The only problem I have with Peter idea is the 50K value. I don't know how that is going to jiggle with the AX gain (ie making it too large).
If this is a problem, it could be avoided by using 2 separate attenuators (one for ea. channel), and paralleling 2 sections together. In this way one could still get 2 gang pot with 25K value. And this should be much easier on Alephs input😉
Peter, do you mean use the attenuator in place of R19 and R29 on the PCB? What about the resistors in series with the input, R18 and R28, leave thease as they are? (all of these are 10K on the PCB)
BTW, the DAC I'm using is by jwb, here's the schematic:
http://atari.saturn5.com/~jwb/43122.png
The CS43122 has voltage output, no I/V, just the IRF610 based follower. Since the follower has a DC offset, I'm using DC blocking capacitors at the output, and the reason I was trying to get a high input impedance on the amp is that I can't fit any Auricaps larger than about 10uF in the enclosure.
So basically I'm trying to figure out the best way to get volume control. Maybe put the attenuators in the DAC, before the follower? But I don't know if the CS43122 can drive them.
I've also considered using smaller caps in front of the follower and biasing the gates with a voltage divider to zero the DC offset, but that could drift over time.
BTW, the DAC I'm using is by jwb, here's the schematic:
http://atari.saturn5.com/~jwb/43122.png
The CS43122 has voltage output, no I/V, just the IRF610 based follower. Since the follower has a DC offset, I'm using DC blocking capacitors at the output, and the reason I was trying to get a high input impedance on the amp is that I can't fit any Auricaps larger than about 10uF in the enclosure.
So basically I'm trying to figure out the best way to get volume control. Maybe put the attenuators in the DAC, before the follower? But I don't know if the CS43122 can drive them.
I've also considered using smaller caps in front of the follower and biasing the gates with a voltage divider to zero the DC offset, but that could drift over time.
Why do I like higher resistances for pots? Easy. I come from a background of messing with tubes. Tube circuit pots tend towards 100k or more, which looks really bizarre to solid state folks. To me it looks normal.
Like so many other things, it's a tradeoff. Lower resistances tend to react less with capacitance and/or inductance in the cables or circuits. Higher resistances place less load on the source. Since tubes do voltage easily and tend to be light in the current department, you use higher resistances.
I have a few more minutes now than when I posted before, so let me fill in some gaps. Speaking personally, I prefer a preamp with gain devices of some sort inside. That doesn't mean that I can't understand the attraction of a passive preamp. (Am I the only one who finds the term "passive preamp" to be an oxymoron?) The pro side of the equation with a passive preamp/volume control is that there are no gain devices to gunk up the signal. Fair enough. The con side is that the output resistance is often higher than you might like, leading to cable interactions.
There's an easy solution, providing that you're willing to forego having a separate box for the volume knob--put the pot in the amp. That will knock out the need to buy one set of cables, one set of jacks, and reduce interactions, all in one go.
Whether to go high or low R is up to the user. Note that I didn't say not to use 10k--it'll work just fine; and yes, you can go lower still if you've got plenty of current capability in your source. The problem is that so many CD/DVD players use lightly biased opamps at their outputs that I'd recommend taking a look inside your player to see what you've got before loading it down with a really low resistance pot. Some players' sound deteriorates if you ask them to generate too much output current.
This is kinda my version of Nelson's preference for efficient speakers. I like "efficient" preamps, crossovers, amps, etc. Something that develops more voltage swing for the same amount of current output from the preceding device.
As I believe grataku said, go ahead and try it. Try both ways, for that matter.
Grey
Like so many other things, it's a tradeoff. Lower resistances tend to react less with capacitance and/or inductance in the cables or circuits. Higher resistances place less load on the source. Since tubes do voltage easily and tend to be light in the current department, you use higher resistances.
