B&S 49-series (810cm3/49ci) - TiPi's conversion for aircraft use

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
It has been suggested to open another thread specific for the conversion of the Briggs & Stratton 49-series of vertical shaft engines to an economical and reliable aircraft engine. So here it is:)

The Briggs & Stratton 49-series family currently (2019) consists of 4 streams of engines:
  • Commercial series (49T, 26 & 27hp)
  • Professional series (49S, 26 & 27hp)
  • Vanguard series (49V and 49R, 24 & 26hp)
  • Vanguard EFI series (49E, 24 & 28hp)
The base engine is the same for all models, the major differences are the air filters and some parts between Vanguard and Commercial/Professional (conrod, exhaust valve, valve cover, starter motor, crankshaft).
Suitable engines are the Professional & the Vanguard series, with 1 1/8" crankshaft (28.5mm).

General specifications:
Bore: 83.81mm (3.3")
Stroke: 73.41mm (2.89")
Displacement: 810cm3 (49ci)
Mixture: 1 twin-barrel Nikki carburetor (28mm throat, 22mm ventury), throttle body injector system on EFI model with O2 sensor
Lubrication: full pressure with full-flow filter and oil cooler
Power: from 24 to 28 (older versions to 30hp)
Weight: generally around 40kg complete, 41.5kg with the big external airfilter

Major advantage over other engines: forged crankshaft, forged alu conrods, built-up camshaft (not cast iron), automotive-style pistons and plated EX valve (Vanguard only). This results in stronger components in critical areas

upload_2019-10-2_20-14-25.pngupload_2019-10-2_20-14-58.pngupload_2019-10-2_20-15-26.pngupload_2019-10-2_20-15-55.png

The B&S 49-series is currently the best value, lowest weight and best quality engine for a 33 to 35hp (maybe up to 37hp with some more work) direct-drive engine conversion to be used in a single-seat very light aircraft. The complete engine without fluids is around 32kg, which is comparable to a Rotax 377, Kawasaki 440 and definitely a lot lighter than a 1/2 VW or the Citroen 2CV. This engine is now the most suitable engine option (called the SE-33) for the Spacek SD-1 Minisport, developped by Igor Spacek.

To achieve a successful conversion from vertical shaft to horizontal shaft, an engineering analysis is required to identify and address the areas affected by turning the engine. Specek has elected to turn the engine with the heads up as that fitted in with the other engine option, a horizontal engine from a B&S competitor (can't name them, their lawyer is watching but the engine model is a CH750).

I have elected to turn my engine up-side-down (after all, I live down-under), meaning heads down. I'm doing this for a number of reasons:
  1. I don't like the bumps on top of the cowling for the cylinders or having the cylinders stick out
  2. the prop axis can be raised to the correct level and the cowling has a clean curve from the spinner to the fuel tank
  3. the oil pump is lower and easier to feed with oil
  4. cooling can be improved by having the air inlet further away from the spinner (higher velocity)
  5. the engine fits within the firewall outline (most streamlined installation)
The main issue with turning a vertical engine to horizontal is the oil pickup. In its original installation, the oil pump is at the bottom of the sump and flooded with oil. Located in the smae axis as the camshaft and driven through a drive adaptor from it, the oil pump inlet is then a considerable distance above the oil level when upright (camshaft above the crankshaft). Oil pressure issues on start-up require a small electric priming pump to prevent any daage to the engine.

In my case, the oil pump is below the crankshaft but the "sump" would be in the cylinder heads (lowest points). The "easy" solution is to install a dry-sump system with the normal oil level at the pump height.

to be continued:p
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
part 2 - What other engines are out there?

The B&S 49-series is not the only suitable V-Twin engine. There are a number of Colomban MC-30 Luciole flying with specially adapted B&S 38-series Vanguard engines. They are 627cm3 (38ci) and rated up to 24hp. The conversion from Michel Colomban is probably the lightest 25hp 4-stroke direct-drive engine available (no flywheel, direct-drive from flywheel end through a special adaptor, custom-made ignition and alternator). 24-25hp is for extremely light aircraft like the Luciole (200kg MTOW) or very light (pilot) SD-1 Minisport (240kg MTOW).
upload_2019-10-3_21-19-31.jpeg

A number of European manufacturers have used 38-series engines as PPG propulsion units, with a belt reduction drive. They get 35-40hp out of this engine at 4,100 to 4,300rpm. This does require many more components to be upgraded from stock (conrod, valve springs, camshaft, carburetor, head porting).


Kohler CH740/750 have also been used in a number of early SD-1 Minisports. The engine is rated at 30hp but will weigh about 4kg more than the B&S 49xxxx (36-37kg). Big drawback is the cast iron crankshaft.
upload_2019-10-3_21-28-0.png

Honda GX690 are a very well built engine with forged crankshaft & conrods but again, quite heavy for the displacement and power (688cm3/42ci and 24hp) and about 4-5kg heavier than the Briggs 49xxxx. This engine is unique amongst the industrial V-twins as it has an aircraft engine style one-piece cylinder and head assembly.
upload_2019-10-3_21-32-22.jpeg

Subaru EH72: this was initially my preferred engine, well built (forged crankshaft, roller bearing) but again, heavy at 46kg. Displacement is 720cm3/44ci with a rated output of 25hp (carburetor) or 28hp (EFI).
upload_2019-10-3_21-39-6.png


Kawasaki also has a few models in the 750/850cm3 range but they are extremely heavy (56kg vs 40kg for the Briggs 49xxxx).

Kohler and B&S engines in the 900-1,000cm3 range (54-61ci) are also too heavy, most of them weigh 56-61kg.
 

Attachments

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
part 3 - which end to use for the propeller?

There are a number of pros & cons for bolting the propeller to the flywheel or the PTO shaft. In the end, the specific installation requirements or personal preference will decide where the propeller goes. Here I'm just listing some information that I have gathered over the years and this is applicable to the B&S 49-series (adjust for other engines as needed):
  • the PTO end has a quite large (41.2mm dia and maybe 50% longer) bearing compared to the flywheel side (35mm dia)
  • flywheel side has a replaceable bearing insert
  • flywheel end has a taper shaft stub (better torque transfer capability)
  • PTO is a standard size shaft of 28.5mm (1 1/8") dia with a keyway cut (stress riser)
  • PTO cover is extended out by about 50mm compared to horizontal engines (and the flywheel end), giving soem extra clearance from the prop
  • exhaust is on the PTO end and it is preferable to face the exhaust into the wind (tractor installation)
upload_2019-10-3_22-20-26.png

In the end, I selected the PTO end as my propeller drive. The main reasons are the extra space between the head and prop face, the exhaust facing the wind and the beefier bearing.

