David W. Macpherson is a PIA Life member, now 83. We talk to him about his long career at Harland, later United Pumps Australia.
I was one of those lucky people born at the start of a decade, 1930, making your age at any historic event easily calculated.
I was brought up in Sydney, with a life of surf, sand and sailing. My father was an Engineer, and he was NSW agent for The Harland Engineering Co of Alloa, Scotland. I was brought up with tools, and he was instrumental in making a workshop, where he encouraged me to hand saw the air-raid shelter 6” x 3” and 4” x 4” ironbark beams, into workshop benches. They are in my workshop here.
I completed a degree in Mechanical and Electrical Engineering at Sydney University, at a time when many of the students were CRTS servicemen after their wartime service, in fact on the same course was my brother, 5 years my senior. I remember in my final year, trying to decide whether to do a thesis on the way-out subject of holes in semiconductors, which in the future was called the transistor; or on the new CSIR development, a computer, plagued with valve failures; or on the latest power amplifier, saturable reactors. I chose the latter, but the developments in the former two makes one wonder! In my first experience of a purchasing failure, the promised Mu-metal cores did not arrive, so I thought of a device that would today be called a spectrum analyser, but with a very low frequency spectrum, made from wartime predictor parts.
In 1950 I gained a scholarship to join Metropolitan-Vickers, Manchester, one of the 40,000 employees at that site, to do a 2 year college apprenticeship, which is the way English graduates receive industrial experience. The ABC program “Foyles War”, shows the scene in England at that time. However in Manchester, not only could I not understand them, but they could not understand me!
How did you get involved in pumps?
After 5 years at M-V, the last few years as application engineer, I joined The Harland Engineering Co, in Alloa, Scotland in 1957, because of family associations, and so started my lifelong experience in centrifugal pump design and manufacture. Harland made large pumps, boiler feed pumps and water turbines for hydroelectric generation and DC motors for variable speed drives (Harland Drives). I also had a short time with Harland Drives, London, who was a subsidiary making control gear for sectional DC electric variable speed drives for paper-making machines.
On return to Australia in 1958, I commenced as contract manager in Sydney, tendering for erection and commissioning of pumping installations using Harland Australia engineered pumps for the Public Works department and consultants. My largest job was for a remote automatic pumping installation with an air tank water hammer suppression system. By reason of many graphical calculations, the only way to do such calculations at the date, I developed a system of critical damping of the water hammer surges, by drilling a hole in the non-return valve flap leading to the air vessels. Unfortunately, I was not present when the installation was completed, and was not informed of the late closure slamming of this NRV, which could have been prevented by limiting its opening range. This installation was unusual I that, to solve problems of a long transmission line, I used synchronous induction motors, where starting current was reduced by rotor resistance starting and voltage drop was reduced by using power factor correction.
This period was appreciated because I got to outback NSW in the course of commissioning water and sewage pumping stations for the PWD contracts. This involved testing the installation to see if pump performance complied with the Contract. One memory and mistake stands out. I volunteered to do the flow measurement. To measure the pumping rate, a measuring tape was nailed to a piece of wood as a float, and the tape held in the finger above the wet well as in fishing, so that the level in the well could be measured against the edge of the wet well. It took a long time. I remember being first offered this duty – it was a sewage pumping station.
During that period I also acted as commissioning engineer for Harland on paper making machines. In those days the variable speed drive consisted of a long motor – DC generator set, a generator for each of some 8 motor sections of the paper machine, kept in increasing speed relationships by a long beam of differential driven carbon pile field regulators. These were installed at APM at Peetrie Qld, Botany & Shoalhaven in NSW and Maryvale in Victoria, amongst others. In Box Hill, Melbourne in 1963, I commissioned the first solid state variable speed drive for a paper making machine, a MG (machine glaze) tissue machine. This was the saturable reactor of student days grown up into a PMA magnetic amplifier AC/DC drive.
In 1964, I left my beloved Sydney, on invitation (i.e. compulsion) to join Harland Engineering Australia, Melbourne, a manufacturer of mass produced end suction pumps and engineered pumps. Here I was responsible for pump development and testing.
