My issue...and I am not sure it's real (is it within specs? I couldn't find it in the manual). It seems I have amplitude modulation even when the VCA input is shorted (frequency is lower than the dialed frequency, maybe some harmonics).
Methodology:
-Using the TDS320' averaging function, read the Vp-p of the waveform
-Using the TDS320' enveloppe function in order to use the cursors to measure the accumulated amplitude differences; Vvar
-Calculating % = Vvar / (Vp-p/2) or 20x log (%)
Here are examples.
At 5x1K, I have 14.81Vp-p and 0.6V of variation, about +/-2.025% or -21.8dB of amplitude noise
At 5x1M, I have 14.04Vp-p and 0.14V of variation, about +/-0.98% or -34dB of amplitude noise.
At 5x10M, I have 10.68Vp-p and 0.28V of variation, about 5.2% or -25.6dB of amplitude noise.
There are no specs in the service manual about this exact information. It seems to me that the way they generated the switching of current direction is highly dependent on the stability of these matched diodes in the bridge, couple that with the choice of the charging capacitors used, proposes these performances.
Any clue?
What is the frequency of the AM?
Most likely, the Wavetek 166 is functioning fine. I think a large part of what you are seeing is due to the oscilloscope. Because of the 8-bit converter's quantization plus any noise, the envelope will generally show a larger voltage waveform. In averaging mode, any of these "outliers" will be smoothed out.
For comparison purposes, I hooked my Wavetek 166 to a Tek TDS520 scope and used the scope's Vp-p measurement in both average and envelope modes. If the average Vp-p was 10V, the envelope Vp-p was about 10.1V. Slightly better results might be obtained by enabling the scope's bandwidth filter.
I tried the same experiment with both an HP 3314A and HP 3335A generator. The former uses a similar RC-type of function generation whereas the latter is a digital synthesizer. In both cases I got just what I saw with the Wavetek 166, namely a slightly larger envelope.
Alright...While one might not be a 'young player' and actually teaching about Shannon-Nyquist, you can still foul your thought process, if not careful.
(Apparently, I can't simply paste a .png file here)
I had these amplitude issue from one cycle to the next in the 50Mhz range.
Every few cycles in the 40Mhz range and as the main frequency lowers, more and more cycles partake in the AM effect.
I decide to put the signal into my SA and expected to see serious spurious noise below the main frequency.
It turns out the signal is rock solid on my HP 8594e.
The source of the AM seem to come from....."the scope?"....Well the scope I use is a TDS320 from Tek and it samples at 500MS/s.
That means that there is only 10 samples per cycle. So the AM is not existant and the result of aliasing.
Thanks to edavid for the question that made me search harder...
Yes, I figured that out Saturday night.
The less than perfect signal from the 166 is mainly phase noise I believe.