What is important when using the PTO end for the propeller is that the flywheel is being removed or replaced by the absolute minimum mass to hold the ring gear, ignition magnet (and counter weight) and possibly the alternator magnets. Having a rotational inertia at both ends of the crankshaft WILL break it at some stage.

There are also many conversions flying with the propeller fixed to the flywheel through an extension shaft. Example of the SD-1 Minisport out of the factory.
CIMG1592.JPG
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
part 4 - which way to turn the engine?

So far, I have selected the Briggs 49xxxx (in my case, the older model 49M977), and will fit the prop to the PTO shaft. Now we need to decide which way to turn the engine. As the original engine is a vertical shaft, the engine needs to be turned horizontal unless you use it to drive a helicopter (Mosquito with governed 4-stroke engine???). There are 2 possibilities:
  1. "right way up" as in cylinder heads up
  2. upside-down (cylinder heads down)
Both have their advantages and disadvantages and will also be influenced by the choice of aircraft for this engine. What are the things to consider?
  • going from a vertical shaft engine to horizontal will require re-working the oil system (reservoir, oil pickup, dipstick etc)
  • the carburetor and intake manifold need to be changed as the carb needs to be horizontal (if you want to use the same one)
  • cooling of the cylinders and heads
In the case of my SD-1 Minisport, I have decided to install the engine upside-down for the following reasons:
  • to keep the engine fully inside the firewall outline (I like the look of the Super Kingair and Q400 turbo-prop cowlings)
  • the prop axis will be at the ideal height, the same as the Verner JCV-360 and Hirth F-23 installations (geared engines)
  • the oil pump is lower (below the crankshaft), making the oil system a bit simpler with reduced suction height
  • this installation will require a dry-sump system (more details to follow)
Sketch with the engine in the upside-down position on the SD-1 fuselage
upload_2019-10-4_11-49-25.png
upload_2019-10-4_11-51-50.png

SD-1 with the engine upright and prop axis at the same height
upload_2019-10-4_11-53-42.png


Verner JCV-360, belt reduction drive with prop axis above crankshaft
upload_2019-10-4_12-6-37.jpeg

SE-33 (Briggs 49xxxx) showing bumps on cowling to fit cylinder heads. The prop axis (engine) has been lowered to reduce the intrusion into the pilots view.
upload_2019-10-4_12-7-26.jpeg
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
part 5 - the oil system

Now we are getting to the juicy bits, the modification of the oil system. In the original configuration, the engine has the oil pump at the PTO end of the camshaft, driven by a a small driver that fits in a slot on the camshaft and the oil pump gear. The pump is of the gerotor style, with a 5 lobe ring (lobes on the inside) and a 4 lobe gear (lobes on the outside). The oil level is above the pump and the pump is fully submerged. The oil goes from the pump outlet to the filter adaptor on the PTO cover. The pump outlet also has a (safety) relief valve and there is a bleed-off for the camshaft bearing. The oil filter adaptor diverts the oil to the oil cooler, then through the oil filter and back into the PTO cover. From there the oil is split to the PTP main bearing (this also supplies the big end bearing) and through a hole, into the main case. The oil gallery in the main case feeds the flywheel main and the camshaft bearings. All other lube points (cam lobes, cam followers, push rods, valve rockers, pistons & rings, small end bearings) are lubricated by the oil spray and oil mist in the crankcase.

This is a B&S video explaining the lube system in a vertical engine to some degree, the V-twin starts at 0:50

So, what do we need change to maintain the original conditions when turning a vertical shaft engine horizontal?
  1. oil reservoir: the oil needs find a new home
  2. oil pump pickup: the oil pump is now well above the oil level and the oil pump inlet needs to be extended to where ever the new reservoir is
  3. oil drain back: where does the oil in operation now collect, how do we get that to the reservoir?
  4. breather: the original breather location is at near the camshaft bearing at the flywheel end, is this still an oil-free high point?
  5. oil drain: where is the lowest location to fit an oil drain?
Some data for the pressure lube system:
capacity: 1.9-2.0lt
pump volume: approx 12lt/min at 3,600rpm (3.5ml per crank revolution)
normal operating pressure: 1-3.5bar (15-45psi), relief valve seems to be 3-3.5bar

In a heads-up installation, there is a possibility to use the space at the bottom of the crankcase as oil reservoir (sump) but the capacity is somewhat limited in order to keep the oil away from the rotating crankshaft. A bolted-on extension might be advisable (see VW sump extensions).
In a heads-down installation, there really is only a dry-sump system that will work correctly. Thsi could also be an option for a headsup installation.

What options do we have for a dry-sump system?
  1. gravity system: won't work due to the distance from the oil level to the pump
  2. pumped system with scavenge pump: standard system for most automotive, motorbike and many aviation applications. Requires an additional (or several) pumps to collect the oil and pump it to the external reservoir
  3. blowby pressure with external tank: successfully used in thousands of Rotax 912 engines
Due to its simplicity, I have chosen the Rotax system for my engine. Here is a diagram and some explantions:

the oil pump draws the oil from the oil tank, through the cooler (which is rather unusual)
from there it flows through the filter and to the lube points
the used oil collects at the bottom of the crankcase and is pushed back to the tank by the blowby gasses from the running engine
the blowby gasses leave the oil tank through the crankcase breather in the oil tank neck, after the oil and air have been separated through a splash plate and a mesh screen

simple, no moving parts other than the pump:)
upload_2019-10-4_20-1-12.jpeg

In the case of the 49-series engine with the heads-down, the drain is from both rocker covers (lowest point of the engine), through a check valve (the 4-cylinder engine has no crankcase pulsing but the V-twin does due to the pistons not moving in unison) to the oil tank. The oil tank employs the same principles as the Rotax tank, splash plate and mesh separation of the oil and air.

The perforated tube is the return line, with the openings facing the back wall in the tank.
The mesh screens removes more air from the oil as it passes through
The marker pen is an additional screen in the breather outlet, filled with SS mesh to trap remaining oil dropplets
Total tank volume is 2lt, oil volume will be around 1.25-1.3lt and 0.5lt in the engine
upload_2019-10-4_20-11-4.jpegupload_2019-10-4_20-16-16.jpeg

The tank will be mounted so that the oil level is at or slightly above the pump. One thing this oil pump is not good at is drawing oil from a lower level after some time of no use (pump is losing prime). This can be prevented by having the oil level at or above the pump or by priming the pump with an auxillary pump prior to starting the engine (manual or electric).

Modification to the oil pump inlet: the PTO cover has a small opening for the oil pump inlet, covered by a clip-in screen. Since there is no oil there anymore, an alternate method of connecting the oil feed line to the oil pump needs to be found.

