Vacuum Facts
Active member
The science describing how all vacuum cleaners clean flooring, regardless of who designed and manufactured them, has been known for over 250 years and is (relatively) easy to understand. As a scientist with knowledge in this area, I’ve decided to produce this summary to complement the full video lecture already available online to help people who genuinely want to understand. The commonly seen measurements of airflow and suction, particularly at an open hose or equivalent, do not directly inform cleaning capability on carpet. The reason for this requires an understanding of this known science, and a brief and hopefully accessible explanation is given below. I appreciate this post might not be for everyone.
The science:
To accelerate a particle out of a carpet, an aerodynamic force needs to be imparted by a moving fluid (air in this case). This aerodynamic drag force (equation 1, figure 1) scales with the square of the airspeed and therefore dominates all other terms. Airspeed is the single most directly influential aerodynamic parameter in determining the applied force.
There are three other factors that affect either the airspeed or the drag force:
1. Cleaner head
2. Dirt properties
3. Carpet properties
The cleaner head is particularly important (point 1. above). It’s designed to be fully sealed to the floor, as this makes use of a well-understood physical phenomenon first described by Daniel Bernoulli, with a special case described by Giovanni Venturi. The edges of the cleaner head create a constriction in bounded fluid flow that results in a substantial increase in the speed of the air on entry—airspeed being the most important parameter, as above in equation 1. Equation 2 in figure 1 quantifies the Venturi effect and shows that the pressure difference inside and outside the cleaner head (i.e. suction) drives the airspeed increase. Equation 3 in figure 1 combines the earlier two equations into a single expression to quantify the acceleration of a dirt particle from a carpet. It is NOT a function of the total magnitude of airflow, but of suction pressure within the head and some other terms that relate to the dirt particles themselves. This relates to point 2, above: at any given suction pressure, particles with a greater drag coefficient—like fluff, or that are larger—like hair relative to microscopic dust, are easier to remove, whereas heavier particles, like sand, rice, and grit, are tougher to remove. The other influence of the cleaner head is the brush bar. This directly grabs loose surface dirt but also can locally separate pile to increase airspeed at greater depth to better accelerate particles out. It also agitates—which doesn’t mean makes particles visually bounce outside the head, which is literally irrelevant since they’re not going anywhere—but repeatedly jostles fibres under the head where the aerodynamic drag force is applied and freeing any that become netted by the fibres to be accelerated in the flow.
Carpet properties (point 3 above) also affect airspeed as a function of pile depth and is a bit more complicated to consider. Most modern fitted carpets have liquid-impermeable, non-porous backing, which reduces the airspeed to zero as you approach the base (see figure 2), preventing particles from being easily removed at the base on deeper pile carpets. Particles rarely naturally get to that depth on such flooring since deeper carpet is actually a filter that physically blocks them; almost all dirt remains in the upper regions with basic housekeeping (plenty of evidence of that here). Loose shaggy rugs often have porous, permeable backing allowing relatively higher airspeed at the base, but typically relatively lower airspeed towards the surface, for a given cleaner and air power—see figure 2, which is illustrative.
To understand this and a number of other points I’ll come onto, you need to understand vacuum dynamics (or basic electrical circuit theory) to help appreciate the physical relevance of suction, airspeed, and air power, and the relationship between them. Suction, airflow, and air power are all functions of the air resistance in the air circuit and the maximum suction pressure the motor can produce, as derived in equations 4–6 in figure 3, where their normalised relationships are also plotted.
There are two important things to understand from this:
1. The type of carpet affects the air resistance, R. Loose shaggy rugs reduce it whereas fitted carpets with impermeable backing increase it, and the plot shows how suction, airflow, and air power respond accordingly.
2. Motors that can provide a greater air power (under resistive load) generate a greater maximum suction pressure (equation 6). This allows suction to be sustained in the presence of higher leakage airflow into the cleaner head on flooring that offers reduced air resistance (equation 4), thereby maintaining airspeed and dirt removal rate.
To expand on these two points, the equations and plot illustrate why a larger airflow into a cleaner head, such as occurs on loose shaggy rugs or with a poor suction seal, is bad. It neutralises suction pressure and cleaning performance and requires more air watts and greater maximum suction pressure to compensate. The total magnitude of airflow increases, but its speed decreases. Furthermore, the equations show that as the air resistance of flooring increases, you need a higher-pressure motor to compensate and sustain suction pressure (equation 4) and maintain airspeed. What this means is that cleaners with low suction motors (e.g. the old-fashioned ‘dirty fan’ machines) don’t have a problem on carpets with permeable backing since there’s little resistance for them to overcome. But as air resistance increases, they can’t draw as strong a vacuum under the cleaner head to maintain airspeed, affecting cleaning performance. There is test data to demonstrate this in this video. There are lots of tricks of the trade to cover up this shortcoming, and they’re discussed in the lecture.
When you understand this science, it’s natural to cringe when you see measurements of suction and airflow from an open hose (or equivalent)—which isn’t representative of a cleaner head that’s well sealed to carpet. What should be measured instead is actual dirt extraction rate under representative, real-world conditions in a way that shows understanding of statistics. Dirt removal from carpets is stochastic, since it’s a first-order system, and any measurements need to capture this and look at removal rates from a carefully controlled initial condition, following a precise real-world representative and reproducible methodology. An example of the data you get by doing this that allows fair relative performance comparisons, and without the aid of a laboratory and robots, can be found in figure 4 and any of my more recent reviews.
