# Unusual variations on two-stroke head design.

### Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Hi Owen,
Yes, I have been thinking about something like an Archimedian screw - with variable pitch as the passage progresses along the pipe.
First I need to understand - what length and diameter of header pipe are you planning? Do you have a design of expansion box already worked out?
If not, this guide seems straightforward to design one.

As gases do not really appreciate cones and tubes, just flat planes that they reflect off, the "spiral" tube should be pretty much OK as would a rectangular tube. In simple terms, the mean diameter of a spiral tube would be akin to the length of the middle of the spiral chamber - that is really just a rectangular tube wrapped into a spiral. But as the cross-section needs to change, the change of pitch of the screw alters the length of one side of the rectangle - thus increasing or reducing the CSA, and simulating the expanding and reducing cones of a twin-cone expansion box.
So I understand you are planning on a 60cc, engine, 6300rpm as the tuned speed. (I may get this completely wrong, but please bear with me?).
So:
we can calculate closely the pipe length. The formula for determining the length is:

Lt = (Eo x Vs) / N English OR (83.3(Eo x Vs)) / N Metric

Where:

Lt = tuned pipe length, in inches OR millimeters

Eo = exhaust open period, in degrees

Vs = wave speed (1700 ft/sec OR 518.16 Meters/sec at sea level)

N = crankshaft speed, in RPM
I.E.
83.3 x (120 x 518.16)/6300 = 822.1mm.
{trim}

K2
here is a possible sleeved folding scheme, where the tail cone is a convergent annulus around the centre cone.
I see the side view shows a much shallower convergent angle for the length, even though the annulus starting
area should be very close at 1/2 r wide.
A fairly long slotted or perforated area is needed for the transfer to the muffler stage.

How do you think this would work?

Will it actually reflect the main sonic pulse?

I will calculate some actual lengths.

If I can keep the centre cone to 72 mm, then the overall diameter will be 110 mm., which is probably OK.

Thinking about return pulses- unless 45 degree cones are used, the return pulses will bounce around quite a bit.

***************** TP60 engine mods are coming along well.

I will turn the barrel and piston around again, to make the Cool-side the combustion thrust side.

The ring land above the exhaust is overloading, and wiping aluminium above the exhaust.

This may be caused by insufficient piston support beside the exhaust, or piston rocking.
Fixing the cold side should reduce rocking.

It will be interesting to see how the inserts stand up, and whether J-B can be used as a piston skirt coating.

The alternate Devcon etc "plastic metal" options seem to be only rated to 120 degrees, vs around 200 degrees C for J-B.
J-B seems to have a gritty surface texture.
It doesn't seem to erode piston rings though.

#### Attachments

• folded pipe version 22-03-22.jpg
105.5 KB · Views: 70
The conic sections look ok.
inlet chosen at 1 inch.

the sticking point is the centre parallel section, of 645mm

front cone = 271mm, tail cone = 226mm, and diameter = 63.5.
This is a very strange looking pipe.
I will compare it with the 66 cc torque pipe, which looks more "normal".
overall length = 58.3 inches.

1) I think this spreadsheet has been 'fiddled" with.
The example calc i did, and an actual similar example, are wildly different.
2) are there any advantages in using shallow angles?
Are 8 degrees and 16 degrees OK?

You were suggesting that a conventional pipe would be just over 1 metre in total.
Mine shows 1.3m just over the cones, disregarding the header pipe and the stinger.
the example pipe shows 580mm over the cones, but he uses a 400 mm header instead of a 250mm one, which is the recommended maximum length.
This 580 dimension would fit nicely into the everted cone layout from the previous drawing.

What effect does a very long header pipe have? This is the size across the bendy bit supplied. about 320mm.
how would this affect the other dimensions?

Is the length from piston to the end of the tail cone the dimension that is maintained, and take the extra length out of the centre
section.

The only difference is the example pipe uses 160 degrees exhaust duration, whereas mine is 180 degrees.
- a difference of + 12.5% on time. - would that make the pipe 12.5 % longer?

My transfers are 154 degrees, which is a little to close to the exhaust for good blowdown.

I checked it on a rod angularity diagram. The rod length is 66 mm.

160 degrees seems quite short in duration.