There are basically 2 options:
  1. an adaptor needs to be manufactured to enable the addition of an oil pump feed line from the new oil tank
  2. block the openening completely and add a suitable fitting to the oil pump cover on the outside. The pump side of the cover would need to be reworked to mimic the case-side passages and the case side passage needs to be closed off (covered).
At this stage, I'm looking at option 1 but haven't really worked out yet how to do that.
upload_2019-10-4_20-38-17.png
upload_2019-10-4_20-45-36.png

Some other considerations:
Lubrication of the valve gear in the cylinder heads: in the upside-down installation, the oil will flow through the heads and provide more than adequate lubrication. In a heads-up installation, there will be a need for some trickle oil as the pushrods are solid and won't transport any oil

Hydraulic lock with heads-down (when parked): this is a possibility but is easy to address with pulling through the engine as part of the pre-flight. This is actually the mandated procedure for the Rotax 912 before checking the oil level. It is called "burping the engine" as a distinct sound from the open oil tank indicates that all remaining oil from the crankcase has been transferred to the oil tank (the level dipstick is in the tank).

Oil in the crankcase when running: while the engine is running, the oil will be thrown around the same as in an upright installation, the only difference is that the oil will run down the inside towards the camshaft valley and then drain through the passages to both rocker covers. There will be a small volume that stays in the camshaft valley that might need to be drained through an extra drain plug at time of oil change.
 

Attachments

Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
Part 6 – intake and exhaust

V-twins with a common crank pin for both cylinders (see crankshaft in part 3) are quite unique. The result of having a single crank pin and cylinders off-set from each other at 90° results in an uneven firing interval. There are 450° of crank revolutions between firing of #1 and #2 and 270° between #2 and #1 (or the other way around depending how you number them). This has consequences for the intake and exhaust systems as they time between the intake events and the exhaust events is uneven.

Essentially, the 90° V-twin has to be viewed as 2 single cylinders and treated in the same way. This means that the intake system (air filter to cylinder head) and exhaust system (cylinder head to tail pipe outlet) need to be completed separated between the 2 cylinders.

How is this achieved:
  • if possible, use 2 air filters or at least 2 pipes from the air filter to the carburettor
  • 2 carburetor or as most V-twins have now, a double-barrel carburettor
  • Separate intake manifold of equal length and volume (cross section)
  • Cylinder heads are mirror image so symmetric
  • Exhaust pipe and any muffler separate to the tail pipe end (a common muffler can work if the tail pipe ends at the muffler inlet eg no continuation inside the muffler)
The dynamics in a common intake or exhaust (pressure waves and pressure pulses) will play havoc with cylinder filling and mixture distribution. To get maximum power and minimum vibration, each cylinder needs to fill with an equal volume of air and with an equal volume of fuel. This is only possible with fully separated intakes and exhausts.

The 49-series engine is fitted with a double-barrel Nikki carburettor of 28mm throat (throttle or butterfly plate) diameter and a 22mm venturi (cast in the body). It has a common shaft for both butterfly plates, a common float and needle valve and float chamber. Each barrel has its own main jet, idle jet, emulsifier tube and bleed air jet. None of them are adjustable and only the main jets are replaceable.

upload_2019-10-9_20-55-12.png

If you look closely, you can see the screws holding the butterfly plates to the shaft are rather large and the venturi has some visible flash at the narrowest section. There is also a step from the venturi to the nominal bore à room for improvements.

upload_2019-10-9_20-56-8.png

Calculations that I have found suggest a 28-30mm throat for the displacement and rpm. The ratio of the venturi to the throat is a bit on the small side (0.78 for dia or 0.62 for area). Ideal would be a 23mm venturi or even 24mm (as this is application is more like a stationary engine, rapid throttle response is not a priority). Testing will show what can be gained.

One compulsory mod is the remove the “anti back-fire solenoid” that is fitted to the bottom of the float bowl and cuts off the fuel supply to the main jets if not powered up. Bad news if you are not on a ride-on mower.

The exhaust should be 25-26mm ID (the ports are 25mm) to maintain sufficient velocity. There are numerous website calculators for intake and exhaust diameter and length calculations but in practical terms, none of them really fits under the cowl of a really small and slow revving engine. It will be mainly a matter of what will fit.

Mixture control: This carburettor has 1 (one) adjustments, the idle stop. No idle mixture or any other adjustment. This limits what could be achieved with proper tuning of the mixture in the idle, part load (cruise) and full load (WOT) operation range. This carburettor will also not lean at lower air densities (density altitude) and will run rich at altitude.

One relatively simple way is the use the Hacman principle to clean the sea-level mixture when at altitude. A similar system sold by ”Holtzman Engineering” was developed for the snowmobile crowd, they also experience large altitude variations.

The principle is to apply a very small (pilot controlled) vacuum to the sealed float chamber that will then reduce the fuel flow at a given throttle opening (air flow), thus leaning the mixture. In the double-barrel carburettor, a single control will equally affect both cylinders.

A possible option for a bit more power is to fit the carburettor off the 54- and 61-series engines (30mm throat), some have adjustable idle mixtures but a leaning device will still be required. It is doubtful if the larger diameter will bring a noticeable power increase at the standard engine rpm and without extensive intake port and valve work.
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
part 7 – ignition and alternator

The Briggs, like just about all other industrial engines, is using a self-energised ignition coil, one for each cylinder. They are mounted on the flywheel side of each cylinder, facing the magnet that is located on the outer circumference of the flywheel.

The coil pack (B&S calls their ignition Magnetron) is a fully sealed, non-serviceable package containing the primary coil, the secondary winding and the trigger electronics. The ignition timing is fixed at around 25° BTDC with no retard or advance.

It is very difficult to find useful information on the Magnetron ignition. The general working principle is known and shown below:
upload_2019-10-11_21-1-45.png
These coils also have a P-lead that is used to stop the engine (grounding it). This lead needs to be shielded and separated from the other coil (via 2-pole switch or a diode in each lead) to stop interference between the coils and stop radio noise. Different manufacturers use different magnetic poles for the trigger eg some use the N pole and others the S pole.

Due to the relatively large ignition advance while cranking, most engines are fitted with a compression release (the exhaust valve is kept slightly open during the compression stroke). This is to prevent kick-back during cranking that can cause starter motor damage and shear flywheel timing keys. Another way to prevent kick-back is to install an ignition switch separate to the starter switch, this way the starter can be engaged with the ignition still off and the ignition switch is only activated once the engine is cranking nicely.

Kohler did have an external ignition retard module on some models that would retard the ignition by about 15° below 800 or so rpm. Honda does the same, built into their 630/660/690 GX engines coils, retarding to 9° BTDC under 1,000rpm and returning to the 23° BTDC above that.

A drawback of the advanced fixed timing is a rather harsh idle at very low rpm. Most industrial engines don’t idle below 1,500-1,800rpm when connected to the governor (my 49M idles at 1,860rpm, factory set). In the aircraft, a stable idle as low as possible is desirable to reduce the sink rate on approach & landing. Slippery planes like the SD-1, Luciole, Pik-26 etc are difficult to slow down and increase sink with high engine idle speeds.

