Normalizing a Spectrum Analyzer

Can I normalize my spectrum analyzer using an external signal generator to calibrate directional coupler and filters as I'm using MDO4000 oscilloscope??
And what are the steps to do so?
p.s I will connect the time bases of the equipment by the timebase of PM6681R Rubidium counter

I think there are numerous ways to normalize the spectrum analyzer. First, remember that neither the spectrum analyzer nor the signal generator have perfectly flat response. For filters, you have to provide a signal using a specific set of cables, and annotate the measured amplitudes at the various frequencies. If your spectrum analyzer has a normalize capability, which is more common on network analyzers, you could use that.

If the spectrum analyzer has dual trace capability, you can make one THRU measurement, and a second measurement with the filter inserted.

With directional couplers, probably not so simple. You need a good procedure, and good theoretic knowledge of the measurements. I think that may be more complex than normalizing a spectrum analyzer.

This was such a specialized question I feared it wouldn't get a response.

Thanks, Jerry. You are very much appreciated!

From what I read here you are asking if you can connect a signal generator with a spectrum analyzer and use it as a network analyzer to calibrate couplers and filters.

In order for that combination to work you must be able to have the signal generator and the center frequency of the spectrum analyzer synchronized throughout the entire sweep, and you are not typically able to accomplish this. That is why they build Tracking Generators into spectrum analyzers, so that you have a synchronized frequency source available. An alternative could be to automate the process to control the frequencies on the analyzer and the generator, this would work.

A quick and dirty way would be to set the spectrum analyzer to sweep the span you need for your test and place the display mode in Max Hold. Then you set the generator to sweep the same span at a slower rate, so the analyzer gets several passes at each frequency to acquire the peak value for each frequency. This will get you the data you need. What is doesn't allow is to "Normalize" the data unless you take the acquired data without the device under test, save it to memory, then use a math function to show the difference when you take the data with the device under test.

And my Thanks to dwperron as well!

I appreciate your explanations, but why most tracking gen. have low frequencies 3 or 6 GHz, i need one up to 18 GHz?

Making the tracking generator for the lower frequency band is fairly simple. In order to go to higher frequencies it involves band switching to harmonics of the LO frequency. The cost of this, the size and weight, the extra RF signals generated internally that must be suppressed, all of these make a high frequency tracking generator less than practical.

That's why a simple Scalar Network Analyzer and a sweep generator would be a better choice of instrument for the measurements you are making.