RH4 To Narley Port Conversion Tubes

acotrel said:

That looks like the ports in my 850 head which taper from 34mm to 30mm in the first 12mm.. The port shape changes the requirement for the taper on the needles. The port's diameter can change the revs at which the port is most effective. So it needs to suit the gearing for the circuit. If you have quick taper needles and use more throttle too quickly, you can end up with less torque.

Click to expand...

I did have to richen the slides, and the needle jet diameter as a result of these Narley port inserts. The richer side required a leaner idle jet. And I also changed the air jet as well.

WIth the Keihin CRs if anyone runs into this and cannot find needles richer than YY5, I ended up taking 0.0005" (half a thousandth of an inch) off of the straight diameter of a YY5 needle. There are 28mm to 33mm FCR needles that will get you close to there as well and are the right length, the difference is the length to the shoulder of the needle which only affects the needle clip position. I used an FCR needle that was close, but slight too rich it was a GKP, which is 0.001" (one thousandth) smaller than a YY5. The clip position in most cases just moves where the needle needs to be and in this case it moved it to the middle position and gives more range so a positive in my book. However, in the end the YY5 was the way to go.

Richer slides were hard to find, I ended up tracking down 2.0 slides and filing them to 2.5, I should have just chucked the 3.0 slides in the lathe and cut them down to 2.5 in retrospect, but this works.

I should probably do a write up on the jetting as the Sudco documentation has some inaccuracies, and the air jet in these works like a Mikuni VM air jet and effects a lot more range, especially with the 2.5 slides.

I have no hesitation off idle, and it pulls strong up to and beyond 7500. I am on stock sized valves and I have not cut into the head, so I suspect there is more to be found there.

My jetting for reference is:

2.5 Slide
60 Idle jet
YY5 (modifed as described above)
200 air jet (i am also playing with 190s)
Main jet is a 130 (i am playing with 135 and 140 as well)

maaseyracer said:

I did have to richen the slides, and the needle jet diameter as a result of these Narley port inserts. The richer side required a leaner idle jet. And I also changed the air jet as well.

WIth the Keihin CRs if anyone runs into this and cannot find needles richer than YY5, I ended up taking 0.0005" (half a thousandth of an inch) off of the straight diameter of a YY5 needle. There are 28mm to 33mm FCR needles that will get you close to there as well and are the right length, the difference is the length to the shoulder of the needle which only affects the needle clip position. I used an FCR needle that was close, but slight too rich it was a GKP, which is 0.001" (one thousandth) smaller than a YY5. The clip position in most cases just moves where the needle needs to be and in this case it moved it to the middle position and gives more range so a positive in my book. However, in the end the YY5 was the way to go.

Richer slides were hard to find, I ended up tracking down 2.0 slides and filing them to 2.5, I should have just chucked the 3.0 slides in the lathe and cut them down to 2.5 in retrospect, but this works.

I should probably do a write up on the jetting as the Sudco documentation has some inaccuracies, and the air jet in these works like a Mikuni VM air jet and effects a lot more range, especially with the 2.5 slides.

I have no hesitation off idle, and it pulls strong up to and beyond 7500. I am on stock sized valves and I have not cut into the head, so I suspect there is more to be found there.

My jetting for reference is:

2.5 Slide
60 Idle jet
YY5 (modifed as described above)
200 air jet (i am also playing with 190s)
Main jet is a 130 (i am playing with 135 and 140 as well)

Click to expand...

Because I use methanol fuel, jetting is much easier for me, because the jets flow twice as much, so the errors are half as significant, as when you jet for petrol. I would never try to machine half a thou off a carb needle. If you read up on Mikuni needles, there are several stages in their tapers. With methanol, I make my own needle jets from brass hex using a mix of metric and number drills. I can change the jet internal diameters in half thou of an inch steps. With methanol the steps need to be that small. That tells me that tuning that close with petrol is almost impossible with a normal carburetor. You would need quarter of a thou steps.
When I started tuning my Mark 2 Amals on my 850, I'd had experience with my previous bikes with methanol. I knew a 0.117 inch needle jet would be pretty close to good. But when I lowered the needle, I could not force the motor to cough. So I made two more jets at 0.1165 inch. And got the cough, then raised the needles one notch. Then I got two 6D Mikuni needles from my brother and tuned with them. A lot of people have mentioned poor throttle response with Commando engines - mine is as good or better than any other bike that I have ever ridden.
Tuning a two stroke on petrol is twice as difficult as tuning a four stroke on petrol, but guys seem to be able to do it using normal carburettors. In the 1970s, the fastest TZ Yamahas seemed to have Lectron carbs where the needles could be turned to adjust the mixture.