This shows the distance required to change the ignition timing. TDC is the white line, then in 5° increments to 30° BTDC.
upload_2019-10-11_21-2-36.png
What about dual ignition???
Dual ignition was necessary when they had simple magnetos with lots of mechanical parts (points) and it has an advantage with larger-bore engines. In these small-bore engines, there would be just about no measurable power gain. More importantly, there is no room to install a second sparkplug.

What people also forget is that you have a two-ignition system as both coils are fully self-contained and independent of each other (provide their P-leads are separated). The engine will run horribly and will only produce about 40% of max output, but as they say with twin-engine planes, it will take you safely to the crash site.

View of the alternator and ignition coils, alternator wiring is the yellow connector (2 AC cables).
upload_2019-10-11_21-3-18.png
The alternator is a 16A permanent magnet alternator with 8 windings and 12 magnets inside the flywheel. The coil is the same for 10A and 16A output, the only difference is the size of the magnets. The voltage generated in the alternator coils is up to 30V AC, that AC is taken to the combination rectifier/regulator where it gets rectified and the output voltage limited to around 14.4V DC. The regulator is of the series-type, meaning that the AC input side will go open-circuit when the battery voltage reaches its upper limit.

The alternator output is highly rpm-dependent, as this graph shows (not from B&S but same style):
upload_2019-10-11_21-3-56.png
Due to the nature of the PM alternator (only single phase, not 3-phase as the automotive alternators) and the limited rpm, there is some ripple on the DC output that might require some additional treatment to reduce noise and prevent sensitive electronic equipment issues. Rotax strongly recommends the installation of a 22,000μF capacitor for the smoothing of the ripples from their PM alternator. It also assist with maintaining correct voltage if the alternator is disconnected from the battery eg battery master switch off and running directly off the alternator.

The correct way for a regulator is to open the circuit on the AC side (series regulator), not to short it (shunt regulator). Shortening the AC side will cause power loss and heating of the coil. Series regulators are more efficient as they simply turn off the supply side rather than waste the surplus as heat.

For trouble-free operation of ignition and alternator, the ignition coils, the alternator coil and the rectifier/regulator need to be kept as cool as possible. Depending on installation, some small cool air might need to be directed to these items.
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
Part 8 – cooling

As an air-cooled engine, the cooling efficiency is greatly dependent on sufficient air flow past the right surfaces. Air is relatively poor in specific heat removal (heat per m3 of air), requiring rather large volumes of air flowing over a relatively small surface area that needs the heat removed. Fluid to air heat exchangers (radiators & oil coolers) have the significant advantage of being able to use much thinner fins, resulting in a much larger surface area for the heat exchange. Air-cooled cylinder heads and cylinders need a certain fin thickness to prevent vibrations from flexing the fins and breaking them off. Sometimes they even need rubber blocks (combs) to stop fins vibrating at certain rpms.

The Briggs (and all other industrial engines) are designed for harsh applications and have some resilience designed into the system. At the same time, they are fan-force cooled, which is a great help.

Download yourself a copy of the “Honda Engine Application Manual”, the “Honda GX240-390 Technical Manual” and the “Kohler Engine Application Guide”. Absolutely essential reading applicable to any industrial engine that is looking for a new home.

Here are some specifications from Honda and Kohler for the installation of their range of engines:
upload_2019-10-12_14-5-24.png
“Converted” means normalised to 40°C ambient temperature eg what the temperature would be at 40° ambient.
upload_2019-10-12_14-5-47.png
upload_2019-10-12_14-6-0.png
These numbers can be used to estimate the cooling air flow for any size of industrial engines (3,600rpm, adjust for extra hp) as this is a linear function.
upload_2019-10-12_14-6-37.png
Then do whatever is required to stay within those limits to ensure a trouble-free and long engine life. This means that there is a minimum of instrumentation and monitoring required during the test phase of any new engine, and preferably keep it as permanent instrumentation.

The people using Jabiru and VW engines have the same battle with their engines, trying to keep them cool enough. The main problem is that the aluminium used in the cylinder heads is getting soft (annealed) at elevated temperatures (the temps vary a little bit between alloys but they start from 300°C). Keep in mind that the temperature at the temps sensor might be quite a bit lower than the hottest section of the head (around the EX valve seat).

Here is a list of problems that will happen if temperatures are exceeded:
  • Valve clearance: frequent need to adjust valve clearance indicates a problem with the valve (stretching), Valve seat (wear or recess in the head), deformation of the cylinder head (VW and Jabiru with heads pulled onto the cylinder with studs)
  • Valve failure: both VW and Jabiru have experienced EX valve failures from running at excessive EGTs, causing material failure of the valve stem
  • Valve seat failure: EX valve seat falling out from recess in head, usually results in the EX valve hitting the piston and catastrophic failure
  • Leaking head gasket: from either deformation directly or loss of pre-load due to deformation
  • Detonation: detonation originating from over-heated parts in the cylinder head (edges), usually ends with broken pistons
All of these failures have occurred on VW and Jabiru engines, and also on industrial engines IF they have blocked air passages.

Airflow on the 49-series heads:
The original cooling is from the top of the engine (IN port) downwards. There are no fins on the valley-side, the air is directed on the outside of the head towards the PTO side and then directed inwards to the EX port. So the majority of the cooling air flows around the outside of the head and cylinder and through the passage in the head between the valves.
upload_2019-10-12_14-8-13.pngupload_2019-10-12_14-8-28.pngupload_2019-10-12_14-8-39.pngupload_2019-10-12_14-8-52.png
The SD-1 with the SE33 engine is proof that this engine can be cooled successfully if properly cowled as they are not exceeding 200°C in extended climb (CHT under spark plug).

The oil cooler is a full-flow cooler and can be placed anywhere.
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
part 9 – can I have some power please (would you like some torque with it?)

The power output from these industrial engines seems to be a very contentious subject. The power itself is not the issue, what causes some confusion is more about under what conditions it is measured.

Power is not measured directly for mechanical devices, it is calculated from the torque and rpm. Torque is the rotational force that the engine is developing and it varies over the rpm range.

The formula is: power [kW] = (torque [Nm] * n [rpm]) / 9550
Power [hp] = power [kW] * 1.341

There are many factors influencing torque:
  • Air temperature and altitude (and humidity)
  • Engine fitted with air filter, exhaust, cooling fan, alternator (EFI), water pump etc
In 2013, as a result of a lawsuit, the small engine manufacturers had to de-rate their advertised power by approx 10% to meet the new requirement that the engine output had to be within 95% of claimed output (previously only 85%).

