I have used one setup in the past but cannot remember the details. Anyhow, I'll see what I can remember...
The viscometer is basically a tank and a long, vertical capillary tube which is connected to the bottom of the tank. Due to the small diameter of the capilliary tube, the substance will flow in laminar flow. Entrance losses can be ignored as they are small compared to the tube friction losses.
By measuring the flow rate out of the capilliary tube (I used a measuring cylinder and stopwatch) at a known substance level in the tank, the viscosity can be calculated from:
hf = (32*viscosity*L*V)/(density*g*d^2)
hf = head loss due to friction
L = tube length
V = bulk velocity
d = pipe diameter
The temperature can also be taken into account. I'm afraid this is all I can remember. I remember using water, so I'm not so sure it would work for high viscosity substances.
If you do a search for 'viscosity measurement' online, you should find lots of stuff. The old viscosity equipment was basically as eayver has used, a container with a fixed opening. The opening was varied depending on the viscosity of the material and for many pieces of equipment, you simply timed how long it took to run a fixed volume through the hole (hence the use of 'seconds' in many viscosity units).
I've tried to determine it with a cylinder, stopwatch and a sphere. I used it on water but because of the low viscoity of water the stopwatch readings weren't that accurate. I've compared it against proper results and my answers were not even close
The formula that I've used were
µ=gD²/18V*(ps-p)
where:
D=sphere diameter
V=velocity (fall distance/time measured)
ps=sphere density
p=liquid density
For quick off line measurements (assuming non hazardous fluids) there are a variety of different "cups" you can use. These are fast and reasonably accurate. However, you may find it difficult to ,match your results with other viscometers if the fluids are non-newtonian. Falling ball viscometers are commonly used for the measurement of viscosity e.g. Hoffler. They are often used, for example, for heavy fuel oil measurement. The original Stokes law experiment was designed to measure the terminal velocity of the ball but most falling ball viscometers just measure the transit time and are fairly compact. Don't neglect temperature. It can have a profound effect. Falling ball viscometers from Kittiwake, for example, have a temperature control system.
What do you mean by quick? and why is tyhe speed critical? Speed can mean a sacrifice of accuracy. A good falling ball viscometer gives its best results when the readings are taken consecutievly until they agree. In polimerisation reactions end point spotting is critical and some reactions have only a 20second end point window. In these cases process measurement is far better than off line.
Falling ball, capillary, rotational technologies are very familiar as laboratory techniches and many have some success as process measurements. However there are many new technologies which are suitable for the process. Speed and accuracy are not two terms i would associate within the laboratory as it is essential to obtain optimum conditions for the measurement. In process they can go together (in fact some process viscometer systems will give instantaneous measurements of viscosity that are as good or better than a laborious laboratory measurement. (catch 22: they are often far more difficult to use in the laboratory)
If you prefer to provide more information on your needs i can offer a better response.
If you've the budget, you could use a good coriolis meter (to get volumetric flow and generate a pressure drop) and use a DP transmitter to measure the flow accross it.
The viscosity is directly proportional to the ratio of volumetric flow and pressure drop.
Ive seen some Emerson (MicroMotion) coriolis meters (elite series) used to measure viscosity also - their transmiters have a 4-20mA pressure input. These were fiscal transfer meters from the petrochem industry.
Or, try the Endress + Hauser website - I think their Promass 83 range has viscosity measurement built in.
Heres a good article on it..