Just calculating the resistor values of a 24-steps 100k DIY ladder type stepped attenuator. The problem is that I don't have any idea about the optimal maximum attenuation - I am building it for my tube headphone amp which I am using at fairly low volume levels.
My questions are:
1. Do I need 75-70dB maximum attenuation for a headphone tube amp or 65-60dB will be enough?
2. Should I expect problems with the input impedance and the amp performance as a whole if I build a 50k attenuator instead of 100k (my source is CD player)?
3. Does anybody know the attenuation range in dB for normal human live speech?
4. How can I calculate right dB steps for a 24 position attenuator - how did the most logarithmic steps look like?
I can modify the dB step for each position manually in my calculator but I don't know which is the right curve.
Thanks - appreciate your help!
My questions are:
1. Do I need 75-70dB maximum attenuation for a headphone tube amp or 65-60dB will be enough?
2. Should I expect problems with the input impedance and the amp performance as a whole if I build a 50k attenuator instead of 100k (my source is CD player)?
3. Does anybody know the attenuation range in dB for normal human live speech?
4. How can I calculate right dB steps for a 24 position attenuator - how did the most logarithmic steps look like?
I can modify the dB step for each position manually in my calculator but I don't know which is the right curve.
Thanks - appreciate your help!
How much attenuation you need depends on the sensitivity of you amp (and headphones). With my amp and speakers I need up to 60-65dBs of attenuation, but many others only need 40-50dBs. I would start be using a pot first. Set that to your normal listening level, and the measure/calculate the attenuation. Then you'll know what YOU need.
About the step size. 2dBs is often used for 24-step attenuators. So make sure you have the range you will normally need covered in the 2dB range, and you can then make 3-6dB steps outside that range.
50k will not cause problems for your CD-player (unless it has a high-impedance tube output stage). 20-25k would be even better.
Best regards,
Mikkel C. Simonsen
About the step size. 2dBs is often used for 24-step attenuators. So make sure you have the range you will normally need covered in the 2dB range, and you can then make 3-6dB steps outside that range.
50k will not cause problems for your CD-player (unless it has a high-impedance tube output stage). 20-25k would be even better.
Best regards,
Mikkel C. Simonsen
Before you build the attenuator, try a pot in the circuit with some clip leads so that you can quickly take it out and measure the values. At a distance, someone can tell you the theory, but not what will actually work for you in the gain structure and preferred listening levels of your system.
First try it with the volume as loud as you can bear it (make test short!). If it's less than full clockwise on the control then you have too much gain, ie 99% of audio systems. Let's say it calculates out to be 12dB or 4x. Then you know that your attn can be built with a series resistor to remove the unneccessary 12dB before it gets to your attn and wastes steps. In my 12dB example, if you used 2dB steps 6 would be useless to you.
You can also do the same at the bottom end of the range if you don't need the attn to be able to attenuate all the way to zero (use a mute switch for that) to save some more unneccessary steps.
A couple of hours experimenting before you build the attn will allow you to optimise it to what you actually need. And I would probably make the top and bottom 3 steps 3dB and the rest 2dB.
First try it with the volume as loud as you can bear it (make test short!). If it's less than full clockwise on the control then you have too much gain, ie 99% of audio systems. Let's say it calculates out to be 12dB or 4x. Then you know that your attn can be built with a series resistor to remove the unneccessary 12dB before it gets to your attn and wastes steps. In my 12dB example, if you used 2dB steps 6 would be useless to you.
You can also do the same at the bottom end of the range if you don't need the attn to be able to attenuate all the way to zero (use a mute switch for that) to save some more unneccessary steps.
A couple of hours experimenting before you build the attn will allow you to optimise it to what you actually need. And I would probably make the top and bottom 3 steps 3dB and the rest 2dB.
reaction: I hope I am not misunderstanding your question. In that case please forgive me. I am not well acquainted with the formulas - something to do with 20log, 1/(-dB/20) perhaps ? - but to clarify , there can be no absolute, fixed lower limit when we are talking about **attenuation** ("damping?" "dampening?" making it quiter, anyway) of sound in an amplifier, simply because it depends on how strong the signal is to begin with, how much it is amplified afterwards and how sensitive the headphones or speakers are, just as mcs said.
It is, however, rather common to use around -60 dB for the first position of an attenuator, unless the first position is used as "mute". Or -48 dB, on those systems with less gain or less sensitive speakers/phones. But if you are concerned about this value just follow the advice of mcs and brett: measure the resistance on a potentiometer volume and use that, after converting to -dB and then back to resistances through whatever spreadsheet you use. Sorry I do not know the formula.
P.S. Speaking of absolute dB:s, of course you could "reverse engineer" with a Sound Pressure Level meter too, at least with speakers. I remember a sound level comparison table in one of the physics books in school (oh, sweet youth!) where it said beside the 1 dB cell: "The sound of a falling leaf. Threshold of hearing" or something like that. Very poetic indeed. So if your SPL meter measures 1 dB at 1 meter from your speakers with the attenuator in its first position, you can safely say your system volume is exactly at the hearing threshold by one commonly used standard in audio.