Hopefully this clarifies the science and helps people to now spot common errors. The full treatment of the science can be found in this video lecture. I can also clarify any questions here.
The science:
To accelerate a particle out of a carpet, an aerodynamic force needs to be imparted by a moving fluid (air in this case). This aerodynamic drag force (equation 1, figure 1) scales with the square of the airspeed and therefore dominates all other terms. Airspeed is the single most directly influential aerodynamic parameter in determining the applied force.
There are three other factors that affect either the airspeed or the drag force:
1. Cleaner head
2. Dirt properties
3. Carpet properties
The cleaner head is particularly important (point 1. above). It’s designed to be fully sealed to the floor, as this makes use of a well-understood physical phenomenon first described by Daniel Bernoulli, with a special case described by Giovanni Venturi. The edges of the cleaner head create a constriction in bounded fluid flow that results in a substantial increase in the speed of the air on entry—airspeed being the most important parameter, as above in equation 1. Equation 2 in figure 1 quantifies the Venturi effect and shows that the pressure difference inside and outside the cleaner head (i.e. suction) drives the airspeed increase. Equation 3 in figure 1 combines the earlier two equations into a single expression to quantify the acceleration of a dirt particle from a carpet. It is NOT a function of the total magnitude of airflow, but of suction pressure within the head and some other terms that relate to the dirt particles themselves. This relates to point 2, above: at any given suction pressure, particles with a greater drag coefficient—like fluff, or that are larger—like hair relative to microscopic dust, are easier to remove, whereas heavier particles, like sand, rice, and grit, are tougher to remove. The other influence of the cleaner head is the brush bar. This directly grabs loose surface dirt but also can locally separate pile to increase airspeed at greater depth to better accelerate particles out. It also agitates—which doesn’t mean makes particles visually bounce outside the head, which is literally irrelevant since they’re not going anywhere—but repeatedly jostles fibres under the head where the aerodynamic drag force is applied and freeing any that become netted by the fibres to be accelerated in the flow.
Carpet properties (point 3 above) also affect airspeed as a function of pile depth and is a bit more complicated to consider. Most modern fitted carpets have liquid-impermeable, non-porous backing, which reduces the airspeed to zero as you approach the base (see figure 2), preventing particles from being easily removed at the base on deeper pile carpets. Particles rarely naturally get to that depth on such flooring since deeper carpet is actually a filter that physically blocks them; almost all dirt remains in the upper regions with basic housekeeping (plenty of evidence of that here). Loose shaggy rugs often have porous, permeable backing allowing relatively higher airspeed at the base, but typically relatively lower airspeed towards the surface, for a given cleaner and air power—see figure 2, which is illustrative.
To understand this and a number of other points I’ll come onto, you need to understand vacuum dynamics (or basic electrical circuit theory) to help appreciate the physical relevance of suction, airspeed, and air power, and the relationship between them. Suction, airflow, and air power are all functions of the air resistance in the air circuit and the maximum suction pressure the motor can produce, as derived in equations 4–6 in figure 3, where their normalised relationships are also plotted.
There are two important things to understand from this:
1. The type of carpet affects the air resistance, R. Loose shaggy rugs reduce it whereas fitted carpets with impermeable backing increase it, and the plot shows how suction, airflow, and air power respond accordingly.
2. Motors that can provide a greater air power (under resistive load) generate a greater maximum suction pressure (equation 6). This allows suction to be sustained in the presence of higher leakage airflow into the cleaner head on flooring that offers reduced air resistance (equation 4), thereby maintaining airspeed and dirt removal rate.
To expand on these two points, the equations and plot illustrate why a larger airflow into a cleaner head, such as occurs on loose shaggy rugs or with a poor suction seal, is bad. It neutralises suction pressure and cleaning performance and requires more air watts and greater maximum suction pressure to compensate. The total magnitude of airflow increases, but its speed decreases. Furthermore, the equations show that as the air resistance of flooring increases, you need a higher-pressure motor to compensate and sustain suction pressure (equation 4) and maintain airspeed. What this means is that cleaners with low suction motors (e.g. the old-fashioned ‘dirty fan’ machines) don’t have a problem on carpets with permeable backing since there’s little resistance for them to overcome. But as air resistance increases, they can’t draw as strong a vacuum under the cleaner head to maintain airspeed, affecting cleaning performance. There is test data to demonstrate this in this video. There are lots of tricks of the trade to cover up this shortcoming, and they’re discussed in the lecture.
When you understand this science, it’s natural to cringe when you see measurements of suction and airflow from an open hose (or equivalent)—which isn’t representative of a cleaner head that’s well sealed to carpet. What should be measured instead is actual dirt extraction rate under representative, real-world conditions in a way that shows understanding of statistics. Dirt removal from carpets is stochastic, since it’s a first-order system, and any measurements need to capture this and look at removal rates from a carefully controlled initial condition, following a precise real-world representative and reproducible methodology. An example of the data you get by doing this that allows fair relative performance comparisons, and without the aid of a laboratory and robots, can be found in figure 4 and any of my more recent reviews.
Hopefully this clarifies the science and helps people to now spot common errors. The full treatment of the science can be found in this video lecture. I can also clarify any questions here.
Last edited:
