while measuring RMS Voltages with our Picoscope 4824s, we have noticed a drop in the voltage in the minutes after startup. We now believe we have chased this behaviour down to the temperature behaviour of the ADC-Chip.
The first image shows the general behaviour at startup (y-axis is RMS Voltage according to the Picoscope, custom Software):
The source here is a (warmed up) Fluke 5520A calibrator generating 1 V at 50 Hz, cross-checked with a 6.5-digit-multimeter. It takes 5-10 Minutes for the voltage to settle to a constant value.
We also gently heated up parts of the circuitry with a hair dryer. The following image shows the same startup process while pointing the hair dryer at various parts of our and the picoscopes PCBs. The big drops in voltage at the 400-s- and the 550-s-mark occur when the ADC-Chip is heated up very gently.
We now assume that what we're seeing at startup is the internal reference of the ADC heating up, which increases its output voltage, which in turn decreases the measured voltage.
Is this correct or at least plausible? Has anyone ever made the same observations? It's probably impossible to do anything about this, right? The only things that came to our mind were active heating and on-chip-temperature measurement with software compensation, both of which sound like "Now you have two problems":)
I realize we are operating at the accuracy limits of the Picoscope here, and also that the operating conditions for the Picoscope are only 20°C to 30°C. The operating conditions of the HMCAD 1102 are given as -40°C to 85°C in the datasheet, although no temperature coefficients or accuracy dependencys are mentioned.
I am thankful for any comment.
Another thing struck us in the HMCAD1102 datasheets: It is capable of 14bits, at 65 Msps. Given the chance, we would take an 8 channel, 2Msps, 14bit scope in a heartbeat. Is there a chance of a driver update or a new version that enables this?
Thank you very much,