I have a few more minutes now than when I posted before, so let me fill in some gaps. Speaking personally, I prefer a preamp with gain devices of some sort inside. That doesn't mean that I can't understand the attraction of a passive preamp. (Am I the only one who finds the term "passive preamp" to be an oxymoron?) The pro side of the equation with a passive preamp/volume control is that there are no gain devices to gunk up the signal. Fair enough. The con side is that the output resistance is often higher than you might like, leading to cable interactions.
There's an easy solution, providing that you're willing to forego having a separate box for the volume knob--put the pot in the amp. That will knock out the need to buy one set of cables, one set of jacks, and reduce interactions, all in one go.
Whether to go high or low R is up to the user. Note that I didn't say not to use 10k--it'll work just fine; and yes, you can go lower still if you've got plenty of current capability in your source. The problem is that so many CD/DVD players use lightly biased opamps at their outputs that I'd recommend taking a look inside your player to see what you've got before loading it down with a really low resistance pot. Some players' sound deteriorates if you ask them to generate too much output current.
This is kinda my version of Nelson's preference for efficient speakers. I like "efficient" preamps, crossovers, amps, etc. Something that develops more voltage swing for the same amount of current output from the preceding device.
As I believe grataku said, go ahead and try it. Try both ways, for that matter.
Grey
Yes, I will put the volume control in the amp, I never intended to have long cables following a passive attenuator. I can drive 10K and above, or even lower if I figure out how to deal with the DC blocking capacitors (see my previous post).
One thing that worries me with the high resistance pot is the statement on the DACT site: "the load capacitance combined with the volume control output resistance makes a low pass filter that cuts away treble". They have a spreadsheet here http://www.dact.com/html/ac_calculator.html
to determine attenuation by frequency band but I don't know the input capacitance for the Aleph-X to put in the spreadsheet, and I can't even use the spreadsheet since I don't have Excel and OpenOffice can't import passworded files...
So is Peter's idea (replacing the 10K shunt to ground resistors with the volume control) the least likely to mess up sound?
One thing that worries me with the high resistance pot is the statement on the DACT site: "the load capacitance combined with the volume control output resistance makes a low pass filter that cuts away treble". They have a spreadsheet here http://www.dact.com/html/ac_calculator.html
to determine attenuation by frequency band but I don't know the input capacitance for the Aleph-X to put in the spreadsheet, and I can't even use the spreadsheet since I don't have Excel and OpenOffice can't import passworded files...
So is Peter's idea (replacing the 10K shunt to ground resistors with the volume control) the least likely to mess up sound?
Prune said:
Yes, you actually don't change anything in the amp, just substitute fixed resistor with a variable attenuator of the same value. Isn't it a very elegant approach?😉
Thanks! I'm not completely sure how to wire the pot, however. One end to ground, the wiper to R18(28), and the other end not connected? And should I use a linear or log attenuator?
You use log pot. One pin to input (RCA in your case), center pin (wiper) to output (amps input) and third pin to ground.
Peter Daniel said:
The original schematic had 68.1K instead of 10K for R19(29). So I don't see how using a more common pot value of say 25K would be a problem.
A different volume control
Probably is out of this discussion, but what do you think to mak a volume control like the D1 analog stage?
If you insert a pot between q5 and q7 sources you can subtract the two signals (opposite in phase) and then decrease the output.
Is it a bad idea?
Thank you for responses.
Ciao!
Probably is out of this discussion, but what do you think to mak a volume control like the D1 analog stage?
If you insert a pot between q5 and q7 sources you can subtract the two signals (opposite in phase) and then decrease the output.
Is it a bad idea?
Thank you for responses.
Ciao!
Faber:
Do you mean to insert a pot between the Q5 and Q7 *drains*, not sources? If so, then in theory this works by decreasing the open loop gain of the amp, but since we're not talking about an open-loop design like the D1, this idea gets messed up by the feedback trying to un-do what you've just done with the pot. Also, when the pot is set to a very low value, you'll mess up the DC operating point in the amp, probably leading to DC across your speakers (!). So, this idea will not work well with the A-X.
