15V phantom power

[Andy Peters]

A voltage tripler may be a solution, at least if only a few channels require phantom power. And yes, the vast majority of new products has correct P48 (except USB audio interfaces that are bus powered). It's one of the things I always measure when I review gear.

48V from 5V USB? It can be done with a switcher. Be mindful of the start-up current.

-a

 

[jp-apb]

If you're looking to create +48V from your existing rail volatges,
XP Power offers some low-cost isolated DC-DC converters that will do the job.
Small, potted modules that switch at >100kHz. Using one in an upcoming product.
Was easier than rolling my own (and I LIKE to build things  )
Newark has some listed here:
http://www.newark.com/jsp/content/printCatalog.jsp?display=single&cat=c127&page=1484

Look for models with +/-24V dual-output and connect them in series for +48V.
Some converters have wide-range inputs, so you can power them with almost any voltage you have available.
A 2W model will give you enough current for 6 to 8 mics.

Andy,
Didn't try a low-voltage version, but they do have a model that operates from 4.5V to 9V.
Might have to get a sample to see if a USB port will work.


JP

 

[Rossi]

I recall talking with at least one guy who used a switcher to make 48V phantom from a USB (5V) drop. As I recall it was difficult because of limited voltage and current but he found a solution that worked. 

I wasn't exactly clear in my statement: Most USB devices do supply a voltage in the 48V ballpark, but are lacking in the current department. A single device is not supposed to draw more than 100 mA at 5V. That's barely enough energy for two P48 inputs at full current, and the actual device, of course, eats some energy, too. So multi-input bus powered USB audio interfaces either violate USB 2.0 or P48 specifications. Usually the latter. Which is okay since few mics will draw more than 5 mA, but that's one of the reasons I wouldn't recommend bus powered devices for anything serious.

 

[owel]

So multi-input bus powered USB audio interfaces either violate USB 2.0 or P48 specifications.

USB spec is 100mA (called a single unit load).... also the default for an "unconfigured state"

But it's possible for the connected equipment to send a request to the USB port and ask for high power mode (assuming the host's hardware can deliver), and be able to draw up to 500mA.

But if you're just using a USB as a "DC jack power source", then more likely it's only operating in it's unconfigured state, and therefore draw is max 100mA.

 

[JohnRoberts]

I recall talking with at least one guy who used a switcher to make 48V phantom from a USB (5V) drop. As I recall it was difficult because of limited voltage and current but he found a solution that worked. 

I wasn't exactly clear in my statement: Most USB devices do supply a voltage in the 48V ballpark, but are lacking in the current department. A single device is not supposed to draw more than 100 mA at 5V. That's barely enough energy for two P48 inputs at full current, and the actual device, of course, eats some energy, too. So multi-input bus powered USB audio interfaces either violate USB 2.0 or P48 specifications. Usually the latter. Which is okay since few mics will draw more than 5 mA, but that's one of the reasons I wouldn't recommend bus powered devices for anything serious.

I didn't mention this before because I felt it was TMI for the OP who isn't power limited. Also take this with a grain of salt because I haven't reduced it to a working design. If you look at the standard 48V phantom supply, at higher current output (more than half) most of the output "power" is being wasted in the 6.81k dropping resistors.

One approach to reduce this wasted power, if we can integrate this into the whole preamp design, is to replace the 6.8k resistors with lower values. Since most preamps use a nominal 1.5-2k input termination that is effectively in parallel with the 6.81k, replace the 6.81k resistors with around 1K. Now sense the current drawn from that common node and regulate the voltage to be the same as it would be if there was 2.9k feeding the common node from 48V. Effectively a precision fold back current limiting. So the mic sees the same DC voltage it would with standard phantom, and also sees the same AC input termination (the following mic preamp is now high impedance to maintain that nominal 1.5k-2k input termination).

