Reminiscences of a life in control

February 2024 SAIMC

When I was in Matric, I was thinking of studying mechanical engineering, but an uncle who was a leading local entrepreneur said that I should rather take up electrical engineering as the future was for the people who ‘pushed the buttons’, and I decided to follow his advice.

I battled initially at Wits University, as everything was very theoretical and described in mathematical formulae. I remember being told in one of our first lectures that: “You think that you are here to learn Electrical Engineering, but you aren’t. You are going to be taught how to think like engineers.”.

This was very true, and our course in Control in final year was pure maths, and words like ‘controller’ and ‘PID’ were not even mentioned. I managed to get a first in control, but I didn’t really know what it was. That was in 1963.

Transistors were just coming into commercial use, and most of our studies were on equipment still using devices like thermionic valves and mercury-arc rectifiers.

I spent the first two years of my professional life undergoing a ‘Graduate Apprenticeship’ at GEC in England, and specialised in telemetering. On one job we had to send control data from a remote plant to a petrochemical refinery over radio links. This was when I ‘met’ real life control loops and the famous PID controllers for the first time, and I was fascinated by the technology.

A year after returning to South Africa, I got a job with AFH Devers who were the South African representatives for the-then leading and famous Foxboro company. My boss was Morris Hill who had been the Chief Control Engineer at AE&CI.; He gave me a book by a revered Foxboro training officer, Norman Anderson, if I recall his name correctly, which brilliantly explained how pneumatic instruments and controllers worked.

My job was to design systems for various plants. A very noteworthy one was the air conditioning control system for the enormous Carlton Centre, which was then in the process of construction. The control system was very advanced and had been designed by Italian Consultants. Electronic instrumentation was still in its infancy and the consultants had specified pneumatics. I had to engineer the system which was really fascinating as it also involved some feedforward controls, and we used pneumatic mathematical computing blocks for these. One really interesting thing was that the building was going to be the-then tallest building in Africa and consisted of 50 floors; with, I think, eight being below ground. The control centre was in the lowest level. A pneumatic temperature transmitter was mounted on the roof of the centre to measure ambient temperature, and it was calculated that it would take about 12 minutes for a step change in the process variable to reach its final value down on the panel.

The panel was amazing with many standalone pneumatic controllers, recorders, and other instruments on it, and was about 6 metres in length, filled inside with literally hundreds of aluminium pneumatic signal pipes. During the construction phase of the building, a special roadway had been made leading from the ground level in the Centre down to the lower levels, to allow large objects to be transported there. However, when they finally called for the panel to be delivered, the roadway had been filled in, and there was no access to get the panel down to the basement. The only solution was to cut the panel in half, which included all the pneumatic pipes inside it, and then to weld it together underground and rejoin all the pipes. It took over three months to do this work, and cost them a small fortune.

I had an interesting meeting a few years later when I was on a skating course in the Carlton Centre which then had an ice-rink above the multistorey carpark. One of the other participants was the maintenance manager of the Centre. He told me that the control system had never really been used properly, and that the method he preferred to control the air conditioning was a system he called ‘Management by Complaint’.

Electronics was starting to replace pneumatics in the industry in the late 1960s to early 1970s. Many people were very pessimistic about the maintenance of equipment, as there was quite a shortage of personnel with knowledge of electronics. However, it was generally hoped by the control experts that feedback control would really ‘come into its own’ as the new electronic systems were magnitudes more accurate and configurable than the old pneumatic systems, where every item in a control loop had an accuracy of about ±2.5% F.S (full scale). Many of the old pneumatic controllers had the parameters set by adjusting a tiny three-quarter turn knob on a pneumatic device with a logarithmic scale.

When you purchased the old pneumatic controllers it was necessary to specify which parameters you wished to include in a new controller, i.e., P, I or D (proportional, integral and derivative). Basically, each term involved quite a few extra pneumatic components, which made it more costly. When electronic controllers came into the market, you still had to specify which parameters you wanted and pay quite a lot of money for each extra term. This was in spite of the actual cost to the manufacturer being negligible, as it just meant adding in a few extra resistor-capacitor networks. However, as more control people came into the field from the electrical engineering side, they felt that they were being ‘ripped–off’, so there were a lot of protests, and before long all the manufacturers only offered full three-term PID controllers.

