Understanding your instruments and their calibration.

About ten years ago and near Johannesburg which is up at an altitude of around 5000ft or 1524m, I came across a recently installed cold store with dimensions approximately 40m by 20m and designed for bulk storage of pre-cooled plastic bottled drinks. The complaint was that this store had never been able to properly maintain the designed 2°C store temperature except occasionally for very short periods during some of the cooler mornings; typical mid afternoon room temperature was seen to be around 4°C. The owner explained how apparently he understood the refrigeration plant was selected with a total capacity very closely matching the very accurate load calculations used by the designer. The refrigeration plant comprised four MT160 Maneurope condensing units with matching blowers running on R22 each with a design capacity of 26.1kW. The design saturated temperatures were -5°C evaporating and 45°C condensing with short pipe runs having negligible pressure drops.

At the time I felt that observing the refrigeration systems operating readings as occurring within design cold store and ambient temperatures would bring me closer to any clues useful. I also wanted to do this without making any prior adjustments. A hindrance here of course with this particular site was that the very problem with the system I was trying to identify would not provide for the coincidence of design room and ambient temperatures. I had to visit site first thing in the early morning when it was said the cold store temperature would briefly be at the designed temperature of 2°C. Then all I had to do was partially restrict the condensers airflow to raise the saturated condensing temperature to the designed 45°C. On the morning I also positioned my airflow restrictor, a swimming towel, to maintain condenser leaving liquid temperatures above the designed ambient temperature of 32°C. If I allowed liquid temperatures to drop nearer the cooler morning ambient temperatures then the resulting increased refrigerant's net refrigeration capacity would have the effect of lowering saturated suction pressure since less refrigerant would be needed in circulation to establish the TEV's superheat setting.

Now that I had eliminated liquid pressure, liquid temperature and room temperature from the list of system performance effecting variables, I could look to the suction side for clues as to the cause of apparent performance loss. Of course, we know very well that lower saturated suction temperatures and higher suction superheats adversely effect system performance, I only wanted to eliminate them next in a positively methodical manner before moving on, if moving on was going to be necessary.

As is routinely done I attached my suction gauge to the evaporator coil header pressure tapping and my surface temperature probe to the same header. I found that saturated suction temperature was nearly 2K below design while suction superheat was about the same too high. Interestingly, this is to say that the evaporators suction header surface temperature was in fact at the expected design temperature, 4K above the design saturated evaporating temperature.

At this point while I was asking myself "Why, despite the many return visits apparently made in an attempt to rectify this problem, everyone visiting seemed to have overlooked the obvious high TEV superheat settings found on all four systems?" But within the same train of thought I remembered that, to the best of my knowledge, I calibrated my service gauges different to every other technician.

The key was that I adjusted my gauges to compensate for altitude allowing me to see the lower than required saturated suction temperatures when others simply could not.

ike every other service engineers system analyser, mine's gauges used the "Bourdon Tube" principle. Within the gauge a 'C' shaped tube, the "Bourdon Tube", will tend to open or close its 'C' shape depending on whether the internal pressure was greater or lesser than the surrounding (atmospheric) pressure. The problem with this principle is that it does not automatically compensate for change in altitude as would say a diaphragm gauge. A bourdon gauge calibrated to show zero at sea level will still show zero when transported unadjusted up to 5000ft altitude. However, relative to sea level the atmospheric pressure at 5000ft is already a partial vacuum of about 4.95" Mercury. Yet, as already mentioned, every other service engineer I knew would leave their gauges at the factory calibration of 0" Mercury vacuum or 0psi atmospheric pressure.

The approximate atmospheric pressure at that altitude is 12.27psi or 84.6kPa. If atmospheric pressure at sea level would hold a column of mercury against a vacuum a height of 29.92" or 760mm then at an altitude of 5000ft atmospheric pressure could only hold that mercury column 12.27/14.7 x 29.92 = 24.97" or 634mm.

The result of this calibration anomaly was that while the visiting service engineers thought they were seeing the systems saturated evaporating temperatures at the designed -5°C in reality I found the systems operating with saturated evaporating temperatures even lower than the predicted un-calibrated offset temperature of -6.2°C.

Where the MT160 condensing units had a design rated capacity of 26.1kW at -5°C they were in fact achieving only 24.5kW, approximately 6% less than the accurately predicted design capacity.

Convinced I understood the reason for this systems underperformance I went ahead and reduced the four TEV superheat settings by the appropriate amount. With the resulting increased refrigerant feed increasing the evaporator duties there was the needed rise in saturated suction temperature and compressor duties. The subsequent rise in saturated condensing temperature needed to meet the increased total heat rejection naturally increased head pressures so that now finally too the valves were seeing better liquid feed pressures allowing them to better meet their designed capacities.

Clearly they were not exaggerating when they claimed plant selection closely matched calculated load.