I'm in the planning stage of designing a pair of USB-powered speakers for my laptop. As of now USB is only going to be supply power, no DAC or anything.
Regardless, as you might know USB1.0/2.0 spec is only up to 500mA @ 5V...which is 2.5W.
I need to make sure that my amplifier does not exceed 2.5W, right? One of my ideas is that I would step up the 5V via a buck boost regulator. So say that I step the 5V to 10V, does this mean that now I can have 10V, but only up to 250mA power draw vs 500mA before? That would still be 2.5W.
I'm thinking dual LM386s, possibly TDA2822. I'm trying to determine which chips to use for maximum power without killing the USB bus. The NJM 386s have lower THD than the TI/National parts...not sure why, or if it's really true, but I might be leaning towards LM386. Seems kind of lame though...
I want to have 2 drivers per channel. Small midrange and a small tweeter, though a full range might cut it.
Any ideas?
Thanks.
Regardless, as you might know USB1.0/2.0 spec is only up to 500mA @ 5V...which is 2.5W.
I need to make sure that my amplifier does not exceed 2.5W, right? One of my ideas is that I would step up the 5V via a buck boost regulator. So say that I step the 5V to 10V, does this mean that now I can have 10V, but only up to 250mA power draw vs 500mA before? That would still be 2.5W.
I'm thinking dual LM386s, possibly TDA2822. I'm trying to determine which chips to use for maximum power without killing the USB bus. The NJM 386s have lower THD than the TI/National parts...not sure why, or if it's really true, but I might be leaning towards LM386. Seems kind of lame though...
I want to have 2 drivers per channel. Small midrange and a small tweeter, though a full range might cut it.
Any ideas?
Thanks.
Yeah that's good but it only comes in BGA...relatively useless for DIY unless I get a rework station.
But that does look like ideal spec..
But that does look like ideal spec..
Oh, it'll run off 5V. Connect it up with an ammeter in the supply circuit, drive it with something while monitoring the current, limit the input so the current draw doesn't exceed 500mA.
You'd do well to study the USB specification in more detail. From my recollection, the 2.5W power is only available once the connected device has negotiated with the host and the host enables it. Prior to this the power availability is lower, maybe just 0.5W.
What I think this means in practice is if you want your laptop to deliver 2.5W to the speakers, you'll need a USB receiver chip in the speakers. Or you could just power the speakers from a USB wall-wart - to my knowledge they have no such restriction on max power.
What I think this means in practice is if you want your laptop to deliver 2.5W to the speakers, you'll need a USB receiver chip in the speakers. Or you could just power the speakers from a USB wall-wart - to my knowledge they have no such restriction on max power.
Aah, didn't notice that, but you may find something similar from another supplier. Or maybe figure a way to deal with the fact that it's a BGA.
Actually after reading, it seems that no negotiation really has to be made at all, since if we leave D- and D+ untouched, the USB port just acts as a 5V power supply. I guess proper protocol IS to ask the computer for permission to draw x amount of current, but it doesn't seem necessary. I don't know under what circumstances the negotiation is required. Perhaps if it is a digital device like an iPod.
It appears that provided the USB port and cable are rated for it, USB can supply up to up to 1A, (not including 1A) since there is a PPTC on every USB port rated and is at 1A. It doesn't have a reason to limit nor to permit higher current flow.
If you read here, you can see that the device must pass all of these specs to meet in order to allow the full 500mA...but I believe this is mostly for digital devices that will use their USB port for other things such as data transfer and it will be forced to make the handshake with the computer. With an analog circuit, theres no MCU to handshake with, and thus I think the handshake thing does not apply.
However, not to cause problems, I think it would be wise to keep it under 500mA for whatever reason.
Am I wrong with my theory?
It appears that provided the USB port and cable are rated for it, USB can supply up to up to 1A, (not including 1A) since there is a PPTC on every USB port rated and is at 1A. It doesn't have a reason to limit nor to permit higher current flow.
If you read here, you can see that the device must pass all of these specs to meet in order to allow the full 500mA...but I believe this is mostly for digital devices that will use their USB port for other things such as data transfer and it will be forced to make the handshake with the computer. With an analog circuit, theres no MCU to handshake with, and thus I think the handshake thing does not apply.