Another unusual project I was given was to build a controller test bed, as clients would sometime complain that a controller was not performing correctly. This knowledge was very useful to me in later years when manufacturers started building computerised controllers. Most of the DCS controllers were fine, as those companies had been building controllers for decades. However, when PLC manufacturers started including PID controllers in their systems, they usually didn’t work correctly, and in many cases scientific tuning was impossible. Even to this day, when I work on plants using PLC controllers, I find that in at least 85% of cases they have not been set up correctly to allow proper control and tuning!

I spent about two years with the company, and during that time most mining companies were developing new extraction processes which required automation, and were starting to use many more feedback control systems than previously. I did a lot of this work for some mining companies, and gained quite a bit of knowledge of control of mineral extraction processes like gold, and the newer uranium and platinum industries.

I was always a bit rebellious, and in 1969 decided that I never wanted to work for a boss again, so I started a company named Brown & Rutter Control Systems with Clive Rutter who was in my class at Wits University. My side was basically manufacturing and supplying control systems for industrial process plants. Luckily, my previous work with one of the mining companies had been appreciated, and they gave us a lot of orders for control systems at a time when a lot of new gold and platinum plants were being opened. The company also owned one of the big paper manufacturers, and I engineered many of the control systems for a new fine paper plant that was being constructed near Johannesburg.

Computer systems were still quite a way ahead, and with the knowledge acquired at GEC, I designed many logic ‘On/Off’ control systems using relays, timers and rotary sequencers. These would be all done today on a simple PLC.

The most complex system I ever designed and manufactured was the control of the converters at a platinum plant in Rustenburg, which involved complicated logic using hundreds of hermetically sealed relays. I commissioned the system on one of the converters in late December in the early 70s. A few days later, at 02:00 on New Year’s Eve, the plant’s control engineer phoned me to tell me that the converter had overrun its bottom limit when turning out of stack, and had tipped the molten matte into the dump pit, which had unfortunately filled with water after a rain storm. (The roof had not been finished on the plant at that stage). The resulting explosion set the plant on fire and caused a lot of damage. Fortunately there were no casualties.

The engineer told me that I would be sued if it was found that the design was faulty. I spent an anxious few days waiting for the result of the enquiry. Fortunately, it was found that a pin in the mechanical braking system had sheared, and I was exonerated.

Many years later the company replaced the relay logic system with a PLC, and I remember the engineer at the mine who did the work telling me it took months to understand all the dozens of complex interlocks that I had incorporated into the system.

After a while at Brown and Rutter I decided that we should also supply instrumentation and control equipment, so I travelled extensively to many countries overseas to establish agencies. On one trip in 1979 I visited the ISA Instrument and Control Expo, which was held in Houston, Texas that year. There I met an engineer called David Ender who was exhibiting a computerised tuning package. I knew very little about tuning, and was fascinated by the product. In those days, the only computers were massive and very cumbersome mainframe machines. PCs didn’t exist, so David and a brilliant mathematician who worked for him had built their own ‘portable’ computer, which could be moved around inside plants to connect to the standalone electronic controllers of the day. This machine consisted of a mass of electronic boards mounted inside a huge aluminum case, which was on castors and was fitted with lifting lugs. David explained the technology to me and it sounded fantastic. The problem was that the selling price was $75 000, and I didn’t think that any company in South Africa would be interested in such an expensive product for tuning, so I did not apply for the agency.

However, I was very impressed with David Ender, and we established quite a good relationship and kept up a correspondence over the ensuing years. In the mid 1980s he phoned me one evening and said:

“Mike, I have fantastic news. IBM has developed a low-cost personal desktop computer called the XT, and soon virtually everyone will have their very own computer standing on their desk. We have managed to fit our tuning package into this machine and we are now selling it at $4500 (excluding the PC).”

I was on the next flight to the USA and a week later I returned with the tuning package installed in Dave Ender’s own IBM XT which I had persuaded him to sell to me as there was quite a long waiting list for them. The following week I started trying to tune control loops on some of our clients’ plants.