To put some numbers to this, image drawing 5 mA from each leg, 10 mA total. The drop with standard 6.81k resistors would result in a DC voltage of 14V at the mic.  With the 1k build outs the phantom voltage now only needs to be 19V. So the switcher only needs to deliver 190mW (0.01 x 19) instead of 480mW (0.01 x 48) of power.  

To reduce that last bit of power wasted in the 1k build-out resistors further requires tricky input impedance manipulations that even I wouldn't try on such low noise inputs.  How much this buys you in practice depends on the nature of the mic loads. For high current loading you are wasting far more power in the build out Rs than delivering to those loads.

This is mostly mental masturbation. It won't work for a stand alone phantom supplies, or appeal to purists (even though the mic shouldn't see any difference). Further we shouldn't try to power multiple 48V mics from USB on general principles (I worry that 5V USB may go away since 5V logic is getting long in the tooth).

This (not that simple) approach can recapture 80% or more of the power wasted in the 6.81k resistors, but yawn, so what.  

JR

 

[jp-apb]

JR,
Yikes,TMI indeed! 
Too bad you weren't around (or maybe they just didn't ask you) when the Phantom-Power Committee was first meeting.... ;D

JP

 

[PRR]

IMHO, the early BiAmp boards are well-built and beautifully musical, with one small flaw.

(That is, the wedge-shape slide-fader boards. There was an earlier flat-box turn-pot 6-in; lot of features but I was never happy with my recordings. I switched to the 2-transistor plug-in mike amps in the Ampex and OH! it was so much clearer.)

I would NOT modify nor "update" the design. The 5534s and TL074s are used well. The EQ was a real butt-saver on an ugly outdoor concert series and also for fine-sweetening.

The small flaw: the input noise-figure is 6 or 7 dB above perfection. With dynamic mikes in live performance, nobody notices. With hot condensers, the NF will be set by the mike, not the board input. However with dynamics on soft instruments in a quiet studio, the hiss is higher than it could be. Not the go-to tool for ribbons on harpsichord.

Get a 48V power-pack. Wire a jack to feed the 2X6K8 resistors. For 16 happy inputs you should allow 160mA. If you own 16 shorted cables, you should allow for 225mA worst-case.

You may jump the jack to dribble +15V when the 48V pack is not plugged. Many fine mikes will work on +15V through 3K4. A few will insist on a lower resistance, but IME this is rare, and not worth catering fors. Some (both $2000 and $99 types) just ain't happy with less than 45V open-circuit.

 

[wesm]

PRR- Thanks! I like your idea. keeps things simple enough but allows for as good as an external PP supply as I want or a wired-in supply for experimentation or even as a worst-case backup plan.

You mention the 5534's and tl074's are used well, but in my board lf353's and rc4558's are used...seem to be inferior to the aforementioned chips as per specs. my board is the 1683, apparently the 1683B uses the 5534's and 72's. To reduce noise, I may just set up a jack and switch so that I can hook up an external PSU for all rails, as the xformer isn't toroidal, and I suspect noise improvements could be made by eliminating it.

Also wondering if the psu and opamp decoup. caps may be expired as supposedly this unit has been in heavy use for a while...would I really notice a big difference in the sound if the caps were done for? I guess theres only one way to really find out...but anyone who has experience feel free to pop in.

 

[JohnRoberts]

Keep in mind if you just add the 6.81k resistors and 48V externally, that you also have DC blocking caps in series with you inputs, with adequate breakdown voltage capability to withstand 48V, and in the correct polarity.

JR

 

[wesm]

Thanks for the reminder. Luckily the boards are pre-drilled with holes and have traces to support all necessary PP components- a bit weird but I can't complain...makes my life a lot easier!

 

[wesm]

actually after more investigation and examination...it appears the inputs are already hooked up with dc blocking caps(10uF 35V with negative towards pre) in series. as the manual's mod only calls for 15V PP, going to have to update those to 63V. thanks JR for getting me to think harder about what I'm doing here and why.