The current standard for engines under 1lt capacity is SAE J1940 - Small Engine Power and Torque Rating Procedure and SAE J1995 - Engine Power Test Code - Spark Ignition and Compression Ignition - Gross Power and Torque Rating:

Max torque is measured at 2,600rpm (or 3,060 for engines without a max torque rpm) regardless whether that is actually the peak torque point or not. Max power is measured at 3,600rpm for all engines.

All manufacturers are also measuring Gross Torque or Power now. This means the engine is running without air filter, exhaust, fuel pumps etc. This is higher than the net torque or power and is not how the engine will run, so this is somewhat misleading. It was adopted as the engine manufacturer in many cases has no control over the installation (filters, cooling, exhaust etc).

Now to the 49-series: the engines have been advertised as 30hp engines under the old standard, now they are 27hp (same engine). In our application, we will:
  • remove the fan (1-2hp)
  • improve the intake system (1-2+hp): the original installation has 2x 90° elbows before the carburettor, 2x 90° bends in the intake manifold and a 30° bend and 30° kink into the head
  • removal of air filter or replacement with a free-flowing filter (1hp)
  • So, without much work, we now have a 31-32hp engine.
How do we get more?
Question: how big/small does the prop have to be? If you have a draggy and slow plane, bigger is better. If you have a sleek little racer, smaller is better at high speed and rpm.

A published limit from Warp Drive for wooden prop diameter is the tip speed should be max 700fps at cruise rpm.
Cruise rpm 3,500 (35hp): max prop dia 1.25m (49”)
Cruise rpm 3,300 (33hp): max prop dia 1.32m (54”)
Cruise rpm 3,100 (31hp): max prop dia: 1.4m (55”)
These are only approximations to show the relationship between prop dia, rpm and power.

If you can run your prop (engine) at 4,000rpm, you will have a 40hp engine if done correctly (if not done correctly, it will be a 4,000rpm time-bomb).
For all other applications, the only way is to get creative with tuning. With the rpm limited the due to the propeller, the only avenue is to improve the volumetric efficiency of the engine in the max power range of 3,300-3,600rpm.

What can be done?
These engines have a few shortcomings due to the mass-produced components with focus on simple and cheap manufacture. Some examples that I have listed in previous parts already:
  • Poor carburettor venturi finish (flash, steps)
  • Large screw heads on butterfly plates
  • Camshaft EX lobes designed for some EGR effect (exhaust gas recycling), profile is quite different to IN lobes (many engines use the same profile for IN and EX)
  • Sharp edges in the ports from the port to the valve head
  • Valve seats not cut and blended correctly (multi-angle valve job)
  • Valves with ridges on back of seat area (on the valve head)
  • Casting flash on heads and block, affecting airflow
  • Ignition timing (find peak power timing by trial & error)
  • Mixture control (peak power is at an air/fuel ratio of approx 13:1, 12:1 for some extra cooling)
upload_2019-10-12_22-39-39.pngupload_2019-10-12_22-39-50.pngupload_2019-10-12_22-40-0.png
There is a lot of very good information about porting, valve and valve seat profiling, camshaft timing, intake and exhaust manifold calculations and many other items. Just keep in mind that we want all of that at low rpm, so no drag and outright race car stuff. The principles still apply but on a lower scale.

As an example, the flow through a port is governed by the port size and shape but there is still a requirement for a minimum velocity of the air to keep the fuel suspended and carry it into the cylinder. Too large a port will actually reduce the volume that will be aspirated.

Once I start modifying my engine, I’ll post the progress (and possible failures) for others to learn from.

My goal is 49-series Briggs with a solid 35hp at 3,600 to 3,700rpm, no cooling issues and a low idle of 800-900rpm.
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
part 9 - other non-mowing uses of the Briggs 49-series

Just to show that these B&S engines have a strong following in a different application, I went to a race meet of the Queensland Ride-on Mower Racing Association. Drivers from all over the state turned up for this event. They are racing 6 heats of 5 laps on race day plus practice the day before.

Stock classes: Out of the 12 stock class racers, 9 were fitted with B&S 44 or 49-series engines (1 Kohler, 2 motorbike). Rules are:
max 32hp in OEM configuration
no internal modifications
no porting allowed
removal/bypassing of governer permitted
stronge valve springs and keepers permitted
exhaust design is free

These engines run 4,500-5,000rpm on the OEM flywheel (cast iron), most engines are of the 49M-style and have been in use for 6+ years. The drivers all state that they are a very solid engine. Output is in the 40-45hp range.

The Outlaw Class rules:
Commercially available mower engine of max 32hp, OEM block & heads
Billet flywheel compulsory

The drivers I spoke to are running 7,000+rpm with ARC rods, race cam, dual valve springs but OEM crankshaft, block and modified heads. Output 80+hp.

Some pictures of stock class engines. As you can see, some of them don't get too much TLC:
upload_2019-10-20_21-5-4.jpegupload_2019-10-20_21-5-54.jpegupload_2019-10-20_21-6-15.jpegupload_2019-10-20_21-6-38.jpeg
I also have a short video but not sure how to get that uploaded?
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
Part 10 – measuring & quantifying operational data and improvements
If you don’t measure it, you can’t manage it!

One of the objectives of converting this Briggs 49 engine for aircraft use is to have a set of verifiable (not certifiable) data for operation and also maintenance. B&S doesn’t supply any useful data to non-OEM users. This means a number of data points need to be captured in the original configuration to be used as a baseline for any modifications. After all, the engine has been designed, and is, operating in many demanding applications.

Data to be gathered in OEM trim (under load):
  • CHT x2 (each cylinder head, currently fitted with under-plug probe and dedicated CHT probe)
  • EGT x2
  • AFR x2 (air/fuel ratio)
  • Oil pressure (return from oil cooler)
  • Oil temp (return from oil cooler)
  • Fuel flow (timed volume)
  • rpm
  • MAP
  • OAT and carburettor temperature (near the butterfly)
How to load the engine?
There are a few different methods to load an engine and measure its output. Engines produce torque that can either be measured in a brake (water brake, electric brake) or the reaction torque to the load can be measured by having the engine mounted on a pivot. Another method used by small engines is the flywheel dynometer, where a significant flywheel mass with a known rotational inertia is connected to the engine and the rpm increases are measured precisely while going from the starting rpm to the high rpm. A very simple setup but has the drawback that the engine cannot be held a certain rpm. It is mainly used in the karting world where the dynamic response (acceleration) is more important than sustained loads (while testing & tuning).

For aircraft engines, there is also the test club. Certified engines can be tested for required output by using a calibrated test club (4-blade example shown). The advantage is that it also provides cooling air while stationary.

