T Attenuator Termination Kluge

[mediatechnology]

-56 dB at cutoff seems high to me.

[Gold]

-56 dB at cutoff seems high to me.

You would expect more attenuation at the fully CCW position right?

[Radardoug]

It could be your board, and the fact you are going balanced to unbalanced and back. But I would have thought the last position of the attenuator would be an off position, with very high attenuation.

[Gold]

I figured it out. The 10K TKD didn't like the 5K termination resistor for the T attenuators. When I bypassed the termination resistor I got over 80dB cutoff. I'll have to think about the wiring now. If I put the TKD before the 5K Bridged T switches they will still see a 5K load, so that probably won't work. I guess I could terminate the 5K Bridged T at the attenuator. Feed the TKD pot after the termination and use a higher value termination resistor on the PCB's.

[Gold]

It turns out it had nothing to do with the termination resistor. There were a few things that needed sorting out. The major thing was a different grounding scheme. Now I get cutoff better than the Portable One can measure.

I had to put the TKD pots before the Bridged T attenuators. The 12Ω-14Ω series resistance was enough to throw the gain steps off by too much. Now with the pot before the attenuators I'm getting gain steps that are within a few hundredths of a dB accurate. Channel matching when using the pot looks like it's within about 0.2dB over the whole throw.

[Radardoug]

If at first you dont succeed, keep at it! Well done!

[Gold]

Thanks. I’m happy with it. Crosstalk -90dBu at 20K. The input is AC coupled. After that everything is DC coupled. Now I have to put the new pot in the other Output Amp.

[emrr]

Good news. T attenuators get screwy so easily if the surroundings are funky. I have measured so many odd things.

[Gold]

I'm seeing something I don't understand. When feeding a 1K 0dBu tone in, with 11db of gain (zero attenuation with the attenuators) and the fader fully CCW I get a kill below the measurement range of the Portable One. With the fader still all the way down I get a kill of -86dBu with the attenuators at -3dB (8dB of gain). When the attenuators are at 0dB on the dial (-11db gain) I get a kill of -79dBu. It stays at that value for the rest of the switch positions. I get more kill with 11dB gain than I do with -11dB. -79dBu isn't a deal killer but if I can improve it I'd like to.

[Gold]

I spent some time trying to improve this. I noticed there was 20KΩ to ground at the Unbalanced Output feeding the Fader-Attenuator. I sub'ed in a couple of values and saw a pattern. Then I put a rheostat in place of the resistor and adjusted for the deepest cut. Curiously the value was 2k5Ω. Half the termination resistance. Then I tried changing the 5KΩ termination resistor to 2k5Ω. Bingo. I have a 6dB improvement now with a cut of -85dBu at the "0" position (11dB of attenuation). Close enough for government work.