 

[Andy Peters]

I recall talking with at least one guy who used a switcher to make 48V phantom from a USB (5V) drop. As I recall it was difficult because of limited voltage and current but he found a solution that worked. 

I wasn't exactly clear in my statement: Most USB devices do supply a voltage in the 48V ballpark, but are lacking in the current department.

A single device is not supposed to draw more than 100 mA at 5V. That's barely enough energy for two P48 inputs at full current, and the actual device, of course, eats some energy, too. So multi-input bus powered USB audio interfaces either violate USB 2.0 or P48 specifications. Usually the latter. Which is okay since few mics will draw more than 5 mA, but that's one of the reasons I wouldn't recommend bus powered devices for anything serious.

As noted, USB host ports provide 100 mA until a device enumerates and requests more. A powered hub's downstream ports are the same as a host port in this regard. Unpowered hubs cannot provide more than 100 mA on a downstream port.

I haven't seen any bus-powered USB audio devices that support more than two mic preamps with 48V phantom, likely because they just won't work.

What's interesting is how various computers handle things. My iMac's ports handled overcurrent (faulty device pulling too much) gracefully. My MacBook Pro (on the power supply, not the battery), dealing with the same device, threw up an exciting error, saying that "A device was drawing too much current and was shut down." However, in the former case, the USB VBUS dropped significantly.

-a

 

[Andy Peters]

Andy,
Didn't try a low-voltage version, but they do have a model that operates from 4.5V to 9V.
Might have to get a sample to see if a USB port will work.

That'd be interesting ... but again be mindful of start-up current, especially if the user has a mic connected before enabling the phantom supply. The little switcher I'm using does just fine without a load (though its start-up current is not insignificant) but with an Audix TR40 connected ('twas what was lying around), it yanked on the USB VBUS enough to go below the drop-out on the 3.3V regulator powering the micro, so that guy fell down and the reset supervisor on the micro did its job. The obvious solution: enable the switcher but use a high-side switch (MOSFET, whatever) to keep the 48V disconnected from the XLR jack until the 48V switcher is stable.

-a

 

[Rossi]

As noted, USB host ports provide 100 mA until a device enumerates and requests more. A powered hub's downstream ports are the same as a host port in this regard. Unpowered hubs cannot provide more than 100 mA on a downstream port.

I haven't seen any bus-powered USB audio devices that support more than two mic preamps with 48V phantom, likely because they just won't work.

What's interesting is how various computers handle things. My iMac's ports handled overcurrent (faulty device pulling too much) gracefully. My MacBook Pro (on the power supply, not the battery), dealing with the same device, threw up an exciting error, saying that "A device was drawing too much current and was shut down." However, in the former case, the USB VBUS dropped significantly.

-a

Of the USB devices I tested, only one requested more current than 100mA. That was a Zoom guitar interface that acutally contained a 12AX7.

Manufacturers obviously avoid requesting more than 100 mA, and I suspect its for good reason. It's asking for trouble. It likely won't work on a passive hub, and I'm not so sure an active hub would always work. The active hubs I have have five or six USB outs and are powered via a relatively small wall wart. I haven't checked, but I doubt that little switching PSU would supply enough current for more than one USB device that draws maximum current. And as you say, many real world PCs and Macs won't supply high current either without giving trouble.

So 100 mA is likely the most you can safely assume a USB port will supply. Just like 5 mA is usually the maximum current a mic manufacturer can safely assume a P48 input will supply. And even with those modest figures in mind, you will encounter problems now and then. I once had an ADAT preamp on my bench that could not even supply 3 mA for more than a couple of its inputs.

 

[JohnRoberts]

Doing some quick back of the envelope math... 48V phantom, with 6.81k buildout resistors will put 7mA into a dead short, per line, so both inputs shorted is 14 mA @ 48V for 670 mW of power... Interesting since 5V @ 100 mA only provides 500 mW of input power, so even a switcher with 100% efficiency won't support that (luckily it doesn't have to).

