Design Idea: USB > 30V Battery Power Supply

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Bo Deadly

Well-known member
Joined
Dec 22, 2015
Messages
3,266
Location
New Jersey, USA
I'm thinking about a new idea for a power supply that could be handy for a few common annoying scenarios.

First I start with a battery pack of 2/3AAA NiMH cells:

  1.2V
  400mAh
  R 10.2mm
  H 29mm

61KKQupsyKL._AC_SX425_.jpg


These are really common because they're used in solar installations and so they're cheap. You can get them for ~0.75 USD each in quantities of 20 or so.

A pack of 24 makes 28.8V and of course there's a center tap to make a +-14.4V supply. The size of this pack is about the same as three 9V batteries.

Trickle charging is ~0.04 capacity so 400mAh * 0.04 = 16mA.

BattPack24.png


Then I use a MeanWell DETN01L-15 SMPS:

  DCDC
  Isolated
  Unregulated
  SMT
  5V / 253mA in (USB)
  30V / 34mA out (+-15V)
  L 15.3mm
  W 10.9mm
  H 7.1mm
DETN01_DSL.jpg

This is one of several 1W +-15V DC converters from MW and they're cheap ($5 USD).

So the overall idea is to do USB power -> DETN01L-15 -> CLC filter -> charging / iso circuitry -> 28.8V NiMH battery -> +-14.4V.

This could be a great supply for small pro-audio gizmos. Many power strips have USB ports now. You could plug it into your computer. Or use one of those ubiquitous Apple mains / USB adapters and use wall power. The center tap is isolated from USB through the batteries so even though USB is notoriously noisey, it might be quiet enough to use plugged in and trickle charging. Or unplug it and run on battery. A 20mA gizmo like a DI would last 20 hours.

So what do you think? Will it work? I don't really care what people think about tapping USB for pro-audio. If it works it works. USB ports are everywhere.
 
I'm not sure what problem you are trying to solve. Batteries are a PIA, I have designed a couple battery powered drum tuners.

My first a 9V powered not very efficient design ate batteries for breakfast, only 250mAh. My latter powered by 4xAA and using a class D audio amp only sips battery power. I use 4x 1.5V (nominal) standard cells in series to realize only 3.3V to run the micro and associated glue, but operate the class D amp from pre-regulated voltage (or tried to). I used a CMOS LDO 3.3V regulator so I could have easily gotten away with only 3 cells, but I was going for crazy long battery life. With 4 cells my unit does not drop out of regulation until <0.9V per cell, but using 4 cells created a new problem. The voltage of batteries is "nominal" (in name only). Four fresh cells at around 1.65V each exceeded the substrate max voltage spec for the small class D chip, that wasn't pulling from the regulated 3.3V rail. I had to add a crude pre-regulator to manage the too much voltage from fresh batteries when new.

The PS design works flawlessly and I just got a tuner back for a software upgrade, the old batteries were a few years old and still measured 1.5V. I always replace the batteries with new during my flat fee software upgrade, but returned these still good batteries to the customer.

======

My solar driveway lamps use one nicad AA cell stock. I have highly modified mine, because I could. 8) I cobbled two solar cells together in series to get more power on cloudy days. I changed the value of the inductor in the charge pump to lower current drain after dark when driving the output LED (still plenty bright). My final upgrade was to use higher capacity niMH cells and they do indeed work better-longer than nicad.

I was pleasantly surprised that the lamp driver IC shrugged off my major modifications (like 2x solar cells) and has been operating without complaint for years. I suspect the solar cell output is loaded down by charging the battery, reducing the possibility of over voltage to the IC. 

JR
 
The "problem" being solved is that, even though almost everything pro-audio is now digital, the end points are and always will be analog and the end points can be hard to get power to. There's always going to be guitar pedals and keyboards and Space Echos and beat boxes and the little portable gizmo that the guy built so on. All of those devices source or sink signal that should be at pro levels. Ironically, with all of the attention paid to noise and distrotion, traditionally these devices have been accomodated with little 9V battery powered DI boxes or phantom powered devices that have low headroom and weak drive.

Another reason for using batteries is that they could act as a buffer almost like capacitors in a way [1]. So a noisy 5V USB supply is stepped up to 30V and buffered / stored in the 28.8V pack. And you get a center tap for ground. 

