Unbalanced pre out to balanced amp in

By 'across' , I mean in parallel with the output.

If the input of the xformer is supposed to be 20K and the output 600ohms, then the xformer will express 20K to the input if there is a 600 ohm resistance across the secondary.

Its the same idea that you see in a conventional tube amplifier: if the tubes are supposed to have a 3000 ohm load, that will only be the case if the output (8 ohm tap for example) is loaded at 8 ohms.

You'll notice that in the test circuits that produced the graphs on the first page, that there is indeed a 600 ohm resistor (or a pair of 300 ohm units) loading the output of the transformer.

The Jensen transformers are excellent! But- like all transformers, if you want to really get the most out of them, make sure that the secondary side of the transformer is loaded to the value specified. For example if the output is 600 ohms, put a 600 ohm resistor across it (if you don't, the transformer will express the inter-winding capacitance, rather than the turns ratio). That way the transformer will be the flattest frequency response, and minimum coloration, which is to say it will be very transparent.

From the data sheet, a fairly wide range of impedances is shown for the load, but they also show that the output impedance of the transformer is 80 ohms. There is a rule of thumb that suggests that the load would be about 800 ohm (IOW 10X the output impedance).

There is an optimal value- it looks to me (based on the specs) as if this unit is designed to support 600 ohm operation. I bet 600 ohms would do the job nicely. Have you tried loading it ever?

With the right load not only will you get broader, flatter frequency response, but the unit will be more transparent too.

The specs call for a source impedance maximum of 2Kohms. There are a lot of preamps that can do that, although some of them may have trouble making bass.

IMO/IME, you are better off with a step-down transformer, maybe 10K to 600 ohms. You loose some voltage gain, but a wider range of sources can drive it.

At the input of the amp could be another transformer that is 600 to 10K... then there would be no loss in gain. The bigger the step down/step up, the more transformer artifacts are introduced. That is one of the reasons I prefer to direct-couple :)