any reccomendations for a fairly compact, low power amp?
Oct 27, 2004 at 10:20 PM Thread Starter Post #1 of 21

nleahcim

100+ Head-Fier
Joined
Nov 27, 2001
Posts
437
Likes
10
First off - sorry if this is in the wrong place... Debated between amps and here - but I thought this would probabaly be best...

Anyways - so I'm building Guzzler's USB DAC - and now I just found out that it's output is a line out, not a headphone out
frown.gif


Anyways - so I would like to build a simple headphone amp. But just because I like to make things complicated, I'd like to have it run off the USB power supply, the same supply the USB DAC will be running off of. Guzzler said this would be a problem - that I would need an AC coupled design... Not that I know what that is
icon10.gif


Anyways - so I should mention my capabilities I guess - I'm a second year EE student and have a large amount of experience with digital circuits - but next to none with analog audio circuits
frown.gif
Guzzler's USB DAC is my first diy audio project. I can solder anything you put in front of me. I can have any dual sided board milled for me (one of the nice things about being a student and having the right connections).

So - does anybody have any suggestions? I'd prefer not to spend over $50 or so on it...
 
Oct 27, 2004 at 10:29 PM Post #2 of 21
I think USB power was +/- 3.3V DC, not AC. If that's the case, you could build a simple CMoy or MINT well within your budget, using an opamp that's not too voltage hungry (Tangent's opamp guide has details on which ones like low-voltage).

You may also want to consider using seperate battery or wallwart DC power for the amp. USB power is pretty noisy, and odds are you'll hear a big improvement if you seperate the amp's power supply.

What headphones are you looking to drive with this?
 
Oct 27, 2004 at 11:59 PM Post #3 of 21
Quote:

Originally Posted by Megaptera
I think USB power was +/- 3.3V DC, not AC. If that's the case, you could build a simple CMoy or MINT well within your budget, using an opamp that's not too voltage hungry (Tangent's opamp guide has details on which ones like low-voltage).

You may also want to consider using seperate battery or wallwart DC power for the amp. USB power is pretty noisy, and odds are you'll hear a big improvement if you seperate the amp's power supply.

What headphones are you looking to drive with this?



USB VCC is +5V... Headphones are A500s, though I'll probably soon be getting some Shure E2Cs soon as well... I suppose when I'm not mobile I could use a seperate power supply - but when I am mobile, I will want to power it from USB...
 
Oct 28, 2004 at 1:38 AM Post #4 of 21
Ahh, laptop, that makes sense.

You're right, USB is +5V. Hmm, not +/-5V -- that means you'll need an opamp that's really happy with ~2.5V, and a rock solid rail splitter. To be honest, I don't know of any opamps that go that low in voltage, though I'm sure some exist. I guess what I would do is find an opamp that fits those needs, then build a simple feedback circuit around it if for some reason it's not pin-compatible with the more common portable designs.

For a rail splitter, you'll want to use a TLE2426 instead of a simple resistor divider like the CMoy uses, because a resistor divider is prone to getting imbalanced easily. Even if it's off by just a few mV, any opamp you find that's good at 2.5V probably won't like it if the supply drops by what's effectively a large percentage.

Actually, I bet there's a way to increase the voltage at the cost of current. I'm not sure what you would need to do that, but it'd give you more choice with opamps. Of course, let's say you double the voltage to +10V, you'd be down to, what, 250mA from the USB supply, right? That's more than enough for a headamp, but I don't know how much the DAC draws.
 
Oct 28, 2004 at 4:52 AM Post #6 of 21
Quote:

Originally Posted by bg4533
If you have an extra USB port available you might be able to use the power from 2 in series to get 10V.


USB ports in the same "cluster" have the voltage rails all tied together, so that wouldn't work, you'd get a short where you tried connecting the two together.

Have you looked into a switching supply to increase the voltage? I know that the Airhead and TAH use a switching supply, and they sound fine, and the power supply feeding the USB is already switching, so it can't hurt any more than it already is.

I'd say that your best bet would be to look on mouser for a nice DIP8 switching DC/DC convertor, and then bump your 5V to something around 12V, and then split that so you're working on +-6V. You can probably make a nice little Cmoy that way, or maybe a mint (don't know if it would be worth the hassle to make a mint if you won't be using a better supply at least part of the time).
 
Oct 28, 2004 at 12:28 PM Post #7 of 21
It occured to me that what you could do is simply use an open-loop buffer. The buf634 will work at 2.25V, and even though it sounds better inside an opamp's feedback loop, you can get away without it. Other open-loop buffers may work better for this purpose, too. I guess it depends on how easy to drive the A500s are (I don't know them), but you really may not need any gain here, which makes things a lot simpler.
 
Oct 28, 2004 at 2:58 PM Post #8 of 21
Quote:

Originally Posted by Megaptera
Ahh, laptop, that makes sense.