Another option, and the one I have chosen, is the square cross section test club (picture shows my raw club, not drilled yet). The invention goes back to Gustave Eiffel in the early 1900, when he got interested in aerodynamics in his retirement. The test club is built with the cross section leg length being 1/15 of the diameter, and a formula that will calculate torque absorbed by the test club at a certain rpm.
upload_2019-10-27_21-35-33.pngupload_2019-10-27_21-35-49.png

upload_2019-10-27_21-36-13.pngupload_2019-10-27_21-36-25.png
The test club is currently sized for approx 3,450rpm IF the engine is delivering 30hp as per label. The test club is behaving like a propeller so I can adjust the load to 100%, 75% and 55% and record all of the above values for each power setting. I will build a few more (larger) test clubs to load up the engine more and get some torque & power values in the 3,000 to 3,400rpm range.

This style of test club is not a calibrated device but it will give a ball-park figure of torque/power of an engine. Anyone anywhere in the world can build one to the same dimensions and verify the output of an engine. The only criteria for the accuracy of this test club are sharp edges on the club (no rounding or chamfering) and an accurate rpm meter.

To be able to run this test club, I will need to retain the OEM cooling system (fan) or replace it with something equivalent to keep the engine cool.

The data collector is a MGL Xtreme with RDAC module and will record all parameters in 1s intervals.

This set of baseline data will then serve to quantify any improvements in engine output, mixture control and in the final stage, validate cooling and oil temperature efficiency during flight.

Quantifying engine output improvements
Rather than running the engine after every small modification, I have built a small flow bench to improve the intake tract in small steps. Flow benches are a valuable tool for engine tuners as they allow a measured assessment of a particular port, manifold, carburettor or the whole tract and document the effect of modifications through flow improvements.

At the same time, flow benches are not infallible. The biggest issue that needs to be kept in mind when working with flow measurements is that the flow bench is a steady-stream flow, which is not how a real engine breathes. In simple terms, the engine has to ingest the cylinder volume in 180° (0.5 turn) of the crankshaft rotation, then “nothing” is happening in the intake manifold for 540° (1.5 turns).
upload_2019-10-27_21-37-55.pngThe flow bench provides some basic flow data that then needs to be verified on a running engine. Important for a V-twin engine is also the symmetry between the 2 cylinders to ensure a low-vibration engine under power.

Well, this is it until I have finished my instrument panel and the test rig and do some actual testing. First tests will be the original engine in vertical configuration before embarking on mods to turn it horizontal and then on to the power-enhancements.
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
OK, finally some progress :)
After a much longer time getting the MGL Xtreme with the RDAC module set-up and recording, I have mounted the engine in the "test safety cell". Some initial runs without the test club to get everything working and also recording the vibration levels of the engine.
After installing the test club, the engine runs ok at 1,700rpm (won't go any lower) but the vibrations increase above comfortable levels when going abpve 2,000rpm. I need to re-balance the test club (with prop hub this time) and also change the throttle arrangement. Currently, the governor is still fitted and I'm operating the governor control lever. This doesn't give me easy control so I will remove the governor and connect the throttle cable directly to the butterfly shaft.

Here is some trivia collected so far:
Ignition signal:
the ignition will cut out with a resistance to ground <70Ohm (put a 220Ohm resistor into the P-leads to prevent ignition failure from rpm input going to ground). The ignition signal on the Briggs coil is from +3.4V to -7.4V (at 2,000rpm), wave form like this:
1586593979675.jpeg
The alternator has a ripple of nearly 3V at no load. The frequency corresponds to 12 poles at ~1,850rpm. This output is not suitable for any electric/electronic instruments without a battery connected (the ripple is barely visible with a gel battery connected). The voltage is well regulated at 14.2-14.3V.
1586594003032.jpeg1586594022480.jpeg1586594052006.jpeg
And this is the test cell:
1586594267001.jpeg1586594289267.jpeg1586594311894.jpeg1586594236913.jpeg
Vibration: the vibration levels without the test club are quite good, in the 0.15ips range both for vertical (to the crankshaft) and rotational (in the crankshaft arc). The data points around (1) are at 1,850rpm, the data points around (15) are at 3,700rpm.
The data points (17/21) are with the test club.
1586594913799.png
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
The test club is balanced well enough to go to full power and I have now completed the first runs at full load. The results are rather disappointing, the calculated power is only 23-24hp and not the advertised 30hp! The mixture in both cylinders is extremely rich (AFR of 11-11.5 on the LH and under 11 on the RH), with both EGTs also quite low (500-550deg C).
Here is the summary table with the key data:
1587009019148.png
And some graphs:
1587009129028.png1587009167860.png1587009201181.png
Next steps are to remove the carburetor, clean up the rough edges and other bits&pieces and install the leaning device to be able to adjust the mixture for best power.
 

Attachments

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
Carburetor has been disected and violated!
1587119576576.jpeg
Work done:
- Steps in venturi to straight bore reduced (there was a step of between 1-1.5mm all around), original on left, reworked on right.
At the same time, I removed some sharp edges & steps at the venturi throat. This also opened up the venturi from 22 to 22.8mm.
1587119929869.jpeg1587119949744.jpeg
- Replaced the butterfly & choke plate screws with lower profile screws and cut them back flush at the thread end.
1587120122435.jpeg
- replaced O-rings on main jets (this is a known problem, the rubber deteriorates and can then by-pass fuel, making the engine run richer). The new O-rings (4x1mm) were slightly fatter and difficult to install. Correct size is 4x0.8mm.
1587120279617.jpeg
- Blocked off the float ball vent holes and added a hose fitting for an external line to the needle valve (mixture control)
- Lengthened the throttle shaft lever for about 25mm of cable travel (original is only 14mm)
1587120397619.jpeg1587120420027.jpeg
All screws were fitted with medium-strength Loctite.

Still need to finish the manifold (blending the ports to the cylinder head holes), add a couple of MAP probes for precise LH/RH MAP pressure pickup, redo the throttle cable to operate the butterfly shaft directly and add a degree disc to the top of the engine so I can play with the ignition timing.

A bit more work done:
When examining the intake duct, I realised that there were 2 tubes that protrude into the air stream. They direct dynamic pressure to the float bowl, which might explain the rich mixture (pressure in the float bowl will increase fuel flow)? As I have blanked off these 2 ports, I decided to remove the tubes and clean up the air path from the filter to the carburetor. Maybe they are running the engine for the power test not just with no air filter but no intake plumbing at all?
1587372412398.jpeg
The new throttle linkage is also done. The red tubes are the individual manifold pressure ports in the flange of manifold to the head. I have a precision differential pressure gauge and should be able to pick up any imbalance in flow (MAP).
1587372436183.jpeg
I should be able to run another series of tests over the next couple of days.
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
A bit more power but still very rich. I had to remove the leaning device connection as the idle was so lean that the engine wouldn't idle and was difficult to start. With the removal of some of the air filter enclosures, the engine was also running quite a bit hotter, so I need to fill in some of the gaps better.
The rpm is also not steady above about 3,200, I need to revisit the signal conditioner.
Power came out at ~26hp by the rpm from the balancer tacho (average of 3,278rpm) over 5minutes at WOT.
EGT, CHT and AFR have a bit of a spread LH/RH (hotter/leaner on LH)
Air filter restriction is about 8mbar (= less than 1%)
MAP is useless at the moment, will have to revisit and create some damper for the strong pulsing
1587453724482.png
1587453786410.png1587453817633.png1587453860853.png
This carburetor has real limitations as there is no idle mixture adjustment. I might bite the bullet and get a new carburetor with idle mixture screws. They are available for the larger engines (54 and 61-series), with 32mm throat vs 28mm on this one.