Peter:
Take a second look at that last schematic posted by Prune above, and ask yourself what happens to the feedback when it's set to either full volume, or zero volume... this is not a good idea! For the same reason, it will not give you proper operation anywhere except very close to the (resistive, not mechanical) center of the pot's range.
Fred's post near the start of this thread perfectly explains why not to use a high-value pot in the conventional sense, at the input to the Aleph-X. The Aleph-X is happiest being driven by a low impedance source without any modifications to it's feedback resistor network, and yes, R18 and R28 are part of that network, as is the source impedance!
Do you mean to insert a pot between the Q5 and Q7 *drains*, not sources? If so, then in theory this works by decreasing the open loop gain of the amp, but since we're not talking about an open-loop design like the D1, this idea gets messed up by the feedback trying to un-do what you've just done with the pot. Also, when the pot is set to a very low value, you'll mess up the DC operating point in the amp, probably leading to DC across your speakers (!). So, this idea will not work well with the A-X.
Peter:
Take a second look at that last schematic posted by Prune above, and ask yourself what happens to the feedback when it's set to either full volume, or zero volume... this is not a good idea! For the same reason, it will not give you proper operation anywhere except very close to the (resistive, not mechanical) center of the pot's range.
Fred's post near the start of this thread perfectly explains why not to use a high-value pot in the conventional sense, at the input to the Aleph-X. The Aleph-X is happiest being driven by a low impedance source without any modifications to it's feedback resistor network, and yes, R18 and R28 are part of that network, as is the source impedance!
Is it then not possible, by adding resistors in series with the pot, to both restrict the range enough to preserve proper operation, and to still have enough range to be useful as a volume control?
BTW is it necessary to use a capacitor in front of the Aleph-X? I would rather have DC coupling and put a DC servo on my DAC output, so I don't have to use large capacitors.
I was thinking of changing the output of my DAC like this (sorry for the messy drawing):
I was thinking of changing the output of my DAC like this (sorry for the messy drawing):
Prune:
If you DC couple the inputs to the Aleph-X, the DC offsets in the amp will be larger (approximately double) than if the inputs are cap coupled.
As for adding some series resistors with the pot.. you'll still be varying with the feedback network as you turn the pot, and it's just not a good idea, IMHO. I think Fred already posted a good explanation. With the fixed resistors, you will also limit the range of volume control you have, and I don't know about you, but I wouldn't be OK with that... mainly I listen at moderate volume, but there are times I really wanna CRANK it, and times I want it to be whisper quiet. With fixed resistors added, you'll lose mostly on the quiet end, since volume is a logarithmic thing.
If you really must add an unbuffered volume control at the input of the Aleph-X, then just do a standard arrangement, with the input feeding one end of the pot, the other end grounded, and the wiper feeding the unmodified 10K input of the Aleph-X. Keep the pot value as low as feasible. I still find this a conceptually substandard idea, especially considering you've got suitable power supply voltages readily available for a simple FET follower or other buffer circuit.
If you DC couple the inputs to the Aleph-X, the DC offsets in the amp will be larger (approximately double) than if the inputs are cap coupled.
As for adding some series resistors with the pot.. you'll still be varying with the feedback network as you turn the pot, and it's just not a good idea, IMHO. I think Fred already posted a good explanation. With the fixed resistors, you will also limit the range of volume control you have, and I don't know about you, but I wouldn't be OK with that... mainly I listen at moderate volume, but there are times I really wanna CRANK it, and times I want it to be whisper quiet. With fixed resistors added, you'll lose mostly on the quiet end, since volume is a logarithmic thing.
If you really must add an unbuffered volume control at the input of the Aleph-X, then just do a standard arrangement, with the input feeding one end of the pot, the other end grounded, and the wiper feeding the unmodified 10K input of the Aleph-X. Keep the pot value as low as feasible. I still find this a conceptually substandard idea, especially considering you've got suitable power supply voltages readily available for a simple FET follower or other buffer circuit.
- Status
- Not open for further replies.