Noob question about stepped attenuators

[ohshitgorillas]

Finally got around to putting this in last night.

It is way overkill. Even the most compressed recordings I have, on high gain, the first half of the steps available on the stepped attenuator are just too quiet. So at any given gain, I am basically limited to the last 11 steps of the attenuator. Maybe 0, -6, and -12 would have been a little more appropriate, but I will give it some time.

On the other hand, I have a lot of steps of really fine volume control, which is pretty nice

[ohshitgorillas]

I've decided that, while this is definitely a step in the r ight direction, I want to keep trying for a solution that has a more even spread of volumes; I'd rather have headroom on the volume control (which I don't have much of now, altho the amp does get loud enough), rather than a lot of quiet steps at the beginning of the attenuator that I'm not using. I also think that instead of -6 dB, a 0dB setting would be more ideal.

Perhaps a 10k attenuator and a different orientation of resistors on the gain switch? Or would that be a step in the opposite direction?

Or a 100k attenuator with 0, 33k and 75k ohm resistors and whatever matches on the ground side to keep load constant? (Once this is settled, I want to try my hand at building a small crossfeed circuit for the Crack, the design of which is apparently source and load dependent) Is the load ideal at 100k or would 175k total be okay?

I'm beginning to have a better understanding of these things but there are still some things about this which are a bit hazy.

[Paul Birkeland]

The easiest thing to do would be to try running a jumper wire across each of those 50K resistors, then see how you like that.

-PB

[neddoge]

Following

[ohshitgorillas]

Does that put the load at ~150k or is it still 100k?

I will give that a shot when I get the chance, thank you

[Paul Birkeland]

It drops the impedance a bit.

[ohshitgorillas]

I used an alligator clip between the two 50k resistors. I hooked them up on the switch side of the resistors. It didn't do really anything to spread out the volume control but nuked the bass. I tried hooking it up to the signal side of the resistors, but of course this just mixed things down into mono and also did nothing to improve the spread of volume control.

[ohshitgorillas]

I am also wondering if there is a good website you know of or a way for a newbie like myself to mathematically understand the relationship between gain and load, and also between resistance and gain e.g. if I wanted to do a 100k pot with 0, -12, and -24 dB, how would I go about selecting resistances to match those gain values and how would I go about selecting

Parts of the schematic you drew me made sense. The resistance increasing from 50k to 75k to 94k as gain increases is easy to follow. But why does this add a small amount of resistance to low gain, a higher amount to medium gain, and none to low gain? This part makes no sense to me.

I want to understand, and it seems relatively simple, but obviously I'm missing some basics.

[diynewbie]

I'm also watching this with interest.

Your posts reads like you ran a wire from the left to right channel on each side of the switch.  I believe the suggestion was to by-pass the 50k resistor on each channel - i.e a connect a wire from the input side of the 50k resistor to between the 50k and 25k resistors.  Is that what you did?

[ohshitgorillas]

That makes so much more sense... I will try that tonight

[Paul Birkeland]

I am also wondering if there is a good website you know of or a way for a newbie like myself to mathematically understand the relationship between gain and load, and also between resistance and gain e.g. if I wanted to do a 100k pot with 0, -12, and -24 dB, how would I go about selecting resistances to match those gain values and how would I go about selecting

You'll have to read up on Ohm's Law and how parallel and series resistors work, then look up the formula for calculating attenuation based on voltage division (in this case resistance can be used as a direct substitute).

-PB