Should phantom power be supplied by a linear supply?

[Bo Deadly]

I have had very little noise from JLM, Five Fish, and others including CAPI, and my own haphazard replications that I cloned from various PSU's linear supplies, they all seemed excellent. Despite the all in one and small profile os it that much of a difference?

If a linear supply is working for you then you shouldn't care.

The only way to know the difference between a linear supply and SMPS is to simply try each and measure it.

But I know for a fact that the SMPS can be completely void of all noise if implemented properly. So linear can only do worse. In particular linear almost always has a little mains hum. It might be a 6 dB bump at -90 dBFS in which case again, you probably shouldn't care.

 

[Lerok]

What? I've never seen a preamp's current draw vary more than 10% between utter silence and clipping.
Wow! 10 000 uF for less than 1.5A. Nothing wrong with it, but a tad over-engineered. Many would be satisfied with 5 times less.
I have come to favour the LM317/337 over the 78/79xx. They are more robust and less noisy. In addition, they simplify inventory.
What high current? The 78/79xx are limited at less than 1.5A. That would suggest 0.5mm traces. Since space is generally not an issue, I routinely use 0.8mm (32mil).

My thinking was given a higher quantity of preamps (ie, 8 or 16) then the variance between all of them being low (for recording drums) and all being high (for something like a classical ensemble) could be larger than what one would typically see.

10k uf is a bit excessive for sure, I'm really not sure why I wound up with that - 2k is probably fine for my purposes.

I will have to try the LM series - I was planning on using larger traces around the transformer and before the regulators, as it doesn't change the cost at all, and I figured better safe than sorry

At this point I'm building with primarily switching supplies, and while I haven't done any preamps yet, I'm inclined to try them first given the feedback of squarewave and other forum members - if noise becomes an issue, I may recognize this plan, but for now, it's archived away in the vaults. Thank you for the help!

 

[Cranehazard]

My opinion is linear supplies are better. You can hear plainly if the 60hz is filtered out. Switching supplies eliminate all kinds of high frequency stuff out into the surrounding space this hits your transistors and ics and makes it's way down to the audio spectrum [aliasing]. This can be seen on an Audio Precision system. Some identical supplies eliminated a lot more interference even with all parts replaced. I believe due to small capacitances in the board. Opinion: high freqency switching and audio dont play well. The high end equipment I worked on with linear supplies was much better. Companies are forced into this tech by law and financial reasons.

 

[mjrippe]

@Cranehazard - Are you sure "eliminate" is the word you want? Perhaps "emanate" instead?

 

[abbey road d enfer]

My opinion is linear supplies are better. You can hear plainly if the 60hz is filtered out. Switching supplies eliminate all kinds of high frequency stuff out into the surrounding space this hits your transistors and ics and makes it's way down to the audio spectrum [aliasing]. This can be seen on an Audio Precision system. Some identical supplies eliminated a lot more interference even with all parts replaced. I believe due to small capacitances in the board. Opinion: high freqency switching and audio dont play well. The high end equipment I worked on with linear supplies was much better. Companies are forced into this tech by law and financial reasons.

A poorly designed product will suffer from EMI/RFI, noisy PSU and poor grounding.
A well-designed product will make all these externally induced problems less discernible.
I regularly use smps, of which I get performance identical to linear PSU's, by using adequate filtering, shielding and grounding.

 

[Matt Syson]

There are pros and cons of either Switchmode or 'conventional linear' supplies which are also influenced by whether it is a 'commercial' design you are investigating or units for yourself. Safety regulations, EMC/EMI legislation etc. Although not an issue for a few channels of preamps, attempting to get power factor correction on a large (300VA or more 'linear' supply) can get fun (and expensive).
The flipside can be that cleaning up switchmode noise may also require a healthy dose of 'extra' filtering if you truly want 'NO' excess noise appearing on your audio signals.
The comment earlier that the 'ground' of your 'phantom' 48 Volt supplyrealy does highlight that it should really be an isolated supply which is referenced only at pin 1 of the mic XLR, and for a multichannel unit sghould arguably have all the mic XLRs physically close together

 

[JohnRoberts]

The technology does not matter as much as the implementation.

Back last century the engineer over Peavey's large 36x8 sound reinforcement console used a switching PS for the +48. The 36x8 AMR recording console I designed concurrently (using common tooling and components) used a conventional +48V PS. I didn't see any benefit from a lighter PS but my consoles tended to not get moved as often as SR consoles. Ironically perhaps I had a microprocessor running inside my master section, capable of generating midi commands.

