It takes a while (~1/2 an hour) for the temperature control to fully stabilize.  It also tends to overshoot at first.

To chase the high frequency splitting, you'll need an external power supply.  The BK precision 1601 (long flat ones) will do the job just fine, set to around 20V.  You can always turn the voltage down until the ERROR LED on the optical pumping console lights up and then nudge it upwards.

5MHz RF signal with 1.5V pkpk, with the sweep range set at ~3.4 and ~20V external H coil power gets reasonable (order 100mV) signal at 100x detector gain.

Use the galvanometer jiggle to locate when you're picking up a signal, it'll twitch in response much more quickly than the scope will capture an image.

I can't replicate the second plot in the Teachspin examples with the same settings.  There's some 60 Hz noise that's probably swamping out the signal.

I probably shouldn't set the power supply on top of the precision measurement box, eh?

Manual notes that there's 80 uV pickup noise on the detector as well.

If you use the zoom feature of a TBS1052 (the little magnifying glass next to the multipurpose knob) you can keep the triggering position constant while you move the display over.

We can get all they way to 10 MHz, but there's a caveat: you can hit the 2A limit on the regulated power supply getting a strong enough B-field to see things.  Either be quick or find another supply.

There's some pretty funky distortion on the signal shapes, I think the regular BNC cables are about at their limit.

Basic absorption dip is about 1V at 10x gain. 

For the detector amplifier, the DC Offset and OFFSET fine adjust the needle in opposite directions.

Got the scope and fct gen to talk to Python!

{ ${/download/attachments/275156712/RFSweep.png?version=1&modificationDate=1624551886000&api=v2}$ { ${/download/attachments/275156712/RFSweep100HzStep.png?version=1&modificationDate=1624557783000&api=v2}$