J. Rypko, M. Urban. LAL. Dec. 2000

The LAL has build in October 2000, 22 slow control boards and will build 23 more in January 2001. We present here the results of the tests on these 22 first boards.

Figure 1 shows the ADC readings, in millivolts, when grounded, versus the board number.

All offsets are below the millivolt and could be assumed to be null for a 10^-3 precision. Recall that the 10k resistors for temperature sensors are already at the 10^-3 level.

We do not know yet how these offsets vary with time and/or temperature.

Below are the offsets both in millivolts and in ADC channels.

Figure 2 shows the ADC responses to 3.3 and 4 Volts, versus the board #. The 4 Volts was injected on channels 1, 2, 4, 5, 16 and 19. The average response to the 4 Volts is —10 10^-3, unfortunately the 4 volts supply is not stable enough. This can be seen on the 3.3 Volts where the average is now —0.4 10^-3. The relative rms. is 2.5 10^-3 for the 4 Volts but as we said the 4 Volts is not stable enough, and the relative rms. for the 3.3 V is 1.5 10^-3. Therefore, contrary to the offsets, we may want to apply gains in order to have a relative rms. below 10^-3.

Incidentally we see on figure 2 and for the 4 Volts data, that the MUX does not introduce dispersion. Each box in the scatter plot is filled with 6 channels and except for the board 22 (0.4 10^-3), they stand well below the per mill.

The multiplying factors to be applied to the ADC readings are in the table below, versus the board #.

We can assume 1.221 millivolts per ADC bin.

We read the DACs with the ADC.

The average over the 22 boards of the writing-readings is shown below:

The ADC offset is on average 0.4 mV. We conclude that the average DAC offset is 1 mV out of 2500 mV total for the DACs. So again it seems that we should not worry about the offsets and the slopes are fine also.

We take in figure 3, the example of DACs 1-3 going into channel 26 of the ADC.

Taking into account the offsets and the gains of the ADCs we bring the response of the DACs below the 10^-3 level.

Starting with 12 bits we end up with 10 bits with no corrections.

On each board there are 10 temperature sensors. The top of figure 4 shows the temperature measurements versus the board #. The bottom part of figure 4 is just the histogram of the temperatures. We see that the temperature was not a constant as measurements proceeded, but that the dispersion of the temperatures inside each board is very small and corresponds to the same outside temperature. Therefore this is giving us the rms. in the absolute temperature measurements. We know that the resistors are at the 10^-3 level so that we cannot expect better than 0.3 °K for the error on the absolute temperatures.

Figure 5 displays the rms. in temperature readings versus the board #. The bottom part is the histogram of these rms. We conclude that we measure the absolute temperatures within +- 0.5 °K.

We monitor 4 voltages: uncompensated, compensated, thermistor and the offset.

The offset voltage should stand between 2490 and 2510 mV for normal operations. In figure 6 we show the offsets we measured with the 22 boards, without (top) and with (bottom) ADC gains corrections. Again we see that we are at a level below 0.1%, but the average value being 2483 mV we are inside the orange warning range (2450 to 2490 mV). We do not know yet if the ranges were defined too tight or what.

Figure 7 displays the nice correlation between the temperature from the thermistor of the cloud detector and the temperature read from the Analog Devices sensors. The horizontal axis is the board # and it represents different times since only one board is tested at a time.

Figure 8 shows an example of the correlation. We took 50 mV per degree Celsius and an offset of 2483 mV, to translate the thermistor voltage into °C. We see that the thermistor is giving a temperature 1.5 °C lower than the Analog Device, and the dispersion is 0.3 °C.

1) The study of the first 22 slow control boards built at LAL indicates that if we accept a 0.1% absolute precision it will not be necessary to correct the ADC and DACs for offsets and slopes . Having tables for offsets and slopes will gain us a factor 2 in precision.

2) The absolute temperatures with the Analog Devices sensors are known to a precision of 0.5° rms.

3) There is an excellent correlation between the temperatures measured with the Analog Devices sensors and the thermistor from the cloud detector. The cloud detector is 1.5 °C lower than the Analog Devices sensor, and the dispersion is only 0.3 °C rms.