How do pumps work?

Syd Mitchell's picture
Submitted by Syd Mitchell on Tue, 01/16/2018 - 14:17

Thanks again to Syd Mitchell for the great article!

Almost every single pump that Koi keepers are ever likely to use to pump water will be what are known as squirrel cage centrifugal pumps. The term “squirrel cage” refers to the way the motor is constructed and is of no concern to Koi keepers. The term “centrifugal” means that water is drawn through the pump by centrifugal action.

In engineering terms, for very obvious reasons, the motor end of a pump is frequently called the dry end and the end that actually pumps the water is called the wet end. See figure 1 which shows how the two parts fit together and are plumbed into a test tank. It is the wet end of the pump that concerns hobbyists because this is the part that does the job that we expect of a pump – moving water from one place to another.

As shown in the simplified diagram in figure 2, the heart of the wet end is an impeller and the housing it fits into which is called a volute.  There are far more complex designs but figure 2 shows the principle of operation of the simplest design that would actually work.  The impeller is spun round clockwise at a high speed, typically 1,600 to 1,800 RPM, and the water in the volute is flung outward by the centrifugal force supplied by the impeller. The vanes are curved backwards which also helps to push water outwards.

The positioning of the impeller inside the volute is crucial; the gap between them increases in a clockwise direction. This is to gradually slow down the water flow from the very high speed that it has within the impeller, to a much slower speed as it leaves the pump.  Slowing down the rate of water flow may seem an odd thing for a pump designer to deliberately do but, by slowing the speed of the water before it leaves the pump the pressure of the water as it leaves is increased.  This is where pump designers make choices. One choice is for volute designs that produce high flow rates but at such a low pressure that any restriction to the flow dramatically reduces it. These pumps only need low power motors. Alternatively, they can have a flow rate that has sufficient pressure to maintain that flow even when pumping against back pressure.  These pumps need higher power motors. They cannot have low power pumps that also maintain the flow rate when lifting water through any significant height.

Stirring or lifting?
Water is easy to stir. You could think about this as a thought experiment or try it for real. The next time you run a bath, dip one hand into the water and use a circular motion to cause the water to spin.  Water moves fairly easily so it doesn’t take much energy to cause a large circular flow in the bath.  If that rate of flow was actually doing something useful instead of just going round and round in the bath every few seconds, you would have an impressive flow rate. The next part of this experiment is probably better to imagine rather than trying to do it for real.  Imagine using a bucket to scoop the water out of the bath and emptying it into the basin. Depending on depth, the water in a typical bath weighs somewhere in the region of 150 kg.  Scooping out that weight of water in buckets and lifting it through a few feet would take much more effort than stirring it round and round.

This is the important difference between a pump that has sufficient energy or power to move a large volume of water against the force of gravity and a pump that can circulate water quickly, but only as long as it doesn’t have to lift it through any significant height.

Testing flow rates
When manufacturers design new pumps and test the flow rates they can deliver, they test them under conditions that will maximise their output.  Figure 1 shows a typical testing set up with the pump suction port (input) directly connected by a short straight pipe to the test tank and with the pump delivery port (output) being returned to the tank by another short pipe with a single swept elbow. This arrangement minimises any losses due to friction in the pipe-work and produces the maximum flow rate that the pump could ever deliver under the most favourable conditions.

Even a low powered motor will be able to produce very high flow rates with such a simple pipe-work layout. It will be doing no more work than stirring a bath full of water.  The real question is whether or not a pump would be able to produce the same flow rate if it were required to lift that water by a few feet before allowing it to cascade back into the test tank. In reality, due to the weight of water, there will always be a reduction in performance when a pump has to lift water rather than just stir it around.  There will also be losses when a pump has to push water through long pipes, narrow pipes or pipe-work that has a large number of fittings.  These losses are due to the friction caused by the water rubbing against the internal walls of the pipe as it travels along it. There are also losses due to the force required to make water that is travelling in one direction suddenly change direction when it goes round a bend in the pipe.  In order to simplify these losses, they are all lumped together under the name “head loss” and their effect is to reduce the overall performance of a pump in that particular plumbing situation.

The performance of both high and low power pumps will be affected by head loss, but a higher power pump will be less affected than a low power pump since it, not only has the power to circulate water, but it has sufficient reserve power to lift water against the force of gravity or to overcome losses in the pipe-work.

