Mumford,

Let clear first with terminology. You post said about “Incorporating active crossovers into DSET” and I think it is not what you are interested; at least it is not what I am practicing. An active crossover implies a use a dedicated active stage/s just for sake of crossovering. I am religiously against it and I am considering that the active approach is a superfluous waste. Generally there is always an opportunity, even with such a simple amp as my Milq is, to present a filter to an amp, and in case of DSET a filter is an intricate part of design.

The filter I use usually called line-level passive. There are no rules how them are used: they might be high or low pass, the might be located before the amplifier stage, between the stages, in plate supply, in feedback or whenever you can imagine. It is all what and by what kind mince you would like to accomplish. Doing it you need to care about the fixed filter load, about an also of your source to drive the filters, about the slope of your filter and the most important the topology of a filter to the given filtering range. A parallel cap, the LR chain or the RL in feedback would write for you’re the very same curve – it is purely up to your intention and your ability to discriminated results that would desecrate what specific solution your would like to go for. Sure, there are some limitation in amps, particularly if you are willing to go for high order filters, but there are also always the opportunities to use subtraction filters and separated filters to writhe shaper curves… The possibilities are truly endless… 

mumford wrote:

Without using a buffer before your filters, does the source before the amp require any special attention?

In my case it is very much so. With 6 channels the input impedance of my 6-Ch Melquiades is 5K, and this frequency-dependant 5K of very heavy to drive. Some there at my site I wrote that what my preams was out the sound was disastrous and only my DC coupled DACs were able to drive my amps.

mumford wrote:

Can you recommend any books, papers, webpages, etc, that you referenced in designing the crossovers?

No I cannot. First of all because I do not know any. The filters design is well know subject well covers in any book. However, I did not see a whole lot of people who use sane empirical ways to evaluate what filter they need and in what kind application. However it is all not very complex and you will not face any problem. Print your amps circuit and look at it creatively. Then you will find a LOT of places where you will be able to stick a filter. To circulate the filter it and then to fine adjust them very simile… BTW, in contrary to speaker filter the line-level passive filters are very predictably behave.

Romy the Cat wrote:

Mumford, Let clear first with terminology. You post said about “Incorporating active crossovers into DSET” and I think it is not what you are interested; The filter I use usually called line-level passive.

I apologize for my mistake with terminology.  It is indeed the "line-level passive" that you are referring to.  Thanks for your advices.