The comment was that he raised the exhaust 3mm, so it probably wasn't a model aero engine.
- you might find that timing on a moped??
Degrees is not very useful in calculating volumetric efficiency, as the rod angularity makes a big difference.
you wouldn't think that 21mm was half way to 36.5mm.
the balance is only 15.5 mm

#### Attachments

• example torque pipe 22-03-22.jpg
175.5 KB · Views: 56
• first cut pipe 22-03-22.jpg
202.7 KB · Views: 50
Last edited:
Hi Owen, Of course, your knowledge is a bit beyond mine, so I am likely to be wrong in a lot of what I suggest, based on your better expertise. But I am enjoying bouncing ideas around, if it helps you work out a good solution. (But please feel free to tell me where to get off, if the journey goes the wrong way! - I'll not be offended at all).
Ignore my dimensions, I am just as likely to be completely "off-beam" with my assumptions, that will have affected the numbers considerably.
• What effect does a very long header pipe have? This is the size across the bendy bit supplied. about 320mm.
how would this affect the other dimensions
? As far as I understand, the reflected wave relies on the total length from exhaust port to reflection point. This has been advised as the joint between the diverging cone and the converging cone.... SO: A Long header makes for a shorter diverging cone, and a very short header makes for a longer diverging cone.
• I don't understand how to calculate the diameter of the large end of the diverging cone.
• In my head, I think the Concentric cone arrangement you have drawn is good. BUT - I think the reflection point would be the end face where the gases double back on themselves.
• I have been trying to consider what difference a flat end makes, compared to a short converging cone. My memory is inaccurate, but in the Motorcycle club at Uni, we had some pretty intelligent discussions (some were studying this in Mech Eng as their projects). The idea was that the diverging cone expands, cools and slows the pressure wave and exhaust gases. This gives a shorter pipe than if the header CSA alone was continuous to the end of the exhaust. (Simple straight pipe exhausts of pre-WW2 racing). The "reverse cone" - the Converging cone - generates the reflection of the pressure wave - to stuff some gases back into the cylinder by getting back to the exhaust port before the port closes. A Cone, will give some variation (caused by revs, throttle, Ambient air cooling of the diverging cone, etc.) so the "tuned pipe" was more functional across of a range of engine speed and load and not so "Peaky" = exactly tuned to a specific load and engine speed.
• Considering the concentric conical chambers: (A thought experiment). The shock wave is travelling at "sonic" velocity for the residual pressure in the exhaust from the previous exhaust gas pulse (I.E. gases that are still exiting via the tail-pipe: => Faster than in "Air"). As it expands into the diverging cone, it slows, as the pressure drops along the length of the cone. When it reaches the reversing end of the concentric cones, there is a clear reflection wave that will travel back to the exhaust port. If the inner conical part is made a bit longer, this will be less peaky in tuning. Because of this, I can see no reason to have the "reversed cone" of the converging outer cone. So I would make the outer chamber parallel OD at 110mm diameter. In effect, this outer annular chamber would then be a continuation of the diverging cone, thus dou=bling the length of the diverging gas column, and giving a second reflection wave when the gases reach the "front" end, where the discharge pipe is fitted. This second reflection wave, should travel back to the exhaust port at something like twice the time of the first.... so if the engine is running at "half speed" the second pressure wave will "tune-in"...(?). Thus giving a possible tuned peak of performance at half of the 6300rpm...? = 3150rpm? - Any use for "cruising" engine speed and load?
Of course, I may be completely off the rails with these ideas...
K2

Hi, based on my knowledge and experience I think the folded cone won't work very well.

To explain my conclusion I'd like to clarify the internal physics of a tuned pipe
First, you have to consider two properties as separated: gas pressure and waves.
You want to get rid of all the exhaust gas pressure before the next cycle begins and that is controlled by the diameter and length of the stinger and muffler (and you can place it almost everywhere without much effect on performance), so basically it only depend on your target rpm. the pressure in the exhaust should be back to the starting point every 360deg.
The waves are a bit more complicate to manage, and travel at the speed of sound which is affected bay both the gas temperature (450-600C ) and pressure (pressure is usually ignored).
The waves are what do the work of extracting more exhaust gas and push back the fresh mixture that get overblown out of the exhaust port, so de duration of the waves effect depend on the time duration of the exhaust port opening

Ex. rpm=6000
exhaust timing=150deg
6000rpm= 10ms*360deg
exhaust opening time = 4.16ms
so all the last positive pressure wave should reach back to the exhaust port 4.16ms after it started from it, using the sound speed of 570m/s (@~530C) or 57 cm*ms the total length would be 118.5cm (3.8ft)
Higher rpm = shorter pipe
wider port timing = longer pipe

Cones and lengths.
The header length determines how long before the first negative wave (generated by the diverging cone) start to have effect in the port area. Short header= faster action / Long header = Slower action
The diverging cone generate a negative pressure wave that lower the pressure pulling the gases out of the cylinder faster.
Steep cone=higher pull / narrow cone = lower pull //
long cone = longer effect duration / short cone = short duration
usually here you use a longer and narrow cone that gives you a bigger working band and more pull duration.