It’s been a few months since I kicked off this project. Originally, my plan was to get the intake port finalized before the exhaust, but that turned out to be more challenging than expected. The exhaust port, being simpler, and able to get more exacting dimensions from the port scan ended up being easier to prototype and refine. The intake, on the other hand, posed some unexpected hurdles, both in terms of complexity and technology.

One thing I’ve learned through this process is that 3D printing has its own limitations—especially when it comes to print thickness and consistency. I’ve mentioned this before, but it’s more like casting than machining in many ways. The key difference, of course, is that 3D printing is additive, so you're building up material instead of removing it. The process involves lasers and a bonding agent to create a “green print,” which is then baked to remove any excess materials before it gets sintered into its final form.

The biggest challenge I kept facing with the intake tubes was getting the top and sides of the tube as well as the entry and exit points thin enough consistently. If they were too thin, the print would deform. Too thick, and I’d have to grind them down by hand. I ran the first set for quite some time, and they performed incredibly well—actually exceeding my expectations. But I still had a pile of intake tubes from at least a dozen different 3D printers that just weren’t quite right.

The exhaust fittings opened the doors to the idea of raising the floor and removing the problem areas in the print entirely and working with the existing 32mm port ceiling. Prototypes came in just the other day, and the first print was nearly perfect. Best of all, it’s repeatable—I've gotten consistent samples from several different printers.



These new ports are the closest I’ve come to replicating the Axtell templates. Plenty to still work out, but I'm really happy with the progress so far. Testing hopefully this coming week assuming I can work through this electrical repair detour.

maaseyracer said:

It’s been a few months since I kicked off this project. Originally, my plan was to get the intake port finalized before the exhaust, but that turned out to be more challenging than expected. The exhaust port, being simpler, and able to get more exacting dimensions from the port scan ended up being easier to prototype and refine. The intake, on the other hand, posed some unexpected hurdles, both in terms of complexity and technology.

One thing I’ve learned through this process is that 3D printing has its own limitations—especially when it comes to print thickness and consistency. I’ve mentioned this before, but it’s more like casting than machining in many ways. The key difference, of course, is that 3D printing is additive, so you're building up material instead of removing it. The process involves lasers and a bonding agent to create a “green print,” which is then baked to remove any excess materials before it gets sintered into its final form.

The biggest challenge I kept facing with the intake tubes was getting the top and sides of the tube as well as the entry and exit points thin enough consistently. If they were too thin, the print would deform. Too thick, and I’d have to grind them down by hand. I ran the first set for quite some time, and they performed incredibly well—actually exceeding my expectations. But I still had a pile of intake tubes from at least a dozen different 3D printers that just weren’t quite right.

The exhaust fittings opened the doors to the idea of raising the floor and removing the problem areas in the print entirely and working with the existing 32mm port ceiling. Prototypes came in just the other day, and the first print was nearly perfect. Best of all, it’s repeatable—I've gotten consistent samples from several different printers.

These new ports are the closest I’ve come to replicating the Axtell templates. Plenty to still work out, but I'm really happy with the progress so far. Testing hopefully this coming week assuming I can work through this electrical repair detour.

Click to expand...

It makes sense -that should work well. I would never have thought of doing that.

Test fit looks good.

You mention this process to be more like casting than 3D printing. I wonder if investment casting (wax) would give you a more consistent final product. Just a thought.

A thing which has puzzled me is the concept of 'reversion'. I have always worked with the presumption that at TDC when the inlet and exhaust valves are both open, at high revs - the whole system resonates as a sound wave. That step at the bottom of the exhaust port would cause a disruption. However I never really thought much about it. I suggest flow benches do not replicate what happens in a motor, their use might improve the performance of low revving motors more than high revving.
One of my friends was a senior research scientist on a trans-sonic wind tunnel. It is a different game. I have been there and seen it. You do not get what you expect.

bmepkid said:

You mention this process to be more like casting than 3D printing. I wonder if investment casting (wax) would give you a more consistent final product. Just a thought.

Click to expand...

3D printing is more like casting than machining. There is also no way I would ever cast this, lead time on casting is a nightmare, compared to 3D print where you send a file and instructions to a printer and they return your product a few days later.