The results of my first attempts at tuning control loops were extremely disappointing. The software was supposed to be able to tune processes with all types of process dynamics. However, the tuning was not always successful, and very often gave poor results. This mirrored the experiences of many control people I have met over the years who have tried tuning packages. The general consensus was that they did work occasionally, but most times the tunings they gave were pretty useless. Even to this day, surveys have shown that almost 95% of people still prefer to do tuning by trial and error.

I did not do much more work on tuning, as I was kept very occupied with other matters in the business at the time, but I still kept in touch with Dave Ender over the next couple of years. My business in the last half of the 1980s was becoming very difficult. South Africa was in the height of the Apartheid era, and more countries were putting trade sanctions in place against South Africa. It was very difficult to import our instrumentation. Also, computing systems were starting to be used in control systems, and one needed a lot of resources if you were to keep up with the competition. It was also hard to find specialists in this field at the time, and if you did get one, they were soon poached by some of the ‘big boys’. The business was also growing to a stage where I had to increasingly turn my attention to managing, rather than doing engineering. I far preferred being an engineer than having to manage other people.

In late 1989, Dave Ender phoned and asked me to come over to spend some time with him in Phoenix. He told me that he had discovered that nobody really understood the practicalities of feedback control, and we needed to train people. He had developed a course and wanted to try it on me. I told him that it was ridiculous to say that no one understood control, and was very sceptical. However, I went over and spent about 10 days being taught the practicalities of control. The course was literally ‘mind-blowing’. It was one of the most remarkable experiences I had in my career in control. I literally learnt much more than I’d known in the previous 20 years, and I realised that Dave was right and that no one I’d ever met in the control field had much understanding of the practical side. This marked a major turning point in my life.

I realised that the knowledge I had gained needed to be spread far and wide, and I decided to sell off my businesses, and go out into the field to ‘spread the new control gospel’.

So, I began a new career teaching people and also doing consulting on optimising control loops. Initially it was a hard and rocky road. I had no actual plant experience in this field and had to rely entirely on everything I had learnt from Dave in the few days I had been with him. As it transpired, not everything was 100% correct, as he was also learning. My first course was in December 1989 at a pulp plant in KwaZulu-Natal. I have subsequently referred to it as a case of the ‘blind leading the blind’. However, it was amazingly successful, and in the practical session held after the classroom sessions we managed to very successfully optimise some control loops that had previously always been in manual.

It is often very challenging sorting out control problems. I remember that the chief control engineer in a large titanium extraction plant once telling me that I was completely mad. He said: “How can you expect to go into a plant where people have been trying to solve problems for sometimes years without success, and sort it out in a day or two?”

In many ways he was right. I certainly have not succeeded in all cases of solving problems, but in general I can tell them what the cause of the problem is, and what needs to be done.

I learnt an amazing amount over the next few years, and gained a lot of self confidence in the technology.

I always pointed out that companies spend huge amounts of money on installing expensive control systems, but unfortunately they generally show little interest in trying to get their systems working efficiently. This is because very few people, including both control and process personnel realise how badly their base-layer feedback control loops are working. Even to this day, when I go into a plant for the first time, I find only a few loops are operating efficiently in automatic. At least 50–75% of loops have some sort of problems in them that, in most cases, the plant personnel are not even aware of. Unfortunately, most plants have been operating in the same fashion for many years, and the management is quite happy with the operation and sees no reason to invest money to improve the performance. Sadly, with no inside knowledge of a plant’s economics, I cannot predict savings that could be made if things were operating more efficiently. However, in several cases I have performed work in plants where I was told that it had resulted in huge savings, and sometimes literally very large increases in production. One of my jobs in a pharmaceutical company in the UK resulted in them increasing annual production by £14 000 in one plant. Another great success was saving a South African mining company literally millions of Rands on a particular gold extraction process.

