Is an xo tuned for valve amp or SS amp?

[tcpip]

I was thinking about this, though it seems pretty obvious from my simple understanding of physics. Does an xo have to be "designed for valve amps" versus "designed for SS amps"?

As far as I can see, the answer is "yes". The two xo's will be different. The reason being that the SPL curve of a driver, when driven by a valve amp (with a high Zout) will be different from the same driver's SPL when driven by an SS amp. This is because the impedance of a typical driver is highly frequency dependent. This will not impact the SPL curve when driven by an amp which has near-zero Zout and near-infinite current driving ability (at normal listening levels), but will definitely impact the SPL curve when driven by the 4-Ohm or 8-Ohm tap of a valve amp. Therefore, the two different SPL curves will yield two different optimal xo's.

The only time when this won't be true is when we use drivers in the regions of their frequency range where their impedances are flat. I can see this happening for ribbon tweeters, but for most other drivers, this seems unlikely, unless I design a 3-way or 4-way where each driver is used only in a small part of its overall frequency range.

Am I right?

[ergo]

You can try it out by adding a series resistor before your Xover in the simulation software (VituixCAD, LspCAD, Soundeasy etc). Xover should not be influenced much as it is more the load impedance that determines the outcome. The bass tuning of the low frequency driver(s) would change more I suspect between such tube and SS amp.

[JonMarsh]

I was thinking about this, though it seems pretty obvious from my simple understanding of physics. Does an xo have to be "designed for valve amps" versus "designed for SS amps"?

As far as I can see, the answer is "yes". The two xo's will be different. The reason being that the SPL curve of a driver, when driven by a valve amp (with a high Zout) will be different from the same driver's SPL when driven by an SS amp. This is because the impedance of a typical driver is highly frequency dependent. This will not impact the SPL curve when driven by an amp which has near-zero Zout and near-infinite current driving ability (at normal listening levels), but will definitely impact the SPL curve when driven by the 4-Ohm or 8-Ohm tap of a valve amp. Therefore, the two different SPL curves will yield two different optimal xo's.

The only time when this won't be true is when we use drivers in the regions of their frequency range where their impedances are flat. I can see this happening for ribbon tweeters, but for most other drivers, this seems unlikely, unless I design a 3-way or 4-way where each driver is used only in a small part of its overall frequency range.

Am I right?

Yes, and no. I'm a stubborn kind of guy, and I use techniques for developing crossover when ever I can that specifically results in a fairly smooth impedance curve. That is a choice, and if implemented properly works just fine with solid state, vacuum tube, and zero feedback solid state, as well as class D amplifiers.

Both the Ardent and the Isiris are large system examples of this; the NatlieP is a small system example. How this works is a function of both the chosen filter topologies and how driver attenuation is managed.





It's my opinion that any tube amp that can't handle that impedance curve with minimum level interaction needs to go back to the drawing board.

[tcpip]

Yes, and no. I'm a stubborn kind of guy, and I use techniques for developing crossover when ever I can that specifically results in a fairly smooth impedance curve.

Wow. From 50 Hz and up, you're keeping the impedance inside a narrow band. Amazing. Can I get this sort of thing by adding an impedance compensation shunt LCR string for the impedance peaks after I design the basic xo?

[JonMarsh]

You can improve things a lot that way- and for some network concepts, that's basically the only thing that will work- for example, classic two ways or 2.5 ways with waveguides often wind up in this situation depending on how the crossover attenuation factor is handled, due the basic crossover concept- which is often mostly a small cap in series, with just an impedance zobel on the tweeter.

Note, the design posted above does NOT use extra impedance compensation shuts- it's more a function of the quasi LR3 network design I use (I call it quasi because there is no such thing as a Linkwitz-Riley 3rd order network, but it's a third order network with the transition point at -3dB, like a typical LR crossover) and placing most or all of the network attention at the input, not the output.