DAC for my desktop computer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Currently I'm using the standard black Dragonfly USB stick to output into an ancient but serviceable Cambridge Audio 2.1 speaker system. I would like something better. I'm also going to upgrade my speakers to something like Paul Carmody's Speedsters and a DIY sub. I've got a few solid state amps I can use, but I'm itching to try a tube amp like the Reisong A10 EL34.

I'm trying to decide between any number of less expensive DAC's. I see folks recommend the really cheap Chinese DACs and there's the Schiit Modi 3.

Are there significant differences in sound between all of these? My primary digital media are FLAC and 320 kbps MP3.

These are the ones on my list so far:
1) L1387DAC 4X Quad TDA1387 Fever USB Decoder HIFI 44.1K 48K 16bit
L1387DAC 4X Quad TDA1387 Fever USB Decoder HIFI 44.1K 48K 16bit | eBay

2) L1387A 8X decode 1969 pure methailles one machine TDA1387 super TDA1541
L1387A 8X decode 1969 pure methailles one machine TDA1387 super TDA1541 | eBay

3) Schiit Modi 3
Schiit Audio: Audio Products Designed and Built in California

4) Topping D30
Amazon.com: Topping D30 DSD Audio Decoder USB Coaxial Optical Fiber XMOS CS4398 24Bit 192KHz Decoder (D30 Silver): Electronics
 
Last edited:
For a diy'er on a budget I would probably recommend something like the "I2SoverUSB v.III, AK4493 and LME4910 fully soldered combo" at: Products - I2S over USB Audio
Of course it would still need a power supply, and a place to mount the boards might be nice. At least you know the dac quality is going to be pretty good for the price point.

By the way, if anyone claims there is peer reviewed published research showing all dacs sound the same, one might ask the person for the citations... since they don't exist... :)
 
Last edited:
Are there significant differences in sound between all of these?
Not according to double blind listening test results. They are posted online. DAC technology has been matured for 20 years or so. Anyone trying to push upgrading are either salesman or don't know what they are talking about. If you have money for upgrade, use it on speakers and room acoustics since those are real bottleneck of electronic audio.
 
For a diy'er on a budget I would probably recommend something like the "I2SoverUSB v.III, AK4493 and LME4910 fully soldered combo" at: Products - I2S over USB Audio
Of course it would still need a power supply, and a place to mount the boards might be nice. At least you know the dac quality is going to be pretty good for the price point.

By the way, if anyone claims there is peer reviewed published research showing all dacs sound the same, one might ask the person for the citations... since they don't exist... :)

Wow that unit has impressive specs. It looks like the specs are superior to any of the units I linked to. I don't have a problem mounting it on a board or case so long as I don't have to do any complicated soldering. Does it need anything else other than a good power supply?
 
Does it need anything else other than a good power supply?

Not that I know of. It probably needs multiple voltages, maybe +-15v, 5v, maybe 3.3v?

You could try reading their documentation and see if whatever they have answers your questions, or if it raises any new questions. If still questions after that maybe we can help some, not sure, have to see whatever issues may turn up.
 
Re: #1, #2.

You are entering R2R territory and you didn't mention what kind of music you are listening, it is important. For hipop or metal you will be better to stay with delta-sigma design, but for a God sake avoid #4 - Topping D30! Mine sounds very bad, probably has fake opamps, but a lot of jitter too, as anything around 96kHz or above calls for a revenge from Heaven. A Realtek chip in a Sony Vaio laptop gives a better sound quality. @Markw4 is on the spot in his first reply and his recommendation is reasonable too.

For any music with acousting instruments you are in the right place. I own a portable USB powered version of the same Lee design 8xTDA1387 made by NobSound $42 on AliExpress. It is a good choice, it can be tweaked for better SQ. It has a passive I/V and output with a minimum filtering. No opamps is a better than poor opamps, but interfacing with a right pre-amp becomes important. The same applies, jugging from photos to a desktop version of 8xTDA1387 you proposed.

Bear in mind that unless you tweak these DACs, there will be no big jump in sound quality, except perhaps R2R sound which is important. To get a big jump and a taste of high-end, you need something like Audio GD R2R-11 $350. I got it recently and I can sign for it with my both hands. It is a pure R2R NOS design, Class A no feedback discrete amplifier, a pre-amp function for both HPA and RCA outputs. Everything is high quality and sound is very real.
 
