Airflow in Peterson System Pipes: A Conversation with Dr. Dunn

Airflow in Peterson System Pipes: A Conversation with Dr. Dwain Dunn

In physics and engineering, fluid dynamics is the field that describes and measures the flow of fluids, including air and other gasses, as they move through various environments. One such environment is that found inside of a tobacco pipe. Air must move efficiently from the bowl, through the shank and stem, and out of the mouthpiece for proper function. The efficiency of that movement is crucial because interruptions can cause problems like moisture buildup and heat that are antithetical to the conversion of tobacco into satisfying smoke.

Airflow in Peterson System Pipes: A Conversation with Dr. Dwain Dunn

Dr. Dwain Dunn

Fluid dynamics are essential to solving issues for such applications as oil or water pipelines, aircraft lift, and other forces, weather prediction, interstellar observations of gas clouds, ocean currents, and myriad other subjects. The mathematical formulas required are breathtakingly complex, and for those of us who are not physicists, they more resemble the Elvish writing found in Tolkien than the geometry formulas we learned in school. But there are some who understand and use them to interpret real-world effects and answer real-world questions.

Dr. Dwain Dunn is one such individual. He understands the field because he spent years earning a Ph.D. in mechanical engineering and more years applying his expertise. His specialization is in turbo machinery like jet engines, and he finds ways, for example, to improve fuel consumption and reduce efficiency loss. He's currently a lecturer at the University of Dublin. Appropriate to his location, he's a Peterson pipe enthusiast.

When enthusiastic pipe smokers apply their professional expertise to their favorite hobby, interesting things happen. Dwain has mapped the airflow through Peterson pipes to see exactly what is happening and just how well System pipes work in terms of airflow. What's particularly amazing is just how well Charles Peterson designed the System more than a hundred years ago and without the advantages we have today. Examined in the clinical light of modern-day science, Peterson pipes are anachronistic, with excellent flow properties and sophisticated design ahead of their time.

Dwain has mapped the airflow through Peterson pipes to see exactly what is happening...

While aviation technology is what Dwain teaches, his interests are more wide-ranging. "Toward the completion of my degree," he says, "I found that fluid flow was more interesting than most other aspects of mechanical engineering, so I started to specialize in that and found that computational work was the most intriguing. It's where you mathematically describe a scenario that you want to analyze, and once you've got that mathematical model into a computer you can run a series of equations that dictate or describe how fluid behaves."

"...once you've got that mathematical model into a computer you can run a series of equations that dictate or describe how fluid behaves."

That's what he did with airflow in Peterson System pipes. "I originally did this just to investigate how the System worked. I've now run a whole series of analyses, including things such as the well depth." We call that the tobacco chamber depth here in the U.S. Dwain also took into account the gap-stem element of the Peterson System, which maintains a tight stem fit even after years of wear. "As the stem wears, obviously the tenon and mortise change their relative locations, so I've looked at a bunch of those variables. I looked at the Deluxe System with the chimney on the end of the tenon, as well as at its cross vents, and I've looked at the airflow in the mouth." He has gathered and examined monumental quantities of data.

Chimneys, Sometimes Known as Extenders or Condensers

For those unfamiliar, the chimney screws into the end of the tenon and was originally made of bone, with vents drilled into its sides. It was part of Peterson System designs from the beginning, according to Peterson expert Mark Irwin on his site,, and was shown in the original 1896 catalog. "The top tier of System pipes," writes Mark, "had a funneled, screw-in bone extension, which was cut according to the size of the mouthpiece, bowl and mortise. It could be quite large in the case of an Oversize or 'House Pipe' and was scaled according to the bowl in question. Bone was used until around 1963, according to the retired craftsmen at Peterson, at which time the factory switched to aluminum."

Mark adds that the chimney's purpose "is to aid in condensing water vapor in pipe smoke and cause the condensate to fall into the reservoir. This dries and cools the smoke as well as prevents solid material (ash or tobacco) from entering the airway of the mouthpiece. If you've used a System pipe you know that the System does in fact work in this way when coupled with the P-Lip mouthpiece and a correctly-drilled reservoir."

