Here is the completed crossover PCB. It is constructed more or less to the original circuit diagram of the 760/770 crossover board. Before building,I redrew the circuit in LTspice and checked its behaviour. This also provided a means of generating the netlist. After making the board, I tested it by injecting a swept sine wave signal, 20 hz to 20 khz into the input, whilst monitoring the output using the freeware HOLMImpulse . This programme is designed for measuring the impulse response of a room but proved excellent in this role too.
The circuit diagram is shown below. This is the output from LTspice. Note that parallel capacitors are used intwo locations to make no longer available values of 33nF. Well I actually settled for 32nF. The simulation showed very little difference due to the error.
Frequency Response from Spice Analysis
Here are the results of the LTspice simulation and the actual measurements. Their similarity is remarkably good so I guess I got the circuit right. The crossover frequency is slightly lower than advertised. It is about 750 hz wheareas the usually quoted figure is 800 hz. However, you can see from the simulation that if the treble gain was dropped by 2dB, the crossover frequency (or at least the point at which these lines crossed) would increase to about 800hz.
Measured Frequency Response
This is remarkably similar to the predicted response. Note the two discontinuities at 50 Hz and 100 Hz. These are caused by mains hum in the measuring apparatus.