Last edited:
frugal-phile™
Joined 2001
Paid Member
I'm also going to upgrade my speakers to something like Paul Carmody's Speedsters and a DIY sub. I've got a few solid state amps I can use, but I'm itching to try a tube amp like the Reisong A10 EL34.

That speaker is likely unsuitable for that amplifier. SE amps tend to have higish output impedance so you want to look for a flat impedance curve, something the Speedster does not have.

An externally hosted image should be here but it was not working when we last tested it.


Speedster TMM - undefinition

reisong-boyuu-a10-el34-bube-amplifier.jpg


And if it is for a desktop, you are likely not sitting far enuff away to allow for the integration of the physically distant drivers. This is were a single driver FR speaker has a big advantage, and given near field one of the gitchas w FR — inability to push SPL — is moot.

I'm trying to decide between any number of less expensive DAC's. I see folks recommend the really cheap Chinese DACs and there's the Schiit Modi 3.

I really like what Shitt is doing. But with DACs based on the typically avaialble DAC chips, the performance will be dominated by the surrounding circuitry and the power supply.

If you are connecting via USB there is also a significant difference in the capabilities of the USB inputs. You do want it to be isochronous asynchronous (most of them). There are only a few solutions for this, the new Schitt version is very good.

Now there are 2 Modis, a cheap one with the AKM DAC chip, and the more expensive MultiBit version. That will be better, and uses a true multibit DAC chip designed for medical & industrial processes — a bitch to use for audio.That is the one i would choose. Neither of the Modis have the new USB tech, they use the C-Media CM6631A implementation. You need to go a DAC up to get the new Unison USB input.

I have the Gungnir Multibit.

dave
 
Re: #1, #2.

You are entering R2R territory and you didn't mention what kind of music you are listening, it is important. For hipop or metal you will be better to stay with delta-sigma design, but for a God sake avoid #4 - Topping D30! Mine sounds very bad, probably has fake opamps, but a lot of jitter too, as anything around 96kHz or above calls for a revenge from Heaven. A Realtek chip in a Sony Vaio laptop gives a better sound quality. @Markw4 is on the spot in his first reply and his recommendation is reasonable too.

For any music with acousting instruments you are in the right place. I own a portable USB powered version of the same Lee design 8xTDA1387 made by NobSound $42 on AliExpress. It is a good choice, it can be tweaked for better SQ. It has a passive I/V and output with a minimum filtering. No opamps is a better than poor opamps, but interfacing with a right pre-amp becomes important. The same applies, jugging from photos to a desktop version of 8xTDA1387 you proposed.

Bear in mind that unless you tweak these DACs, there will be no big jump in sound quality, except perhaps R2R sound which is important. To get a big jump and a taste of high-end, you need something like Audio GD R2R-11 $350. I got it recently and I can sign for it with my both hands. It is a pure R2R NOS design, Class A no feedback discrete amplifier, a pre-amp function for both HPA and RCA outputs. Everything is high quality and sound is very real.

My musical preferences are all over the place. I listen to almost everything - rock and roll, pop, folk, jazz, classical, etc etc. However in front my computer I tend to spend more time listening to "quieter" more introspective music since I'm multitasking. Good recordings are tantamount. There's so much out there that just garbage recordings. I love intimately recorded jazz. It's so real you can almost see and feel Miles Davis behind the computer screen (remember this is for my desktop system).

Rock and Roll and pop are almost impossible to judge because my reference for them is my experience in clubs and stadiums. It's multi-mono amplified garbage sound usually at ear splitting levels.

I read that the USB powered units may be preferable for desktop computers because the USB connection minimizes jitter. Is that true or is more green marker snake oil?

I'm also told the software makes a difference. Right now I'm listening to Miles Davis and switching back and forth between VLC and Foobar2000. I'm not sure I can hear a significant difference although the bass on the Foobar2000 may seem a tad more solid and defined. It's probably my ancient 2.1 Cambridge system. As soon as this crazy pandemic is over, I'm running out and grabbing everything I need to build a set of good nearfield monitors and a sub!

I read great reviews of the Topping products. But I also read there are fake versions of these imports.

Schiit gets good reviews although some folks seem to hate them for one reason or another.

The bottom line it seems, is there are probably small differences between any of these units in the $100 range. I'd consider a sub $1000 unit for my "good" system in the theater room, but not for a desktop pc.
 
