There was a post on the QRP-tech yahoo group about the NorCal S9 Signal generator
This is very close to the circuit for the Elecraft XG2 Receiver Test Oscillator
This looked like something that might be useful for testing the Spectrum Analyzer after I get it finished. I decided to build a similar unit, using parts from both design. Looking at the NorCal circuit, I saw that they used 20 and 14 db pads instead of the single 34 db pad used in the Elecraft model. I went with the NorCal circuit and added switching around both pads to have options of 50, 10, 5, and 1 uv. outputs.
I want to put the signal generator in an Altoids tin, so I used the same template as my AD8307 power meter as the basis for the board layout. I have found that I get more consistent etching of the ground pour if I do it as a hatch instead of solid. I plan on using 2 row .1" header strips with jumpers instead of switches, so I placed extended solder pads on the board. This way I can just bend the bottom part of the header pins out to the side and solder directly to the top of the board.
Now to etch the board after I get back from a hamfest tomorrow.
Well it was a little damp and chilly at the Hamfest, so I didn't stick around as long as I usually would. This gave me plenty of time to etch the circuit board. For SMD boards I have been using the blue press-n-peel transfer paper to eliminate etch through, and I get a cleaner pad for the SMD components. I also let the boards sit in some "Liquid Tin" solution for about a minute to give a nice tin coating to the traces. This seems to solder a little better than the bare copper.
I built the oscillator portion of the board to test how much output I was getting. From the Elecraft document it looks like the oscillator should put out about -50 dBm before the attenuators. Hooked mine to my home-brew AD8307 power meter and got a reading of -48dBm. Checked the voltage to the oscillator and had about 1.6 volts, I might try adding another diode in series to drop the voltage and see what that does to the output.
Now on to the attenuators, I am missing a one of the values for the 20 db pad but will try with the closest value I have for now. Will replace with correct value when the assortment of low value SMD resistors I have comes in.
Since it is raining out I decided to finish the board and add the attenuator components. With the attenuator pads out I measured -68.2 dBm on my power meter. I added another diode in series with the existing diode, this dropped the voltage to the oscillator to 1.07 volts. Measuring the output with the additional diode I measured a value of -72.8 dBm. This is more than close enough to the desired -73dBm output level of the Norcal S9 generator.
Only have two crystals in the circuit for now (3.561 and 7.030) will add others as needed. Listening to the output on a receiver. The 50uv gives about a S7 reading, readings for other levels are below S1 but I can hear definite changes in the signal as I change attenuator settings.
Unfortunately I do not have any calibrated test equipment to verify the values I found. But I believe the accuracy should be enough for testing the home-brew equipment I am working on.
Link to dropbox for Eagle files and toner transfer image.
I finished drilling some holes in an Altoids tin and and mounted the circuit board. I also printed a simple instruction card showing the jumper settings for the different signal frequencies and attenuator settings. I sized this to fit inside the top cover of the Altoids tin.