The converging cone work the same way but generate a positive wave that push the gases back, so you want its action to be of a shorter duration and at the last moments of port opening (short and steep cone at the max distance from the port)

The cylindrical middle section is only used for space requirement as you could have the two cones back to back but usually as the cone widens the wave intensity decrease so after a certain point is not as effective. It also help to reduce the sharp change in wall angle

--------
Hence my conclusion is that a a folded back cone will act almost as a straight wall giving a very powerful wave but of an extremely short duration (that could be totally ineffective due to gas inertia).
But maybe some experimentation would prove me wrong

Hope this would be of help
Cheers, Riccardo

The conic sections look ok.
inlet chosen at 1 inch.

the sticking point is the centre parallel section, of 645mm
You can get some idea of construction by looking at 2 stroke dirt bikes. Many use flat sheets TIG welded all around then inflated hydraulically hydro formed it’s really tough to edge weld thin sheets especially stainless . Most of my pipes were aluminum purchased basic parts then made adjustable length slip joints so I could adjust various sections once it worked I just tacked them together. Some scale models I used copper fittings carefull trimmed to reduce weight. You can take off 50% or more if you are careful . Copper eels easily or brazes easy too do not use solder it mels way too low. My new steamer has 1/4” copper fitting from heat and air cond parts they are tiny prices just went up almost over night 45 deg elbows dot seem available excep special order. If you use low melt temp bismuth material you can bend without kinks McMaster Carr is where to go . Look for cero bend metal
I’ve not seen a big Rc flying wing for a long time they used to be popular in Rc combat .

Byron
front cone = 271mm, tail cone = 226mm, and diameter = 63.5.
This is a very strange looking pipe.
I will compare it with the 66 cc torque pipe, which looks more "normal".
overall length = 58.3 inches.
This particular engine may suit a semi-tuned pipe.
The conical expanding part is ok out to around 60-70 mm in diameter, but a 6000 rpm pipe would be too long.
If I got a good torque build from 5000 rpm, I could spin up to possibly 7000 rpm, depending on the torque curve.

Prop torque climbs very quickly. I think static thrust is proportional to the cube of the power, and
has a square relationship to rpm and velocity. T = dens x effy x V sq x A (thrust)
- similar format to drag, only no 0.5 factor.

A pipe on this engine may give 30% more torque, = 30 % more power at the same rpm,

and possibly up to 10% more thrust..

To get from 6000 to 6500 is am 8.3 % increase, 1.083^2 = 1.17 or 17% more thrust.
at say 10% more thrust, velocity increases by 1.05, or 6300 rpm.
This is not exactly correct, but gives some idea.
if I increase speed by 1.05, then power is x 36.5 % more.

Is there any way I can get a correct tuned response into a pipe that is only 320-400 mm long?
The idea is that this length would fit into a typical model aircraft, which is maybe 800mm long in the body.
If a short tailless design is considered, the body length would be less.
I am considering wingspan from 60-80 inches, or 1.5 m to 2m.

Hi, based on my knowledge and experience I think the folded cone won't work very well.

To explain my conclusion I'd like to clarify the internal physics of a tuned pipe
First, you have to consider two properties as separated: gas pressure and waves.
You want to get rid of all the exhaust gas pressure before the next cycle begins and that is controlled by the diameter and length of the stinger and muffler (and you can place it almost everywhere without much effect on performance), so basically it only depend on your target rpm. the pressure in the exhaust should be back to the starting point every 360deg.
The waves are a bit more complicate to manage, and travel at the speed of sound which is affected bay both the gas temperature (450-600C ) and pressure (pressure is usually ignored).
The waves are what do the work of extracting more exhaust gas and push back the fresh mixture that get overblown out of the exhaust port, so de duration of the waves effect depend on the time duration of the exhaust port opening

Ex. rpm=6000
exhaust timing=150deg
6000rpm= 10ms*360deg
exhaust opening time = 4.16ms
so all the last positive pressure wave should reach back to the exhaust port 4.16ms after it started from it, using the sound speed of 570m/s (@~530C) or 57 cm*ms the total length would be 118.5cm (3.8ft)
Higher rpm = shorter pipe
wider port timing = longer pipe

Cones and lengths.
The header length determines how long before the first negative wave (generated by the diverging cone) start to have effect in the port area. Short header= faster action / Long header = Slower action
The diverging cone generate a negative pressure wave that lower the pressure pulling the gases out of the cylinder faster.
Steep cone=higher pull / narrow cone = lower pull //
long cone = longer effect duration / short cone = short duration
usually here you use a longer and narrow cone that gives you a bigger working band and more pull duration.