The only reason these work at all is because microphones don't draw the full 14 mA, however even 5 mA per line leaves precious little spare current to run the rest of the electronics.  Perhaps my power saving trick is more useful for the general solution than I thought, if it is desirable to run from 100mA.

JR

PS: A classic design issue is sizing the phantom power supply for a console full of inputs. In premium designs it is not uncommon to size for all inputs but one shorted, and the last one still deliver adequate voltage. In value mixers there is a certain amount of statistical guesswork re typical draw and typical numbers used. Years ago there was little chance of a value mixer being loaded up with a lot of phantom powered mics, now these are more common.   

 

[Rossi]

IIRC, my old Behringer MX8000 (long out of use) had a 48V rail capapble of about 240 mA for 24 mic inputs. People say it's a crappy mixer, but the mic inputs were actually pretty decent (they used 2SB737 input transistors in early revisions). 10 mA per channel seems about adequate, there are no mics that actually use 14 mA, because that would mean zero voltage, and you're not likely to have all channels shorted. It's enough for 24 real world high current mics at once.

In actual use with maybe 8 phantom power mics connected at the most, the PSU is oversized. But it's a good idea to oversize a little if it helps you to stay out of trouble. As I said before, P48 is meant to be convenient. Scrooging on P48 seems like a bad idea as it undermines the convenience concept.

 

[JohnRoberts]

IIRC, my old Behringer MX8000 (long out of use) had a 48V rail capapble of about 240 mA for 24 mic inputs. People say it's a crappy mixer, but the mic inputs were actually pretty decent (they used 2SB737 input transistors in early revisions). 10 mA per channel seems about adequate, there are no mics that actually use 14 mA, because that would mean zero voltage, and you're not likely to have all channels shorted. It's enough for 24 real world high current mics at once.

In actual use with maybe 8 phantom power mics connected at the most, the PSU is oversized. But it's a good idea to oversize a little if it helps you to stay out of trouble. As I said before, P48 is meant to be convenient. Scrooging on P48 seems like a bad idea as it undermines the convenience concept.

You may not appreciate the sharp pencil nature of mass market design. When I first started at Peavey in the mid 80's I had the luxury of relatively low manufacturing cost compared to direct competitors so I could engineer in more robust margins for such things. By the end of my tenure, some 15 years later the low end market had dramatically changed with several new competitors and Chinese manufacturing becoming commonplace. I no longer enjoyed the luxury of designing in hidden costs (JIC) that were unappreciated by the end user and only I would know are there, since I would then lose at point of sale due to higher price point. Behringer had a reputation for cutting corners in knock offs of other's successful products. 240mA seems like a generous current budget for 24 inputs in a value product. I am pleasantly surprised. 
======
Your favorable impression of behringer mixer preamps invites another of my personal peeves.  Every time there is more than two mic preamps in a single chassis, consumers ASSume they are inferior to stand alone preamps. There is no scientific basis for this belief. 

Yes, the 2SB737 is more than quiet enough to not be the weak link in a SOTA preamp. I personally used the 2SB737s (s=graded for something like 600-800 beta) before working at Peavey, and once there I used the 2SD786 (npn complement to 737) already in their system. We actually put the 786s in the low cost unity 300 mixers (cheap for it's day), but nobody would believe me when I challenged customers to compare them to big dollar preamp channels. 

In consoles, execution matters as much as the components used. So while you can't make a good console using crap parts, good parts don't automatically make a good console. I never personally investigated the Behringer stuff for sound quality, so I'll take your word for it.  YMMV.

JR

 

[Rossi]

Personally, I wouldn't buy a newer Behringer mixer. The older stuff was pretty sturdy, but when all their competitors went to China as well, Behringer kept lowering their prices below the competition. The newer stuff looks and feels so flimsy, I don't even care to check out how it sounds anymore. Plus, I simply outgrew their product line. The only piece of B-Gear I still use is a headphone amp, and I'm not happy with that one, either.

