Frustrated with understanding power supplies... DIY solutions

[Sythrix]

I am trying to come up with two effective power supplies, one for the RME ADI-2 DAC and another for the upcoming Massdrop X THX AAA 789 Linear Amplifier.

RME ADI-2 DAC: Idle power consumption: 7 Watts, Max. power consumption: 18 Watts (9 - 15V supply accepted).

Massdrop X THX: 24V at 1.8 Amp.

Benchmark says it's all a myth and that linear power supplies are actually noisier than (proper) switching supplies.

As far as I can tell, the entire audio world says otherwise.

So I've been looking at building something that can supply the high demand, like the AMB σ11 or σ22. but that's another thing... I don't understand single vs dual rail. I would imagine that single would be appropriate for these two, but since I don't understand it, I can't verify.

I'm also absolutely terrible at fabrication. I can solder and assemble PCBs, but If there isn't a box where it fits and I have to drill more that a couple holes, I'm screwed. Making things fit to exact specifications is usually beyond me. I have access to a 3D printer, but no skill or programs to create designs, which leaves me dependent on existing designs... which means I usually can't find what I need. Which leads me to the next problem...

In order to actually build the σ11 to accommodate one or both of these devices, it would seem to require a heatsink solution beyond what the maker intended, which means modifications, which means more headache for someone who's terrible at fabrication, because I don't really know how to implement an effective heatsink solution.

I've seen this Linear Audio Silent Switcher before, but looking at the voltages it accommodates, it would only be useful for the DAC (EDIT: Maybe nevermind? I can't find much on the available amperage, but it might not be able to deliver on overall wattage).

Otherwise it seems like every time I google for linear power supplies they don't look anything at all like I'm expecting and their application in audio is dubious, with little explanation about the characteristics that are pertinent.

The Jameco power supplies look like I'd be able to find one for the DAC (18 W), but apparently would have high voltage ripple (200 mV max)? I don't know.

I found these DIYINHK power supplies from searching the forums, but they don't look like they'd be able to accommodate any of the devices in terms of constant power delivery.

So yeah, that's where I'm at. Sorry for the giant block of information. This was half a rant and half a call to help, so even if no one replies, I feel better for venting my frustrations. -->

 

[RayDunzl]

They both come with wall warts.

What's the problem again?

Afraid of making your noisy AC power noisy?

 

[Arnold Krueger]

I am trying to come up with two effective power supplies, one for the RME ADI-2 DAC and another for the upcoming Massdrop X THX AAA 789 Linear Amplifier.

RME ADI-2 DAC: Idle power consumption: 7 Watts, Max. power consumption: 18 Watts (9 - 15V supply accepted).

Massdrop X THX: 24V at 1.8 Amp.

Benchmark says it's all a myth and that linear power supplies are actually noisier than (proper) switching supplies.

As far as I can tell, the entire audio world says otherwise.

Not the entire world, not even a major part of it.

Your typical wall or cord wart power supply is just a power transformer, a rectifier and a capacitor, provided that it has a DC output. Many of them are just a transformer if they are rated for AC output. Most of the pro audio gear I have on hand use the transformer-only versions. These are all major sources of audible electrical noise. A common 18 vac power supply puts out approximately 18,000 millivolts of EMI centered on a clearly audible frequency of 50-60 Hz.

Most quality audio gear of the grade of RME has one or more power filters and built-in DC regulators, thus treating the inbound power very skeptically. Note that they specify a nearly 2:1 range of acceptable voltages. I'll bet money that if you provided 12 volts DC with 3 volts of ripple (horrendous!), good gear would still sound brilliant.

Switchmode adds several kinds of protection. They are inherently regulated because this is actually the easiest way to make them. The ripple frequency is so high that the passive components to supress it are tiny and inexpensive. The ripple frequency is octaves above the range of human hearing, sometimes at twice or more the frequency of sampling at 192 kHz.

Basically this means that expecting some magic power supply to help sound quality is just of an unrealistic fantasy as worrying that an external power supply with nominal ratings will hurt it. Obsessing over wall warts is a clear symptom of audio nervosa, not relevant technology.

