Yup, I have a 6W SET and about 600W of woofer amp. I use the output from my voltage amp, right before it hits my EL34s, so it's kind of halfway in between.
In general, if using a SET output I'd double check the OPTs are capable of good LF extension.
Along with concerns of getting the VLF extension from the tube amplifier output, we also have to consider output impedance of the amplifier as this is all that is damping the wired connection between the main speakers and the input of the subwoofer amplifier. It's easy to forget that if there is acoustic energy in the room and you don't have anything connected to the speaker terminals, the speakers, especially the woofers, act as an inefficient microphone which produces a Voltage across the open speaker terminals. With a very low impedance solid state amplifier connected and powered on, this will be damped rather effectively assuming you aren't using excessively long, tiny gauge speaker wires and separately run wire back to the amplifier. I have observed the rare case of bizarre behavior with low-watt tube amps driving speakers with powered subwoofer sections where the signal is jumpered at the speakers. The higher you have to set the gain of the subwoofer amp, the more of an issue this can be.
Personally I'm very resistant to the concept of connecting a loudspeaker in parallel with the input of a powerful subwoofer amplifier. I understand the convenience and conceptually common signal to both, but I would also point out that adding subwoofers gives us the freedom to separate our selection of amplifier, speakers, and the speakers' placement in the room from the need for powerful deep bass. If the subwoofer is a great performer, I don't see the benefit of adding another level of distortion if the preamplifier is capable of driving the amplifiers directly.
I think it's also important to realize that many of these suggestions came about when most were setting up subwoofers with their ears and possibly an SPL meter in hand. While still a common practice, there's no arguing it's hunting in the dark vs actually connecting a PC & microphone to address the major issues before fine tuning. It's harder to take suggestions of worthwhile differences or improvements seriously when very significant and highly audible integration issues have been skimmed past. Decades ago we thought most all ported speakers had to be sloppy and inarticulate, while a significant portion of flagship speakers are now ported. Things improved greatly when we were able to measure woofer parameters more accurately, simulate enclosures on a computer rather than a calculator, and it became more practical to measure the frequency response of speakers we design. In-room, in-system, subwoofer integration is still a bit of a black art in the hi-fi world, and is likely ripe to see some new tools and raise the level of expectations.
