Look forwards to it Franco, he he. If you're right, you may cost me a lot of money in future.
Well, I use digital volume control plus (in one rig) digital crossover filters so the extra bit depth may be more important in my setup than in yours.
Speaking of items to achieve the top 0.1% of fidelity, if you want to control the buffalo via I2C from the BBB, then your variable grounding scheme could be beneficial. Best if that input is isolated as much as possible. The 9038 memory registers are not as stable as the older 9018s.
I can comment on this.
My case is sealed. I was using Placids sinked into the aluminum case. Heat was no problem. IIRCC, DAC was using 30 watts. Recently I changed DAC power supply to Centuar (not sinked to case). Now using 6 watts. DAC power supply is on 24/7, case is barely warm to the touch.
Ah quite possibly that is why. I'm old school and use a preamp (input selector+volume pot+buffer) as use several sources.
I will still do the test again however as bought a couple of albums in CD and high rez formats for this purpose. You can certainly hear a good remastering vs orginal on Redbook (e.g leftfild leftism original vs leftism 22 or Massive Attack Blue Lines, vs remastered; 2 of my favourites). Norah Jones and Moby on redbook vs 96khz or 48/24bit not appreciably. So I would spend the money on another album instead.
I will still do the test again however as bought a couple of albums in CD and high rez formats for this purpose. You can certainly hear a good remastering vs orginal on Redbook (e.g leftfild leftism original vs leftism 22 or Massive Attack Blue Lines, vs remastered; 2 of my favourites). Norah Jones and Moby on redbook vs 96khz or 48/24bit not appreciably. So I would spend the money on another album instead.
Not planning on controlling the Buffalo via I2C but good to note anyway.
I like things simple and robust; give me a button over a touch screen any day, dont give me access to information or controls that arent essential to operation.
I like things simple and robust; give me a button over a touch screen any day, dont give me access to information or controls that arent essential to operation.
So what usb card are you fellows using to connect the Buffalo Dac for streaming? Amanero?
I have been using the JLSounds card, and it seems that I will have to order my third one. It sounds pretty good but I am thinking that it is kind of finicky or unreliable now. My MacBook is not recognizing the card. Was working fine last time I used it 3 days ago. I’ll check the connections to the dac card. Anyway is the Amanero recommended or which other…. Thanks.
I have been using the JLSounds card, and it seems that I will have to order my third one. It sounds pretty good but I am thinking that it is kind of finicky or unreliable now. My MacBook is not recognizing the card. Was working fine last time I used it 3 days ago. I’ll check the connections to the dac card. Anyway is the Amanero recommended or which other…. Thanks.
Quite agree, and you are 100% in charge of complexity since you have the honor of writing the startup and/or I2C control script(s). 😀 Basically, you avail yourself of more operational modes, but no need for dinking around after you ID the best configuration. But no worries - the default onboard firmware is very good for general use.
Manolo, I would suspect the Mac as much as the i2soverusb card.
After (at least one) cold re-boot, go to the upper left apple -> About this Mac -> More Info -> System Report -> USB. Do you see the audio system listed as connected? If not, try unplugging the USB and re-connecting to the Mac while watching the information about the USB interface.
After (at least one) cold re-boot, go to the upper left apple -> About this Mac -> More Info -> System Report -> USB. Do you see the audio system listed as connected? If not, try unplugging the USB and re-connecting to the Mac while watching the information about the USB interface.
FWIW, I’ve had two Amaneros. One purchased in 2013 went bad. I Replaced it last year and it’s been working well.
Sorry, I have no Amanero experience. But I'm curious how you are powering the i2soverusb board. In four builds I've had no issues when powering them separately from the internal USB 5v. I just use a linear regulator on the input side, and a separate one on the isolated output side.
Meanwhile, I hope CED's project is going well...
Works as anticipated, so I'm pleased.
Fast boot feature works: ramps up to ~4v within.. 1/10th second..? on powerup before disengaging and letting the regulator noodle up to 5V operating point.
Trimpot sets output voltage with high degree of accuracy; 20 turns to set a 10v range.