The converging cone work the same way but generate a positive wave that push the gases back, so you want its action to be of a shorter duration and at the last moments of port opening (short and steep cone at the max distance from the port)

The cylindrical middle section is only used for space requirement as you could have the two cones back to back but usually as the cone widens the wave intensity decrease so after a certain point is not as effective. It also help to reduce the sharp change in wall angle

--------
Hence my conclusion is that a a folded back cone will act almost as a straight wall giving a very powerful wave but of an extremely short duration (that could be totally ineffective due to gas inertia).
But maybe some experimentation would prove me wrong

Hope this would be of help
Cheers, Riccardo
I tend to agree with you, from my limited understanding of what is really going on. The thoughts were that the doubled-back gas flow would enable the overall length to be effectively halved. - The >1m designed system is just too long for the fuselage. so 1/2 ~1/3rd of that physical length is what Owen wants to design.
Having read your explanation, what concerns me a bit is the idea of a small tail-pipe (as shown by Owen's diagram?)? Considering that the gas will have expanded from (say 60ccs at 3 bar at exhaust port) to less than 0.4bar above atmosphere, then surely it needs a big hole to escape in a short time through such a small hole? I would have expected a larger hole that the header pipe? - even something like 4~5cm dia? - Then the pipe would be able to get to atmospheric before the next pulse of exhaust when the port opens after the next bang?
I'll admit to never having designed IC engine exhausts, but I am curious, as I have designed an effective silencer for a 6l cylinder evacuating to atmosphere from 28bar. (A deep "Boom!" at 136dBa at 1m from the exhaust with the previous silencer/expansion box - my silencer achieved below 125dBa. at 1m from the exhaust, with a higher frequency "crack" sound.). But that was for a High Voltage circuit breaker that was 1.5 miles from the nearest habitation...
K2

1) The offset double-reverse-cone reflector is intended to reflect a plane-like sound wave into the outer chamber.
The main gas flow is moving a lot slower, but may create a weak reflecting wave at that point.

2) I have come up with a double-fold with a centre section that has a constant cross-section.

3) the very narrow "V" between the 2 shells probably makes a very poor reflector, but may be good at accelerating the body of the main gas flow.
The outer chamber of 3 should end with a short cone about 32-45 degrees to the inner shell.

4) the gas will be getting quite cool by the time it gets to the outer shells, so velocity of sound will be slower, and effective length will be more.
The current cone-part length of the double- folded system is around 750mm.
It wouldn't take much to get it to 1000mm.

Predicting what it should be is tricky, but probably less than 1330 mm.

5) I have found that the shell temperature of a sudden exhaust expansion settles around 140 degrees C, so the bulk of the chamber with a controlled expansion will be much cooler.
What temperature and pressure combination is contributing to 1700 ft/sec speed of sound?

6) The exhaust chamber has a full half-cycle to evacuate. With an controlled expansion of 20x, the chamber pressure
and density should already be pretty low, plus it retains a fair amount of velocity, so it should shoot out of a 14mm hole ok.
The velocity disappears if the gas is not channelled out, but extended and offset stingers seem to work, so
it is probably not a big problem. It will reduce evacuation efficiency.

I calculated that the discharge pressure from 10:1 compression should only be around 20 to 30 psi.

From the amount of noise it makes you would think pressure would be over 100 psi- rather like a .22 short machine gun??

A 400cc 4-stroke has no problem with about a 20mm outlet choke point, and the velocity is quite high.

A 60cc two-stroke should be quite happy with a 14mm outlet tube.

I don't know why they want a 6 inch stinger tube, though. Maybe there is some gas slug inertia effect,
helping to get a lower evacuation pressure.

Here is the new version of the folded pipe. this gives about 830mm over the cones, plus the 320mm header pipe.
any ideas whether this would work at 6300 rpm, 180 degrees duration?
Any reverse pulse just before transfer close would be beneficial, as long as the pulse lasts for at least 30 degrees, until the exhaust closes.
I presume a standard reverse pulse is either very precise, or fairly long duration.
Maybe worth a try?
I will check for jobbing sheet rolling services.
I have some 0.2mm stainless , but I couldn't roll it up.

#### Attachments

• new version folded pipe 23-03-22.jpg
133.2 KB · Views: 50
Improved reflector angles to stop back-reflection from initial wave front.
The previous designs allowed some of the wave to reflect back instantly.
looking at it, two stages in the outer cone would be better- 45 degrees, then 30 degrees.
What do you think?
Can any direct reflections escape back down the centre cone?

