Affordable balanced DAC?

Arcygenical

Fully [H]
Joined
Jun 10, 2005
Messages
25,066
I'm looking for an affordable balanced DAC. The entirety of my rack is all balanced, so why shouldn't the audio source be balanced... But, I'm not a huge fan of spending 200-300$ for a USB Balanced DAC, which may run into the same issues, see:

I'm getting a LOT of noise across the PA system (2x generic 750wrms Behringer PA's, and a decent 2 way crossover, along with the old Monoprice 31 band Parametric EQ), whenever power usage changes on my Gen 1 Ryzen PC - aka during games, cutscenes/vs menus, encoding video, watching video, hell, even moving the mouse causes a slight whistle around the 3000hz mark (Can you imagine how annoying this is - I've seen almost a 60w fluctuation as well, which is 10db or so over background usage of 6w @ low volume). Clearly, the 3.5mm solution on board is absolutely terrible. I've tried lifting the PA grounds, tying them to earth directly (via copper water piping to earth) and I've even gone as far as ungrounding the PC... Alas, there's still noticeable whine with ANY power usage change. And yes, I've tried around 3 PSU's now.

I do have an EQ that can take unbalanced inputs and "convert" them to balanced - but obviously we're losing the other half of the signal by doing so, simply pseudo-balancing - Other audio sources, say, a phone, or a macbook pro, sound absolutely wonderful, even with unbalanced input. It's just the damn PC.

I do *not* have a ground loop, as that's impossible with my power delivery scheme.

So... Balanced DAC? USB (ugh). Sound card that does TRS balanced? Anything affordable? I only need stereo output, as I run a 2.2 stereo setup. I don't mind building it myself, either. The last few TOSLINK and USB dac's that I've built ftrom Aliexpress have been incredible, at around 30$, albeit, not balanced. I don't care about TRS or XLR.

Thanks :/
 
I'm looking for an affordable balanced DAC. The entirety of my rack is all balanced, so why shouldn't the audio source be balanced... But, I'm not a huge fan of spending 200-300$ for a USB Balanced DAC, which may run into the same issues, see:

I'm getting a LOT of noise across the PA system (2x generic 750wrms Behringer PA's, and a decent 2 way crossover, along with the old Monoprice 31 band Parametric EQ), whenever power usage changes on my Gen 1 Ryzen PC - aka during games, cutscenes/vs menus, encoding video, watching video, hell, even moving the mouse causes a slight whistle around the 3000hz mark (Can you imagine how annoying this is - I've seen almost a 60w fluctuation as well, which is 10db or so over background usage of 6w @ low volume). Clearly, the 3.5mm solution on board is absolutely terrible. I've tried lifting the PA grounds, tying them to earth directly (via copper water piping to earth) and I've even gone as far as ungrounding the PC... Alas, there's still noticeable whine with ANY power usage change. And yes, I've tried around 3 PSU's now.

I do have an EQ that can take unbalanced inputs and "convert" them to balanced - but obviously we're losing the other half of the signal by doing so, simply pseudo-balancing - Other audio sources, say, a phone, or a macbook pro, sound absolutely wonderful, even with unbalanced input. It's just the damn PC.

I do *not* have a ground loop, as that's impossible with my power delivery scheme.

So... Balanced DAC? USB (ugh). Sound card that does TRS balanced? Anything affordable? I only need stereo output, as I run a 2.2 stereo setup. I don't mind building it myself, either. The last few TOSLINK and USB dac's that I've built ftrom Aliexpress have been incredible, at around 30$, albeit, not balanced. I don't care about TRS or XLR.

Thanks :/
What's wrong with USB? For starters you can drop the input sensitivity of your amps and crank up the volume on the PC. That should dampen your PSU interference nicely. Your Behringers should be equipped with input sensitivity adjustments. A USB solution will isolate your audio source from your computer nicely so I would just get a USB DAC on the cheap and forget about balancing. You only need balanced interconnects when you do very long pulls (i.e. PA solutions) or use very low level signals such as microphones.
 
Heh, horrible site design. Have to register just to see the content.
 
Heh, horrible site design. Have to register just to see the content.

They do quite a bit of cool stuff. Currently have the DT177X Go, which is a flatter DT770, and the Elex, which is a tamed Focal Clear. Oh, and the HE4XX, for a cheap entry to decent planars.

And their amps? World-class.
 
They do quite a bit of cool stuff. Currently have the DT177X Go, which is a flatter DT770, and the Elex, which is a tamed Focal Clear. Oh, and the HE4XX, for a cheap entry to decent planars.

