Post Go back to editing

ADL5903 Aging


We use this RMS power detection for L-band (1-2 GHz) level monitoring.

We built a look up table (calibration) per unit. 

Can Analog Devices supply aging data for ADL5902?

Do we need to recalibrate power detector after some years?

Each how many years?

Operating temperature range is from -40C to 70C, also calibrated for each unit.

Can Analog Devices advise anything that we can do in our side to avoid or compensate for the aging issue?

Please advise.

  • Hi fnoriega,

    Which part are you using, ADL5902 or ADL5903? Both are mentioned above.

    Can you tell us, what is the magnitude of the long term drift that you observe? How does that compare to your requirement?

    We always recommend in-system calibration for detector ICs such as these. It's clear that you do this now, which is good. 

    Kindly note, however, that for these two devices, we generally would NOT recommend an individual in-system calibration for temperature drift. Instead, we would recommend that the application should accept the temperature drift of the part, which in this case would be quite good, +/- 0.2 to 0.4 dB, depending on the part and other specifics. The reason is that the devices are already internally compensated for temperature variation. Applying an external compensation 'on top of' the internal compensation could be making the problem worse rather than better. 

  • Hi Bruce,

    Sorry, I'm using ADL5903.

    We didn´t start testing yet, but our client ask us about the neccessity of recalibrations after some years.

    Could you clarify abour this topic?

    Thanks in advance

    Have a nice week

  • Hi fnoriega,

    We do have Long-term drift data on some of the RF detector devices; not all. Luckily, ADL5903 is one of them. 

    Looking at the data, we see drift of a few hundred'ths of a dB, over a contiguous time period exceeding 900 Hrs (about 37 days).  Because the file size is quite large, I'll send in separate email to the address registered with your username. 

    Also, please understand that we don't really know what the few hundred'ths of a dB might be due to: maybe the IC, or maybe the PC board, or the test setup, or the solder flux on the PCB, or maybe solvents used to clean the board, SMA connectors and cables, maybe the multimeter, etc. etc. 

    I once worked on a system where we were able to reduce drift over time by first baking the boards around +80 deg C, to help 'burn-in' the boards, immediately prior to calibration. In that case, it really helped a lot. 

Reply Children
No Data