Many Peterson smokers over the years have discarded those removable chimneys, primarily because they become gunky and discolored if not maintained. Yet they are part of the System as Charles Peterson originally designed it, and the chimney contributes to the smoking character of System Petersons. Dwain Dunn's analyses of airflow using the chimney, however, reveal superior airflow when they are used. "Any negative issues I saw in the analyses," he says, "were fixed by the chimney. From what I can see, the big difference is that the airflow is a little bit cleaner on the Deluxe with that chimney."

The Gathering of Evidence

Airflow in Peterson System Pipes: A Conversation with Dr. Dwain Dunn

Illustration of System airflow, courtesy of Dr. Dwain Dunn

How Dwain accomplished his observations is interesting. "I didn't model smoke explicitly," he says. "I modeled air and as closely as possible the properties of smoke. But to analyze smoke explicitly, yeah, I'd still be running the first set of simulations because it's extremely complex. The chemical reactions are immense and you need to capture all of them to get it right, so I decided to simplify it and instead use air at about 50% humidity, so there is moisture and it takes that into account. But the little particulates, the oils, the aromatic compounds, and things like that that provide the flavor, but I decided to neglect all of that because, as I said, trying to get that would be very, very difficult. And then it would change from tobacco to tobacco and pipe bowl to pipe bowl. If you pack a little bit more or a little bit less, if the day is a little bit more humid, all of those things change, so I decided just to eliminate those variables and look at everything using a set standard, and for that, I chose humid air."

Aside from humidity, the density of the smoke changes as a function of temperature. "So as the temperature of the smoke changes, the density changes as well. Some of the analyses revealed the density changes through the pipe start to finish."

...the density of the smoke changes as a function of temperature.

He simulated an ember, which he calls the "cherry," as the continual ignition source within the bowl. "I made that 500℃, which is low when you're looking at the burning temperature of tobacco because it can get up to over 1,000℃, but we don't want it to be that hot because we lose a lot of the flavor. So I chose a lower value of about 500℃ and you can actually see the temperature change as the air moves through the pipe. It heats up through the cherry and then as it's in contact with the briar it slowly but surely cools down, and as it cools, its density changes as a function of temperature.

"As the air moves through the tobacco, obviously if the tobacco is more moist than the air, the air will pick up some of that moisture, but the change in characteristics of that air isn't drastic enough to cause a noticeable difference puff to puff. It might be noticeable from pipe to pipe, which is why you need to figure out the different humidity level of the tobacco for each pipe that you have. Some pipes like a dryer tobacco, some like it more wet. But once you've got the pipe going, it should be relatively constant, and the air will pick up humidity or lose it based on the state of the tobacco and the ambient air."

The humidity level of the tobacco makes an important difference. "When you dry your tobacco," says Dwain "you're changing the burning point. The dryer the tobacco, the hotter it burns, the more moist the tobacco is, the more difficult it is to burn, and the harder you're going to puff to keep things going. You have to keep stoking that ember, and that is eventually going to cause chaos. You're now going to be drawing steam through, and if you have condensation points, you're going to have more condensation and a wetter smoke, which is not ideal. But this is where the balance comes in because a lot of the flavors are in the moisture. So, if it's overly dry, you lose a lot of the flavor. You need to get to the appropriate balance, and once you know what that balance is — and again, it changes from tobacco to tobacco and pipe to pipe, at least in my experience — some tobaccos need to dry more if you're smoking them in a different type of pipe; the shape of the bowl plays into this a fair bit as well."

With a larger bowl comes a larger ember and more heat generated for the next layer of tobacco, but also, more flavor becomes available. "The cherry itself is going to carbonize some of the aromatic components of the tobacco. So, you're picking up the aromatic characteristics from slightly lower down, not necessarily in the cherry itself, which is why when you get to the bottom of the bowl, all of that is now evaporated and there's very little flavor at the bottom. The reason for that is because you're drying up those volatile aromatics."

With a larger bowl comes a larger ember and more heat generated for the next layer of tobacco, but also, more flavor becomes available.