I get a lot of external power supplies from Jameco and eBay, with no instances of measurable violence adversely affecting the performance of good gear. In fact, I'd go so far as to say that any audio gear that is exceptionally sensitive to power is simply defective. Junk!

 

[Sythrix]

They both come with wall warts.

What's the problem again?

Afraid of making your noisy AC power noisy?

I'm not quite sure why you're commenting.

Is it to say that you are in the "linear is not better" camp?

Why not just say that?

 

[Sythrix]

Not the entire world, not even a major part of it.

Your typical wall or cord wart power supply is just a power transformer, a rectifier and a capacitor, provided that it has a DC output. Many of them are just a transformer, if they are rated for AC output. Most of the pro audio gear I have on hand uses the transformer-only versions. These are all major sources of audible electrical noise.

Most quality audio gear of the grade of RME has one or more power filters and built in DC regulators, thus treating the inbound power very skeptically. Note that they specify a nearly 2:1 range of acceptable voltages. I'll bet money that if you provided 12 volts DC with 3 volts of ripple (horrendous!), it would still sound brilliant.

Basically this means that expecting some magic power supply to help sound quality is just of an unrealistic fantasy as worrying that an external power supply with nominal ratings will hurt it. Obsessing over wall warts is a clear symptom of audio nervosa, not relevant technology. I get a lot of external power supplies from Jameco adn eBay, with no instances of measurable violence adversely affecting the performance of good gear. In fact I'd go so far as to say that any audio gear that is exceptionally sensitive to power is simply defective. Junk!

See I've never really thought too much about it, but there is a significant amount of talk, both on this forum and elsewhere, stating that noise can be introduced from switching supplies. I haven't really committed to one or the other, just getting a feel for what others think is worthwhile. If I get the consensus that this is a waste of time, then by all means, I do not want to spend any more on it!

 

[Arnold Krueger]

I'm not quite sure why you're commenting.

Is it to say that you are in the "linear is not better" camp?

Why not just say that?

OK, "Linear is not better".

Please quote me, early and often!


The fact that some of the cleanest audio gear in the known universe comes out of gear with switchmode power supplies should be sinking into the generally hidebound whacky world of high-end audio.

 

[RayDunzl]

Is it to say that you are in the "linear is not better" camp? Why not just say that?

I'm not in a camp.

I suppose I have both types here. Everything audio and video is sourced to one wimpy wall socket, except the PC, which is on the same breaker but across the room on another outlet and UPS.

Cables are a tangle and there are 8 wall warts and maybe 18 other power cords total.

I haven't noticed (audibly nor measurably as much as I can measure) any practical problems from my own visually (viewed with oscilloscope) noisy mains.

Nothing hums or crackles or makes itself known.

If you do have problems, then by all means, look for a corrective measure.

 

[RayDunzl]

but there is a significant amount of talk, both on this forum and elsewhere, stating that noise can be introduced from switching supplies.

Noise can be introduced by "linear" supplies too.

They are switching, just not as quickly (often). The rectifier diodes turn on approaching the peak of the AC power wave (noise spike occurs on the mains) and turn off when the voltage of the filter cap matches the output voltage of the transformer secondary (smaller to non-existent glitch) usually as the wave still approaches the peak, though depending on the load, could occur past the peak but almost always at a higher level on the AC wave than the turn-on point.

 

[Sythrix]

OK, "Linear is not better".

Please quote me, early and often!

Will do!

 

[Sythrix]

Noise can be introduced by "linear" supplies too.

They are switching, just not as quickly (often). The rectifier diodes turn on approaching the peak of the AC power wave (noise spike occurs on the mains) and turn off when the voltage of the filter cap matches the output voltage of the transformer secondary (smaller to non-existent glitch) usually as the wave still approaches the peak, though depending on the load, could occur past the peak but almost always at a higher level on the AC wave than the turn-on point.

See, that's some interesting information.

Benchmark made mention that they will introduce magnetic "hum" into the signal, which can approach audible levels.