It is true that some of the lower priced external preamps are pretty similar to console preamps. The only things that make them worthwhile are extras such as variable hi pass or polarity reverse etc. You have to invest some money (usually $ 1000+) to get something that is different and sounds better than a console pre. And even those often aren't lower noise (those early Behringers and Mackies for instance were EIN -129 dBu).

 

[Andy Peters]

As noted, USB host ports provide 100 mA until a device enumerates and requests more. A powered hub's downstream ports are the same as a host port in this regard. Unpowered hubs cannot provide more than 100 mA on a downstream port.

I haven't seen any bus-powered USB audio devices that support more than two mic preamps with 48V phantom, likely because they just won't work.

What's interesting is how various computers handle things. My iMac's ports handled overcurrent (faulty device pulling too much) gracefully. My MacBook Pro (on the power supply, not the battery), dealing with the same device, threw up an exciting error, saying that "A device was drawing too much current and was shut down." However, in the former case, the USB VBUS dropped significantly.

-a

Of the USB devices I tested, only one requested more current than 100mA. That was a Zoom guitar interface that acutally contained a 12AX7.

Of the handful of USB audio devices that support phantom power that I've tested, the descriptor dump indicates that they request the full 500 mA.

Manufacturers obviously avoid requesting more than 100 mA, and I suspect its for good reason. It's asking for trouble. It likely won't work on a passive hub, and I'm not so sure an active hub would always work.  The active hubs I have have five or six USB outs and are powered via a relatively small wall wart. I haven't checked, but I doubt that little switching PSU would supply enough current for more than one USB device that draws maximum current. And as you say, many real world PCs and Macs won't supply high current either without giving trouble.

The whole point of the USB current-consumption field in the Device Descriptor is so that the device WILL work. The device reports its actual in-use current requirement. The host, which keeps track of everything attached to the bus, can then make a decision about whether the device should be allowed to operate. So if a device that requests 500 mA is connected to a bus port for which only 100 mA is available, the host will report that the device needs more current and will not be configured.

If the device requests only 100 mA and it requires 500 mA, then it's likely that the host will enable the device, but things will not work because perhaps only 100 mA is actually available. And this is at cross purposes with the USB spec -- it breaks the bus. This is why crap like those dual-type-A cables for hard disks (one for data and power, the other for extra data) are a disaster -- the device reports 500 mA current consumption but it actually uses up to 600 mA and the host driver that handles power management doesn't know that extra 100 mA.

Now in the case of passive hubs, assuming that the upstream port is providing 500 mA, it's obvious that there's no way this hub's downstream ports can source 500 mA. Some current is used by the hub electronics. As for powered hubs, all of the powered hubs floating around the house and at work have supplies that do 2.1 A or 2.5 A at 5V. This is sufficient to provide 500 mA to four downstream ports. And for shits and grins, I've tried plugging in a bunch of high-power bus-powered things to a powered hub connected to my laptop and they all worked. I suppose that if I was worried that the wall warts weren't doing as they say, I could put a load on them and test them.

So 100 mA is likely the most you can safely assume a USB port will supply.

No, that is not true. You can assume that a host port (built into the computer) will supply up to 500 mA. Poorly-designed laptops cheat and do not provide 500 mA, but that's OK assuming that they are honest about their sourcing capability and as such disable a high-power device.

The problems occur when devices do NOT adhere to the spec, and unfortunately that's all too common, since the USB Implementer's Forum (the group that assigns Vendor IDs and promulgates the spec) is quite toothless indeed.

-a

 

[JohnRoberts]

Thanks for the USB 411...

New question.. Do you feel the 5V and 500ma capability are safe? As in safe to depend upon for future products. 5V is already getting long in the tooth for digital logic, so I would expect some pressure to move USB or a USB like interface to 3.3V or lower.

JR