And their amps? World-class.
HE4XX are on sale for $130 even currently.
 
Just answer everybody all at once, I having issues with USB and most first generation ryzen motherboards... especially when it comes to audio, latency and timing. I know there are fixes but I really would like a plug-and-play solution. As im only ever using stereo, I was hoping for a solution that avoided USB however I realize it is quite likely impossible.

I run my amplifiers at -25db, or the Max sensitivity of 0.77v as is. However unfortunately, the whine that I experienced through exerting my processor, does increase in volume as output volume is increased equally.

I might try a couple more USB Dacs and forget about balancing but Focus perhaps more on proper power delivery and isolation that way.
 
You need to try optical / TOSLINK.

Find a cheap Topping DX7, for example. This gives you optical input and XLR output, as well as decent headphone amplifier options that will power just about anything well, and cleanly, as well as plenty of other connectivity.
 
You need to try optical / TOSLINK.

Find a cheap Topping DX7, for example. This gives you optical input and XLR output, as well as decent headphone amplifier options that will power just about anything well, and cleanly, as well as plenty of other connectivity.


Something like this is exactly what I was looking for. I was hoping to avoid copper connections.

But 699$ here in Canada. Fml.
 
Ouch -- I got mine from Drop.com, but that may not be any better. Topping does make a D70 that ditches the headphone amplifier, however they use a pricier DAC and it costs just as much as a DX7.

Try finding the SMSL SU-8 locally or from China, might be more palatable.
 
Ouch -- I got mine from Drop.com, but that may not be any better. Topping does make a D70 that ditches the headphone amplifier, however they use a pricier DAC and it costs just as much as a DX7.

Try finding the SMSL SU-8 locally or from China, might be more palatable.
Going for budget I quite like SMSL, I have their M3 combo dac/amp, and pushes the entry level HE-35X I have great (didn't want to spend more than $150 for both cans/dac at the time).

I've got the itch to upgrade to the shiit magni/modi stack next, undecided on cans to pair them with.
 
Here's another one, pure SPDIF (optical and RCA) to balanced and unbalanced:

Enog 2 Pro

It's not really a common use case in and of itself, but there are products for it.
 
Have you considered doing this over the network via Dante and one of Audinate's Ethernet-to-XLR adapters?
 
I had no idea about Dante, I just looked it up and I can't believe audio over ethernet hasn't come around sooner thats cool as shit.
No more fat shitty specialized cables running all over the place for audio just run some cat 6 and all the receiver/transmitter devices.
 
I had no idea about Dante, I just looked it up and I can't believe audio over ethernet hasn't come around sooner thats cool as shit.
No more fat shitty specialized cables running all over the place for audio just run some cat 6 and all the receiver/transmitter devices.
Oppo Blu Ray players have been using DLNA since at least the Oppo 105, perhaps before. I got mine in 2013.
It controlled my PCs DLNA server playback over ethernet, it was great and had very good isolation.
Unfortunately Oppo no longer make BD players.
Look up DLNA devices on google.
Set your PC or a NAS up as a DLNA server and you are golden.
 
I had no idea about Dante, I just looked it up and I can't believe audio over ethernet hasn't come around sooner thats cool as shit.
No more fat shitty specialized cables running all over the place for audio just run some cat 6 and all the receiver/transmitter devices.

You'd be relying on their DACs and ADCs at each stage; pushing a single run would mean adding an additional D/A conversion.

This isn't really superior to running XLR around, from a sound quality perspective, because you're adding a conversion that can only degrade the sound, and you're adding latency at each conversion point.

For a larger setup with real routing needs between distant audio 'nodes' it starts to make more sense, but even then, you'd want to be sure that the D/A conversions aren't doing anything adverse for the use cases in mind.

Oppo Blu Ray players have been using DLNA since at least the Oppo 105, perhaps before. I got mine in 2013.
It controlled my PCs DLNA server playback over ethernet, it was great and had very good isolation.
Unfortunately Oppo no longer make BD players.
Look up DLNA devices on google.
Set your PC or a NAS up as a DLNA server and you are golden.

Yeah, damn near everything that plays music and has network connectivity speaks DLNA these days. For locally stored music it's a decent dedicated solution.
 
Last edited:
  • Like
Reactions: Nenu
like this
I had no idea about Dante, I just looked it up and I can't believe audio over ethernet hasn't come around sooner thats cool as shit.
No more fat shitty specialized cables running all over the place for audio just run some cat 6 and all the receiver/transmitter devices.