Dwain measured the speed at which smoke travels through different parts of a System pipe and found that the velocity reduces as it reaches the stem, indicated by dark blue in his diagrams. "When the smoke enters the System at the reservoir area, the first thing it has to do is move around the tenon, and then travel down toward the well, change direction, and then get sucked straight back in. Now in engineering terms, this is what's referred to as a momentum trap. Any particulates that you have in the air, any smoke particles, unburned tobacco, or moisture droplets, as soon as they reach the stem and slow down, they don't have the energy to carry on being drawn up into the stem and will settle into the System reservoir."

" engineering terms, this is what's referred to as a momentum trap."

Dwain says that it's the same way a vacuum cleaner traps dust. "It's a similar concept. By slowing the airflow you're reducing the drag forces on the particles, making them less likely to be drawn up into the stem. Another way of looking at this is terminal velocity. If you jump out of an aircraft you travel nice and fast, you open up a parachute and start to slow down. And the slow-down is what you want with the parachute, but here by slowing the air down all the particles now start to lose the energy required to carry on moving with the air so they will fall out of it. So you'll get the particles that'll now start to separate and fall downward. So if you've got little droplets of moisture they'll fall down into the well, any particles of ash, anything like that."

That effect is visually apparent on the chimney itself. "There's always a little ring of discoloration, a little ring of dirt and ash and little particles. The reason for that is because the air has slowed down so much that those particles get trapped. The chimney obviously cools the air down a little bit which causes some of the moisture to condense and stick to the surface, which acts as a further trap for airborne particles. Those that don't stick to the chimney will then slow down in a circulating flow and then settle."

Airflow in Peterson System Pipes: A Conversation with Dr. Dwain Dunn

Illustration of System airflow, courtesy of Dr. Dwain Dunn

The chimney provides two important features. "The first is the rounded aspect of the inlet on the outside. What that does is allow the air to enter into the draft hole more slowly, which helps reduce whistle. There's separation of the flow, and you get a little recirculating region at the entrance without it. That separation will cause condensation to sit there, and that's where any gurgle will start because you now have this little recirculation. Whereas on the Deluxe System, because it's got the rounded inlet, you don't see that little recirculation region forming, which is an advantage because it will eliminate the gurgle." That's all due to the shape of the chimney.

"If we look at the Standard System, you don't have a rounded corner, and what happens is the air comes across causing a recirculation region. By having the chimney rounded, the air moves around more smoothly; it goes up into the extender a lot smoother, and you don't have that recirculation region on the inside."

The second feature that Dwain notes is the internal curvature of the chamber (or reservoir) around the tenon. "That curve directs the smoke to move up and around and then back down, again slowing it down, giving it more time in contact with the briar. Longer contact with the briar means more moisture is drawn out of the smoke and into the briar, with more temperature transfer to cool it down more. And because it's slowing down for longer, the smoke is less likely to carry as many particulates. Those are the two main features I found that put the Deluxe a step above the Standard, but there's an associated cost difference between the two, so it's justifiable. The Standard System is standard; if you want deluxe performance, the Deluxe System is a better way of doing it."

"Longer contact with the briar means more moisture is drawn out of the smoke"

The Uselessness of Stingers

Dwain notes that the stingers found in many pipes from the 1940s-50s are not efficient, which perhaps explains why so many smokers immediately removed them from pipes. "They're bad, in my opinion. Of course, it depends on the design of the stinger. One of the problems is that as the smoke approaches the stinger, it has to speed up because it's a blockage in the airflow. You don't have a recess like in a Peterson; it's just stuck in the stem of the pipe. The pipe itself isn't accommodating the change in area, the smoke has to speed up, and any particles in the smoke are being energized and are less likely to fall out. They will condense a little bit, but because they're moving so fast they can't condense enough.

"The biggest drawback with stingers is that the entrance to the stem itself is always reduced, always. I've got some stingers where it takes a two-and-a-half, maybe three-millimeters stem hole, and it reduces it to less than a millimeter. So now you're really speeding up the air, and the faster it goes the more likely it's going to cause tongue bite because it hasn't had enough time to cool down. On some of the pipes that may work — for instance, the Kaywoodie Drinkless — I think the way that they've done it is not bad, with the ball at the end with a series of holes in different positions. I think that's probably the best way to do it, but you'd need to make sure that before the smoke actually gets to those holes, it's had a chance to cool down. Stingers do the exact opposite of what you would hope by restricting the airflow. They create bottlenecks where moisture will accumulate. And obviously, it's more prone to whistling. Two of the different stingers that I've got created terrible whistling. So I just take them out."