I also was on a team commissioning a large board (for cardboard boxes) machine at APM Fairfield. I always remember the operators using children’s scooters, handle streamers flowing, to get from one end of the very long machine to the other end. This commissioning was not without incident, because, although no damage was caused, the flexible drive connecting motors to the machine had transient oscillation during acceleration. This was caused by the machine manufacturer using the usual spring flexible couplings, which caused torsional oscillations excited by a high gain speed control, a recently discovered phenomenon for these machines. However this was easily solved without costly changes. Harland PMA AC/DC drives did not survive the solid state AC power transistor (IGBT) motor drives, dating from 1983.
A pump manufacturer’s test facility is essential to its integrity and quality control, so there has always been a large dedicated area referred to as the ‘test bay’. Difficulties in the test bay can really run away with money and time.
The comments which follow will be of interest for companies who have to test pumps, particularly as a large water storage tank is not necessary, except for testing wet pit vertical pumps or as a heat sink.
Unfortunately, the Harland Test bay had flow measurement by weirs, a system recommended by their Alloa ancestors, but which was effective in entraining air in the recirculated flow. In addition, Harland was now manufacturing under license United API610 pumps which required every pump to be tested, frequently with NPSH testing. This had to be done on a suction lift, with the pump above water level. Not only was this rather hazardous, but acceptable NPSH results were difficult to achieve.
I started experimenting with closed loop testing to remove these 2 difficulties, initially without a storage vessel. This was hard to control, so I obtained a horizontal LPG vessel, and arranged a coaxial inlet and outlet through the manhole cover at one end. This was extremely successful, and was used until the Company became United Pumps Australia and moved to their Sunshine factory in 1985. One of the important requirements, is to use a small diameter horizontal vessel as this determines the minimum NPSH available, and run it nearly full, otherwise the change from positive pressure to vacuum becomes too long for each test point. For large flows or large powers, an auxiliary closed loop would be built, part of the flow being bypassed to the permanent closed loop for de-aeration and cooling. Many hydraulic turbines were tested in the auxiliary loop, the pressure pump discharging into the power recovery turbine, with its driver motor loaded as an induction generator.
I remember Bill Aitken as a brilliant pump designer, and have often thought how the difficulty of testing his large pumps for performance, would have been avoided with this closed loop testing. Unfortunately, he had left before closed loop testing started. I have a saying that a manufacturer’s Pumps are only as good as its test bay; or rather in this case, a better test bay would have prevented delays.
When I came to Harland Australia, I gave the test bay operator a 24” slide rule, which in spite of the length, was easier to use than a shorter one. In order to reduce the time taken on repetitive Test Bay calculations at every test point, I bought every new Hewlett Packard programmable calculator as it became available. To make this practical with dirty fingers, the program sequential keying guide and later the magnetic program card were enclosed in a thick transparent envelope. I made the test bay operator’s day when I showed him the newspaper article saying the HP 65 programmable calculator he was using, was the standby computer for the NASA moon shot!
The advent of the PC
In 1975, we bought our first personal computer, which was a Tektronix 4051 with a refresh graphics screen and 39Kb of memory, an unreliable tape drive, and no hard disk. This computer came with a comprehensive math and graphics library, written in Tektronics Basic, rather like HB Basic. I then started to teach myself computer programming, mostly in private time, and produced a program to reduce repetitive hand calculation in performing centrifugal pump calculations and graphically draw curves on a flatbed plotter using cubic splines. Through various changes of language and PC computing power, I have continuously developed this program for engineering and sales purposes into a large program called UPA Affinity.
Today, United Pumps Australia uses a test bay program I wrote in VBA for Excel, to record hydraulic, vibration and sound test data, which can be read into the Affinity program. Pump performance curves of tests, API 610 format test curves, or complete Iso-efficiency “Standard Curves” can be produced automatically from test data at the touch of a few keys using this program.
Pump hydraulic and mechanical design
With regard to hydraulic design, all companies have their own methods, because hydraulic design can only be based on parameters extracted from test results. Having been close to test results for a long time, I always start with the power/ flow rate curve for a basic understanding of pump characteristics, and the first point of investigation for the reason for an underperforming pump.
In this business of fixing pumps on test, one feels like a medical physician, you only know what the patient tells you, and you cannot see inside!
The effect of changes of volute with the same impeller is dramatic for small pumps, and should be part of Sales training when quoting pumps:
• The impeller dimensions determine the power curve.
• The volute casing sets the flow rate and value of best efficiency and the shape of the HQ curve.
• The curve flattens as the minimum volute area increases.
• The power taken remains the same at the same flow rate.
