Eppley PSP – Design Notes
The Eppley PSP (or simply Eppley Pyranometer) is an instrument for the measurement of radiation in the range approximately between 285 and 2800 nm. As noted earlier, this wavelength range is commonly referred to as shortwave radiation, and it’s primary source is the sun.
The Eppley PSP is shown in diagram 1. The important components of the PIR are the
The thermopile of the PSP is designed to be sensitive to shortwave radiation. It is painted black to maximise heating, and has a sensitivity of approximately 9 µV/Wm-2. Betts, Clarke, Cox and Larkin (1985) examine the infrared reflection properties of the coating used on radiometric detectors (ie the coating on the thermopile). The coating used on Eppley PSP’s is Parsons’ black optical lacquer. Although this coating was not investigated, Nextel 3M black paint, which is widely used to coat the sensitive elements of radiometers according to Betts et al. (1985), was investigated. It was found that the Nextel 3M black paint has negligible values of reflectance in the near-infrared region. The thermopile surface can be considered an ideal black body.
The transmittance filter of the Eppley PSP is comprised of two optical glass domes above the thermopile. Diagram x shows the transmittance of these domes. As can be seen from this diagram, the transmission spectrum is very flat and has a cutoff for wavelengths below 300nm. The upper cutoff is not defined in the literature available.
Saunders, Foot, Kilsby and Seymour (1991) note that the "clear dome" filters (WG295) transmit radiation from 0.3-3.0 µm, and as nearly all of the radiative energy is inside these limits the exact wavelengths of the filter cutoffs is not critical.
Bannehr and McFarland (1992) asses the temperature characteristics of the temperature compensated Eppley PSP used on NCAR aircraft. They state that a small error is caused by temperature gradients within the pyranometer, however the error should be small because the instrument is equipped with temperature compensation electronics. During experiments the pyranometer sink temperature was monitored, as was the instrument output under varying temperature conditions. There was no obvious correlation between the sink temperature and the instrument output. This is attributed to the built-in temperature compensation electronics. The overall variation of the output was approximately 5 Wm-2 in the temperature range -40° C to +25° C.
The official Eppley literature states that the Eppley PSP has a temperature dependence of ± 1 % over the ambient temperature range -20° C to + 40° C. It is worth noting that, for temperatures outside the upper limit, the sensitivity decreases rapidly with increasing temperature according to graphs supplied with Eppley calibration certificates. Foot, Hignett and Kilsby (1986) state that for a standard Eppley PSP operating below -20° C an error of 1-2 % can occur.
Saunders, Foot, Kilsby and Seymour (1991) make no temperature corrections to data collected, as the environmental temperature rarely falls outside the range for which the instrument is compensated.
Saunders, Foot, Kilsby and Seymour (1991) state that differential heating of the dome due to kinetic heating may introduce changes to the pyranometers apparent sensitivity or zero offset. Measurements suggest an offset of 2 to 3 Wm-2 should be added to the derived flux at normal (90 m/s) for Eppley PSP’s, to account for differential heating.
Saunders and Barnes (1991) analysed two sets of "box pattern" flights to determine the effect of airspeed on pyranometer fluxes. They found that the pyranometer fluxes reduced by 0.5% for the faster box pattern at 132 m/s. They also state that the results are sensitive to the correct pitch/roll angle offsets being applied. They have "some confidence that these [results] are reasonably accurate".
Foot (1982) deals very effectively with the problem of thermal lag of the pyranometer. He notes that the purpose of the double dome on the Eppley PSP is mainly associated with ground use where the outer dome will heat up in direct sunlight and radiate in the infrared; the inner dome will then reduce the transfer of heat to the thermopile. On an aircraft, where the outer dome is well ventilated, the errors arising from the dome being at a different temperature are usually small.
Foot (1982) discuses experiments to examine the thermal lag of the instrument. During these experiments the upward facing pyranometer and pyrgeometer thermistor outputs were logged. The pyrgeometer thermistor was logged because it has a thermistor on the silicon dome which gives an indication of how the temperature of outer glass domes (which cannot be measured) are changing.
After rapid descents the normal Eppley PSP recovers to a normal signal after about 5 minutes. The sink temperatures stabilise after about 1 minute. It is clear that the response of the instrument after rapid ascents or descents is greatly influenced by the thermal inertia of the inner glass dome which is not forcibly ventilated like the outer dome.
We might expect that on a descent the dome would warm more rapidly than the bulk of the instrument and give rise to a positive offset on the instrument. A negative offset is observed however. This is probably because glass has a thermal conductivity approximately 1/200th the value of aluminium (which forms the bulk of the body of the instrument). Therefore, in rapid descents the sink of the instrument warms more rapidly than the inner surface of the glass. The net flow of heat in thermopile is therefore from the sink to the blackened surface, and thus the signal is negative. Foot (1982) notes that silicon (as used on pyrgeometers) has a thermal conductivity close to aluminium, therefore the competing processes are likely to be very different in the two instruments.
In conclusion, Foot (1982) states that during rapid profiles when the pyranometers are not in thermal equilibrium large errors (>14 Wm-2) are to be expected from Eppley PSP’s. These errors are larger than those experienced with pyrgeometers (less than 10 Wm-2) because glass is a much poorer conductor of heat than silicon. He recommends that after profiles, at least 5 mins should be allowed for the temperature of the inner glass dome to stabilise. It is not sufficient to note when the sink temperature stabilises.
Hignett (1987) made use of the thermistor buried in the body of each pyranometer to monitor the temperature of the thermopile heat sink. He states that when the pyranometers are not in thermal equilibrium the readings are unreliable, thus data were taken from straight and level runs with sufficient time allowed for equilibrium to be reached following any manoeuvring.
The Eppley PSP’s used by Airborne Research Australia are fitted with an amplification circuit designed at either Flinders University or NCAR. For details of these circuits please see appendix xx. It is assumed that the change in amplification factors are negligible over time, thus no attempt has been made to measure the amplification seperately, other than several "spot" checks over the last four years.
The amplifiers are powered by an external 28V source and have shown through many years of use that they are reliable and provide consistent amplification.
![[To ARA homepage]](/images/up_ara_home.gif)
(Last update of this page: 2OCT99 by dgp)