Need help w/ in-line attenuator values

[Mat86]

I have an issue with pre-amp and EQ noise on a couple setups and thought in-line RCA attenuators would be the solution to try. I'm building some cables and thought DIY'ing my own or building them into the interconnect cables would be the perfect solution. I figured either -6db or -10db might be the sweet spot, as muting is -20db and a bit overkill and I just want to lower the noise floor and compensate by boosting signal from my DAC and/or volume pot because on both I'm basically stuck at 9-o'clock

One online guide says 6.9k ohm + 4.7k ohm (to ground) would work for -10db, and another guide on youtube suggested 6.8k ohm + 4.7k as well (or 5.6k & 10k for -6db), however, a guide on the Hoffman forums says use 56k & 22k(!) and an old thread here said 47k + 47k (?)

What are the right resistor values to go with?

One amp is coming between an EQ and power-in, while the other is after the EQ in the tape loop. The second amp has a tape-input sensitivity of 150mV/47k ohm, and the EQ has an output impedance of 75ohms (stats)

 

[cademan]

I don't think that anything you are trying to do or propose will work or do any good. If you have an unusual louder than normal noise floor (hiss), it is probably due to bad regulator transistors or diodes in the power supply.

Unfortunately, any inline attenuators would also lower the overall signal.

 

[Mat86]

I don't think that anything you are trying to do or propose will work or do any good. If you have an unusual louder than normal noise floor (hiss), it is probably due to bad regulator transistors or diodes in the power supply.

Unfortunately, any inline attenuators would also lower the overall signal.

Lowering signal is what in-line attenuators do, fortunately.

 

[cademan]

Right, but I thought you were wanting to lower the noise floor which is hiss and usually caused by possible bad parts? Lowering hiss and signal output using resistors is like putting a band aid on it.

 

[elnaldo]

If you have an input impedance of 47K, I wouldn't wire a 10K attenuator in parallel. It would drop the input impedance too much. That's why some web sites suggested approx 100K (47+47K)

There is a relationship between voltage vs dB, so it's not difficult to calculate the desired attenuation using a voltage divider.

Or, solder a small 100K preset, and try different attenuations. Then you can replace it with 2 fixed resistors.

 

[rothwellaudio]

I don't think that anything you are trying to do or propose will work or do any good. If you have an unusual louder than normal noise floor (hiss), it is probably due to bad regulator transistors or diodes in the power supply.

Unfortunately, any inline attenuators would also lower the overall signal.

Right, but I thought you were wanting to lower the noise floor which is hiss and usually caused by possible bad parts? Lowering hiss and signal output using resistors is like putting a band aid on it.

Not true. If a piece of gear has a high noise floor it's possible to improve its signal-to-noise ratio by increasing the signal level (if there's enough headroom). Conversely, it's possible to degrade the S/N ratio by decreasing the signal level.
For example, a preamp with poor noise which has its volume control at the input can have its S/N ratio effectively improved by using the volume control at a higher setting and using a pair of attenuators at the inputs to the power amp. Having the volume control 10dB higher and offsetting that with 10dB attenuators at the power amp inputs improves the S/N ratio by 10dB simply by having 10dB more signal in the preamp.

 

[cademan]

Interesting!

But why not try and find the problem especially if there is noticeable hiss when volume is at zero?

Noisy transistors? Diodes?

Isn't this what the OP was talking about? If not then my mistake.

 

[rothwellaudio]

But why not try and find the problem especially if there is noticeable hiss when volume is at zero?
Noisy transistors? Diodes?

All electronic circuits produce noise. Even a resistor just lying on a bench produces a noise voltage. Getting a good S/N ratio requires good circuit design, though that shouldn't be too difficult, but a circuit doesn't have to be "faulty" to be noisy.
Take as an example a simple line level preamp. You could choose to use an op-amp in inverting mode and have a potentiometer as a volume control at its input, something like the circuit below.

You might decide it would be a good idea to have a high input impedance, so you choose Rin to be 1 meg. You might think it would also be a good idea to have plenty of gain, so you choose Rf to be 20 meg. You might also want a low current drain so you can run it from a battery, so you choose a low current op-amp.
The circuit would work, but without anything being "faulty" it would produce audible hiss.

 

[cademan]

Not always. I had a dumpster mixer board find a few years back that had bad white noise issues with any amplifier I used. All the amps were dead quiet with any other preamps and mixers used.

You only pointed out a general hiss problem that everything is supposed to have. Not buying it. Things are supposed to be quiet by design and if they are noisy by design, stay away from that particular brand or design. Nothing I have ever owned had a noise floor issue like that old mixer board I had. Tossed it right back into the dumpster I found it in.

Why does the OP have this issue where I have never seen this issue here before?

 

[rothwellaudio]

Not always. I had a dumpster mixer board find a few years back that had bad white noise issues with any amplifier I used. All the amps were dead quiet with any other preamps and mixers used.

You only pointed out a general hiss problem that everything is supposed to have. Not buying it. Things are supposed to be quiet by design and if they are noisy by design, stay away from that particular brand or design. Nothing I have ever owned had a noise floor issue like that old mixer board I had. Tossed it right back into the dumpster I found it in.

Why does the OP have this issue where I have never seen this issue here before?

Not always? Not always what?
Are you saying that most audio equipment has no noise, ie a S/N ratio of infinity?