DC output impedance is measured at a high-ish 0.8R but I can live with that. Worst case scenario, in this specific application and my planned configuration, is c. 225 mA max load for the Vd rail, feeding Vd, Vdd & Dvcc. That equates to 0.2v output drop from offload to max load.
The AD7151 regs are LDO, with a dropout at max load of ... 0.6v IIRC? So I can set these pre-regs to 5V offload to be safe and still be within a comfortable operating margin. Preregs themselves require about 5v in-out but my transformers are running at ~12-14v so again comfortable margins.
Additonally, dynamic output impedance is much better- falls rapidly above DC to ~ 0.03R by 10Hz (IIRC) where it levels off. Its based on a capacitance multiplier if the above didn't suggest that topology but i've refined the design over probably 20 years. Most schematics you see on this topology fail to really leverage the benefits and eliminate the weaknesses IMO.
Pros:
very stable- no oscillation even with no output capacitance, indifferent to the ufs if there are some.
good line and load rejection for a single stage regulator, acts pretty close a brick wall all things considered.
Very high bandwidth, far beyond any feedback-based design ive encountered.
very low noise.
Goodish dynamic output impedance
Cons:
slow rise time, measured in seconds.
highish DC ouput impedance
can drift a little under heavy thermal effects
As long as you understand what compromises are involved and what your performance requirements are, it can be an excellent solution. To my mind it should do what I want it to here; act as 'domestique' to the tridents:
block line noise getting to the AD7151 and stop digital hash getting back into onto line while providing a good low impedance low noise supply for the single Trident to work off. And most importantly wont oscillate when stuck in front of the trident's low impedance input caps.
I've tried the topology directly on AVcc in the past and it didnt work at all, totally dulled the sound. Seem to remember Russ mentioning very low impedance is required in this location. So its good where it is IMO.
next step: Build another 4 of the things. Takes about 40 mins to assemble and test each but I'm looking to be time poor over the next 2 weeks.
Fast boot feature works: ramps up to ~4v within.. 1/10th second..? on powerup before disengaging and letting the regulator noodle up to 5V operating point.
Trimpot sets output voltage with high degree of accuracy; 20 turns to set a 10v range.
DC output impedance is measured at a high-ish 0.8R but I can live with that. Worst case scenario, in this specific application and my planned configuration, is c. 225 mA max load for the Vd rail, feeding Vd, Vdd & Dvcc. That equates to 0.2v output drop from offload to max load.
The AD7151 regs are LDO, with a dropout at max load of ... 0.6v IIRC? So I can set these pre-regs to 5V offload to be safe and still be within a comfortable operating margin. Preregs themselves require about 5v in-out but my transformers are running at ~12-14v so again comfortable margins.
Additonally, dynamic output impedance is much better- falls rapidly above DC to ~ 0.03R by 10Hz (IIRC) where it levels off. Its based on a capacitance multiplier if the above didn't suggest that topology but i've refined the design over probably 20 years. Most schematics you see on this topology fail to really leverage the benefits and eliminate the weaknesses IMO.
Pros:
very stable- no oscillation even with no output capacitance, indifferent to the ufs if there are some.
good line and load rejection for a single stage regulator, acts pretty close a brick wall all things considered.
Very high bandwidth, far beyond any feedback-based design ive encountered.
very low noise.
Goodish dynamic output impedance
Cons:
slow rise time, measured in seconds.
highish DC ouput impedance
can drift a little under heavy thermal effects
As long as you understand what compromises are involved and what your performance requirements are, it can be an excellent solution. To my mind it should do what I want it to here; act as 'domestique' to the tridents:
block line noise getting to the AD7151 and stop digital hash getting back into onto line while providing a good low impedance low noise supply for the single Trident to work off. And most importantly wont oscillate when stuck in front of the trident's low impedance input caps.
I've tried the topology directly on AVcc in the past and it didnt work at all, totally dulled the sound. Seem to remember Russ mentioning very low impedance is required in this location. So its good where it is IMO.
next step: Build another 4 of the things. Takes about 40 mins to assemble and test each but I'm looking to be time poor over the next 2 weeks.
Missing standoffs and headers also came yesterday.
Annoyance to deal with: spotted a pcb error on the Trident mounting board, polarity of the Vd header is back to front, gah.
Typical. But at least I spotted it and everything else checks out. its not ideal but I'll cludge a workaround for now. Next time I do a board run I'll add a replacement board. it shouldnt matter but I know from experience if I have to service the unit in a (few) years time I'll not know what that was about and probably fry something.
Annoyance to deal with: spotted a pcb error on the Trident mounting board, polarity of the Vd header is back to front, gah.
Typical. But at least I spotted it and everything else checks out. its not ideal but I'll cludge a workaround for now. Next time I do a board run I'll add a replacement board. it shouldnt matter but I know from experience if I have to service the unit in a (few) years time I'll not know what that was about and probably fry something.
Might be able to grab half an hour tomorrow and solder up the mercury/Buffalo/trident stack as I'm working from home.
Did the reading on the mercury thread and have the salient points on assembling the board now as not been able to find a guide.
This is my dac project. A few years since completed. Started with the dac 18 board and then upgraded to the 38, as well as the IVY upgraded to the Mercury. The upper right hand small board is the JLsunds I2s to usb converter which I was alluding to in a previous post. Ordered a new one and everything is well. And, I was powering the 38 board with the HD 5v supply but I was not able to Upsample to DSD 256 without noise. So now the power supply board is the Centaur which gives the capability for DSD 256 upsampling. Hope this pictures and ced’s project motivate others to build this amazing dac at least until TPA stop selling the boards, etc.
Attachments
Thanks Franco, I just read through the mercury thread on this forum.
Really boiled down to which resistors and capacitors to use and what the power rails were.
I probably have OCD or something, but decided to go for 'in between' what TP offered.
So two +/- 15v rails is fine. Pretty standard.
output swing I actually want around 3v RMS as that matches my current setup so ordered some 90 ohm 0.1% resistors.
The analogue output filter I want an F3 of around 30khz, not the 20khz or 40khz options offered by the TP supplied parts.
I think TP have implemented a single pole filter but its not actually that clear to me. I'm guessing one set of caps to filter the balanced output and 1 for SE ouputs? A schematic sure would be nice here.
Ive ordered 3.3 and 1.5nf caps for anF3 roll off of 30khz anyway. Main reason being I have RFI filters on the input of my preamp and poweramp both set to 25khz IIRC. While my hearing tops out at 14khz I still don't want the tonal balance of my system messed with too much by attenuating the treble noticably, I spent a lot of time getting it right.
On the other hand I want the high frequency output hash of the dac filtered out like hell. My preference would have been for something like a 3 pole Bessel TBH but will see how it sounds and work from there.
Ive made allowances to add an analogue output filter after the dac too if necessary.
Really boiled down to which resistors and capacitors to use and what the power rails were.
I probably have OCD or something, but decided to go for 'in between' what TP offered.
So two +/- 15v rails is fine. Pretty standard.
output swing I actually want around 3v RMS as that matches my current setup so ordered some 90 ohm 0.1% resistors.
The analogue output filter I want an F3 of around 30khz, not the 20khz or 40khz options offered by the TP supplied parts.
I think TP have implemented a single pole filter but its not actually that clear to me. I'm guessing one set of caps to filter the balanced output and 1 for SE ouputs? A schematic sure would be nice here.
Ive ordered 3.3 and 1.5nf caps for anF3 roll off of 30khz anyway. Main reason being I have RFI filters on the input of my preamp and poweramp both set to 25khz IIRC. While my hearing tops out at 14khz I still don't want the tonal balance of my system messed with too much by attenuating the treble noticably, I spent a lot of time getting it right.
On the other hand I want the high frequency output hash of the dac filtered out like hell. My preference would have been for something like a 3 pole Bessel TBH but will see how it sounds and work from there.
Ive made allowances to add an analogue output filter after the dac too if necessary.