can't explain why.

Fuel pressurization in model AC is mostly driven by a need to overcome, or at least stabilize, varying hydrostatic head changes of the fuel to carb inlet via the tank + line. If the nose points up, the tank is effectively lowered, which leans AF mix to carb system. If the nose points down the opposite: tank is effectively higher which richens AF mix. If you bank & yank or roll or spin... basically any maneuver where G-forces are involved, that potentially magnifies the hydrostatic variation effect because the fuel sees these G forces. So boosting the tank pressure helps to a degree & is simplest to implement. Pressure regulation is a better way, but more/different parts & tuning.

A fuel system for a stationary engine doesn't have to deal with most of these factors. A float bowl takes care fuel level hydrostatic changes for example. But obviously there can still be further tuning significant complications.

The RC engine analogy came to my mind too, maybe for a different reason. There are plenty of examples of multi-cylinder 4-cycle (usually aircraft related) gasoline spark ignition engines with what amounts to a Walbro style carb. Many have enviable idle/transition/full speed. Those carbs look kind of crude, but there is actually a lot happening behind the scenes. It wasn't until I watched a bunch of detailed fuel circuit explanation videos did I appreciate (something less than 'understand' haha) what was going on. They are part pump, part float chamber, low speed, high speed & transition circuit all in a little cube. Anyways none of this helps the current situation & I'm pretty sure Terry has implemented Walbro style carbs on other engines so is already familiar. I'm just a bystander but it kind of looks like one of those design issues with 5 different knobs to turn. Somewhere among the setting permutations is an ideal setting, but getting there could be a journey. And 'knob' is a simplistic euphemism because its probably related to orifice size, or needle taper or.....

Anyways, beautiful build Terry. I'm confident you will get there in the end or close enough for your high standards. I have learned a lot just watching your builds & explanations along the way!

 

[mayhugh1]

In preparation for the flywheel retro-fit, the bell housing was removed from the engine while it was still on the display stand. The flywheel's workpiece was a 4-1/2 lb Stressproof disk originally purchased as a back-up workpiece for the ring gear. The flywheel itself was designed to be a bolt-on to the sleeve on the clutch's inner race which in turn is keyed to the crankshaft. It was also located as close as possible to the ring gear so its o.d. could be maxed out inside the confines of the bell housing.

There was room for an additional 1.33 pounds of steel inside the bell housing, and SolidWorks' estimate for the finished flywheel's moment of inertia was 2.82 lb-in^2. In comparison, the moment of inertia of the ring gear is 0.82 lb-in^2 although it doesn't rotate when the engine is running.

The plan was to make just one change at a time, and so the carburetor's third air bleed was put on hold until after the engine was tested with its new flywheel. But since the engine had accumulated an hour or so of running time, I decided to inspect the bottom-end for any signs of wear or fasteners that might had been loosened by vibration.

The oil and coolant were drained, and after unbolting the engine from the display stand it was returned to the engine stand. Upon removing the oil pan, I was surprised to see how much the rubber pan gasket had grown. One side was nearly a quarter inch longer than when it was installed. Conventional engine oils usually cause rubber gaskets to swell, but I'd used a synthetic blend and wasn't expecting much of a change. When it came time to reinstall the oil pan, the old gasket was too distorted for reuse, and a spare gasket set made up earlier was used instead.

The oil was pretty dark for just an hour of running. Some discoloration was to be expected from combustion chamber products getting past the rings during break-in. However, the oil's carbon black color was an indicator of too much time spent running too rich while playing with the carburetor. The reasonably colored plug insulators seen earlier seemed to say that my tuning had been wildly see-sawing between rich and lean.

The bottom end looked good. Rod clearances didn't seem to have changed, and the rod and main bearing bolts were still tight. A couple main and rod caps were opened for inspection, and their bearing surfaces and corresponding journals were shiny and scratch-free.
While the pan was off, a new magnetic drain plug was machined to replace the original one which had a thread problem and had been seeping. The engine was returned to the display stand and refilled with new fluids including 270 ml of 5-W20. Heavier oils can help consumption problems, but since I'm not using an oil pump and there's no sign of ring problems, a thin oil that will more easily find its way into bearings seemed more appropriate. Finally, the clutch assembly was reinstalled including the new flywheel.

After reassembly, the engine started up easily and ran with no signs of vibration. It was immediately obvious that I was gravitating toward needle settings were now leaner by 3/4 turn. After a 30 second warm-up, just 3 turns open provided a smooth throttle response between 850 rpm (finally!) and 3200 rpm with the air bleed vent 3/4 covered. Opening the needle another quarter turn moved the top end up to 4500 rpm but raised the minimum sustainable throttle to 1200 rpm. Spark timing was revisited, and as before the engine seems to be happy with 15-20 degrees BTDC throughout its operating range.

The flywheel cured a couple ills. My guess is that a lack of significant mass on the end of the crankshaft to absorb torsional oscillations excited by the engine's power pulses allowed them to resonate at unfortunate speeds in the engine's midrange. The resulting engine vibrations died away when the engine was revved up, but they were always present in the rpm range where the engine would spend most of its running time. Besides being annoying they took away a lot of the fun of running the engine. At 1/3 scale I think the crankshaft was long enough and the power pulses were strong enough to create a problem that wouldn't have been noticed at smaller scales.

The flywheel significantly lowered the minimum idle rpm from 1300 rpm to 850 rpm. Some of the running improvement was inertial, but a portion resulted from the carb tuning trending leaner after the addition of the flywheel. The previously richer settings overpowered the air bleed(s) even with the air vent fully open and also raised the minimum idle speed. My transitional bleeds which also supply a bit of fuel tended to complicate the tuning since they were also supplying a bit of fuel.

With a nearly 4:1 throttle range for a single needle setting, the 289's carb is working as well as any of the other air bleed carbs I've built. I'm not sure if the third air bleed below the throttle's centerline that I've been considering is going to make any additional improvement. If it does, I'll post the results. In the meantime I've made a Youtube video of the engine running as it currently does with the flywheel. If the third air bleed makes a significant difference I'll also make a new video. - Terry

 

[Sparky_SC]

Fantastic ! I have been very anxious to see/hear it run and it sounds great. I especially like how you approach things like a scientist. No doubt I will read this thread several times to digest the information within. The build certainly proves that the whole is equal to the sum of its parts.

Thanks for posting so much detail !

 

[michael-au]

What a fantastic engine, looks like a full size one, its has been a great thread to follow, thank you for sharing the journey

Michael

 

[pileskis]

As usual, just outstanding craftsmanship!
Like others, I appreciate the time and methodical effort you put into documenting everything.

Sid