D.S.
It is, however, rather common to use around -60 dB for the first position of an attenuator, unless the first position is used as "mute". Or -48 dB, on those systems with less gain or less sensitive speakers/phones. But if you are concerned about this value just follow the advice of mcs and brett: measure the resistance on a potentiometer volume and use that, after converting to -dB and then back to resistances through whatever spreadsheet you use. Sorry I do not know the formula.
P.S. Speaking of absolute dB:s, of course you could "reverse engineer" with a Sound Pressure Level meter too, at least with speakers. I remember a sound level comparison table in one of the physics books in school (oh, sweet youth!) where it said beside the 1 dB cell: "The sound of a falling leaf. Threshold of hearing" or something like that. Very poetic indeed. So if your SPL meter measures 1 dB at 1 meter from your speakers with the attenuator in its first position, you can safely say your system volume is exactly at the hearing threshold by one commonly used standard in audio.
D.S.
Hi,
Since most attenuators are no more than simple voltage dividers it's quite easy to add an extra divider (or several ones) to an existing setup by simply wiring up a switch with an extra 10 or more dB of attenuation.
Put it either in front or behind the attenuator, it will of course put an extra set of contacts in the direct signal path but I think it's the best solution for use with a headphone of such high sensitivity.
The first position of the attenuator (24 pos only gives you 23 pos of attenuation, 24 is often full out) shouldn't be sacrificed for muting duties, a simple shorting switch to pull output to ground should be adequate.
If you're concerned about signal integrity losses of the extra switch you can parallel contacts on a double switch and use one of those for each channel...
Or if you want total luxury and peace of mind you can throw in a few relay switches too.
Cheers,
Since most attenuators are no more than simple voltage dividers it's quite easy to add an extra divider (or several ones) to an existing setup by simply wiring up a switch with an extra 10 or more dB of attenuation.
Put it either in front or behind the attenuator, it will of course put an extra set of contacts in the direct signal path but I think it's the best solution for use with a headphone of such high sensitivity.
The first position of the attenuator (24 pos only gives you 23 pos of attenuation, 24 is often full out) shouldn't be sacrificed for muting duties, a simple shorting switch to pull output to ground should be adequate.
If you're concerned about signal integrity losses of the extra switch you can parallel contacts on a double switch and use one of those for each channel...
Or if you want total luxury and peace of mind you can throw in a few relay switches too.
Cheers,
Cliplead the Bourns out into the circuit. Adjust it until it is the loudest you can bear; be careful and respectful of your hearing so keep this short. Remove pot without adjusting the setting. Use a DVM to measure between the wiper and the top terminal (Rwt) and the wiper and the bottom terminal (Rwb).reaction said:
Attenuation as a ratio at that setting is Rwb/(Rwt + Rwb) = answer
In dB = 20 * log(answer)
OK, so lets assume you did the 'max volume' test and then measured the pot at Rwt = 70k and Rwb = 30k.
Attn ratio is 30/(30+70) = 0.3 = - 10.5 dB
Add a 330k resistor in series with the top terminal of the pot; at max cw rotation of the pot, you're now at the max volume you can use.
Listen to a few discs (and other sources you use) and for each one measure the Rwt and Rwb for each. Also try it, and measure at the lowest volume you could ever see yourself using. Measure Rwb; lets say it's 2k. Beacuse we've alreadt added a 330k to the top of the pot effectively making it a 430k pot, just multiply Rwb x 4 and solder an 8k resistor between the bottom terminal and ground.
Now you have a 440k composite 'pot' that at each end of the 100k variable part gives you your max and min listening levels.
So, what value do you want the whole pot assembly to be? To simplify work for yourself, I would suggest just using the 10k values that you can get from the Goldpoint site and add a 33k series resistor to the top and a 200 ohm to the bottom. Otherwise scale all the values you come up with in your measurements so that you get a 10k, 100k, 500k or whatever value load you wish the previous stage to see.
Don't waste a position on a 'zero' or off level, add a switch that shorts the VC output to ground when you want to mute the system.
This will give you a good idea of the 'range' of settings of the volume control that you'll actually use. Most people, most of the time actually only use a few of the max number of settings. For example, on my HT receiver with a 0-100 scale, I tend to use 30-36 for movies at night and 45-55 for annoying the neighbours type loud. Under 30 is too soft for me to use most of the time, and over 55 just means lots of distortion (too loud anyway), ie, I only use approx 25% of the available settings. The other thing I find way to often is that most people have too much gain in their system, and their constantly attenuating it away, and also that many are afraid to turn the volume control past about the 10 o'clock position. I've set mine up with 'enough' gain so that when I want it really loud, full rotation of the volume control gets me that. My VC in the music system only has 24 steps too (S&B TX102) so I needed to get the gain structure right if I wanted to have a useful VC, ie not just use two or 3 settings.
It will disappear in the off setting; where is 'too soft' depends on you, your preferences and ears as well as your 'phones and amp.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.