Dante is one of several second generation audio-over-IP technologies (AVB, BluLink, Q-Sys and AES67 being the others) which have replaced several older first generation technologies (Ethertalk, Cobranet). They tend appear in profession gear and for awhile every manufacturer was trying to push their own solution. Dante is currently seen as the 'winner' in this format war with AES67 being newer and interoperable with Dante. Personally I'm more of a fan of AVB since I have a few Macs that can output that audio stream directly to some DSPs I have which in turn have balanced IO connectors for what analog connectivity I need. Audio decoding is done on the computer as the network audio is multichannel PCM (with a 10 Gbit card, I have the potential of a thousand 24 bit 96 kHz audio channels to use).

The thing about Dante is that is wide spread enough that you can find inexpensive IO boxes or even amplifiers that'll ingest a stream. There are a couple of ~30W PoE+ based speakers that'll communicate over Dante. I'm eager to see what the ~60W or even ~90W 802.3bt enabled speakers will bring. I recently picked up some Klipsch R51M speakers recently which their rated power would be within the power budget of a good class D amp using the highest PoE delivery. Then scaling upward in speaker count basically becomes a function of how large the networking switch is and how the audio is mastered. For gaming where sound sources can be generated anywhere in the environment, there really isn't conceptual limit to how many audio channels can be used given enough processing resources on the host.

I've been optimistic that a company like MS, Google or Apple will push one of these audio-over-IP technologies to the masses so that real audio-over-the-network receivers and PoE based speakers become place.
 
The thing about Dante is that is wide spread enough that you can find inexpensive IO boxes or even amplifiers that'll ingest a stream. There are a couple of ~30W PoE+ based speakers that'll communicate over Dante.

So therein lies the question: how is the quality of the DACs being used on the end of a twisted-pair cable providing power being measured?

I do hope that the solutions in play are working with better quality power than your average ethernet switch pushing out to WiFi access points and cameras and so on :).
 
The DACs are professional grade with a very high signal to noise ratio. Dante transport is typically 96 kHz, 24 bit so the source is good with some gear supporting all the way up to 384 kHz. The other thing worth noting is that the amount of analog circuitry is kept to a bare minimum so the likelihood of inference there is also minimized to generally not be a concern. Pretty much you can paste in the discussion of powered vs. passive speakers here with big difference being that audio transport is digital to the speaker itself. And to double that debate, there are of course powered audio-over-IP speakers that contain an otherwise normal self contained amp.


The PoE switches I currently use for audio-over-IP installs are Cisco or Extreme networks with healthy, over provisioned PoE budgets and redundant power supplies. The other nice thing about networking based audio transport is that redundancy there is a solved problem with both Dante and AVB supporting dual transport with fail over support. One PoE speaker I've since supports fail over and PoE across both ports so that meets some European safety standard to be used in emergency scenarios. Disabling green Ethernet is a must for the speakers and generally advised any audio-over-IP gear. That does cause problems with sudden increases in power draw.
 
The DACs are professional grade with a very high signal to noise ratio. Dante transport is typically 96 kHz, 24 bit so the source is good with some gear supporting all the way up to 384 kHz.

I'm sure they're rated well, but the implementation is at least as important as the components used; that's why I'm more interested in how they measure.

The other thing worth noting is that the amount of analog circuitry is kept to a bare minimum so the likelihood of inference there is also minimized to generally not be a concern. Pretty much you can paste in the discussion of powered vs. passive speakers here with big difference being that audio transport is digital to the speaker itself. And to double that debate, there are of course powered audio-over-IP speakers that contain an otherwise normal self contained amp.

This problem is already solved another way: balanced cables. Obviously there's a point where distance and EMI become an issue, but that's true for both, really.

The PoE switches I currently use for audio-over-IP installs are Cisco or Extreme networks with healthy, over provisioned PoE budgets and redundant power supplies.

Having enough power isn't really the issue so much as that power being clean. The concern is that power supplies designed for datacenter use are at the very least not at all optimized for powering speakers.

Obviously in an ideal environment these aren't real issues, but they can become issues when challenges are introduced.


Also, thank you for the reply and general education!
 
  • Like
Reactions: Nenu
like this
IHaving enough power isn't really the issue so much as that power being clean. The concern is that power supplies designed for datacenter use are at the very least not at all optimized for powering speakers.
This.
Either an economic basic type or switch mode power supply will be used.
Neither are optimal for high quality hifi.

Another issue, PoE DC is supplied over a long distance making it prone to voltage modulation at the receiving end.
ie as the power required (current flow) changes, the voltage drop over the cables resistance changes (larger with length), making power supplied to the device less stable, adding more noise.
As pointed out, longer wire lengths are also prone to picking up more EM noise freqs.
I wouldnt choose this method for audio unless it wont be used for critical listening.
ie for a garden / house party or pa system it could work very well.
 