Finally some success with the rpm signal to the MGL! I tried all sorts of contraptions to condition the P-lead from the coil but nothing gave a stable readout. Then I got a Hall sensor module (Arduino module) which gives a beautiful squarewave signal, mounted about 20mm away from the ignition magnet on the outside of the cover:
1588401117370.jpeg1588401145438.jpeg
Only problem is the MGL didn't understand the signal, too much ON and not enough OFF. Searching the internet for a signal inverter revealed a very simple circuit using 1x NPN transistor and a couple of resistors:
1588401303669.jpeg1588401321118.jpeg
Now the rpm reading is rock-steady.

The final run in the OEM configuration should happen early next week, can't make too much noise over the weekend.
 

Attachments

Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
OK, I think the final results are in.
This run had the following changes from the previous runs:
  • Ignition advanced to 29deg BTDC (from 21deg)
  • Mixture at WOT leaned to 13.5-14.0 (a wee bit lean for max power, but couldn't make it richer)
  • EGTs up to 645/635degC and much more even
And the outcome: about 27hp

I believe that this is about the maximum this engine will deliver in an unmodified state. The only extra gain is the removal of the fan, my guess is about 1-1.5hp.

The rpm reading started to play up again, either from the vibration (temporary wiring) or temps. The RH O2 sensor signal also disappeared towards the end.
I was unable to measure the fuel burn, the fuel in the measuring tube was just too much disturbed.
1588647237731.png1588647390069.png1588647421323.png
1588647462969.png
Next steps are stripping the engine and starting on the conversion to horizontal crankshaft as well as the power mods (head porting, valves etc). I also have a larger carburetor on the way (from the 54-series, 32mm, with idle mixture adjustments) to try out.
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
Post Mortem (non-terminal & voluntary) :)
Well, the engine is stripped down and here are the highlights:
  • My crankshaft is cast, not forged (cast steel)
  • Crankcase and crank cover are very roughly finished (still some flashing about, lots of sharp edges)
  • Conrods might be forged, difficult to tell (they do have the smooth surface that forged items generally have)
  • Pistons are very light and low (low friction type)
  • Both big end bearings showed signs of off-centre contact with the crankshaft journal
  • The cylinder head dates back to the 40ci engine series (656cm3), the combustion chamber diameter is for a 75mm bore. The 49-series has 83.8mm bore.
  • Heads are imptric! The hold-down bolts for the rocker arms are M8, IN & EX manifold bolts are imperial, valve stems are 6mm. The rest of the engine is imperial.
  • Valves are 34.5mm IN and 31mm EX
  • IN valves are pretty much a stamping (one valve has the stamped head off-centre)
  • 1 IN valve had no valve stem seal fitted
  • Ports in the head are unfinished, simply 2 drillings intersecting at 90deg
  • Crankshaft main bearing on the PTO side is 68% bigger than the flwheel side (dia 41.2x40mm vs dia 35x28mm)
  • There was quite a bit of debris in the filter, consistent with aluminium flashing (slivers) and very little wear debris
Here are some pictures:
CIMG2262.JPGCIMG2263.JPG
The bore (outer edge of black rings) is considerably larger than the combustion chamber diameter.
CIMG2264.JPG
Intake port, sharp edges on the intersection
CIMG2271.JPG
PTO journal, only problem is the sharp edge around the oil hole
CIMG2272.JPG
Big end journal, has chamfered edges
CIMG2273.JPG
Oil pump inlet from the sump side, lots of casting flash left
CIMG2274.JPG
Oil inlet passage to pump, small screen at pump inlet.
CIMG2277.JPG
very thin rings and low surface area of piston, rod might be forged
CIMG2278.JPG
Not much actual filter!

Next steps:
  • Flow-testing the cylinder heads and gradually improve flow
  • Design and build prototype intake manifold
  • Working out the oil supply and drain arrangements
  • Get some spares (Vanguard IN and EX valves, maybe a Vanguard crankshaft and con rods, EFI cyl head gasket)
  • Remove all the flashing and smooth out any visible imperfections in the crank case and cover (stress risers)
  • Reassemble and make some more noise
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
Right, onto the performance enhancements Setting the baseline :)

I managed to get my hands on a set of Hyper-Performance Bee-Hive valve springs. They are the best springs available for the B&S engine and should be good for an extra 20% of power.
CIMG2279.JPG

Just kidding, they are a set of springs from an old shower tap and are the perfect match for the B&S for the flow bench set-up. The normal spring is too strong to open the valve while measuring the pressure drop and flow, so weaker springs are required.

Here is the set-up with the head on the flow bench, dial gauge on the intake valve to measure valve lift. The parts in red are the mechanism to operate the valve to any desired valve lift (the black knob is the valve lift adjuster) and base for the dial gauge holder.
CIMG2281.JPGCIMG2282.JPG

I still have to plug some holes and check for leaks, then on with measuring the flow of the original heads at 0.5mm increments of valve lift. I might need to liberate the household vacuum cleaner depending on what my little unit can deliver.

Update 10/5/2020:
Here are the preliminary results. My vacuum cleaner run out of puff at 5.5mm valve lift, need to change the plumbing to add a second vac.
To put these numbers into perspective: the engine needs ~11 lt/s at 3,600rpm and WOT (per cylinder, steady flow). This volume has to be ingested in approx 1/4 of the time (during the intake stroke), so the dynamic flow is roughly 4x 11 = 44 lt/s (or 93cfm).
EDIT: graph with correct flow rates (well, "uncalibrated TiPi numbers")
I redid the flow calibration of the flow bench and found a couple of errors. The set-up I have now produces repeatable results. The flow numbers are relative to my flow bench, they are not absolute numbers and can't be compared to other flow bench results.