That speaker is likely unsuitable for that amplifier. SE amps tend to have higish output impedance so you want to look for a flat impedance curve, something the Speedster does not have.

An externally hosted image should be here but it was not working when we last tested it.


Speedster TMM - undefinition

reisong-boyuu-a10-el34-bube-amplifier.jpg


And if it is for a desktop, you are likely not sitting far enuff away to allow for the integration of the physically distant drivers. This is were a single driver FR speaker has a big advantage, and given near field one of the gitchas w FR — inability to push SPL — is moot.



I really like what Shitt is doing. But with DACs based on the typically avaialble DAC chips, the performance will be dominated by the surrounding circuitry and the power supply.

If you are connecting via USB there is also a significant difference in the capabilities of the USB inputs. You do want it to be isochronous asynchronous (most of them). There are only a few solutions for this, the new Schitt version is very good.

Now there are 2 Modis, a cheap one with the AKM DAC chip, and the more expensive MultiBit version. That will be better, and uses a true multibit DAC chip designed for medical & industrial processes — a bitch to use for audio.That is the one i would choose. Neither of the Modis have the new USB tech, they use the C-Media CM6631A implementation. You need to go a DAC up to get the new Unison USB input.

I have the Gungnir Multibit.

dave

Yeah, I can see what you mean about the separated drivers and being nearfield. So should I instead consider a Transmission Line design instead? I've just only started to learn about these. Is there a good design you would recommend?

You're right, I don't need huge SPLs. Imaging, definition, air, extension etc., are what I'm looking for.

It sounds like the "safest" best is the Schiit Modi 3 -name brand, known source, warranty, customer service. I love the idea of the really advanced designs like the AK4993, but I'm not sure if I'm technically able to take advantage of its features. A simple plug and play box would be.....well....simple.
 
frugal-phile™
Joined 2001
Paid Member
I read that the USB powered units may be preferable for desktop computers because the USB connection minimizes jitter. Is that true or is more green marker snake oil?

Flobby Dust. It really dependsn the execution of the DAC and how good the power from the PC is.

You see special USB cables tha remove the power from the USB and allow you to plug in an external power supply.

dave
 
frugal-phile™
Joined 2001
Paid Member
So should I instead consider a Transmission Line design instead?

That is a completely seperate thing. A FR or a multiWay can be TL loaded. A TL small enuff to fit on a desky=top won’t go all that low. A FR also means you are buying 1 driver per box, or for the same money 2 drivers & an XO. Quality vrs quantity.

The type of box loading is different thing than 1 or 2 drivers (+XO) for the box.

Her eis a set of good Desktop TLs. Woden Design 3" Box Designs, BabyLabs & more

Many others (not TLs) too. Including miniOnkens (uFonken, uMar-Ken, others), Martello, XRK has a few.

These are the 1st gen uFonken.

microFonken.jpg


dave
 
If you are connecting via USB there is also a significant difference in the capabilities of the USB inputs. You do want it to be isochronous asynchronous (most of them). There are only a few solutions for this, the new Schitt version is very good.

Now there are 2 Modis, a cheap one with the AKM DAC chip, and the more expensive MultiBit version. That will be better, and uses a true multibit DAC chip designed for medical & industrial processes — a bitch to use for audio.That is the one i would choose. Neither of the Modis have the new USB tech, they use the C-Media CM6631A implementation. You need to go a DAC up to get the new Unison USB input.
Dave, what do you mean that CM6631A is not a new tech?

I am aware of a very long thread dedicated to this chip, but have no time to read all. It seems to support the latest UAC 2.0 standard and the isochronous asynchronous transfer mode you require. This is all we need for the sample rates above 96kHz.

Which new tech stuff do you recommend? A-GD R2R-11 has well respected low jitter Amanero Combo 384 module, the other one is a popular XMOS U208. I am not aware of any other new tech available.

These cheap Chineese 8xTDA1387 DACs use an older interface chip CM108. It only support UAC 1.0 in isochronous adaptive mode, so jitter is expected higher. The interface is limited to 44.1/48kHz, so jitter should be no problem at that speeds. On my laptop it looses synchronisation at times generating purr, purr, pur, when using WASAPI push mode. When switching to the WASAPI event, it just breaks a stream for one second, the same as XMOS U208 on D30 or Amanero on R2R-11 do, no difference. It happens when laptop is overloaded and runs out of memory, so it is related to the Windows. I don't complain about jitter on the NobSound DAC either, but D30 gives lot of jitter despite of having high speed interface XMOS chip. So it is a matter of implementation rather than a chip or transfer mode.
 