Stingers do the exact opposite of what you would hope

The P-Lip Mouthpiece

For his initial investigations, Dwain did not have reliable measurements for the geometry of the human mouth but has since extended his analysis, and it turns out that the Peterson graduated bore and P-Lip stem function admirably in conjunction with the human mouth. "From the analyses, it actually looks like the P-Lip is doing a fair amount of good stuff." The graduated bore helps accelerate the smoke as it rises to the lip button. "The reason it speeds up is because of that narrowing. You're getting a change in diameter, the cross-sectional area is reducing, and that reduction is speeding up the airflow." It's in the mouth at the P-Lip that the highest velocity is attained.

"By speeding up the air," says Dwain, "or speeding up the smoke as it enters the mouth, it'll now saturate the palette by recirculating around the entire cavity." The smoke more efficiently contacts the back of the palate and spreads across all of the mouth's contours and covers the entire tongue. The images he provides indicate that the distribution of smoke is complete. The adjusted velocity of smoke generated by the P-Lip allows it to fully encompass the back of the mouth and remain in contact with the tongue longer. "Obviously, you can do whatever you decide to as part of the puff. I know one guy on YouTube who says, 'You need to chew the smoke.'" By approximating chewing movements, with the mouth opening and closing, the smoke moves around even more and distributes flavor across all of the taste buds.

Another variable that Dwain examined was the horizontal angle of the stem in the mouth. His results indicate that center position is superior for flavor distribution. Clenched to one side, distribution is reduced. "Your cadence will change this as well. If you puff a little bit harder, it'll behave differently." Dwain's measurements reflect his own average cadence. "What I did to get these values was look at the speed with which I draw, and I calculated what that was, and I imposed that on the simulation. So for someone who doesn't necessarily sip, but takes very short, large puffs, the airspeed will increase and that'll change the distribution. So you have to have a holistic approach. The front could give you better flavor, but if you change the way that you draw on the pipe, if you change your cadence, then the front might not work very well for you. If you start puffing very slowly, and then come to a peak, and then slow down your puff towards the end so it's a smooth transition, you'll get a different effect than if you take short, sharp puffs. So all of these things will play a part."

Peterson: More than the Sum of its Parts

Airflow in Peterson System Pipes: A Conversation with Dr. Dwain Dunn

Illustration of System airflow, courtesy of Dr. Dwain Dunn

All of Dwain's investigations, all of his hundreds of simulations, started because he was curious about Peterson pipes. "I wanted to see what the System actually did because all of the literature makes a whole lot of claims, but some of them just sounded a little bit too impossible. So I thought, 'Well, maybe there's something going on here I don't understand. Let me run a couple of simulations, have a look, see what's going on, and maybe I can understand what's happening.'"

"I wanted to see what the System actually did"

What he found were considerable advantages that perhaps even Charles Peterson failed to anticipate. "There are a lot of unintended benefits in all of this. Some of the literature says that it improves this and it changes that. And yes it does, but it also does so much more."

As an example, Dwain refers to the graduated bore of the P-Lip mouthpiece. "I think the inlet starts at 5.5mm, or something like that, and then the very tip of the P-Lip is closer to about one-and-a-half millimeters. I have forgotten the exact dimensions at the moment, but it's a graduated bore. What that allows is an acceleration of the air as it moves through the stem. It changes the terminal velocity of the particles, so as the smoke cools and you start to get the aromatic particles that are condensing in the smoke because the smoke is speeding up, you're more likely to catch those particles and take them through to the mouth. So by changing the diameters, you can change the size of the particles that you want to let through.

"Because the stem is bent, it creates a vortex, which is why I say that the bend in the stem is quite important. As airflow moves through a bend, it acquires rotation. A vortex is created. Because of that vortex, when the smoke exits the pipe and enters the mouth, we've already got this spreading out of the smoke, instead of just a thin jet with the same orientation. Because it's spinning, you now get a more even spread of the smoke. I think that's an unintended and beneficial consequence of the tapered bore."