What needs to be right on these and consistent is not always what we typically think of when it comes to casting or machining. One of the biggest overlooked issues in 3D print is the transition from a Green print through the sintering process where there is a specific shrinkage to the final product. This can take refinement and is not always the percentage the printer claims. This is where it helps to learn about the material and process from each printer. However, once you know how well your printer controls their process you can get very replicable results. By getting rid of the very think walls at the top and sides I have eliminated a problem area, and captured a feature I wish I had in the first place.

acotrel said:

A thing which has puzzled me is the concept of 'reversion'. I have always worked with the presumption that at TDC when the inlet and exhaust valves are both open, at high revs - the whole system resonates as a sound wave. That step at the bottom of the exhaust port would cause a disruption. However I never really thought much about it. I suggest flow benches do not replicate what happens in a motor, their use might improve the performance of low revving motors more than high revving.
One of my friends was a senior research scientist on a trans-sonic wind tunnel. It is a different game. I have been there and seen it. You do not get what you expect.

Click to expand...

Flow benches cannot capture reversion well, and the best thing you can do with a bench is find how high you can raise the floor until you start to impact flow or exit air speed.

My flow bench parts should show up in the next week or two and I should be able to capture data on these. Also if any one has a broken RH4 head with at least one good intake port and exhaust port let me know, I am in need of a flow bench test head that is not the one attached to my bike.

One of the things which happened when I was a kid really impressed me. Three of my friends went to the old Fishermens' Bend airstrip to ride their bikes. One of them had a new Velocette Thruxton. He fitted a megaphone to it and almost immediately burned a piston. How does a megaphone cause a motor to run lean ?

acotrel said:

One of the things which happened when I was a kid really impressed me. Three of my friends went to the old Fishermens' Bend airstrip to ride their bikes. One of them had a new Velocette Thruxton. He fitted a megaphone to it and almost immediately burned a piston. How does a megaphone cause a motor to run lean ?

Click to expand...

So when it burned a piston was it running flat out? If so was it on the same day it'd been running flat out with the standard exhaust?
Was the air temperature the same ? Was it running exactly the same revs for exactly the same distance ?

Memory of events from the distant past is a tricky thing.

baz said:

So when it burned a piston was it running flat out? If so was it on the same day it'd been running flat out with the standard exhaust?
Was the air temperature the same ? Was it running exactly the same revs for exactly the same distance ?

Click to expand...

The three guys rode their motorcycles about 10 miles to the Fishermen's Bend airstrip where they did high speed runs. I think the megaphone would have been fitted at the airstrip. The bike would not have been ridden across Melbourne with a megaphone exhaust. At any rate, when my mate and I met them. Bob Lewins was being towed home. A 350cc Manx Norton, and a 7R AJS usually have reverse cone megaphones - how do they work differently to an open cone megaphone ? I suggest it is not simple gas flow. With two-stroke expansion chambers, the diameter of the centre section or the header pipes can change the torque characteristic. I suggest most of this stuff is found by trial and error - not design.
I once had a set of expansion chambers made which gave extreme torque - they made the bike really stupid. I didn't have the brains to raise the gearing high enough.

acotrel said:

A thing which has puzzled me is the concept of 'reversion'. I have always worked with the presumption that at TDC when the inlet and exhaust valves are both open, at high revs - the whole system resonates as a sound wave. That step at the bottom of the exhaust port would cause a disruption. However I never really thought much about it. I suggest flow benches do not replicate what happens in a motor, their use might improve the performance of low revving motors more than high-revving.

Click to expand...

Of course, flow benches can’t replicate what happens in a motor, because a flow bench is merely a machine that measures airflow through a duct. The induction system, the exhaust system, the cam, the compression ratio, etc, are all interrelated and central to how air flows in an engine, so how could a flowbench possibly address all these variables? However, what a flow bench can tell you is the amount of air moving through an inlet duct, and how near, or far, from entitlement airflow the specific duct being tested is.

As an example consider a 30 mm RH10 port, which has an entitlement airflow of ~ 160 cfm @ 28” WC. If we measure a stock RH10 head on a flow bench it would flow ~ 125 cfm at 28” WC (this has been shown here on the forum ad nauseum over the decades). So, despite how “good” an RH10 head performs, it is far from achieving entitlement airflow. Assuming your engine is cammed, compressed, and piped optimally, if you can’t get more than 125 cfm through your 30 mm port, you’re stuck concerning increasing HP and you’ll never catch Buddy with the big-valve 30 mm port that flows 150 cfm. Although the flow bench can’t replicate a running engine for all the reasons cited above, it can certainly characterize an engine’s potential based on airflow numbers. And no, flow benches are not more useful to improve the performance of low-revving engines rather than high-revving engines, because airflow is what dictates the rpm regime at which an engine can operate – no airflow, no rpm.

