Gas Dynamics and Carburetors

Contributed By: John Brewer

This will be in two parts. Part 1 will consider the operating theory and practice of a carburetor. Part 2 will talk about mixture requirements and air capacity of Internal Combustion engines.

The topic under discussion is very complicated and several issues are involved. It must at all times, during a discussion such as this, be remembered that an Internal Combustion Engine (ICE) is a highly non-linear mechanism and any control system designed to operate with it, must take this into account. In English, an engine behaves in different ways at different operating points (rpm, temperature, load, throttle opening, fuel composition, etc.) and all of it's systems must be designed with these factors in mind. Also, a venturi, the primary fuel metering component of a carburetor, is non-linear and thus operates differently for different air flow rates. With this in mind, lets forge ahead.

A discussion must first be given on carburetion to make sure we are talking about the same thing. Basically, a carburetor controls the power output of an engine by controlling the air flow into the engine, most spark ignition engines operate like this as opposed to diesel engines which control power by regulating fuel flow. When the throttle is closed (usually a butterfly valve) very little air flows into the engine. When the throttle is open, there is a much larger orifice and much more air can flow. Now, as air flows through the throttle, it flows through the venturi. A cross section of a venturi (as implemented in a carb) looks like this:

       ---------------------------------Carb Body----------------------------
       XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
       ----------------------------------------------------------------------

- ---> Air Flow        (outside venturi)
                           b-1---->
       ---->a /XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX/c---->
              /XXXXXXXXXXXXXXXXVenturiXXXXXXXXXXXXXXX/
                /XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX/
                  /XXXXXXXXXXXXXXXXXXXXXXX/
                     /XXXXXXOXXX/
                            b-2----------->
- ---> Air Flow
                              (inside venturi)
                     /XXXXXXOXXXX/
                  /XXXXXXXXXXXXXXXXXXXXXXX/
                /XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX/           
              /XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX/
              /XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX/

- ---> Air Flow        (outside venturi)

      ----------------------------------------------------------------------
      XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
      ---------------------------------------Carb Body----------------------

A Venturi is just a tube with a cross section of a varying width, and in fact, is really just a lifting airfoil rolled into a tube. The venturi operates on a principle called continuity which just means that the air is a continuum and there can't be any places where there is nothing. When air flows into the carb, there are two places it can go, outside the venturi, and inside the venturi. The air flowing outside the venturi flows along pretty much unaffected. Lets consider two streams of air, very close together which impinge upon the lip of the venturi at point 'a'. One stream goes outside the venturi, the other goes through the venturi. Now continuity states that, at point 'c' (at the end of the venturi) the two streams must come back together. (In a simplistic sense, molecules of air next to each other at point 'a', will be right back together at point 'c', having taken differnt paths.) Notice then, that air flowing through the venturi will have a little bend to go over- this is the key to both venturis AND airplanes. Since the air must join up at the end of the venturi (because of continuity), and the air flowing through the venturi has -farther- to go (because of the little hill) than the air outside the venturi, then the air flowing through the venturi must flow -faster-. (Air at 'b-2' is moving faster than air at 'b-1' even though the two streams were moving the same speed at point 'a'.) Since there is the same 'linear amount' of air spread over a greater distance inside the venturi, the molecules must be farther apart and the pressure is then lower. This low pressure inside the venturi (compared to atmospheric) pulls fuel out of little holes inside the venturi (marked as 'o' in the throat of the venturi, usually 4 in each.) This relationship between the speed of air flowing through the venturi, and the amount of fuel that gets pulled out of the little holes as a result of the low pressure generated by the air flow, determines the mixture that goes to the cylinders. Simple huh?

Think about what is important then and what the engineer has to consider to get the mixture correct (and why EFI is much better.) The size of the bore. The width and length of the venturi and the shape and height of the curve inside the venturi. The size of the holes inside the venturi and their number. The size of the tube feeding the holes and it's length. The size of the main metering jets which determine the maximum flow rate. The depth of the sump (bowl) inside the carb. Consider that the whole thing has to work across a huge range of temperatures of incoming air, fuel, humidity, engine temperature, angle of the vehicle, etc. Wow! Unfortunately it gets worse because the mixture demands of an engine are not directly related to the performance curve of a venturi. As more air flows through a venturi, more fuel gets pulled out. However, since all these things are made of metal and are rigid, once the thing is built and all the holes are drilled, the mixture is pretty much set across the rpm range. (We are not considering the adjustable idle mixture here since at idle, the venturis are not passing enough air to pull -any- fuel out of the bowl.) What you end up with is a device that is very poorly suited to the fuel demands of an ICE, far to lean at idle and at wide open throttle. That is why we have to hang a bunch of extra stuff on carburetors like acceleration pumps, economizer valves, power valves, and idle systems. Not to mention startup systems like a choke. Talk about a compromise!

One other factor that must be mentioned before the next discussion. The overall size of venturis and carburetor bores is very important. As venturis and bores become larger, the behaviour of the venturis, and the predictability of how they meter fuel, becomes more difficult to control. This is typically called the range of linearity. When you hear about linearity in this sense, it is a control system term that -basically- separates easily controllable and mathematically describable phenomenon (linear) from super complex, mathematically intractable, and often uncontrollable phenomenon (non linear). For example, the manner in which a rock falls to the ground is easily describable and mathematically predictable. It can be described by -linear- differential equations. Release a flat sheet of paper from over your head, and every time you do so, it will behave differently. It will always hit the gound, but predicting the path it will take is truly an impossibility. This is not a perfect anlogy but it will do for now. Interestingly, the behaviour of a falling rock becomes increasingly difficult to predict as the speed is increased. Go fast enough, and you have to invoke non-linear differential equations which must take drag, viscosity, and shape into account. (For rocks falling we are looking now at speeds approaching that of sound.) Basically, go fast enough and you exceed the range of linearity. Stay slow and you can neglect the effects of these forces even though they are always present- if very small.

Engineers like linear. Linear is easily describable. Linear is controllable. Pretty much most of designing a control system is determining the range of linearity, matching a control system to it, and then staying within the range of linearity. Carburetors are no different. Keep them linear and they will function much better. Small venturis, small bores, and small orifices to flow fuel are much easier to deal with than their large counterparts. So, more venturis and more bores of small sizes can handle more fuel more efficiently than a single large barrell. Thats not the only reason but it is about half of it. More later.

Next- Mixture requirements and air capacity!

John Brewer

"Any man who would trade liberty for security, deserves neither."
Benjamin Franklin