#### Attachments

• improved angles for end reflector 23-03-22.jpg
160.7 KB · Views: 44
I tend to agree with you, from my limited understanding of what is really going on. The thoughts were that the doubled-back gas flow would enable the overall length to be effectively halved. - The >1m designed system is just too long for the fuselage. so 1/2 ~1/3rd of that physical length is what Owen wants to design.
Having read your explanation, what concerns me a bit is the idea of a small tail-pipe (as shown by Owen's diagram?)? Considering that the gas will have expanded from (say 60ccs at 3 bar at exhaust port) to less than 0.4bar above atmosphere, then surely it needs a big hole to escape in a short time through such a small hole? I would have expected a larger hole that the header pipe? - even something like 4~5cm dia? - Then the pipe would be able to get to atmospheric before the next pulse of exhaust when the port opens after the next bang?
I'll admit to never having designed IC engine exhausts, but I am curious, as I have designed an effective silencer for a 6l cylinder evacuating to atmosphere from 28bar. (A deep "Boom!" at 136dBa at 1m from the exhaust with the previous silencer/expansion box - my silencer achieved below 125dBa. at 1m from the exhaust, with a higher frequency "crack" sound.). But that was for a High Voltage circuit breaker that was 1.5 miles from the nearest habitation...
K2

I understand the limit that Owen have, as space requirement has always been an issue for 2stroke motorcycles (look at the crazy design of some multi-cylinder motorcycles)
As for the tailpipe I didn't explained myself very well, given also that I never gave to much of a thought about that, making the stinger diameter always between 1/2 to 3/4 of the header diameter only basing it on "feel".
But the idea is to keep a backpressure higher than atmospheric but lower than the exhausting pressure, permitting an almost constant flow keeping the pressure and temperature of the gasses inside also more constant (helping also with silencing).

1) The offset double-reverse-cone reflector is intended to reflect a plane-like sound wave into the outer chamber.
The main gas flow is moving a lot slower, but may create a weak reflecting wave at that point.

2) I have come up with a double-fold with a centre section that has a constant cross-section.

3) the very narrow "V" between the 2 shells probably makes a very poor reflector, but may be good at accelerating the body of the main gas flow.
The outer chamber of 3 should end with a short cone about 32-45 degrees to the inner shell.

4) the gas will be getting quite cool by the time it gets to the outer shells, so velocity of sound will be slower, and effective length will be more.
The current cone-part length of the double- folded system is around 750mm.
It wouldn't take much to get it to 1000mm.

Predicting what it should be is tricky, but probably less than 1330 mm.

5) I have found that the shell temperature of a sudden exhaust expansion settles around 140 degrees C, so the bulk of the chamber with a controlled expansion will be much cooler.
What temperature and pressure combination is contributing to 1700 ft/sec speed of sound?

6) The exhaust chamber has a full half-cycle to evacuate. With an controlled expansion of 20x, the chamber pressure
and density should already be pretty low, plus it retains a fair amount of velocity, so it should shoot out of a 14mm hole ok.
The velocity disappears if the gas is not channelled out, but extended and offset stingers seem to work, so
it is probably not a big problem. It will reduce evacuation efficiency.

I calculated that the discharge pressure from 10:1 compression should only be around 20 to 30 psi.

From the amount of noise it makes you would think pressure would be over 100 psi- rather like a .22 short machine gun??

A 400cc 4-stroke has no problem with about a 20mm outlet choke point, and the velocity is quite high.

A 60cc two-stroke should be quite happy with a 14mm outlet tube.

I don't know why they want a 6 inch stinger tube, though. Maybe there is some gas slug inertia effect,
helping to get a lower evacuation pressure.

I'll try to give my view on your points.
1)the double reflecting cone will cause a lot of diffraction nullifying the effect of the waves (the interaction is the same of waves on a river, that reflect and move in a completely different way than the water flow).
the gas flow will easily double back into the annular passage even with a flat end as it follow the pressure difference and inertia.
2)the constant cross section could be useful for helping the gas expand and slow but wont affect the wave
3)as before the flow would have slowed considerably in the first cone but the local flow speed is not much influent on the pumping effect
4/5) 1700ft/s is at around 740F and is considered as average gas temperature. Obviously the pipe wall will be at a much lower temperature because while the gas it pretty hot is not very dense and will be expanded and expelled before having enough time to heat up the exhaust (same as inside the cylinder, the burning gas will be exceeding 2800F and aluminium melt at 1200F but the time is very short limiting the amount of heat transferred)
6) I think it might interesting to try eliminating the stinger altogether and just have one hole (or many small holes), but the sound might be too loud

I put here a few link of expansion pipes for Italian Vespas and moped so it might give you an idea of space constrained tuned pipes
the original are just baffled pot that give a lot of backpressure (10mm final diameter for a 50cc)
vespa original
this is for the 200cc and the header diameter is 40mm while the stinger is just 16mm
"tuned" original
this is an original modified to give a little more power but look inconspicuous
vespa tuned
racing exhaust for the same engine, and is all bent over itself in order to fit it under the frame

ciao moped original
pot muffler for 50cc header 22mm stinger 10mm
ciao moped tuned
same engine but "snake" pipe to fit under the frame

Here are the "before run" photo of the J-B weld changes.