As most many pumps are cast from patterns, the use of throat area gauges should be part of quality control, particularity for multistage pumps. Previous test records are also a most valuable resource because future impeller diameters should be calculated from previous tests, so they are devalued by bad records or just taking one point of a change. The process of fixing a pump on test starts with an examination of the power curve compared with previous tests, so the integrity of test results is doubly important. You can tell if someone has turned the wrong impeller diameter, or used the wrong pattern or the volute throat is too small, or inaccessible passages are partly blocked, all from the power curve.
It is also interesting to note that some advances are made by accident or by forced adoption of a design feature of other fluid handling industries.
As part of hydraulic design, it is important that new Standard curves are drawn from the first test results, because I am of the opinion that sales data for some licensed products were based on actual throat areas different from the licensor’s drawings. Some NPSH curves are wrong, and this is not discovered because small pumps are rarely tested by the licensor, at least in USA. The same applies to vertical turbine pumps which are very difficult to reproduce from drawings due to the methods of drawing bowls. The throat area is impossible to determine from a drawing and the effective throat is critically dependent on patternmaking practice.
With regard to mechanical design, the movement into the manufacture of API 610 heavy duty pumps completely changed our understanding of pump design. No longer was the pressure casing design just ‘designed’ to pass pressure test as in a general purpose pump, but stresses, bolting and construction had to comply with the ASME pressure vessel code. Excel programs were written for the computationally intensive and rarely calculated Section 8 of this code for full face flanges, the closure joint of the pressure casing. The difference for high pressures is substantial, and hydro test failures were stopped. Maximum allowable shaft deflections were specified, as was the deflection of pump and baseplate under specified pipework forces and moments. For certain installations, proprietary Rotodynamic analysis for lateral or torsional critical speeds and calculation of damping at running speed was required, which led to a better understanding of the dynamics of pumps, particularly multistage and vertical turbine.
With the exception of lateral Rotodynamic analysis, I wrote extensive Excel spreadsheets instead of custom programs because of advantages of spreadsheets. I mention this because Excel and similar spreadsheets with their add-ins (VBA for Excel) are easily understood and can be set up by anyone, including those involved in sales, and written to meet the user’s needs. The spreadsheet is the way to take advantage of computers in your organisation, even to track your customer’s preferences if it is not already part of your manufacturing software.
The growth of the computer industry from an expensive mainframe where you submitted a stack of punched cards for processing overnight, to one where there is a networked computer on every desk available to save time, is as great a change as in telecommunications. It has gone into all phases of design, manufacture and company organisation, and totally affected the way an engineer works. Where do you see drawing boards now?
API610 pump testing also extended the manufacturer’s responsibility to acceptance testing for guaranteed vibration and sound pressure limits, which had to be achieved in the artificial environment of testing. Why artificial? Because on test the majority of the driver power is dissipated in the test bay by pressure drop across the discharge throttling valve, causing cavitation, noise and vibration whereas in use the power is dissipated over the whole site. For high powers, the difference is huge. After many attempts to make quiet and vibration-free throttling valves, including using inflexible hydraulic turbines, I found the most expensive Drag valve was successful, because in this valve, cavitation is limited by having multiple parallel flow paths of multiple pressure drops in series. This reduces the flow velocity in each path and so reduces cavitation. This type of valve is required for high power testing.
For vibration investigations, a spectrum analyser, the modern version of my student day’s project, is essential. The resonant frequencies of bearing housings or the test piping can be determined with this analyser by a tap of an instrumented hammer, which is a special hammer fitted with an accelerometer. A specialised spectrum analyser instrument is preferable, but once again, software is available for the PC.
Harland Engineering Australia was sold to Industrial Engineering in 1973, becoming Indeng Pump Division, but some years after the whole large group fell to a takeover. Selected staff was formed into United Pumps Australia at a new location in Sunshine, in order to manufacture spares for the large population of United pumps in Australia. From that revival, UPA has progressed, concentrating on API 610 pumps. The pump test bay luckily remained unsold and was moved when adjacent land at Sunshine became available.
Today, UPA has a test bay with a low level closed loop able to test to below 0.5 m NPSHa, and a high level closed loop for top suction pumps. It has a wet pit for vertical pumps and cooling. It has a considerable direct on line starting capacity at voltages up to 6.6 kV with provision for 11 kV, with earth leakage protection and good electrical safety and is serviced by 20T and 5T overhead cranes and adequate headroom.
