Hi this is my attempt to create a cheap but effective 1-60MHz measurement marker generator and simple tracking generator for a Tektronix 491 vintage spectrum analyzer. The nominal levels of the input of the analyzer are shown in the right.

Calibration, a must for old machines, should be of no importance with my plugin, since the DDS generated marker is set to the same frequency and level on the screen, with the actual signal under measurement and the level of the marker is measured afterwards separately (signal and marker switched on/off) with an accurate and sensitive power meter. Since both the marker and the signal under measurement present at the RF port of the analyzer, sny amplitude/frequency mis-calibration issues on the 491, will affect both the signal under measurement and the marker the same amount, so no problem in theory.

The power meter will be a MAR8 MMIC, followed by an AD8307 to measure down to -100dBm or so. Hopefully the MAR8 will have a flat response on 1-30MHz?
The DDS will be this http://qrp.gr/dds and no DDS amplifier will be used to avoid introducing harmonics and level variations and because it is not needed as the DDS outputs a much greater signal alone than needed. I found the level of this DDS to be exceptionally flat in the 1-30MHZ and I do not want to spoil this.
The simple tracking generator (no sweeper, as I need to manually introduce and measure the marker in this configuration) allows a DUT (eg an amplifier) to be examined in the analyzer screen. The PWR meter is connected with a separate connector to the rest of the system and can be patched to be used with the system or separately. A DUT can be connected there also so that the plugin can be used as a manual analyzer (portable) but it is no selective and no harmonics measurements can be performed in this point.

This should make the analyzer a usable machine again, as, without the plugin, it can only do relative amplitude measurements by removing/adding dB from the IF attenuator. Despite based on a vintage analyzer, I believe it will do a much better job than the FFT that I have used before, in the frequencies of interest.

What do you think of my setup?
Are there any points you should consider improving?