I bought one of these 1kH signal sources as a result of macboy's post in another thread:
These are the opposite of spending $20k on an AP unit:
http://www.ebay.com/sch/mickevich/m.html
At only ~$50, maybe it's worth a try. You have access to an AP unit from anther lab, so you could at least use that to measure the distortion from one of these to see if it will cut it.
This isn't a review as I've only used it to test my Keithley 2015 THD so I can't say how low distortion it is, only give an upper limit.
First, a couple of things about the board - one it doesn't have holes of mounting screws though it does have a single screw at the front to mount it in a case.
Secondly it requires a rather awkward 35V power supply but it only draws 26mA as shown in the first photograph.
My lash up is rather messy as shown in the second photograph but good enough for my tests. It is a two channel source so I used the second channel for scope output.
The Keithley 2015 THD test procedure calls for a 0.95V rms signal at 1 kHz. Adjusting the pot to give the level is rather awkward as there is a delay in response and 0.94V was as close as I could get. The frequency is about 1Hz low (999Hz rather than 1kHz) but that seems good enough to me.
The % THD test showed a reading that went between 000.000 and 000.001% well within the K2015 THD's spec of < 0.004%.
The measurement in dB showed -99.5dB (though the decimal point figures move around). This is a measurement of the K2015's noise floor rather than the oscillator but shows it is at least as good as -99dB (the claim is less than -145dB but of course I can't measure this).
(Edit : the spec for the 2015 is for a residual distortion of -87dB.)
Anyway - as an oscillator to test a 2015THD it works very well and it might well work very well for more serious THD measurements.