Also there is a convenience to batteries and especially ones that can be recharged from USB which is everywhere from your entry table, your laptop, the car, an airplane, etc. I'm not sure your PITA point applies to this idea because, unlike your drum tuner, I'm talking about using NiMH. Being a one-off design it is a PITA to make new packs but presumably they would last a few years (espcially if they live an easy trickle charging life) and I could make them 2-3 at a time.
 
Here's an example of what one of these battery powered gizmos would look like:

SmEncl0.png


So the cable clamp would have a USB cable and 4 signal cables w/ XLR or TS or whatever. The data port is audio over ethernet. Two dual op amps could make 4 sturdy outputs. Now you can put this thing at the end of a potentially very long cable for a plate reverb or in some industrial application where the power isn't reliable or maybe just connected to your laptop for analyzing gear in the field. There's lot's of uses.

[1] I could be wrong about the batteries being like a buffer because I have not yet devised the circuit to ensure that the batteries are always getting 0.04 C to trickle charge regardless of the load being driven. And I might have to connect the center tap to the SMPS to ensure that the batteries are not discharged unevenly. So it's not crystal clear how well I can isolate the downstream supply from the SMPS which no doubt will be a noisy source. Forturnately it's only 34mA (or maybe I use the 2W SMPS and get 67mA) so chokes in the >10mH range are available.
 
"Dude sorry I forgot to buy 9v batteries" becomes "Dude sorry I forgot to charge my pedalboard".  ;D

I like the concept, especially for things like field recording.
 
Speedskater said:
What about using electric drill batteries?
Because they're huge. 2/3AAA are tiny. I thought about using A23 but inside they're actually button cells and so A23 rechargeables are not that common. 2/3AAA are everywhere and inexpensive. But pre-fabbed 14.4V packs (probably used in power tools actually) are pretty common so 2 of them would work with a scaled up charging / filtering circuit. This whole idea scales pretty well incidentally. But for a small gizmo, it's worth making little packs out of 2/3AAA. I was thinking about making a PCB with big holes but I suppose I should just make them like everyone else by soldering them together with flat bus wire, shrink wrapping it and adding some leads with a JST connector.
 
squarewave said:
I'm thinking about a new idea for a power supply that could be handy for a few common annoying scenarios.

First I start with a battery pack of 2/3AAA NiMH cells:

  1.2V
  400mAh
  R 10.2mm
  H 29mm

61KKQupsyKL._AC_SX425_.jpg


These are really common because they're used in solar installations and so they're cheap. You can get them for ~0.75 USD each in quantities of 20 or so.

A pack of 24 makes 28.8V and of course there's a center tap to make a +-14.4V supply. The size of this pack is about the same as three 9V batteries.

Trickle charging is ~0.04 capacity so 400mAh * 0.04 = 16mA.

BattPack24.png


Then I use a MeanWell DETN01L-15 SMPS:

  DCDC
  Isolated
  Unregulated
  SMT
  5V / 253mA in (USB)
  30V / 34mA out (+-15V)
  L 15.3mm
  W 10.9mm
  H 7.1mm
DETN01_DSL.jpg

This is one of several 1W +-15V DC converters from MW and they're cheap ($5 USD).

So the overall idea is to do USB power -> DETN01L-15 -> CLC filter -> charging / iso circuitry -> 28.8V NiMH battery -> +-14.4V.

This could be a great supply for small pro-audio gizmos. Many power strips have USB ports now. You could plug it into your computer. Or use one of those ubiquitous Apple mains / USB adapters and use wall power. The center tap is isolated from USB through the batteries so even though USB is notoriously noisey, it might be quiet enough to use plugged in and trickle charging. Or unplug it and run on battery. A 20mA gizmo like a DI would last 20 hours.

So what do you think? Will it work? I don't really care what people think about tapping USB for pro-audio. If it works it works. USB ports are everywhere.

I have a better circuit that fully utilized the 5V USB port, and if there's no USB port to tap into, you can use your phone charger which also puts out 5V but packs more current. Alternatively, you can also use a single Lithium 4.2V battery in the thousand mAH, and if you don't want to buy new batteries, just salvage some good ones from weak laptop batteries that people threw away. I have used this circuit in all of my recent portable/non-portable pro audio projects.

It takes in 3V upto 30V DC, filtered with a small 10uF cap and a ferrite bead. Output is adjustable symmetrical (+/- and GROUND) DC voltage.
Main IC is XL Semi XL4009E1 (4A) or XL4008E1 (3A) with a switching frequency at 400kHz.
Pretty quiet. I don't even need to add linear regulators at the output, but if you want, you can do so.