Case in point, Axtell provided Norton heads with correctly sized ports and higher airflow (via a flow bench) than others were achieving at that point in time. This high airflow supported a higher rpm operating regime, such that it exposed the mechanical limitations of the Norton crankshaft, cases, and valve train. Some suggested his engines weren’t durable. In my opinion, it was a simple case of Axtell providing airflow potential capable of blowing up a well-prepared engine.

WZ507 said:

Of course, flow benches can’t replicate what happens in a motor, because a flow bench is merely a machine that measures airflow through a duct. The induction system, the exhaust system, the cam, the compression ratio, etc, are all interrelated and central to how air flows in an engine, so how could a flowbench possibly address all these variables? However, what a flow bench can tell you is the amount of air moving through an inlet duct, and how near, or far, from entitlement airflow the specific duct being tested is.

As an example consider a 30 mm RH10 port, which has an entitlement airflow of ~ 160 cfm @ 28” WC. If we measure a stock RH10 head on a flow bench it would flow ~ 125 cfm at 28” WC (this has been shown here on the forum ad nauseum over the decades). So, despite how “good” an RH10 head performs, it is far from achieving entitlement airflow. Assuming your engine is cammed, compressed, and piped optimally, if you can’t get more than 125 cfm through your 30 mm port, you’re stuck concerning increasing HP and you’ll never catch Buddy with the big-valve 30 mm port that flows 150 cfm. Although the flow bench can’t replicate a running engine for all the reasons cited above, it can certainly characterize an engine’s potential based on airflow numbers. And no, flow benches are not more useful to improve the performance of low-revving engines rather than high-revving engines, because airflow is what dictates the rpm regime at which an engine can operate – no airflow, no rpm.

Case in point, Axtell provided Norton heads with correctly sized ports and higher airflow (via a flow bench) than others were achieving at that point in time. This high airflow supported a higher rpm operating regime, such that it exposed the mechanical limitations of the Norton crankshaft, cases, and valve train. Some suggested his engines weren’t durable. In my opinion, it was a simple case of Axtell providing airflow potential capable of blowing up a well-prepared engine.

Click to expand...

Give us some details about the big valve 30mm port head with what sized valves? and 150 cfm flow

jseng1 said:

Give us some details about the big valve 30mm port head with what sized valves? and 150 cfm flow

Click to expand...

30mm port, 3mm oversized inlets, work by Comnoz. Not quite 150, but close…

The velocity with which the incoming charge moves across the top of the combustion chamber when both valves are open at TDC is probably much higher than can be achieved with a flow bench. With jet aircraft swept wings are much faster than straight - that was not predicted when the first jets flew. I work on the premise that there are also Kadency effects in four stroke motors. Some of what passes through the two open valves into the pipe, gets stuffed back into the combustion chamber before the exhaust valve closes, depending on the exhaust system. It is not purely a matter of gas flow through the combustion chamber.

An expansion chamber might work on a four-stroke motor, but it's dimensions might be absurd.

Fast Eddie said:

30mm port, 3mm oversized inlets, work by Comnoz. Not quite 150, but close…

Click to expand...

That is a lot of work for 20cfm, any idea how much air speed you gained? 145 CFM would be around a 90% VE for a 41mm valve, air speed would be interesting to see as we can then get an idea of air density going through the port which is what really matters. Air cooled engines are very tough to chase flow gains out of.

I am setting up my flow bench right now and I don't want to comment on flow numbers that I have as they are from a bench that I most likely will not be using any more and I want to share consistent change over numbers. I'll share once I get up and running.

maaseyracer said:

That is a lot of work for 20cfm, any idea how much air speed you gained? 145 CFM would be around a 90% VE for a 41mm valve, air speed would be interesting to see as we can then get an idea of air density going through the port which is what really matters. Air cooled engines are very tough to chase flow gains out of.

I am setting up my flow bench right now and I don't want to comment on flow numbers that I have as they are from a bench that I most likely will not be using any more and I want to share consistent change over numbers. I'll share once I get up and running.

Click to expand...

This is the velocity graph below.

What I can tell you is that this work yielded a 9 (yes NINE) rwhp increase on this particular engine set up.

Old news though, well worth reading through Comnoz’ head flow testing thread if you’re interested in this stuff.

Link to the thread: