Interesting construction this, makes a lot of sense to wire it like you do. I am wondering a bit about the values selected for the resistors. Normal pots comes as you say most often in the range of about 20k to 100k, especially if there are tubes involved. This volume control will have as low as 1k and be below 2k for much of it's normal use range I would guess. So the source will see a load of 1k to 2k most of the time except on full where it will see 48k. I suspect 1-2 K input impedance could be problematically low for some equipment? Your tube riaa preamp for example , how does that fare with 1k load? Possibly quite well I guess since it has a nice buffer in form of a cathode follower as output, but it's output impedance would be in the vicinity of 18k I suspect? So any particular reason for that low resistance in the control, any downside to have a bit higher resistance in this network? Higher resistance could pick up a bit more noise I suppose...
It's great that someone is actually thinking about my design choices! First, yes the Stepped Attenuator may not work in every setup or play nice with every preamp (but that's common for dedicated auxiliary equipment). It's primary design goal was as a great sounding master volume control for our own Kit Systems. When I have time i'll design other resistor sets for a broader application. To achieve the best sound possible, I started with the theory that the resistor in the signal path should be as low a value as can possibly work (I always approach design work with this premise), and that value is 1k. What that means is the load impedance that the preceding stage sees is 1k + the stepped resistor (typically 40 - 100 ohms). Yes depending on how everything is wired up other resistor values are in parallel, but because they're all much higher (220k, 100k, 470k) and they effectively don't matter. Our preamp output impedances are as follows, E80cc 360R, Universal 6or12SN7 800R and the Phono Preamp 910R. So all of our Preamps have lower impedance and are basically impedance matched with our Stepped Attenuator. The resulting sonic improvement was simply stunning, it was like a veil had been lifted from the music and we had thought our reference system was close to sounding it's best . I credit the improvement on 4 design attributes. First a much lower resistor in the signal path (typically 70-100k), second very closely matched pairs of stepped resistors (0.1 - 0.5%), close matched impedance (though this i'm not certain about and would need to do some more bench and listening tests to determine if in fact it matters that much) and lastly no resistive trace (pot), which clearly colours the sound (in a good way for a digital source, not so for an analogue signal).
@@mellowtoneamps4932 Thanks for the great reply! Very interesting this design with such low impedance on the input with basically matched output and input impedance. I believe the conventional rule of thumb is to have at least 10 times higher input impedance than output impedance. Thus we seldom see less than 10K for volume pots. I guess you use 50 or 100K the alps pots in the preamps? I have actually built a version of your Phono preamp using a 7N7 I got from you but building it as a point to point construction, just done the first listening test and it sounds great, I just have a bit of hum pickup in one of the channels I need to work out a resolution to. Anyways I made a LTSpice simulation of it and tested the frequency response with various load resistances. Seems output impedance around 910 is right at least around and above 1kHz, but when we move down in the bass department (where I live, hehe, bass player you see) I found the output impedance increases , since the 2.2uF output cap will start to make a proportionally larger impedance contribution. At low B (38Hz) the impedance of 2.2uF is alone 1.9k so you get quite a bit of bass drop off (9 to 12 dB) gradually below about 100Hz according to the spice simulation at least. I think it would be good to test and listen a bit extra in that area? In simulation this effect was there with load resistance up to around 20k where it was just about less than a dB. I have a couple of very nice E80CC tubes i got from you also that I will build a pre of (currently used in first stage of my Crack hp amp) and I suspect I will build a stepped attenuator for it but I suspect I will opt for a higher series resistor to start with than you have settled for here, but must test if I can hear somthing that may convice me otherwise! Keep up the great work and service you provide!
One further note, be very careful of computer modelling programs, they can be less than accurate. We use them often, but one thing they are very bad at is modelling the correct electrical properties of specific tubes. When we are nearing completion of a new design we leave the computer completely out of the picture and switch to bench testing and critical listening. The bench testing will document the "actual" performance and of course the listening tests will determine if we're in the right path to great sound!
@@RaviNewfarm Set "B" is tested and the 10k input resistor tests good and makes for a good middle compromise between a low value of 1k and a high one at 100k. And best of all the Stepped Attenuator can basically go in any position in our system, though at the very end of the preamp chain and feeding the mono blocks gave the best overall performance. We still need to do more critical listening, but the first listens were really good, with excellent clarity and low distortion, and of course the lower frequencies are not rolling off in any significant amount. Testing showed the flatest signal response will always be with the alps pot in circuit and no stepped attenuator, but given the sonic improvement with the stepped attenuator in and the alps out, i think a small roll off in the very low frequencies is a good trade off.
No, it's proper title is a "Shunt Stepped Attenuator". IMO this is by far the best type for high quality audio. The one downside is you need to carefully balance the input resistor (which establishes most of the input impedance) with the output impedance of the connecting gear. To do that we designed a low impedance resistor "Set A" and a medium impedance "Set B".
Interesting construction this, makes a lot of sense to wire it like you do. I am wondering a bit about the values selected for the resistors. Normal pots comes as you say most often in the range of about 20k to 100k, especially if there are tubes involved. This volume control will have as low as 1k and be below 2k for much of it's normal use range I would guess. So the source will see a load of 1k to 2k most of the time except on full where it will see 48k. I suspect 1-2 K input impedance could be problematically low for some equipment? Your tube riaa preamp for example , how does that fare with 1k load? Possibly quite well I guess since it has a nice buffer in form of a cathode follower as output, but it's output impedance would be in the vicinity of 18k I suspect?
So any particular reason for that low resistance in the control, any downside to have a bit higher resistance in this network? Higher resistance could pick up a bit more noise I suppose...
It's great that someone is actually thinking about my design choices! First, yes the Stepped Attenuator may not work in every setup or play nice with every preamp (but that's common for dedicated auxiliary equipment). It's primary design goal was as a great sounding master volume control for our own Kit Systems. When I have time i'll design other resistor sets for a broader application. To achieve the best sound possible, I started with the theory that the resistor in the signal path should be as low a value as can possibly work (I always approach design work with this premise), and that value is 1k. What that means is the load impedance that the preceding stage sees is 1k + the stepped resistor (typically 40 - 100 ohms). Yes depending on how everything is wired up other resistor values are in parallel, but because they're all much higher (220k, 100k, 470k) and they effectively don't matter. Our preamp output impedances are as follows, E80cc 360R, Universal 6or12SN7 800R and the Phono Preamp 910R. So all of our Preamps have lower impedance and are basically impedance matched with our Stepped Attenuator. The resulting sonic improvement was simply stunning, it was like a veil had been lifted from the music and we had thought our reference system was close to sounding it's best . I credit the improvement on 4 design attributes. First a much lower resistor in the signal path (typically 70-100k), second very closely matched pairs of stepped resistors (0.1 - 0.5%), close matched impedance (though this i'm not certain about and would need to do some more bench and listening tests to determine if in fact it matters that much) and lastly no resistive trace (pot), which clearly colours the sound (in a good way for a digital source, not so for an analogue signal).
@@mellowtoneamps4932 Thanks for the great reply! Very interesting this design with such low impedance on the input with basically matched output and input impedance. I believe the conventional rule of thumb is to have at least 10 times higher input impedance than output impedance. Thus we seldom see less than 10K for volume pots. I guess you use 50 or 100K the alps pots in the preamps? I have actually built a version of your Phono preamp using a 7N7 I got from you but building it as a point to point construction, just done the first listening test and it sounds great, I just have a bit of hum pickup in one of the channels I need to work out a resolution to. Anyways I made a LTSpice simulation of it and tested the frequency response with various load resistances. Seems output impedance around 910 is right at least around and above 1kHz, but when we move down in the bass department (where I live, hehe, bass player you see) I found the output impedance increases , since the 2.2uF output cap will start to make a proportionally larger impedance contribution. At low B (38Hz) the impedance of 2.2uF is alone 1.9k so you get quite a bit of bass drop off (9 to 12 dB) gradually below about 100Hz according to the spice simulation at least.
I think it would be good to test and listen a bit extra in that area? In simulation this effect was there with load resistance up to around 20k where it was just about less than a dB.
I have a couple of very nice E80CC tubes i got from you also that I will build a pre of (currently used in first stage of my Crack hp amp) and I suspect I will build a stepped attenuator for it but I suspect I will opt for a higher series resistor to start with than you have settled for here, but must test if I can hear somthing that may convice me otherwise! Keep up the great work and service you provide!
One further note, be very careful of computer modelling programs, they can be less than accurate. We use them often, but one thing they are very bad at is modelling the correct electrical properties of specific tubes. When we are nearing completion of a new design we leave the computer completely out of the picture and switch to bench testing and critical listening. The bench testing will document the "actual" performance and of course the listening tests will determine if we're in the right path to great sound!
@@mellowtoneamps4932 Duly noted! Off to testing then 🙂
@@RaviNewfarm Set "B" is tested and the 10k input resistor tests good and makes for a good middle compromise between a low value of 1k and a high one at 100k. And best of all the Stepped Attenuator can basically go in any position in our system, though at the very end of the preamp chain and feeding the mono blocks gave the best overall performance. We still need to do more critical listening, but the first listens were really good, with excellent clarity and low distortion, and of course the lower frequencies are not rolling off in any significant amount. Testing showed the flatest signal response will always be with the alps pot in circuit and no stepped attenuator, but given the sonic improvement with the stepped attenuator in and the alps out, i think a small roll off in the very low frequencies is a good trade off.
This is a ladder attenuator?
No, it's proper title is a "Shunt Stepped Attenuator". IMO this is by far the best type for high quality audio. The one downside is you need to carefully balance the input resistor (which establishes most of the input impedance) with the output impedance of the connecting gear. To do that we designed a low impedance resistor "Set A" and a medium impedance "Set B".