Tetra, based on your troubleshooting steps in your last post, it does sound as though the problem is in the 950.

In your description of hiss and hISs, the 770, when it's inputs are disconnected, is slightly noisier that when the inputs are connected to a low impedance low noise source (the 950 when the volume control is at minimum). This part seesm normal to me.
The amplifier input stage generates noise, and when the input is connected to a low impdance source (such as the 950 output) this noise level is lowered.

I am not familiar with the 950 internal design, but there may be a muting circuit that senses that when the volume control is at minimum, inserts a mute state (off) somewhere in the signal path. Similarly, some CD players and DACs, when sensing digital zero (all data bits in the word are zero) will mute the output, giving an artifically low noise reading in this state.

I would not be suprised if the noise increase a little as the volume contrl was increased to maximum, but I would not expect a large increase. If the amp has high gain and/or the speakers very sensitive, then levels of hiss (hiss, hISs, and HISS would be louder than the case of lover gain amps and lower sensitivity speakers.

Your description of your system that produces the hiss (950 connected to 770 with 950 volume control at min.) and HISS (volume control of 950 off of the min. setting) seems to point the finger at the 950 being the problem. To be more positve, does the HISS level change when all inputs to the 950 are removed? If possible, try the same test with the 950's inputs shorted.

Paul

p.s.

I looked at the specs for the 950, particularly the singnal to noise ratio specs. The digital preamp section has a worst case spec. of 98dB relative to 2v output. This corresponds to an output noise voltage of about 25.2microvolts, something you are not going to see on the vast majority of o'scopes or meters. The analog preamp secion has no signal to ratio spec. at all, suprisingly. I wonder if this omission is intentional.

[This message has been edited by Paul J. Stiles (edited May 27, 2002).]