Airflow in Peterson System Pipes: A Conversation with Dr. Dwain Dunn

Illustration of System airflow, courtesy of Dr. Dwain Dunn

Even with all of Dwain's measurements and conclusions, he emphasizes that personal experimentation is still necessary. "Find what works. If you don't like a tobacco, put it in a different type of pipe. Put it in a different-shaped bowl. Put it in a different-sized bowl. This is just another level of experimentation." His experimentations have made Dwain a convert, though. "I'm probably going to start loading up on System pipes from now on. It was only three years ago that I got my first Peterson pipe. Before that, all of them were relatively straightforward pipes. Nothing fancy. And I thought I had everything down pat. Then I smoked a System pipe three or four times and I thought, 'No, this can't be the ultimate thinking man's pipe. Something's not right here.'"

He went back to the Peterson literature. "They've a document titled 'How To Smoke A Peterson Pipe.' I read through that and realized I was smoking it wrong. As soon as I swapped over to Peterson's recommendations, I realized, 'Wow, this is a completely different animal."

Since that epiphany, Dwain gravitates to his System pipe 90 percent of the time. Especially important is the advice to tuck one's tongue under the indentation of the P-Lip button, which promotes smoke distribution. "The shape of your tongue during the draw also makes quite a difference, depending on where on your tongue it's hitting. So by changing where the stem is in your mouth, it changes the impact zone on your tongue. And then if you flex your tongue, it'll change that altogether again. It's reflected in the simulations." He started with a very simplified, flat-tongue simulation, mainly because the amount of data associated with a three-dimensional model slowed his computer simulations. "Just by including mouth geometry, it takes three minutes to load the file, as opposed to 30 seconds without it, because it's so detailed. That mouth geometry was something that I found on a dental website where they used it to make samples for dentures, braces, and those sorts of things. So it's very, very detailed, and as a result, it just takes a really long time to do anything with it."

However, when he ran the simulation, it became evident that mouth geometry is a crucial factor. "That's especially true with System pipes and the P-Lip, because it changes the way that you experience the smoke. So now all the simulations take quite a while because of that mouth. But it's interesting, so I don't mind." became evident that mouth geometry is a crucial factor

What's particularly intriguing is that Dwain started as a skeptic regarding the claims for the superior smoking characteristics of Peterson System pipes. They didn't smoke right for him until he accepted Peterson's advice on smoking technique. "Ultimately, the System works. It may not work exactly as Charles envisaged, but it does definitely work. The way that the System is designed is similar to what you see in a lot of vacuum cleaners nowadays. You'll have the part that comes into a reservoir, and there's a part that points down into a trap, and then a part sitting a little bit further up to draw the air out again. As the air comes out of the vacuum cleaner, any particulates and moisture sits in the bottom of the trap, and you have clean air that comes out the other side. So the Peterson design is very common in the engineering world, as a particulate filter, specifically." How Charles Peterson managed that design is a mystery, but he was ahead of his time.

Modern calculations have proven the Peterson System to be remarkably efficient. "Throughout the whole System pipe," says Dwain, "all the way up to the very edge of the P-Lip, before it enters the mouth, the turbulence level is what wind tunnels would classify as smooth laminar flow. No turbulence. So if you're getting turbulence, you're not sipping. You're definitely not sipping. If you smoke normally, you shouldn't have turbulence. Now, that's not to say you're not going to have those strange voices because you can still get a flow disturbance, but it's not turbulence. Turbulence is a different animal."

Dwain Dunn's work has illuminated how the Peterson System works and demonstrated what may be accomplished by combining one's avocation with recreation. We've known that Systems smoke differently. Now we better understand why.


Category:   Pipe Line
Tagged in:   Interview Peterson Pipe Making Pipe Science


    • Rick on September 23, 2022
    • Have you ever heard of a blend, Norris Gold? I use to smoke it years ago and I still can smell that aroma at times when I'm remembering how enjoyable it was.

    • Joe on September 24, 2022
    • This was a really great article and while I'm not a System smoker it was fascinating to read through. Has made me rethink giving a System pipe a shot.

