XLR and Alepph-X

[russbryant]

Originally posted by Nelson Pass
Noise issues from the environment aside, there is not
much difference in sound between balanced and unablanced
on these circuits as long as the diff pair is constant current
sourced. Having said that, the difference is that the balanced
does sound better.

Nelson,

Were you laughing out loud or quietly to yourself when you hit the send button?

Russ

[jarek]

Don't use too high value on a pot. Probably 10K would be a good value, if not don't go higher than 50K.

Is it nessesary to add a cap at the input on the PCB?

[Peter Daniel]

If your preamp doesn't pass any DC, or has output coupling caps you don't need input caps in Aleph X.

[jarek]

Thank you Peter for your reply and your time. My CD's has output coupling caps. But still one thing is questionable for me. If we don't add any cap at the input on the Alephs’s PCB we will bring about parallel connection input resistor R19 through R18 to resistance of the pot for DC (I assume work without XLR). Then, the input DC current will be divided between these resistances. If we change volume on the pot, the summary resistance will change too. Notice, that such a resistance R29 remains constant. Is that situation doesn't cause any DC problems at the output?

[Peter Daniel]

Whe you build the circuit, you'l be able to check it out and if needed install coupling caps. Although my BOZ has output coupling caps I still had the 500mV of DC offset when no input coupling caps in my Aleph X.

[jarek]

OK., I'll try both ways. BTW, 0,5 V with my 4 Ohm speakers would give about 3,75 W to the heat sink extra.

[jarek]

I have just run my AlephX. My setup is like Grey’s design with ‘magic resistors’ and two 4,7 uF capacitors at the front end added. As I use RCA connection one input is shorted to ground before the capacitor from the source side of view (when I shorted after this capacitor there was 4 V differential DC at the start!!!).
Yesterday I hooked up my scope and took same measurements. I was surprised there was differences between two sides of the amp. One channel of scope was connected to +output and ground and the other to –output and ground. The branch with input grounded has more gain and 1 kHz square wave looks pretty good in opposite to ‘active’ branch which has less gain and is undercompensated. The summary output signal is undercompensated too.
Is it normal for unbalanced driving AlephX? Should I increase value of correcting capacitors (4.7 pF) in both sections?

[Nelson Pass]

This isn't normal. You should be seeing the same amplitude
value on both sides. As you decrease the value of resistance
from output to the Sources of the differential pair, you can
start seeing something like this, and I recommend that you
raise the resistance values for a start. (or balance the inputs)

This setup is not necessarily bad sounding, but it certainly
won't give you maximum power.

[jarek]

Thank you for the reply.
Hmm.., but I do not understand exactly what resistance you mean. Is it the problem with 'magic' resistors which are 4.7 kohm? If so, what value to start?
BTW, I have 18 V peak before clipping without load my PSU gives 14.4 V rail and only 10 V peak with 4 ohm load (my bias current is lowered to 1.6 A - 0,33 ohm bias resistors).
Regards

[Nelson Pass]

Specifically I mean the resistors (2 of them) each of which attaches from one amp output node to the Source pins of the input diff pair. They are used to reduce absolute DC drift, but they also load these Sources, so that some drive asymmetry is introduced with a single-ended input. This effect is reduced as you increase the values of these resistors.

I don't regard this asymmetry as much of a problem until the output amplitude is seriously different on both halves, cutting into the maximum output.

If increasing the value of the resistors gives you more absolute DC offset than you want after trimming, then think about output loading resistors, which load both outputs to ground through 20 to 100 ohms, also helping to stabilize the absolute DC.

Remember that absolute DC on the order of a volt or so is not a problem, as it doesn't appear across the load.