This also shows the new head.

The small steel pins seem to have escaped, but top end damage was minimal. A few slight piston dings.
The swaged bush seems to be holding well.
I put a new ring in, so it will take a while to settle down.
It got through 7 minutes running with nothing alarming happening.
I will strip it after 5 x 7 minute tankfulls., and see how things are going.

The edges of the rear wall insert look sunken, but they are flush to the touch.
The side transfer length increase has given no problems.
You can go up to 90degrees of port if there are sloping edge details in place.
The slopes I am using seem to be enough.
The profile around one end of the transfer is larger, but that part of the port is blocked with J-B weld.
the slight chamfer there is good for ring anti-snag.

The compression release hole has opened up again. It was only blocked with zap fast glue.
The engine has a slight sound of 4-stroking at 5,000 rpm, but I think the odd sound could be due to the
compression release hole.
This 2mm hole probably doesn't waste a significant amount of power around 5-6000 rpm.
I checked the area for heating, and it seemed steady around 155 degrees C . Not unusually hot.

It runs a little oddly at present, but I expect it to be more consistent and tuneable once the ring is run in.
It needs a counter weight on the offset muffler mount.
It was quite steady before I cut the fins out of the head, making it about half the weight.
the muffler is angled up about 60 degrees to horizontal, and has its own board.
The previous muffler mount had a counter weight as well- about half a pound,- a spare panel beating backing block??
the twin didn't need a counterweight, but its vibration is in a different plane.
I expect this one has mainly a rotating imbalance. around the propeller axis.

#### Attachments

• IMG_20220324_1.jpg
127 KB · Views: 59
• IMG_20220324_2.jpg
216.9 KB · Views: 56
• IMG_20220324_3.jpg
228.7 KB · Views: 57
• IMG_20220324_4.jpg
222.6 KB · Views: 54
• IMG_20220324_5.jpg
221.4 KB · Views: 55
• IMG_20220324_6.jpg
233.1 KB · Views: 57
• IMG_20220324_7.jpg
102.1 KB · Views: 56
I'll try to give my view on your points.
1)the double reflecting cone will cause a lot of diffraction nullifying the effect of the waves (the interaction is the same of waves on a river, that reflect and move in a completely different way than the water flow).
the gas flow will easily double back into the annular passage even with a flat end as it follow the pressure difference and inertia.
2)the constant cross section could be useful for helping the gas expand and slow but wont affect the wave
3)as before the flow would have slowed considerably in the first cone but the local flow speed is not much influent on the pumping effect
4/5) 1700ft/s is at around 740F and is considered as average gas temperature. Obviously the pipe wall will be at a much lower temperature because while the gas it pretty hot is not very dense and will be expanded and expelled before having enough time to heat up the exhaust (same as inside the cylinder, the burning gas will be exceeding 2800F and aluminium melt at 1200F but the time is very short limiting the amount of heat transferred)
6) I think it might interesting to try eliminating the stinger altogether and just have one hole (or many small holes), but the sound might be too loud

I put here a few link of expansion pipes for Italian Vespas and moped so it might give you an idea of space constrained tuned pipes
the original are just baffled pot that give a lot of backpressure (10mm final diameter for a 50cc)
vespa original
this is for the 200cc and the header diameter is 40mm while the stinger is just 16mm
"tuned" original
this is an original modified to give a little more power but look inconspicuous
vespa tuned
racing exhaust for the same engine, and is all bent over itself in order to fit it under the frame

ciao moped original
pot muffler for 50cc header 22mm stinger 10mm
ciao moped tuned
same engine but "snake" pipe to fit under the frame
I worked my way through the adiabatic calcs, based on 70 bar at 10:1 cr,
and the peak temp seemed to be around 1300 deg c,
discharge temp 140 deg c,
discharge pressure 20 psi. (gauge).

Where do the higher pressures and temperatures come from?

You won't get excess heating and erosion of four stroke exhaust valves at these conditions.

I am sure you won't get 393 deg C in the muffler unless you are running on the limiter, and have a lot of unburned fuel,
or excess fuel is burning in the pipe.