Yeah, that's basically the concern I'm seeing Nenu.

One mitigating factor, which is something I'd like to see shown, is that a 'power cleaning' stage could be used to feed the DAC and amp. This isn't any different than say the active PFC and so on used in decent UPSs or even the products used for home theaters, which I use for my PCs and home theaters myself just to ensure that PC, monitor, and audio power supplies themselves receive cleaner power than what would typically be provided by the apartment complex grid.

Power's just electrons, and if you can get them flowing correctly in a consistent fashion, even the datacenter-grade PSUs used by the likes of Cisco would be fine for getting consistent current so long as the distant end is prepared to 'condition' that power for audio use.

That's the question I'm looking at: if you're running a PoE connection from an ethernet switch out to a speaker, providing power and data, what provisions have been made to ensure that the audio that has been digitized somewhere else is what the speaker is reproducing?
 
I'm sure they're rated well, but the implementation is at least as important as the components used; that's why I'm more interested in how they measure.

Fair point but being more professional in nature, there tend to be less exaggeration than on the consumer side of things.


I
This problem is already solved another way: balanced cables. Obviously there's a point where distance and EMI become an issue, but that's true for both, really.

For analog, balanced is nice because it is indeed more difficult for interference creep in, but not impossible. The more likely scenario is that interference from the end points gets carried over the cable (ie 60 Hz ground hum) than introduced midway. For line level transport, there is a reason why balance still is in use despite being decades old. The main advantage of networked audio here is that it replaces the old multiheaded snake cabling with far cheaper/easier to acquire category cabling and is bi-directional. Not to dive into the analog vs. digital debate as the network solutions mimic the same quality as AES3 digital audio which leverages the same basic cabling as XLR.

However, balanced audio cabling isn't used for actually power a speaker. The comparison would be a powered speaker with it own internal amp being fed by a balanced cable because normal speaker cable can suffer from interference like any other unbalanced analog cable. The big benefit here for PoE is the single cable nature for use cases with less than 90 W of power needed. That is a huge advantage in terms of installation due to the need for only low voltage category cable and the sheer reduction in cables generally (this also adds up as some PoE speakers also add various sensors, beacons that are are useful in a corporate installed scenario).


Having enough power isn't really the issue so much as that power being clean. The concern is that power supplies designed for datacenter use are at the very least not at all optimized for powering speakers.

Obviously in an ideal environment these aren't real issues, but they can become issues when challenges are introduced.

Ah, I think we were looking at this from slightly different angles. There is another DC-to-DC transformer to convert from the PoE supplied voltage for what the speaker actually uses and some additional capacitors to maintain their RMS rating with decent maximum peaks. I presume the context was regarding just the quality of the power supplies for the switches themselves. However, even with a good power supply in the switch, PoE power at the receiving end needs to be seen as dirty due voltage drop over distance and how notorious switches in generally behave while handling near maximum rated PoE load. To your point, PoE power is dirty, very dirty hence the need to clean it locally before being supplied to the speaker. This is especially noteworthy as more cheap consumer switches are including PoE support as those don't have the best data center grade power supply. This is one of the driving reasons why PoE speakers carry such a premium now.

Being able to scale down the technology for the consumer mass market will be a challenge due to costs without cutting corners. I do get the concern that corners may be cut down the line for the electronics inside of the PoE speaker but considering their relatively new professional niche, the quality of the electronics overall has been good (sound quality is another subjective matter).

Not to split into too many tangents but PoE based LED lighting is the same way so they don't flicker when PoE power also changes. The professional/corporate installation side is moving toward one cable to rule them all.
 
That's the question I'm looking at: if you're running a PoE connection from an ethernet switch out to a speaker, providing power and data, what provisions have been made to ensure that the audio that has been digitized somewhere else is what the speaker is reproducing?

Are you referring to how the audio is routed on the network? IE how the audio source selected makes it to the destination? If so, there are various routing applications used that can change which transmitter goes to which receiver on a per channel or per bundled channel stream basis. This effectively makes the network based solutions a 4096 x 4096 audio matrix. Going larger is conceptually possible as well but that dives deep into the networking side to setup various domains and how they're patched into each other akin to how bridging two analog matrix switchers would work. Various mixing, delays etc. have to be handled at the end points appropriately.
 
Fair point but being more professional in nature, there tend to be less exaggeration than on the consumer side of things.

I really expect it to not be a problem, but it's important to look at not just in a testing environment but also under stress in a real deployment.