I read that the USB powered units may be preferable for desktop computers because the USB connection minimizes jitter. Is that true or is more green marker snake oil?
USB powered units give less jitter, as there is no ground loops between devices. In R1R-11 Audio GD has removed 5V supply on USB connection in the Amanero module to reduce influence of ground loop, but a ground USB wire is still there. The higher priced A-GD models have USB isolators. You can always add a separate USB isolator, but be aware that all isolator chips have an inherent jitter, like S/PDIF Coax or optical and not all devices do a job as specified. It is better to remove ground loops, from your system then a direct USB interface gives the best performance.

I had a ground loop in my system cause that the first attempt of playing R2R-11 was negative. A remedy was very simple: to reverse Euro two-pin plug of my laptop PSU. All polution, grainy sound disaperared. If you have a grounded power plug, then try the same on the amplifier.

Your music preference tells me that you should definitely try R2R DAC. My Nobsound 8xTDA1387 gave me an idea what I should be looking for and R2R-11 is the final. It is 8-times more expensive, but worth it. I am not going to replace it very soon.

As for a Topping (again) there is a website heavily sponsored by this company, all units measured are specially prepared for testing. Commercial models significantly differ from a testing unit. Mine has a cheap 50c oscilators instead of many times more expensive SiTime oscilators in the reviewed unit. Once people started complain regarding D30's they received, a guy do not do hardware teardown anymore, so we don't know what was reviewed. I have purchased D30 from a Topping authorised reseller for a full price, it is not a fake cheap copy.
I'm also told the software makes a difference. Right now I'm listening to Miles Davis and switching back and forth between VLC and Foobar2000. I'm not sure I can hear a significant difference although the bass on the Foobar2000 may seem a tad more solid and defined. It's probably my ancient 2.1 Cambridge system.
I have VLC installed on my laptop, but I haven't figured out how to configure it to use WASAPI exclusive mode. This is the mode you should use for bit-perfect transfers. There should be no difference between players. If you see a difference, it means a sound goes through Direct Sound mixer, it is bad.

There are two places you should set in Windows sound device properties to enable WASAPI exclusive mode:
1. Volume and Balance 100%
2. Allow application to take exclusive control of this device (set on).

After you do it, Foobar2000 will use it. You should also give application priority (a one tick below of exclusive control).
 
Last edited:
The bottom line it seems, is there are probably small differences between any of these units in the $100 range. I'd consider a sub $1000 unit for my "good" system in the theater room, but not for a desktop pc.
Claims of audible difference between DACs have been made often but without supporting evidence. There have been listening tests done between cheap DAC and expensive DAC under double blind condition with levels matched. No audible differences were found. They are posted online.
 
frugal-phile™
Joined 2001
Paid Member
Dave, what do you mean that CM6631A is not a new tech?

It is relative. The 1st current gen USB interface software came from Wavelength. Then a couple more including the C-Media. There are a few others, and i fully expect there are a some proprietary ones (i can’t see the megaBuck DACs using anything but tech rolled in-house). These are all small dedicated uProcessors with software.

Mike Moffat — a digital audio rock-star) — was never really happy with how well USB worked as an audio interface (USB was not designed for audio transfer — a shame FireWire never caught on with mass consumers, it was specifically designed for Audio/Video, maybe Thunderbolt will, but it has a huge amount odf USB momentum to overcome, not likely). So he decided to start from scratch using a PIC32. Just released in the last few monthsin the latest Bifrost (and now comes on Gungnir and Yiggy (don’t ask me to spell that).

It happens when laptop is overloaded and runs out of memory

Any good music layback software should not allow that.

dave
 
...Audio GD R2R-11 $350...

Never heard of that one. Googled it to see, found ASR review: Review and Measurements of Audio-gd R2R11 DAC & Amp | Audio Science Review (ASR) Forum

Personally, seems to me that measurements don't exactly correlate with human perception. However, both measurements and listening should be used by good audio designers, again IMHO. Both things have some value.

None the less, I am kind of struck by the magnitude of the ASR graphs and measurements. I would be curious to hear the dac, but honestly I would be very hesitant to buy one based on the review.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.