If you want, I can post the schematic and the working board design (2" x 2", 2 layers all SMD)
 
Sure thing. Here's a first photo of the circuit in one of my newest custom pro audio project (AIO talkback mic+studio monitor controller+4 channel headphone distribution amplifiers) that I'm building for a friend.

VIN = 5V from a DC power supply adaptor, not from my computer USB port because I lend my USB male to DC male barrel jack cable to another friend. I'm using a DC barrel jack female port instead of a USB female port for this project because that's what I had at the time I was designing the mainboard, lol.
 

Attachments

  • IMG_20201026_164122.jpg
    IMG_20201026_164122.jpg
    307 KB · Views: 17
Here's the schematic taken from the original circuit I found on Banggood.
The Banggood version uses additional linear regulators at the output.

To get a different output voltage, use this formula.
VOUT = 1.25 x (1 + R2/R1)

So, for 15V R2 is ((15/1.25) - 1 = 11K) not 12K
 

Attachments

  • USB power.png
    USB power.png
    45.5 KB · Views: 27
The board design. Yep, I have JLCPCB manufactures all of my boards for me, because they're fast, reliable, and cheap.

IC is XL Semi XL6009E1, the 4A version of XL6008E1.

Power inductors are Sumida CDRH127NP-330MC

All electrolytic capacitors are low impedance, long life (Chemicon GXE/LZX/LZH/KZM/KZH) (Nichicon HE)
Chip capacitors are 1uF 50V 1206 X7R 10%

Rectifiers are Vishay SS34.

Resistor is *suggested thin film 1% 1206* but I just used what's available in my bin which is a thick film 1% 1206, and noticed no difference in performance.

Trimpot is Bourns 3296P 50K.

Ferrite bead I have no idea what the brand is or the operating frequency is, the only info I have is that it has a max current rating of 1.1A. Bought it from a local online parts shop.
 

Attachments

  • USB power.jpg
    USB power.jpg
    187.3 KB · Views: 11
I had the boards made awhile back, didn't use trimpot for R2, so they have a fixed output voltage.
 

Attachments

  • IMG_20201026_54627.jpg
    IMG_20201026_54627.jpg
    67.3 KB · Views: 15
There ya go. Powered from my laptop USB port ;D
 

Attachments

  • IMG20201026173904.jpg
    IMG20201026173904.jpg
    79 KB · Views: 22
Cool. Nice work. But it's not isolated so the ground of your device will be connected to the ground of wherever USB is plug into which could potentially be very bad (sometimes you can get DC currents that cause horrendous noise).

Also, note that the power of my circuit is not limited to 34mA. That's just for charging the battery. The battery can supply much more current. Although on average, any more than about 20mA and the battery will start to drain. So it is intended exclusively for small device like DIs.
 
squarewave said:
Cool. Nice work. But it's not isolated so the ground of your device will be connected to the ground of wherever USB is plug into which could potentially be very bad (sometimes you can get DC currents that cause horrendous noise).

The only one time I did experienced a nasty ground loop was when I used it to power a Bluetooth audio transmitter/receiver together with my audio circuits.

Then again, you are free to modify it by adding a ground isolator to the input like the B050SS (5V @1A) for an example.

Also, note that the power of my circuit is not limited to 34mA. That's just for charging the battery. The battery can supply much more current. Although on average, any more than about 20mA and the battery will start to drain. So it is intended exclusively for small device like DIs.

I totally get it man. Sorry that I hijacked your thread. I was merely offering an alternative that gives you more options and flexibility.
 
L´Andratté said:
Proof! ;D
Thanks for uploading! Have to think about it...
Why do you think it is "better" design? Bigger input range, lower noise?

Wider input and output range. More options for input source, like I stated before, you can utilize:

computer USB port,
AA/AAA battery pack,
single Lithium 4.2V battery,
9V battery,
phone charger,
DC power supply adaptor,
SMPS, or
whatever that can supply at least 3V.

As for lower noise, so far I have not experienced any unwanted noise (ground loop, or switching noise) in my audio signals. I suppose if you want more supply noise rejection, you can either add a DC ground isolator at the input (or output), or add linear regulators to the input (or output)

To add a 7815/7915 pair combo at the output, you'd need 18.5V (R2 = 13.8K, R1 = 1K)
 

Latest posts

Back
Top