For analog, balanced is nice because it is indeed more difficult for interference creep in, but not impossible. The more likely scenario is that interference from the end points gets carried over the cable (ie 60 Hz ground hum) than introduced midway. For line level transport, there is a reason why balance still is in use despite being decades old. The main advantage of networked audio here is that it replaces the old multiheaded snake cabling with far cheaper/easier to acquire category cabling and is bi-directional. Not to dive into the analog vs. digital debate as the network solutions mimic the same quality as AES3 digital audio which leverages the same basic cabling as XLR.

XLR isn't perfect, but it's a damn sight better than most unbalanced solutions that are more vulnerable no matter the effort put into them (and cost to the end user). The main issues I'm seeing here are power 'dirtiness' from the source or introduced in line, as well as EMI that could introduce distortion into the balanced signal via overloading the correction or cause packet loss in the data connection the same (which is, incidentally, also 'balanced'). I can't state which would be more susceptible to different types of EMI though, that's out of my wheelhouse ;).

However, balanced audio cabling isn't used for actually power a speaker. The comparison would be a powered speaker with it own internal amp being fed by a balanced cable because normal speaker cable can suffer from interference like any other unbalanced analog cable. The big benefit here for PoE is the single cable nature for use cases with less than 90 W of power needed. That is a huge advantage in terms of installation due to the need for only low voltage category cable and the sheer reduction in cables generally (this also adds up as some PoE speakers also add various sensors, beacons that are are useful in a corporate installed scenario).

I'm with you on this. PoE and cheap twisted-pair is absolutely a draw when runs lengthen. Flexibility of installation increases while cost decreases, and yes, the potential for the speaker 'node' at the end of the run to serve additional functions is certainly attractive!

Ah, I think we were looking at this from slightly different angles. There is another DC-to-DC transformer to convert from the PoE supplied voltage for what the speaker actually uses and some additional capacitors to maintain their RMS rating with decent maximum peaks. I presume the context was regarding just the quality of the power supplies for the switches themselves. However, even with a good power supply in the switch, PoE power at the receiving end needs to be seen as dirty due voltage drop over distance and how notorious switches in generally behave while handling near maximum rated PoE load. To your point, PoE power is dirty, very dirty hence the need to clean it locally before being supplied to the speaker. This is especially noteworthy as more cheap consumer switches are including PoE support as those don't have the best data center grade power supply. This is one of the driving reasons why PoE speakers carry such a premium now.

Yeah, if they're accounting for the dirtiness in the power delivery from the PoE DAC/Amp to the speaker, then we're probably good. My concern is that this is both going to add cost until economies of scale and proper competition kick in, and as well, this is most certainly a point in the architecture where manufacturers will be looking to cut costs as the technology proliferates. It's definitely something to keep an eye on.

Being able to scale down the technology for the consumer mass market will be a challenge due to costs without cutting corners. I do get the concern that corners may be cut down the line for the electronics inside of the PoE speaker but considering their relatively new professional niche, the quality of the electronics overall has been good (sound quality is another subjective matter).

The challenge is that while the idea is fairly novel, the different technologies used aren't, just the aggregate. This means that there's very little to stop players from jumping in that generally would have no business producing audio equipment, and yet, they'll likely be doing so and targeting consumers first. Hell, consumers don't even need PoE switches, though those are ideal; if they're going to deploy the switch core somewhere noise would be a concern, like in an A/V cabinet, PoE adapters are cheap and could be used on the distant end.

Not to split into too many tangents but PoE based LED lighting is the same way so they don't flicker when PoE power also changes. The professional/corporate installation side is moving toward one cable to rule them all.

Ubiquity brought these out to a collective head turn and sigh. Obviously they're trying to get their own knowledge sink primed, and they do also have PoE access points that have speakers (and microphones?) included, just more targeted toward public address than quality audio reproduction -- for now.


So perhaps we'll have USB-C and Cat6a and call it done at some point in the near future, eh?
 
Are you referring to how the audio is routed on the network? IE how the audio source selected makes it to the destination? If so, there are various routing applications used that can change which transmitter goes to which receiver on a per channel or per bundled channel stream basis. This effectively makes the network based solutions a 4096 x 4096 audio matrix. Going larger is conceptually possible as well but that dives deep into the networking side to setup various domains and how they're patched into each other akin to how bridging two analog matrix switchers would work. Various mixing, delays etc. have to be handled at the end points appropriately.

Somewhat, but more looking at the idea that there is at least a DAC on the distant end for the speaker amp and likely ADCs elsewhere. I think you've already covered that pretty thoroughly above though, thanks!
 
Back
Top