Later last century at Peavey we designed a nice monitor console built into a flight case, using a universal input mains voltage switching power supply for all rails. I am unaware of any customer complaints about PS noise from either of those mixers.

JR

PS: My first technician gig back in the 60s was working on a DC to DC switching supply... It is remarkable to see how much the technology has improved in over half a century.

 

[abbey road d enfer]

The comment earlier that the 'ground' of your 'phantom' 48 Volt supplyrealy does highlight that it should really be an isolated supply which is referenced only at pin 1 of the mic XLR,

Many a commercial product is testimony to the fact that it's not an absolute necessity. What is necessary is understanding the effects a piece of wire or sheet metal has on current circulation. The proven need to provide a direct path to EMI/RFI to the skin of the equipment does not imply that the negative of the phantom supply must also be referenced to the same point; it just means that the rail must be clean enough in reference to pin 1. This is achievable with the 48V negative referenced almost anywhere to a relatively clean ground, and filtered in reference to pin 1. This is usually done with a resistor and a capacitor whose negative is referenced to pin 1, not to audio ground.

and for a multichannel unit sghould arguably have all the mic XLRs physically close together

Although desirable, mainly for practical reasons, it's not a necessity.

 

[Matt Syson]

Playing 'devils advocate' I would argue that getting it 'right' is getting more important than ever with the increasing use of 'cheap and nasty' switchmode supplies for phone chargers and other random non audio gear and of course mobile microwave transmitters that are almost a necessity these days.
One task I had in 1983 was to prevent noticeable artifacts when a UK legal radio walkie talkie was operated while it was lying on the surface of an audio mixer. I suppose the use of 'dithered' SMPS can make ultimate silence (basic white noise only) from a piece of audio gear more fun when you may have several 'transmitters' of noise. Thus needing to 'reject' several 'bandwidths' of interference.

 

[JohnRoberts]

I submit it is always important to get it right. Using a switching supply introduces another variable to manage (HF noise), but that is the work of engineering. Weight the cost/benefit and make your choices.

JR

PS: I have joked that the 1970s textbook discussions about noise reduction and shielding did not anticipate today's RF soup... but the physics remains unchanged.

 

[dyamakuchi]

but the physics remains unchanged.

The physics of the EMI is definitely one factor. Reliability is another animal altogether. More parts equals less reliable...in general. I can tell you from first hand experience, you don't want to be driving the bus when that mixer's supply goes out in front of a client. Much less at a live show. Even if you brought a spare console, it's still a shipwreck. Very, very bad optics.

Also, although the SMPSs generally do ( _mostly_ ) move the noise out of the audio band, if you've got an A/D, there's always a chance to alias it back down again. In fact, any RF engineer will tell you, all you need to demodulate AM is a diode and cap. Plenty of nonlinear processes even in an opamp or simple transistor stage to give you a headache.

The SMPSs are probably the better solution all things considered, but there's a lot of factors involved. Good luck!

 

[Bo Deadly]

The SMPSs are probably the better solution all things considered, but there's a lot of factors involved. Good luck!

I'm not going to challenge everything anyone claims about SMPS being inferior but there are many people here other than myself who have found SMPS to be superior to linear in almost every way. And one way in which they are definitely superior is in reliability. The reliability of electronic components depends mostly on heat and SMPS are much more efficient than linear supplies which means you dissipate a lot less heat for a unit of power. Some of the details about how to best utilize SMPS for minimizing noise are yet to be finalized. There are important rules for using SMPS that must be followed. If these rules are followed carefully, it is virtually guaranteed that the resulting supply will exhibit lower noise and be more reliable than the equivalent linear supply. I know we have a lot of new people around here and that's great but it's going to take some time to bring people around to the fact that SMPS is one of the few clear winners of classic pro-audio disputes.

 

[dogears]

+1 for SMPS over linear.

 

[Brian Roth]

One thing to think about is the total lack of any schematics for a SMPS. If it craps out, the best you can hope is that you can find a replacement from the OEM of the equipment. Sometimes I discover the OEM used an off-the-shelf SMPS once I inspected the bad PSU card/module for markings.

After decades in this biz I can almost always do a field repair on a linear supply....exceptions being a crap design where everything was toasted due to poor heat management.

Of course, REAL pro audio manufacturers included schematics in a service manual. Think of Ampex, SSL, MCI, Amek, etc.