    • Doc Brown on September 25, 2022
    • Great Scott! What an amazing article, and the illustrations are pretty cool too. I have a few system pipes that I need to dig out and show some attention to; reading this revived my interest. We've got to get back, back to the art and science of pipe smoking. A very intriguing and entertaining article, thank you.

    • Mark Irwin on September 25, 2022
    • It’s fascinating and gratifying and almost mysterious to see Dr. Dunn’s science validate Charles Peterson’s design. I’ve often wondered what a qualified scientist would make of the System and the claims made for it, and having one based on Dublin do so just makes the whole thing more amazing to me. Thinking man, indeed!

    • Mark Irwin on September 25, 2022
    • BTW—the Peterson book is still available at SPC—the link is just to the first edition.

    • Joseph Kirkland on September 25, 2022
    • Chuck, you always exceed expectations! You’ve done it again. Translating such difficult concepts for a lay reader is absolutely amazing!!

    • Sam P on September 25, 2022
    • As a confessed pipe gear head I found the article enlightening and hugely satisfying. Thanks Chuck,well done Dr. Dunn.

    • Rick Newcombe on September 25, 2022
    • This was so interesting. Dr. Dwain Dunn is confirming scientifically many of the conclusions I have reached through 45 years of trials and error as a pipe smoker. The only point I questioned is his assertion that the bend of the mouthpiece is key to achieving the maximum airflow. I have pipes that were made many decades ago that are straight and have an orifice mouthpiece just like a Peterson. I have opened the diameter inside the shank to 4.5 mm and the airflow is ideal -- starting with a very wide opening that is balanced by a restricted airflow before the smoke enters the mouth. One of those straight pipes is a 1920s Peterson, and it is a superb smoker. Great article!

    • Chris Streeper on September 25, 2022
    • What a fascinating exploration of the system pipe! I first became intrigued by fluid dynamics when learning about carburetors and ship board/ truck based fire fighting systems. I've always wondered what an engineer would make of the Peterson System Pipe. Thanks to all who contributed to this article.

    • Laurentiu M on September 25, 2022
    • Great article! Fluid dynamics chewed me up and spit me out, but I still like the topic and love visualizations. These are great, especially the animated gif. Getting an expert's well founded opinion is wonderful. As a previous poster put it, we need more on the science of pipe smoking.

    • Ken on September 25, 2022
    • I was totally unaware of the screw in chimney! This article provides encouragement for me to give it a try. I thought "DeLux" meant fancier materials. I still do not car for the shape of the PLip stem; it's too uncomfortable to hold, regardless of where it sits on my tongue. I usually go for fishtails on a Peterson and a saddle bit on any other pipe I purchase. Perhaps a PhD can engineer a bit with fishtail shape, but PLip advantages!

    • Rick Newcombe on September 25, 2022
    • After commenting on this article, I took photos of the "1920s Peterson" that I made reference to in my comment above. I sent them to Peterson expert Mark Irwin to confirm that the pipe really was made 100 years ago. He said no, it was made 70 years ago, in the 1950s. No matter -- the point about a wide open shank with a P-lip opening in the stem really does improve the smoking quality, regardless of whether it is a bent or straight pipe.

    • Astrocomical on September 25, 2022
    • Interesting. I think MORE research at pipe factories should be done like this and I don't think it has ever been done before.

    • Dwain Dunn on September 27, 2022
    • Thanks for the kind words everyone. It was a fun bit of research. There is more that I covered, but I think Chuck sorted the wheat from the chaff quite nicely.Rick to comment on your comment, I was merely indicating that the Peterson System would work better with a bent system pipe as opposed to a straight system pipe (the bend induces a larger vortex that spreads further in your mouth). Again though this would most likely depend on how you smoke it (cadence, mouth shape and p-lip location and orientation). Ultimately what is really important is that you have figured out what works for you and you enjoy it. It is possible the changes you made would not work for the next person.

    • JACK SCHEIBLE on October 9, 2022
    • Very interesting article. Are you using PIC models or SPH? I think if you used SPH you could change the particle kernels used for the tobacco itself as it burns.(Used to do CFD long ago. Switched to Computational Astrophysics. Used to read and write in Tengwar, too!)

Join the conversation:

This will not be shared with anyone

challenge image
Enter the circled word below: