Sorry to say, but that can still be anything. I had a buzzing that changed with the volume of the power amp (only when the preamp was connected, didnt matter If something or nothing on the pre connected) drove me crazy. Turned Out to be my interconnect was the culprit. But only in that combination of preamp and Power amp. That just as an example. Also:
Ground loops can be tricky although the concept seems easy. I would try all the different tricks to identify a Ground Loop to be sure. Or try cheapo transformers/caps first, it will be heartbreaking if you get the good stuff and then the hum isnt gone ^^
Greetings
Oli
Ground loops can be tricky although the concept seems easy. I would try all the different tricks to identify a Ground Loop to be sure. Or try cheapo transformers/caps first, it will be heartbreaking if you get the good stuff and then the hum isnt gone ^^
Greetings
Oli
If this headphone amp is not true balanced and only has a XLR input and after that a volume pot directly to ground this will happen.
so what headphone amp is this? (circuit?)
We use close to ideal clocks in real devices too
All I wrote here is about PCM, 64fs DSD has a different noise spectrum, depending on implementation, maybe resampling can really bring something here.
As of a few months ago, I have been using Holo Audi May L2 DAC and HQ Player. I was a DSD non -believer, but this combination changed my opinion completely.
There are 2 modes of operation that sound truly fantastic: no processing of any kind, i.e. just raw PCM fed to this R-2R DAC. The other is DSD.... which also sounds exceptional; very relaxed, detailed, natural, extended, spacious - a true bliss.
So, apart from commenting on your observation re DSD, another reason why I am posting this here is an interesting fact that I noticed when I started to compare (heavily modified) DDDAC to Holo May DAC.... they have identical sound signatures; the unmistakeably nice sounding (to me) NOS signature. The reasons why I decided to move away from DDDAC are:
- (due to its true NOS nature and no digital/analog filters) high-frequency roll-off
- distortions that (since I built a very revealing Aleph J) started to really bother me big time.
Also, with the amount of money I poured into making my DDDAC sound good (all in power supply and voltage regulation), I was approaching the point where I started to consider a great-measuring, nice-sounding R-2R NOS DAC, so the May was a no-brainer for me. It is worth its weight (sound
) in gold!!!Yes, I wouldn't want to do that! Thanks for sharing your experience!
It's a Burson Soloist 3XP, I chose it based on my earlier posts as it's apparently AC-coupled. I dont believe it is 'true' balanced design. The pre-amp / volume is based on the MUSES72320 chip and uses an opamp input buffer around it. So it's not a traditional volume pot, and I obviously cant get hold of a circuit diagram to check the grounding. If anyone has any ideas, Id be really grateful!!!
You clearly wrote what I only warily mentioned
For DSD HQPlayer (upsampling+high precision filtering)+FPGA+resistors (for DSD not R2R really) is the way to achieve a level of ultra-upgraded DDDAC. At least two ideas from DDDAC can be used too, so as DIY it's possible to make it better than May DAC. The reason why I did not write about it, is a problem with this solution. Personally, I think HQ Player interface is far away from the best choice, but it's not the biggest problem here.
Well, even as a DIY project you need to spend a lot of time and money, and your project will be completely dependent on one person. Sources are closed, and I think you can't buy sources from the author with a reasonable price too. And no commercially available alternative of HQ Player. At least, I don't know of it. If you know an alternative solution, it's exceptionally interesting.
Sorry, Doede, but I write it here.
If one day Doede or Lucian or Ian will say what they are not interested in DDDAC (WaveIO, FIFO, etc) anymore (I hope this day will never come
, the community can support it, small modules can be replaced, we have a choice. Many people here, including me, invested a lot into this project, and we will never lose our investments. It’s great!With a project dependent on one software developer you can lose all in one day. And the question is: are you really ready to accept this risk, or not.
Returning to the topic, DDDAC is a great SAFE investment
I don't see why NOS would involve high-frequency roll off. The DAC just converts the digital input data to an analog signal. There is no high-frequency roll-off involved in NOS.
Also, the DDDAC non-linear distortion is waaaay lower than the Aleph-J. Also, I wouldn't call the Aleph as "very revealing". It's a good amp, but on the softer side of neutral.
Since you seem to move away from the DDDAC, can I ask you to buy your DDDAC stuff from you? If you're interested, please send me a PM.
I agree with Matthias, and I don’t understand the final comparison too. If DSD -> on-fly to PCM-> DDDAC vs DSD->HQ Player->FPGA->multi-bit paralleled R segments, it's one thing. But it's not a problem of DDDAC in any case.
If NOS PCM -> multi-board DDDAC with a good resistor and transformer vs NOS PCM -> FPGA -> single R2R -> some sort of active+capacitor output (even though according to the photos of the May DAC, it is many times better than "some R2R DIY DAC"), it's really strange for me how the latest can win.
Hello,
It seems like there are a few people beating around the bush concerning the DDDAC.
Doede conceived this little baby with a multitude of ideas to arrive at its present level.
As far as mainboard and the little tower there are very few tricks one could do. Change IV resistors, replace some regulators, MAYBE Tent shunt for 3,3 volt as well.
The power supplies can be improved with a moderate investment.
I just saw Audio Creative added a new output transfomer!!
Doede himself also added some things to supply a '' nicer input signal'' to the mainprint.
To me this is still a can of worms so i am still using Wave IO plus Aurender.
The DDDAC like i constructed it was just mainboard, 4 boards on top, two choke input supplies, better resistors, Sowter, Wave IO and some belleson regulators.
The list of things to get in order to replace the Aurender is longer and i would need a dummies book or a well written manual to actually make it. Just one little mistake and it wont work at all or it will work but without all the benefits you should get.
Greetings, Eduard
It seems like there are a few people beating around the bush concerning the DDDAC.
Doede conceived this little baby with a multitude of ideas to arrive at its present level.
As far as mainboard and the little tower there are very few tricks one could do. Change IV resistors, replace some regulators, MAYBE Tent shunt for 3,3 volt as well.
The power supplies can be improved with a moderate investment.
I just saw Audio Creative added a new output transfomer!!
Doede himself also added some things to supply a '' nicer input signal'' to the mainprint.
To me this is still a can of worms so i am still using Wave IO plus Aurender.
The DDDAC like i constructed it was just mainboard, 4 boards on top, two choke input supplies, better resistors, Sowter, Wave IO and some belleson regulators.
The list of things to get in order to replace the Aurender is longer and i would need a dummies book or a well written manual to actually make it. Just one little mistake and it wont work at all or it will work but without all the benefits you should get.
Greetings, Eduard
Are you talking about replacing the WaveIO with a Raspberry Pi and a reclocker? I know that may sound like a big can of worms, but if you stick to the tried and tested solutions it will work without too much hassle. I can provide some pointers or even step-by-step instructions if you want.
This is what I measured with my DDDAC
fS=44.1kHz: 0dB @ 20Hz = -3.2dB @ 20kHz
fS=88.2kHz: 0dB @ 20Hz = 0.8dB @ 20kHz
Others have explained it a number of times, so I will not waste time... Ken Newton explained it here:
https://www.diyaudio.com/forums/dig...4-frequency-response-nos-dac.html#post5614454
"That roll-off is normal for NOS DAC operation to exhibit. Theoretically, it should be -3.16dB @ 20KHz rather than -2.5dB, assuming a 44.1KHz NOS sample rate. The audible issue with the roll-off isn't so much the drop at 20KHz, but the fact that it colors down to 5KHz, two full octaves.
Technically, the roll-off stems from the SINC function response mask created by the sample-and-hold (stair stepped) output of the DAC chip, which then mildly suppresses the highest frequencies. This supression varies with the sample rate sent to the DAC quantizer unit. The higher the sample rate, the less audible the suppression, because it consequently gets shifted higher in frequency. By the way, digitally filtered DACs also exhibit this roll-off, but oversampling increases the sample rate, making it less audible. Any remaining roll-off is then EQ'd to flat within most digital audio oversampling filters.
Equalizing the roll-off is correct to do and improves the tonal balance. There are several ways to perform the EQ in the analog stage, depending on your specific stage implementation."
Re: DDDAC distortions, it has to process the upsampled material to address the above roll-off; the high sample rate material exhibits higher distortions when processed by DDDAC, which has been a problem (for me at least) since I built it.
Re: Aleph J.... the original design and implementation can be vastly improved to get truly great 2-gain stage amplifier; try searching the Aleph J thread for Aleph J MKII; I have provided detailed instructions and the modified circuit diagram.
fS=44.1kHz: 0dB @ 20Hz = -3.2dB @ 20kHz
fS=88.2kHz: 0dB @ 20Hz = 0.8dB @ 20kHz
Others have explained it a number of times, so I will not waste time... Ken Newton explained it here:
https://www.diyaudio.com/forums/dig...4-frequency-response-nos-dac.html#post5614454
"That roll-off is normal for NOS DAC operation to exhibit. Theoretically, it should be -3.16dB @ 20KHz rather than -2.5dB, assuming a 44.1KHz NOS sample rate. The audible issue with the roll-off isn't so much the drop at 20KHz, but the fact that it colors down to 5KHz, two full octaves.
Technically, the roll-off stems from the SINC function response mask created by the sample-and-hold (stair stepped) output of the DAC chip, which then mildly suppresses the highest frequencies. This supression varies with the sample rate sent to the DAC quantizer unit. The higher the sample rate, the less audible the suppression, because it consequently gets shifted higher in frequency. By the way, digitally filtered DACs also exhibit this roll-off, but oversampling increases the sample rate, making it less audible. Any remaining roll-off is then EQ'd to flat within most digital audio oversampling filters.
Equalizing the roll-off is correct to do and improves the tonal balance. There are several ways to perform the EQ in the analog stage, depending on your specific stage implementation."
Re: DDDAC distortions, it has to process the upsampled material to address the above roll-off; the high sample rate material exhibits higher distortions when processed by DDDAC, which has been a problem (for me at least) since I built it.
Re: Aleph J.... the original design and implementation can be vastly improved to get truly great 2-gain stage amplifier; try searching the Aleph J thread for Aleph J MKII; I have provided detailed instructions and the modified circuit diagram.
Gaming with math and digital equalizer to achieve -0.00db at frequencies close to Nyquist freq looks great, except 1 fact: no one tried to see not only DAC, sine-wave signal and math, but the whole signal processing chain. Clearly, there are two options:
1. Old native 44.1k records. In my point of view, the only reason of CD-quality records stays in collection. At the time of recording, both recording and mastering was checked in the studio w/o oversampling, and it already contain best possible correction. Not by math, oversampling and so on, but by people.
2. Downsampled from hi-res. It depends on.
Not sure what someone even seriously checking the results of downsampling to uncompressed PCM for digital downloads. So, it depends on software used for downsampling.
And, eventually, hi-res recordings at the time of late CD-era. It depends on studio equipment. Maybe you can achieve something on some tracks if studio mastered CD on equipment close to your, but this FR correction comes with its own price.
In any case, chain looks like this: hi-res (studio) -> downsampling (with DF) -> 44.1k -> upsampling -> several DF -> EQ -> DAC...
And with upgrade collection to hi-res it will be: hi-res->NOS DAC.
1. Old native 44.1k records. In my point of view, the only reason of CD-quality records stays in collection. At the time of recording, both recording and mastering was checked in the studio w/o oversampling, and it already contain best possible correction. Not by math, oversampling and so on, but by people.
2. Downsampled from hi-res. It depends on.
Not sure what someone even seriously checking the results of downsampling to uncompressed PCM for digital downloads. So, it depends on software used for downsampling.
And, eventually, hi-res recordings at the time of late CD-era. It depends on studio equipment. Maybe you can achieve something on some tracks if studio mastered CD on equipment close to your, but this FR correction comes with its own price.
In any case, chain looks like this: hi-res (studio) -> downsampling (with DF) -> 44.1k -> upsampling -> several DF -> EQ -> DAC...
And with upgrade collection to hi-res it will be: hi-res->NOS DAC.
Nothing new from me.
It depends on parts you have in hand. Try to rotate plug in a wall socket (DAC and amp), try to power amp from battery or another PSU, try unbalanced connection with capacitor. Try to check DDDAC ground, chassis, amp with screwdriver with neon lamp (or any other sensitive AC-line tester). Check DAC internal grounding, check how you really connected DAC and amp. If your interconnection cable is long and/or its quality is unknown, try to connect DAC and amp with very short twisted wires (good CAT5e/6 network cable is a excellent inexpensive source of quality cooper pairs for experiments). Also, in some cases wall sockets may don't have a really connected safety ground, in that case any chassis connection to fake "safety" ground only makes things many times worse.
Also, it depends on the transformer used, if it has an electrostatic screen (Talema doesn't have it), it many times better, but they must be connected.
In any case, try to identify a problem before buying some expensive stuff.
Thanks, man. I'm pretty sure its to do with the fact the amp isn't truly differentially balanced. I don't have the issue on my Hypex NC400 power amp for my speakers so I presume there is a difference in how the XLR inputs are set up. Ill try with RCA cables first and then just try to make my way through your list!
It could be the noise coupling through the ground, from the PC / laptop... If you use a laptop, try unplugging the charger, i.e. try to run the laptop of the battery.
Also, WaveIO has a ground isolator chip... which definitely helps with ground noise isolation, I am not sure which USB card you are using.
Also, WaveIO has a ground isolator chip... which definitely helps with ground noise isolation, I am not sure which USB card you are using.
I'm using the DDDAC version of the WaveIO but hadn't tried disconnecting the USB yet to take the PC out of the equation. Will also try that!
My AC is grounded to chassis, with the WaveIO and DDDAC both powered by independent secondaries from the transformer (via super regulators) so WaveIO and DDDAC power are both grounded to the transformer. The DDDAC and WaveIO are connected to each other via the isolated I2S output of the WaveIO through a ribbon cable, so all 5 ground pins are connected.
For the output of the DDDAC to the XLR output plugs I have used a 20cm length of shielded interconnect cable. Therefore, XLR pin1 is connect to the shield of the interconnect but is not connected to the common pin of the DDDAC (only pos and neg are connected to XLR pin 2 and 3). So I guess this means pin1 is effectively not connected to anything within the DDDAC itself. HOWEVER, the connection to my NC400 power amp, has pin 1 connected to the power amp chassis. I never have had a ground loop buzzing from the NC400. I don't know how pin1 on the Burson Soloist XLR input is connected.
You had no problems before, but now they have appeared. My opinion is that Pin 1 should be connected to the chassis (or to the GND of the DDDAC supply), the GND of the DDDAC supply and the GND of the WaveIO should be connected to the chassis separately via 10A 600+V greatz (or antiparallel diodes 10A 600+V) and 10ohms 2W MOX + 100nF X2 capacitor.
Direct connection of the signal GND to the chassis can cause a ground loop. If the signal GND and the metal chassis are not connected in any way, various problems are possible, like buzz, hum an noise.
This is how I connected the signal GND-s to the chassis of the two DDDACs and so far there have been no problems with either balanced or unbalanced connections. Pin 1 is connected to the GND of the DDDAC supply.
The safety ground must be always connected to the metal chassis of the device, which is protection against high contact voltage according to all regulations.
Direct connection of the signal GND to the chassis can cause a ground loop. If the signal GND and the metal chassis are not connected in any way, various problems are possible, like buzz, hum an noise.
This is how I connected the signal GND-s to the chassis of the two DDDACs and so far there have been no problems with either balanced or unbalanced connections. Pin 1 is connected to the GND of the DDDAC supply.
The safety ground must be always connected to the metal chassis of the device, which is protection against high contact voltage according to all regulations.
Last edited:
Would connecting pin1 to the common output pin help or would it need to specifically be the ground of the DDDAC psu? Do you have a picture of your set up as an example? Sorry, Im not an expert I have just tried to set up the connections per the manual and by studying the audio-creative set up.
Safety ground is definitely connected to chassis
