I need a way to attenuate the differential signals from op amp I/V, I would rather digital volume control but this is not possible.
A good quality, low tolerance stepped attenuator would be the only option for accurate attenuation of differential signals. The issue is that this I/V will also be used to drive HPs directly and I want output impedance as low possible.
The only way I can think of achieving this is using a rotary switch like those used for stepped attenuators to change the value of feedback I/V resistors and manipulate output voltage.
I also like that this will not add anything extra to the signal path, similar to a digital attenuator .
While this would be done in a way that keeps feedback paths from op amp output to inverting input as short as possible it feels iffy having them connected in such a way...
Is this generally safe or likely to cause issues?
DAC will be AD1862
I/V will be OPA1622
A good quality, low tolerance stepped attenuator would be the only option for accurate attenuation of differential signals. The issue is that this I/V will also be used to drive HPs directly and I want output impedance as low possible.
The only way I can think of achieving this is using a rotary switch like those used for stepped attenuators to change the value of feedback I/V resistors and manipulate output voltage.
I also like that this will not add anything extra to the signal path, similar to a digital attenuator .
While this would be done in a way that keeps feedback paths from op amp output to inverting input as short as possible it feels iffy having them connected in such a way...
Is this generally safe or likely to cause issues?
DAC will be AD1862
I/V will be OPA1622
Last edited:
You'll want to look out for the parasitics (stray C, series L) of your stepped attenuator as they're quite likely to compromise the stability of the I/V opamp when the attenuator's used as the feedback path. Also make sure the attenuator's of the make-before-break type as you never want the opamp left open loop, even for under a mS as it'll swing to a rail and make a tremendous 'pop' in your ears.
Having a switched attenuator in this spot sounds like a bad idea.
You have not established the desired volume control range which will translate into resistor value range. Only then will you have sensible input data to discuss. It may turn out this fits within the range of an LDR or a switchable resistor matrix like the one inside Muse.
You have not established the desired volume control range which will translate into resistor value range. Only then will you have sensible input data to discuss. It may turn out this fits within the range of an LDR or a switchable resistor matrix like the one inside Muse.
make or break types seem to be common for these switches at least, the question is how long a cheap switch will last before the mechanism starts to fail and starts producing those pops
It wasnt fully decided if would use the AD1862 and I think now it wont be due to various factors.
The alternate option was building Ak4499 DAC in pin control mode, and these same requirements would apply there since pin control does not offer digital attenuation.
For the AK4499 the resistor values would be from ~300 ohm down, based on the DS using 360 ohm for OPA1612 in the recommended I/V circuit.
It wasnt fully decided if would use the AD1862 and I think now it wont be due to various factors.
The alternate option was building Ak4499 DAC in pin control mode, and these same requirements would apply there since pin control does not offer digital attenuation.
For the AK4499 the resistor values would be from ~300 ohm down, based on the DS using 360 ohm for OPA1612 in the recommended I/V circuit.
Doing that with AK4499 seems problematic. Why pin control mode? Its easier to use serial control mode, almost trivial, rather than to mess around with the output stage in a way that is unlikely to work well.
For the most part serial control seems to involve setting fixed 1/0 values for different options which probably wouldnt be too hard to learn, but the digital attenuator requires the DAC to be actively updated with a large range of values ... im imagining that would be pretty difficult for a beginner to implement, would it not?
Even though that would solve the issue for PCM, the most exciting feature of serial mode is DSD bypass mode, which faces the same issues of controlling volume as pin control mode.
An alternative to digital and typical analogue attenuation is still required
Pin control is also more foolproof when it comes to building your own PCB, I will design the PCB in pin control but make sure to pay attention to pins used or shared by serial control so it can be reconfigured for serial control if needs be.
Last edited:
Looking at page 41 of the AK4499 data sheet, it appears DSD is not available at all in pin control mode, never mind DSD volume bypass. Why would anyone want to give up DSD, that's the best part?
Also, using the digital volume control is simple enough. A rotary encoder or a pot could be used to set the volume level. Not much to it. There are existing Arduino libraries that simplify rotary encoder use.
Besides, if you need help setting up and using serial control mode, all you have to do is ask for some help right here. I would be happy to assist with how to set up AK4499 to sound quite good using some power supplies, serial control, and USB input. There may be others who have some experience with AK4499 that would be willing to help too.
I was never impressed with DSD on Sabre DACs, it sounds good (''full'' ''lush'' etc.) compared to PCM but seems coloured or a bit fake and that ultimately PCM was more transparent. Even software PCM upsampling had a similar issue compared to native PCM, and i noticed this on a wide variety of other DACs.