Bri

 

[abbey road d enfer]

More parts equals less reliable...in general.

This is largely debatable. Give me 100 resistors against one electrolytic capacitor, give me 100 transistors against one power xfmr; give me 100 regulators against one connector.
And think of it, modern SMPS have less and less parts. I've had my share of duff console PSU's, with blown transformers, fried power regs/transistors and corroded connectors.

 

[FIX]

I've used SMPS (switch mode PSU) since the Tonelux days, and no one ever knew... I designed a dual stage common mode filter board that got the noise down to about a half a m/volt at 60KHz, and then the cable to the racks eliminated the rest. One trick is to treat the PSUs like balanced outs and run them into to CM filter and then join the grounds after them. I still use them today. The best thing that ever happened to SMPS was the CE directives, as they couldn't just dump the noise into the AC ground, so the output right off the bat is quieter now than before. Some use pulse width for regulation and some use sine waves and vary the frequency in a resonate transformer. You can't beat the efficiency and weight.

Some studios that had an SSL 9K console went to the Atomic brand SM PSUs and dropped their electric bills by thousands per month.

 

[Trouble Report]

Hi, designing a power supply for a modular preamp box I'm putting together - I'm currently running +/-15vdc out of a center tapped transformer for power for the pres themselves, and was a little curious as to your guys' thoughts on the impact running phantom power through a cheaper switching supply would have on the sound - I'm assuming any small sonic difference from a switching supply on a mic's internal circuit would be amplified once it hits the pre stage, but I know Douglas Self's small signal audio design recommended a voltage tripler hooked up to the positive rail.

Any thoughts?

While working at Avatar Studios in NYC, a very high profile, very expensive switching supply was bought from a company that shall go nameless, to take the place of a linear supply system that powered a Neve VR (it was getting a little wonkee). Very shortly after installing the supply there were complains of "whistling" in the noise floor. We switched back to the original Neve linear supply and the complaints stopped. Techs are a curious bunch; we dragged out the spectrum analyzer and got very revealing results. With the linear supply the console output displayed basic quietude, except for the 120Hz component way, way down in the noise floor - more or less what you'd expect. Then we connected the shiny new switching supply, and the display exploded with sonic artifacts, including a fairly prominent (and audible) band right around 1K. Turns out, it's not the switching fundamental that is the real problem - that's way above the range of human hearing. It's the harmonics that it spews, and those frequencies combine and alias (beat) against each other, to produce artifacts at lower frequencies, like, dead in the middle of our audible range. This phenomenon also has implications for music (duh). The upper harmonics in music also combine with the power supply harmonics, to their determent. So...NO! No switching supplies on the audio rails, thank you.

 

[abbey road d enfer]

You know what they say, don't throw the baby with the bath water...
Because you've had a bad experience with a poorly designed (although apparently expensive) SMPS, you must not condemn them all.
I don't know when this happened, but until a few years ago, SMPS were designed for industrial applications where rail noise is not an issue.
Using SMPS for audio is a relatively new concept and many mfgrs have taken that into account.
When I started my career, the company I worked for used 24V PS. Many were not very good in terms orfnoise and/or had trouble starting with the highly decoupled audio circuits.
It turned the Philips produced the best PS in that respect; I would think because their engineers were very much aware of pro audio.

 

[JohnRoberts]

Perhaps share the name of the "expensive" PS that didn't work for you so others here don't repeat the mistake. In my experience with studio gear, more expensive does not automatically correlate with good design.

I ASSume you contacted the PS manufacturer to confirm that it wasn't faulty.

As I have shared probably too many times I worked on switching PS back when they were new, and probably before some here were born (1960s). It is a mature technology now and properly implemented not the weakest link in those consoles.

JR

 

[Bo Deadly]

... there were complains of "whistling" in the noise floor ... a fairly prominent (and audible) band right around 1K.

That is almost word for word exactly how I describe the result of not loading an SMPS enough (Rules for SMPS see #2).

Although that is not what I would expect from a supply for a very large desk. For large supplies you need to take extra precautions. In particular I recommend capacitance multipliers followed by a large filter cap to cover emitter resistance. Someone here did a 5A supply for an Allen Heath desk and we found it was noisy until adding the final cap. We made hi-res spectrums and got noise free results.

So again, the details are in the implementation. Don't roll your own. Use some MeanWell modules with CMs and input filters. It can be done. Whoever did that supply just whiffed. It happens. One data point is not conclusive.