One of the other sad things is that many plants are investing huge amounts in APC (Advanced Control Systems) without properly optimising the base layer control systems first. The APC sits above the base layer controllers and send signals to the setpoints of those controllers. If the base layer controls don’t work properly, the APC cannot operate efficiently. The APC practitioners invariably argue that they do sort out the base layer first, but I have found in numerous cases that they do not know how to perform such optimisation properly, and generally just try and adjust the tuning. An example of this was in a particular chemical plant that had installed APC and was not getting the desired results. I worked with them in optimising nearly 200 base layer control loops. A few weeks after the end of the exercise they reported back to management that an 80% improvement in the performance of the APC had been achieved by getting the base layer controls to operate efficiently.

There is no doubt that although one can be taught the principles of practical optimisation; practitioners do need to gain quite a bit of experience. Once they really get into it, it is remarkable to see how much self-confidence they gain. I have had a lot of feedback over the years from past delegates on my courses on how the knowledge has helped them. Some have claimed that using optimisation has added millions to the bottom line of their companies.

I am one of the few people who can say that they actually benefited from the

COVID-19 pandemic. Just before it struck, I was reaching the stage of life where it was getting difficult for me to go into plants. In fact, some plants would not let me in due to my age, even though I was still very fit, and I was seriously considering retiring. However, with the lockdown along came Zoom, and I found a method where I could carry on teaching and indeed also performing optimisation from my home.

This has proved very successful, and although I really am ‘semi-retired’ I still provide services of this nature to quite a few companies. In fact, even better, some of them are in far-away countries like Japan, Panama, and Ghana, to name a few.

I hope I can carry on doing this as long as I’m able, as I really enjoy the challenges of this type of work.

About Michael Brown

Michael Brown.

Michael Brown is a specialist in control loop optimisation, with many years of experience in process control instrumentation. His main activities are consulting and teaching practical control loop analysis and optimisation. He now presents courses and performs optimisation over the internet. His work has taken him to plants all over South Africa and also to other countries. He can be contacted at Michael Brown Control Engineering CC, +27 82 440 7790


Share this article:
Share via emailShare via LinkedInPrint this page

Further reading:

From the office of the CEO: Be warned, you are doing engineering work if …
This is not a new law. The only thing that is new is that ‘engineering work’ is now defined. You need to study the definition for yourself.

SAIMC: Johannesburg branch
At the meeting of the Johannesburg branch of the SAIMC held on 10 April 2024, Sagadevan Kanniappen, sales application specialist for WIKA South Africa, delivered a presentation on the ASME Performance Test Code 19.3.

SAIMC: Durban branch
The topic of integrating an IoT-enabled sensoring system and cloud technology for predictive maintenance of a buried steel pipeline drew record attendance at the April technology meeting of the Durban branch of the SAIMC.

SAIMC: Secunda branch
The SAIMC Secunda branch held its monthly Technology Evening on 10 April 2024 at the Proconics office in Secunda.

From the office of the CEO
When South Africa clinched victory in the Rugby World Cup for the fourth time, the nation collectively rejoiced. However, amidst this optimism lies a pertinent question: is South Africa truly a land of hope, or is it merely a fool’s paradise?

SAIMC: Johannesburg branch
The local Johannesburg Branch of the SAIMC hosted a technology evening at the Bryanston Sports Club on 13 March 2024. It was sponsored by Phoenix Contact and the technical presentation covered the full gambit of industrial wireless technology in the automation field.

SAIMC: Durban Branch
The SAIMC Durban technology meeting in March drew a record attendance, with an old friend of the SAIMC, Gary Friend presenting on the very topical subject of ‘Risks, rewards, advantages and disadvantages of new technologies and the impact of legislation’.

SAIMC: Secunda Branch
The SAIMC Secunda branch held its monthly Technology Evening on 6 March 2024 at the Proconics, Secunda office. Sagadevan Kanniappen from WIKA presented on Thermowell stress calculations as per ASME PTC 19.3.

Loop signature 22: How cyclical disturbances affect a control loop
Michael Brown Control Engineering Editor's Choice
When tuning noisy loops, we recommend in our courses that one should eliminate the noise by editing it out, so the tuning will be done only on the true process response, free of any noise. The controller is controlling the process, and is not controlling the noise.

Case History 191: The weakest link.
Michael Brown Control Engineering Valves, Actuators & Pump Control
Which is the weakest link in a control loop? The answer, without any doubt is that, in most cases, the final control element is the weakest link.