Under adverse conditions, temperature can get over 650 degrees C - a nice red glow for steel.

Other factors causing non-adiabatic behaviour- condensation of ionised mixtures, formation of compounds at lower temperature and pressures, strongly delayed flame front- usually only seen in Wankel engines at very high rpms.

Does this effect actually reduce output power, or just spread chamber pressure over a wider range of piston motion?
- rather like a steam engine.

I have grabbed you references, and will have a look.

It looks like I will have to buy my own rolling machine- fairly cheap, but the centre roller looks too large.
How do I make cones to such small opening diameters?

Well, still adiabatic, but the heat capacities given are not so good for actual combustion.
Cp and Cv are not constants.
There is probably a lot more heat energy than indicated by the peak pressure.
This could be found independently by measuring the amount of petrol in grams and using the calorific value.
I think around 10:1 is the usual mixture in mass ratios?
I could then compare the values for adiabatic work, torque, and theoretical efficiency.
Does the 30% power, 30% cooling, 30% out the exhaust, really hold in this case?
Specific fuel consumption comes into this, too. 2-strokes use about 30% more fuel than 4-strokes for the same power.

Back to the current engine mods.
Maximum RPM seems to have settled at 5000 rpm, so I will reinstate the boost ports in a different way.
I think I can feed them from the transfer chests, in passages under the barrel internal surface.
The rear of the side transfer chests can be extended a bit, as well.
The actual bore of the boost ports outlets will be restricted to 5mm diameter, angled 60 degrees.

I will strip the engine tonight, cut the new passages, J-B on some new passage covers,
and photograph the J-B barrel and piston wear tomorrow morning.

My Notebook screen seems to be failing, so I am using an old notebook and my one-drive hookup
for now.

Here are the "after" photos.
There are no problems with adhesion with the J-B weld.
The main problem is that aluminium wears too quickly sliding over the J-B Weld.
I will sand back the affected parts of the piston, and build it up with many coats of Caliper paint.
This may stay put, is hard, solvent proof, and will protect the aluminium.
This is a matte ceramic-based coating, and may wear badly on the J-B weld as well.
I really need a spray-on adhesive dry lube coating.
I will look for some.

#### Attachments

• IMG_20220325_1.jpg
39.5 KB · Views: 58
• IMG_20220325_2.jpg
39.2 KB · Views: 56
• IMG_20220325_3.jpg
47.7 KB · Views: 58
• IMG_20220325_4.jpg
51 KB · Views: 59
Now here is a totally speculative mod:

I have embedded 3 pieces of raw PTFE sheet in the piston thrust face.
This stuff is quite soft and easy to cut with a scalpel blade.
Dimensions 3mmx13mm.
These are not enough to take the full piston thrust, but I am hoping for a "Burnish-in"
effect, where the PTFE gets burnished into the rest of the piston face and the cylinder wall, including a bit of
J-B Weld surfacing.

"Before" photos to follow, tomorrow, after I shape and finish the piston skirt.

I will arrange for slight insert protrusion, and a light push fit for the piston into the barrel.

These should stay in place much like o-rings in a groove.

The piston recess holes have 45 degree sides, and the inserts are an arc of a circle.

I noticed that they have low bridging strength, so I filled in behind then with J-B weld as well.

Here are the "before" photos.

I ran it, and it didn't blow up in 7 minutes running, so this mod appears to be safe.
Lets see if the piston mods are good for multiple hours running!

The engine didn't run so well, but symptoms are of poor ring to cylinder seal, so that should improve on the next run.
(inconsistent mixture requirements, not wanting to rev continually).
The photos also show how I am feeding the boost ports.
There are a couple of oil leaks through the J-B weld edges, but no serious air leaks, so I will leave that alone for now.
There are places on the piston inserts where I can add some more PTFE strips.
I will refresh the strips on the next rebuild, see if it made any difference to wear, and add more strips.
If the lower cylinder wall on the cold side was more flush and not built up with J-B, then that would cut
down on wear.

However, just refreshing the PTFE strips each rebuild may keep this engine running well enough for more mod tests.
I want to see if higher transfers combined with reed valves makes any difference, and if it will make enough power
at the 6000 rpm point.
<edit>
After run 2, it seems to be running tight and seizing, from about 4000 rpm continuous.

That is bad for the exhaust side skirt wear. I can feel tightness on the first turn, then it frees up.

Seizure may be caused by the expansion of the PTFE with heat.
Also, possibly the tightish initial fit is being made worse by piston expansion as well, and a lack of "give" in the PTFE.