Flow rates in practical situations
What does loss of flow rate mean in practice?  I haven’t covered flow rates in relation to biological filtration yet but, in an ideal situation, the turnover rate of a Koi pond should be at least once every two hours.  This means that, in a 1,000 gallon pond, the pump must be capable of pumping 500 gallons of water per hour through the filter system.  For a 2,000 gallon pond, it must pump 1,000 gallons per hour and so on. I once tested a pump, which according to the manufacturers, would pump 11,300 litres (2,480 gallons) of water per hour, but it only achieved 7,100 litres (1,560 gallons) per hour in a well designed filter system that had the minimum possible number of bends in the pipe-work. That represented a loss of over 1/3 of its rated performance.  What would be the effect of losing 1/3 of the pump’s performance?  Quite simply, if you were expecting a two hour turnover, the pump would only be delivering a 2 hour 40 minute turnover.

It would be unfair to single out any particular manufacturer and put their pumps under scrutiny, so without naming names, these are the actual flow rates that can be expected from a typical popular pump under various conditions. Pumping water straight out of a pond and back again, similar to the test set up in figure 1, it will deliver its rated performance of 12,000 litres per hour.  If it has to lift water by 1 metre, it will lose nearly 20% of its flow rate and only deliver 9,700 litres per hour.  If it had to lift water to 2 metres above pond water level, it would lose nearly 40% of its flow rate and only deliver 7,400 litres per hour.

Since pump flow rate losses don’t just involve those due to lifting water, but also include losses in pipe-work and bends, it is quite possible that a filter system that returns to a pond via a shower that is only 1.3 m above water level could also have a 40% lower flow rate than expected. This is especially true when the loss of flow due to the restriction imposed by the spay bar is taken into account.

This does not mean that low power pumps are inferior.  In the simplest of filtration systems, a pump that can pump, say 10,000 litres per hour, will perform virtually as well as a higher power pump with the same flow rating. The difference between the two pumps will become obvious when they are plumbed into more complex pipe-work layouts or when they are required to lift water. A low power 10,000 litre per hour pump may not deliver 10,000 litres under those conditions and it will be necessary to choose a higher rated pump in order to achieve 10,000 litres after accounting for the losses.

Too high a flow rate
Whilst it is rarely the case that a pump with a high flow rate could actually cause the pond turnover rate to be too high, there are cases when restricting the pump flow might be desirable. There are different reasons why this may be necessary but one example would be if a pump has been fitted with a flow rate that is too fast for the filter system.  This would mean that the rate at which the pump is drawing water from the filter would be faster than the filter can refill from the pond.  The result would be that the filter system would slowly empty and the pump would draw air and deprime. Pumps can only run for a couple of minutes without a proper flow of water through them before the internal seal that separates the wet end from the motor is damaged. If pumps and pipe-work have been correctly chosen, this situation should never occur but since, in the real world, accidents sometimes happen, rather than buy a new lower power pump, it may be preferable to restrict the flow rate of the one already fitted.

In this, or any other situation where it may be necessary to reduce the pump flow rate, simply fitting a flow restrictor such as a valve immediately after the pump is all that is needed.  There is a prevalent myth that restricting pump flow will cause it to “burn out”, this isn’t true.  If the restriction was so great that the flow was reduced to a trickle, the internal shaft seal would overheat and be damaged but, with the exception of extreme cases, the flow rate can be varied over a very wide range without causing damage.

The flow rate should never be reduced by placing a restriction on the input to the pump. This will indeed reduce the flow but risks causing an effect called “cavitation” which causes tiny bubbles to form around the impellor. The processes going on under these conditions are far too complicated to explain here so you will have to trust me on this one but, as these bubbles keep forming and collapsing, they become very abrasive and can cause irreparable damage to the impeller by literally wearing it away.

So far this series has covered the basic design and operation of all the equipment necessary to run a successful Koi pond. The pond design so far has been simple but effective but there are many variations. Before covering some of these and also more advanced filtration techniques it would be well to discuss water, water quality and how this affects the health and development of the Koi we keep.  Next month I will do just that and explain the importance of the phrase: “We don’t keep fish, we keep water”.

Or as I prefer to say: “We don’t keep fish, we keep dihydrogen monoxide” (The correct scientific name for water).

!If you're not havin' FUN, you're not doin' it right