What do you like about the pump industry?
The above shows some of the unknowns in manufacturing pumps, and the time required gaining experience. The challenge and lack of knowledge is why I have stayed in engineering in this industry and not moved on. Rather like the physician, you cannot be dogmatic that you are correct. It has to be proved by test.
One important challenge is that of the basic bid process – to get the job, the company must bid cheaper than other equally qualified company. This is the disappointment in being in manufacturing engineering.
The pump installation in NSW above is the one case I have had where the most expensive bid won.
What is your most memorable moment?
When testing a 10 stage 900kW pump with skeleton staff on a Saturday, throttling valve noise was in excess of 103 dB. Suddenly the noise increased enormously. Everyone dived for the emergency stop pushbutton. The valve trim had eroded through!
The biggest error was on a 48” split case pump, in being persuaded go to full hydrostatic test pressure when it was already leaking. The resulting casing failure washed the underside of the factory roof and some people. This was a most expensive back – to – the – drawing board event. The sequel during commissioning I will leave to another to tell.
Did you have a mentor or other memorable people you worked with?
No, I did not have a mentor as such; I was self-taught, especially as manufacturers you may visit are consistently somehow unable to explain the process of fixing pumps on test. .
However, I thank all those managers who enabled me to stay in employment in troubled times, including David McLeish.
The companies I have worked with seem to have populated a large part of the industry, and are all fondly remembered.
What are the most significant developments you have experienced in the industry?
The most significant change is undoubtedly the decline of Australian manufacturing. This has been due to removal of tariffs and floating of the dollar, which are all good economic changes.
The decline in API610 pump manufacturing is not just cost, but the fact that major projects are designed and erected by Overseas engineer-contractors, who in the case of USA, specify their own manufacturers and do purchasing in the package, so everything is sourced outside Australia. They have even eliminated land-based structures which cannot be imported, by the development of ship-based production facilities. Usually, all that is left of a project is the hope of locally based pump repair, which rarely occurs.
What are the greatest challenges facing the industry?
Not being currently involved in sales, I leave that to others.
However, I have some thoughts to contribute on the role of the Association. In the days of APMA, I used to go to meetings. I envy other professions such as medical research where cooperation between professionals is encouraged. I found in this profession a spirit of intense competition, and a them – and – us attitude, so much so that a conversation with your opposite competitor, would dissolve into ‘anything you do we can do better’.
Now that manufacturing pumps here is about dead, the current Association has to think of ways to service cooperation between competing sales people.
What advice would you give young people in the pump industry?
Protect your hearing! UPA monitored employees for hearing and we were aware of my hearing loss on one side. Without going into details, I am now severely deaf, having gone through 3 sets of most expensive top of the range hearing aids.
You can just as easily get hearing loss from an iPod, so test yourself by reversing headphones with one side disabled, before it is too late. If you do have hearing loss, go to an audiologist early. He will probably recommend hearing aids for both ears to prevent the deafer ear becoming lazy.
Seriously, this industry services a basic, universally required need and you will find that application engineering in pumps is fascinating. Look at the irrigation schemes, the latest one on the news was in the middle of an iron ore mine area.
Are you still involved in the industry today?
I have continued rewriting Affinity from Tbasic to Visual Basic 6 language, to have a full Windows graphical interface, which is the program United Pumps use.
By reason of having written a program whose source code others would find difficulty in mastering, I am answering questions and doing maintenance on this program. Although there are indications that current staff is using it less than I used to, this program is essential for sales bids. I am also translating it into a program language that Microsoft has not planned to eliminate, although that is uncertain with the rise of HTML. It is a job which I enjoy, and keeps me mentally active.
David now spends his time with his grandchildren, keeping up with his son’s inventions, and computer programming. ■
An original ad for the Tektronix 4051 which Harland Australia bought in 1975.
David (left) with Neal Mclean-Brown from Shell in front of the 1.6 MW Newport Pipeline Pump in Geelong in 1986.
Sydney University BE. Mech & Elec, BSc, Fellowship Diploma of Management R.M.I.T., M.I.E.Aust, Chartered
Professional Engineer (retired)
Contract Manager, McKinlay Fletcher P/L, Sydney
Development Engineer, Harland
Technical Engineer, Indeng Pump Division
1985 – 2005
Engineering Manager, United Pumps Australia,
Served on APMA technical committee
Served as Chairman, SAA Committee ME/30, pumps