It will develop very high internal pressure before extruding into the clearance gap.
This does show that the PTFE is staying put, and acting as a bearing.
The rotation resistance will be coming from excess pressure on the exhaust side skirt.
Very likely there will be seize marks.

I will try running in more gradually around 3500 rpm for several runs, before trying for more rpms.
Hopefully it will settle down, then I can do a check-dismantle.

#### Attachments

• IMG_20220327_1.jpg
38.6 KB · Views: 52
• IMG_20220327_2.jpg
43.6 KB · Views: 59
• IMG_20220327_3.jpg
45.9 KB · Views: 49
• IMG_20220327_4.jpg
39.3 KB · Views: 50
• IMG_20220327_5.jpg
46.7 KB · Views: 49
• IMG_20220327_6.jpg
44.8 KB · Views: 48
Plan for a modification to the piston PTFE sliders.

1) PTFE on steel, lubricated ; Cf is around 0.04.( static and dynamic).
This compares well with the aluminium on steel-style lubrication of the piston skirt.
This is mixed boundary, dynamic, and squeeze lubrication, which varies from 0.1 to 0.001
for the dynamic coefficient of friction.

Maximum piston side thrust, assume 15 degrees rod angle, 50bar, 45mm diameter piston
sin 15 = 0.26
area = 1.6 x 10^-3 sq mt.
F = PA = 50 x 10^5 x 1.6 x 10^-3 x 0.26 = 21x 10^2 N, or 2 kN
PTFE area = 26 x 3 x 10^-6 = 78 x 10^-6
PTFE pressure = F/A = 2 x 10^3/ 78 x 10^-6 = 25 MPa
this is pretty high. Recommended maximum is 3 MPa.
<> With 4 strips, this drops to 12.5 MPa

I would suppose that PTFE on steel is only limited by excess heating and the pressure needed to start extruding the PTFE .
This example would be towards the low end, say 2 MPa compressive yield strength,
for general specs.

However, to extrude the material into the clearance gap would be similar to the failure of a rubber o-ring,
which would be into the 100s of MPa, I think.
They are quite happy at 5000 psi, or 35 mPa.
The UTS of nitrile rubber is probably a lot higher, but general softness or resistance to indentation, is lower for the rubber. Modulus of elasticity is also higher for PTFE.

This is sort of apples and oranges, as PTFE takes a permanent indentation, and rubber does not,
with similar pressure applied.

I will run the engine for 5 more cycles, and check for permanent deformation , wiping,
and wear.
(I can increase actual support by 2 more inserts 3x13 area.)

On the exhaust side, the inserts need to be oriented vertically, with square ends.
Otherwise they will fall out into the exhaust port.

I can fit 4 vertically, plus one as a skirt bottom support.

This gives maybe about 6MPa, as it is the low thrust side of the piston.

the J-B weld will take ult = 27 MPa, but not for normal loading.

I will report back on the "after" state of the inserts.

If it is looking good, I will take it another stage.
.
The current problem is that the exhaust side skirt is not lubricating well with a constant pressure on it.
It needs a little clearance to bring in lubricating oil.

This raises the problem in that clearances are fairly ideal without making room for the inserts.

Both flanks need to be reduced by 0.3 or so mm to allow the inserts to stand out a bit, and still
fit in the bore.

However, the outer perimeter of the exhaust side skirt is important in sealing crankcase pressure, so
should be left in place. A flexible rim would be nice, to take load off the lip, but
I can't figure out how it should be attached.
However, excessive lip wear can probably be repaired.

An o-ring format for the inserts is not good, as they need to be vertically restrained.
I presume that end location of 3x2mm area is enough given potential side forces of
1 kN (say) x 0,04 =40N, area = 6 x 10^-6, <> x 4 ;P = F/A = 1.8 MPa.

Shear stress in the insert at 3 x 13 = 39 x 10^-6 , F/A = 40/39 x 10^6 = 1 MPa, <> /4 = 0.25 MPa.

I shall just have to see how the inspection goes.

Drag engines seem to get away with 3 round buttons of a harder material in the thrust face, which must have quite high stress in the insert holes. These are larger pistons, too, -around 75mm diameter.
and greater con-rod angularity, plus running nitro and/or supercharging.
These are not intended for continuous output, however.
<> means a later edit.

Last edited:
I Found the cause of the engine problem:
a damaged ring land and a tiny crack in the piston in the same area.
Just enough to overheat the piston.

Seizing is on the cold side, where the damage is.
The exhaust side skirt looks good.
It doesn't look like the PTFE inserts have any effect.

This engine is retired until further notice.

Replies
14
Views
5K
Replies
44
Views
7K
Replies
9
Views
3K
Replies
5
Views
3K
Replies
115
Views
34K