However, not to cause problems, I think it would be wise to keep it under 500mA for whatever reason.
Am I wrong with my theory?
Last edited:
I just found this Maxim article which seems to imply that in practice, the specs aren't necessarily implemented in full:
Charging Batteries Using USB Power - Maxim
So while it states : All devices that plug into a USB port must start out drawing no more than 100mA. After communicating with the host, the device can determine if it can take the full 500mA.
it also, down at the bottom has a couple of paragraphs under the heading
What Your Mom Didn't Tell You About USB
which look a lot more promising for your application.
Guess 'suck it and see' ? 😕
Charging Batteries Using USB Power - Maxim
So while it states : All devices that plug into a USB port must start out drawing no more than 100mA. After communicating with the host, the device can determine if it can take the full 500mA.
it also, down at the bottom has a couple of paragraphs under the heading
What Your Mom Didn't Tell You About USB
which look a lot more promising for your application.
Guess 'suck it and see' ? 😕
Very interesting. Maxim being funny I see. 😀
Maxim says that USB ports are not current limited.
Okay.
But then all throughout the article it says how they are limited. What I've deduced is that current is limited based on the host's limitations and the device's needs.. So if you are on a unpowered hub, and then the device asks for 500mA, it will say "no, sorry I am an unpowered hub. You get 100mA max". However, if you plug it into a computer and the device asks for 500mA it will allow it 500mA. I guess this is a good safety precaution to prevent hubs from melting. 😛
That's what I got from this. Tell me if I'm wrong.
But I still don't get how all of that exists but yet I can simply bypass that? Perhaps the more professional way to do it is with the fancy USB chips to prevent things from going tragically wrong?
Well one solution I can see is to make the speakers digital in so you have a reason for putting that USB receiving CPU in there. Do you really want to listen to your laptop's on-board DAC??? 😀
<edit> If you do that, then you'll have an excuse to use this uber-cool chip : http://www.nxp.com/documents/leaflet/75017060.pdf
<edit> If you do that, then you'll have an excuse to use this uber-cool chip : http://www.nxp.com/documents/leaflet/75017060.pdf
Last edited:
Are you sure its birdies on the line (VBUS) itself, and not common-mode birdies owing to ground loops? Differential mode birdies can be filtered out (ferrites, chokes and caps) at the expense of some voltage headroom.
I don't think a little bit of noise will be too much of an issue. The whole setup is budget minded.
That's NXP chip looks nice. Too bad I'm not good with I2C.
That's NXP chip looks nice. Too bad I'm not good with I2C.
DC, which translates to a maximum AC output power of 0,312 W into 10 Ohm not counting losses. Into any other load you get even less output power. You can connect several USB outlets in parallel. That practice is often used to feed external drives with higher current consumption than 0,5 A.
Hmmph. I just found an old set of speakers in my basement that are USB powered. Piece of shite Cyber Acoustics- with a TDA2822 chip...One USB port for power...hmmm....
Which is a good indication of how little power is necessary to listen to music, the quality issue aside.
Some of the newer USB speakers are doing power harvesting (think F1 race cars with KERS) see:
http://www.theaudiocritic.com/plog/
http://www.theaudiocritic.com/plog/
Its interesting but not really power harvesting:
...the Olasonic increases that to 10 + 10 watts on peaks with a proprietary circuit called the Super Charged Drive System (SCDS). The SCDS stores USB power as an electrical charge in a high-capacity condenser during low-level outputs and releases the power on dynamic peaks.
The idea is to try to maintain a constant power draw on the USB so that peaks are handled from internal storage rather than the PC.
...the Olasonic increases that to 10 + 10 watts on peaks with a proprietary circuit called the Super Charged Drive System (SCDS). The SCDS stores USB power as an electrical charge in a high-capacity condenser during low-level outputs and releases the power on dynamic peaks.
The idea is to try to maintain a constant power draw on the USB so that peaks are handled from internal storage rather than the PC.
- Status
- Not open for further replies.