I'm not in a camp.

Neither am I at the moment. Was just trying to lay out everything I've been doing and want to learn more about it. My inclination before reading about all of this was to not care...

 

[RayDunzl]

Benchmark made mention that they will introduce magnetic "hum" into the signal, which can approach audible levels.

DAC1 - Internal linear power supply:

DAC2 - Internal Switched supply:

I wouldn't call the hum in the DAC1 "audible", but it was there.

The first graph has a low cut at 10Hz, so the 0Hz rise shown with the DAC2 isn't in the display

 

[Dimitri]

As far as I can tell, the entire audio world says otherwise.

This "audio world" comrpises of DYIers, manufacturers, commercial reviewers (aka advertizing money fueling the cause), non-commercial reviewers ( aka here ) and last but not least ....listeners

Unfortunately each claim needs to be judged on its own merits....

Cost, efficiency,weight,ease of design, ease of repair , noise
Pick any five

 

[Palladium]

I'll take proper evidence from Benchmark (or from anyone else) over sighted testing, hearsay, fantasy-based engineering and banhammer tactics from the "audio world".

Still -150 dB vs -160 dB noise floors demonstrates just how the "audio world" is making mountains out of molehills. Of course atrociously designed equipment excluded.

 

[Sythrix]

I'll take proper evidence from Benchmark (or from anyone else) over sighted testing, hearsay, fantasy-based engineering and banhammer tactics from the "audio world".

Even though I like putting together PCBs, the level of frustration this was causing, coupled with the general consensus that it's not worth it, means I am probably not going to go forward with my build plans I've listed...

 

[amirm]

RME ADI-2 DAC: Idle power consumption: 7 Watts, Max. power consumption: 18 Watts (9 - 15V supply accepted).

I will test this with its own power supply and my lab one and see if there is a difference.

For now, the problem I have measured with switching power supplies is that to suppress their high frequency emissions (over the DC cable), they connect the output to input using a capacitor. This capacitor works both ways: it reduces the conduction emissions, but then passes some of the 60/50 Hz of the main onto the ground lead of the DC output. That in turn causes all the subsequent audio devices to "see" the same mains leakage. It is not an audible concern in measurements I have made but it does show up there as regression from linear supplies.

There are (more expensive) ways to make the emissions low without use of the above capacitor but it is not common. Benchmark gets there with a custom power supply that someone made for them or they changed it.

 

[Sythrix]

I will test this with its own power supply and my lab one and see if there is a difference.

For now, the problem I have measured with switching power supplies is that to suppress their high frequency emissions (over the DC cable), they connect the output to input using a capacitor. This capacitor works both ways: it reduces the conduction emissions, but then passes some of the 60/50 Hz of the main onto the ground lead of the DC output. That in turn causes all the subsequent audio devices to "see" the same mains leakage. It is not an audible concern in measurements I have made but it does show up there as regression from linear supplies.

There are (more expensive) ways to make the emissions low without use of the above capacitor but it is not common. Benchmark gets there with a custom power supply that someone made for them or they changed it.

Thanks Amir!

It will be interesting to see the difference, but I'm actually hoping that it's relatively minor between the two, because this whole linear supply hunt was getting on my nerves anyways.

 

[Wombat]

Thanks Amir!

It will be interesting to see the difference, but I'm actually hoping that it's relatively minor between the two, because this whole linear supply hunt was getting on my nerves anyways.

Power supply quality is not switching vs 'linear' but in the design, application and performance. Cost and size are now major considerations.

 

[Sythrix]

Power supply quality is not switching vs 'linear' but in the design, application and performance. Cost and size are now major considerations.

Yes, well I can see that now, but my hunt was mainly for "linear", because "switching" I had been told, was bad. I see now that it's far too complicated for such generalizations.

 

[Wombat]

I like an isolating transformer in plug-packs(wall- warts). I am plain old-school.

 

[Theo]

Question about the OP: when turning the volume up with no input signal how bad is the hum if any?
Would there be a change in the noise content when power is drawn from a switching supply?