Native DSD did trump everything easily but its so expensive and limited.
AKM's DSD bypass mode is the only thing sustaining any interest in DSD for me.
Native DSD did trump everything easily but its so expensive and limited.
AKM's DSD bypass mode is the only thing sustaining any interest in DSD for me.
Last edited:
The sound of upsampled and converted DSD256 or DSD512 is dependent on a number of factors. It isn't all one sound, lush or whatever. Done with the best conversion quality available today it sounds better than native DSD I have here for testing. Takes a very powerful computer to do it at best quality, unfortunately.
However, could be if we had a native DSD dac (1-bit, with no remodulation) then fully-native DSD would always sound best. Don't know, since one is not here to compare.
I explored HQPlayer extensively, with a gaming PC I could go all the way to DSD512. DSD128 was best rate and default settings of polysinc filter and ASDM7 modulator were best (best as in least distortion/ most transparent).
The DAC was a topping D50 dual 9038Q2M.
When you say PCM to DSD is better than native DSD,
was this just with the Ak4499 or the es9038 aswell?
Anyway an Ak4137 would be required for DSD conversion as Im not using a PC at all anymore (hence not having software volume control).
With clean power and good clocks I think the AK4137 might be the key to getting superior sound from PCM to DSD conversions, especially if the Ak4137 can perform conversion to the similar level to HQP as you said before.
It would be wasted potential to do all that without having Ak4499 in bypass mode.
The DAC was a topping D50 dual 9038Q2M.
When you say PCM to DSD is better than native DSD,
was this just with the Ak4499 or the es9038 aswell?
Anyway an Ak4137 would be required for DSD conversion as Im not using a PC at all anymore (hence not having software volume control).
With clean power and good clocks I think the AK4137 might be the key to getting superior sound from PCM to DSD conversions, especially if the Ak4137 can perform conversion to the similar level to HQP as you said before.
It would be wasted potential to do all that without having Ak4499 in bypass mode.
Last edited:
The best HQ Player DSD results here are with ADM7 512+ and with the most computationally demanding filter I can get to run. That would be 40M Closed Form, or just Closed Form. The newer version of HQ Player has even more demanding filters that sound better yet. It is said that no computer can run them at DSD512 yet. Polysinc filters sound bad to me, but they are easier on the PC.
Seems reasonably close to how Chord DAVE with Prescaler works, except all its computation is done in FPGAs. About $25k for that setup. Can't afford anything like that here, but would sure like to hear one sometime.
AK4137 can sound pretty good although not as good as the best from HQ Player. As I have said before it was hard to get it sounding its best in combination with ES9038Q2M.
Maybe easier with AK4499, don't know. Would like to find out at some point. There is an AK4137 on AK4499 eval board, but only for use with SPDIF and TOSLINK inputs. Would have to hack into the PCB a little to bring in I2S.
Seems reasonably close to how Chord DAVE with Prescaler works, except all its computation is done in FPGAs. About $25k for that setup. Can't afford anything like that here, but would sure like to hear one sometime.
AK4137 can sound pretty good although not as good as the best from HQ Player. As I have said before it was hard to get it sounding its best in combination with ES9038Q2M.
Maybe easier with AK4499, don't know. Would like to find out at some point. There is an AK4137 on AK4499 eval board, but only for use with SPDIF and TOSLINK inputs. Would have to hack into the PCB a little to bring in I2S.
Last edited:
Not a good ideal for insert a VR in a high speed feedback loop.
More information in
Best opamp for I/V conversion? (DAC)
More information in
Best opamp for I/V conversion? (DAC)
Ok, there is a better way to do this at the cost of being kind of inconvenient as a volume control.
Instead of a switch I will use a 2x4 pin header socket joined to the 4 op amp outputs and 4 DAC outputs, for a given volume level there will be a 2x4 pin header plug with the I/V resistors in place to link each pair of pins.
As mentioned already the max value I/V resistor will bypass the pin header to avoid pops when changing resistors, only music must be paused to avoid playing at max volume.
Instead of a switch I will use a 2x4 pin header socket joined to the 4 op amp outputs and 4 DAC outputs, for a given volume level there will be a 2x4 pin header plug with the I/V resistors in place to link each pair of pins.
As mentioned already the max value I/V resistor will bypass the pin header to avoid pops when changing resistors, only music must be paused to avoid playing at max volume.
Last edited:
What about adjusting the opamp input offset voltage (Vref) for minimum distortion for each I/V resistor? Another jumper for that?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.