Do S-parameters exist for the HMC1110?
Questions regarding whether or not Analog Devices has S-parameters on various legacy HMC multipliers like the HMC1110 is a common question. Unfortunately the answer isn't as straightforward as one might hope.
In some cases, actual S-parameters exist. If they are true *.sNp files, these may be in various formats (Mag / Ang, dB / Phase, Re / Im) and some may be S11 / S22 only (*.s1p). In other cases, only return loss data is available (no phase information) or perhaps there is no data.
Next, are they de-embedded and if so to what reference plane? Great question! Most of these products have been around for awhile and unfortunately, short of re-taking the data, there's no way to discern this. We certainly would hope they are de-embedded to the lead for packaged devices and the bond pad for die but that's not a guarantee.
Well, o.k., are these an average of several devices or for tied to a single device? Another great question for which there is typically no way to know for sure unless the file name or contents contains these details.
Finally, what temperature, drive level and bias conditions were used? Typically the file name or contents include this information but again there is no guarantee.
There are many other questions regarding the data for which there is no answer unless the data is re-taken.
So how useful is this information? In my opinion they are of little practical use, especially for those desiring to improve the input or output matching. The above reasons only tell part of the story. The legacy HMC frequency multiplier designs almost always include an input buffer amplifier, multiplication stage and an output amplifier. Due to this configuration, unlike a single device, there is no easy way to achieve a simultaneous conjugate match. This means that there is some risk of degrading electrical and / or temperature performance due to external matching. If users implement external matching ,the responsibility of vetting performance over supply voltage and temperature is not only highly recommended but also their responsibility.
In my opinion, the best way to use these legacy HMC multipliers, especially since they generally are well matched already, is to treat them like 50 ohm gain blocks.
So yes we have HMC1110 S-parameters and they are attached below; the input S-parameters are based on an average of 9 devices while the output is based on the average of 11 devices. Bias voltage and drive levels are contained in the files.
Retrieving data ...