1589198719047.png

Update 13/5/2020:
I had to re-configure a few things on my flow bench. The electronic differential pressure gauge wasn't as accurate as it should be so I reverted back to the old fashioned inclined H2O manometer.
The plumbing has been changed to allow for a second vac to be connected when needed. I also added a large needle valve to aid in the fine control of the suction source.
CIMG2291.JPGCIMG2290.JPG
And this is the result: a flow graph from 0 to 8mm of valve lift, at 250/500 & 711mm of H2O (vacuum below the cylinder head). 711mm is the US equivalent of 28" H2O, a standard in the performance industry.
1589359523604.png1589425599141.png
When comparing the flow between Cyl head #1 and #2, there is very little flow difference at 711mm H2O (28" H2O). This will make it easier to balance the heads for equal flow (reduce vibrations under load).
1589426137255.png
What does it mean???
As is clearly visible, there is no increase in flow from about 5.5mm of valve lift! This means the restriction is no longer the valve opening (valve curtain) but the port size, port shape and/or valve shrouding.
I have expected a result like this, I just didn't know at what point the curve will go flat. The design and finish of the head suggested a result like that.
The good news is that there should be some easy to find little ponies in there :)

Next is a step-by-step improvement process, measuring the port flow after each step. I'm still waiting on some special tooling before I can get stuck in the port and head reshaping.

Out of curiosity, I mounted the intake manifold complete with carburetor, air filter adapter and air filter and checked the flow again at 711mm H2O. The resulting flow is about 20% less at valve lifts above 3mm (2x right-angle and 1x 180 bends from the air filter to the head).
1589540317696.png
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
Performance Improvements - Staged approach
After reviewing the current status, I have decided to do the performance improvements is the following stages:
  1. Intake valve polish and clean-up & seat (3-angle cut)
  2. Smooth the short-turn radius and remove any sharp edges in the IN port bend (side walls)
  3. Smooth the valve guide post and valve guide
  4. Unshroud the valve
  5. Restore CR to 8.8-9.0:1
This will be mainly for the intake valve, stages 1, 2 & 4 also for the EX valve.

Stage 1 has been completed on cylinder #1:
Valve is polished, seat and back face blended & sharp edges rounded. The seat has 3-angle cut (70/45/15deg), seat width 1.2mm
CIMG2298.JPGCIMG2299.JPG
The result: 5-9% more flow in the 1mm to 4mm lift range. This flow improvement should carry through the full lift range once the port & shrouding restrictions have been addressed (OEM+: with manifold, carburetor & air filter)
1589628479214.png
On head #2, I will reverese stages 1 & 2 to see what influence each has without the other.

And here is the final result after stage 3 (unshrouding the valve):
The overall flow improvement is only a rather disappointing 10%:
Stage 1 (valve seat & valve) added about 3-4% from 1-5mm lift
Stage 2 (cleaning up the port intersection) added about 4% from 1 to 8mm lift
Stage 3 (unshrouding the valve) added about 4% from 2.5 to 5mm lift
1589850428964.png
Port shape before & after:
CIMG2301.JPGCIMG2300.JPG
This is the unshrouded valve at full lift (8mm). The clearance is 5mm, not quite the 20% of valve diameter yet (waiting for the correct die burr).
CIMG2310.JPG
One thing I want to try next is filling the far corner in the port junction (right angle at the moment) and blend it to a nice bend. Any other work would be major surgery and beyond the average builder.
One I might try is to open up the long port. It tapers from 27.7mm to 25.5mm over a length of 80mm (casting mould). The valve port is 28.2mm. Just need to find a way of "drilling" this long port out to maybe 27mm (12% larger area).

Well, the rounded outside corner didn't work, the flow rates were down a couple of % across the full lift range. It was a little bit rough but should have made a difference if this corner is the culprit.
1589869219549.jpeg
Last effort would be to open up the port near the bend to 27mm if I can find a method at the end of a long stick. That is a bit drastic and non-reversible, so I need to make sure I have enough meat all around.

Well, I buggered up :(
After looking at the various data and re-reading some of the literature that I have collected, I realised that I "forgot" to add the port intake mouth piece. The restriction (choking) that I measured consistently at nearly the same flow rates was caused by the sharp edge & angled port opening (no manifold fitted).
A few hours later, I had 2 versions of intake aids:
- a SS tube of 25mm OD that fitted neatly inside the port and made a good seal at the valve end of the port (same cross section)
- a cobbled together adaptor and heat-formed PVC fitting that bolted to the port face
The second version delivered more flow so I'm now re-tracing my steps (where I can) and measure with this adaptor fitted.

The key outcome of this test process is to find the best flowing intake port from the port opening into the combustion chamber, so the flow into the port itself needs to be maximised.
CIMG2313.JPG
 
Last edited:

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
451
Location
Peeramon (AUS)
The final results are in :)

After re-doing #1 head with the trumpet attached, these are the results form the flow testing:
1590051929062.png
Stage 1 (3-angle seat cut): improvement of 6% from 1 to 5mm of valve lift
Stage 2 (cleaning up the port area and bend): ~3% from 1 to 4mm lift, 12% from 4 to 8mm valve lift
Stage 3 (unshrouding the valve): 3% from 2 to 5mm lift, 1% from 5 to 8mm of valve lift
All % are relative to the previous condition. All up, the flow improvements are in the 13-15% range.

The restriction (flat line) from 5mm to full lift is caused by the port diameter. The port starts as a 28mm bore and tapers down to about 25.5mm (casting plug), the port under the valve seat is ~28mm.

Ideally, the flow curve would continue to increase to about 6-6.5mm of lift.
Unshrouding the valve would show a better improvement with a better flowing port.
As all of these measurements are with a static flow, the dynamics of a running engine could change a few things.

I’ll assemble my engine with the current head mods and then load test it to see what output I get. If I think it is not enough yet (I’m looking for 35hp), then I’ll open up the long port to 28mm and try again.

It will take me a few weeks, I need to make a dummy firewall, engine mount, oil system mods, new intake manifold etc.

Update 22/5/2020: Exhaust port
I forgot to document the exhaust port measuremnts. Here they are:

The EX port has a very pronounced turn from laminar flow to very turbulent flow (noise & vibration), at 3.0-3.5mm of lift. It is visible in the flow graph from the flat part of the flow curves and the much flatter increase in flwo with increasing valve lift.
Flow improvement is around 10%.
1590122155987.png
Stage 1: valve seat 3-angle cut
Stage 2: bend and port blending/smoothing
Stage 3: valve unshrouding

CIMG2318.JPG
Setup on the flow bench. The flow is reversed by attaching the vacuum cleaner hose to the exhaust port and measure the flow into the cylinder. Plastic header and a wooden flange on an exhaust :)
CIMG2320.JPGCIMG2319.JPG
Port before and after.
CIMG2321.JPG
Valve unshrouded. On the EX valve, there is very little difference in the flow rates as the affected area is on the short-turn radius and doesn't flow much air anyway.
After all mods, the EX flow is approx 70% of the IN flow which is probably not too bad for this type of engine.
 
Last edited:
Top