You're right, USB is +5V. Hmm, not +/-5V -- that means you'll need an opamp that's really happy with ~2.5V, and a rock solid rail splitter. To be honest, I don't know of any opamps that go that low in voltage, though I'm sure some exist. I guess what I would do is find an opamp that fits those needs, then build a simple feedback circuit around it if for some reason it's not pin-compatible with the more common portable designs.

For a rail splitter, you'll want to use a TLE2426 instead of a simple resistor divider like the CMoy uses, because a resistor divider is prone to getting imbalanced easily. Even if it's off by just a few mV, any opamp you find that's good at 2.5V probably won't like it if the supply drops by what's effectively a large percentage.

Actually, I bet there's a way to increase the voltage at the cost of current. I'm not sure what you would need to do that, but it'd give you more choice with opamps. Of course, let's say you double the voltage to +10V, you'd be down to, what, 250mA from the USB supply, right? That's more than enough for a headamp, but I don't know how much the DAC draws.



I think the usb dac draws under 50ma - so it shouldn't cause much of a problem

Quote:

Originally Posted by The_Mac
USB ports in the same "cluster" have the voltage rails all tied together, so that wouldn't work, you'd get a short where you tried connecting the two together.

Have you looked into a switching supply to increase the voltage? I know that the Airhead and TAH use a switching supply, and they sound fine, and the power supply feeding the USB is already switching, so it can't hurt any more than it already is.

I'd say that your best bet would be to look on mouser for a nice DIP8 switching DC/DC convertor, and then bump your 5V to something around 12V, and then split that so you're working on +-6V. You can probably make a nice little Cmoy that way, or maybe a mint (don't know if it would be worth the hassle to make a mint if you won't be using a better supply at least part of the time).



I haven't yet looked at switching supplies yet - I hadn't actually expected 5V VCC to be a problem - I mean that's what *everything* I make runs off of... So what kind of voltage do I need? About switching supplies - any reccomendations? I would think I'd need to be fairly careful in selecting one - as a bad supply would ruin a perfectly good amp...
 
Oct 28, 2004 at 3:28 PM Post #9 of 21
no no! You can't use a rail splitter design if you want to run the amp off the same supply as the DAC is from. The ground on the DAC is 0V, so if you have a virtual ground type amp, you short virtual ground to the 0V (-ve) rail. An example of an AC coupled amplifier is here:

http://headwize.com/ubb/showpage.php...=coupled;CMoy;

It is the design by PRR.

As an alternative, search for jamonts thread about running a MINT with his DAC. As for opamps, I can suggest the AD823. I very much like the sound of it, and it runs well down to about 3.3V or so.

g
 
Oct 28, 2004 at 4:19 PM Post #10 of 21
Ooh, yeah, if the DAC isn't splitting the voltage, splitting it in the amp would be a problem. Didn't realize that about the DAC's design.
 
Oct 28, 2004 at 11:10 PM Post #11 of 21
(This is my first post, so please be gentile...
rolleyes.gif
)

Maybe this could be of some use...

I have built a USB Dac with a mint powered by USB power. I used a max743 regulator which takes +5V (USB) and puts out +-12/15V with the same ground potential. It puts out 125mA in +-12V mode, so you can power to much with it, buts enough for my mint with ad8620 and 2 buf634, no biasing and low-bandwidth. This is also on the limits of the USB supply, but it works fine for me on all computers I've tested. I also don't have any noise issues.

I don't know if this will work for others, but I like it a lot with my hd580 and hd25's. If you want I can post some pics (I don't have schematics, just implemented the datasheet.)
 
Oct 28, 2004 at 11:52 PM Post #12 of 21
Quote:

Originally Posted by guzzler
As an alternative, search for jamonts thread about running a MINT with his DAC.


It's easy to do this with the right power supply, I have a +5V, +-12V Elpac from ebay that works well for this purpose. I've been using my DAC+MINT for a couple of months now and I'm very pleased with the sound.
 
Oct 29, 2004 at 5:51 AM Post #13 of 21
Quote:

Originally Posted by Jeroen
(This is my first post, so please be gentile...
rolleyes.gif
)

Maybe this could be of some use...

I have built a USB Dac with a mint powered by USB power. I used a max743 regulator which takes +5V (USB) and puts out +-12/15V with the same ground potential. It puts out 125mA in +-12V mode, so you can power to much with it, buts enough for my mint with ad8620 and 2 buf634, no biasing and low-bandwidth. This is also on the limits of the USB supply, but it works fine for me on all computers I've tested. I also don't have any noise issues.

I don't know if this will work for others, but I like it a lot with my hd580 and hd25's. If you want I can post some pics (I don't have schematics, just implemented the datasheet.)



It may be running near the limit of the official USB power supply specifications, but I've never seen or heard of a motherboard or hub that can't provide much more amperage than you would need.
 
Oct 29, 2004 at 9:30 AM Post #14 of 21
We discussed USB power limitations a while back. According to the official spec, the USB controller has to enumerate everything that wants to draw over 100mA (1 unit load) from a single port. However, in practice this is rarely implemented. Indeed, Maxim even published an engineering guide about it and said they had never any limited ports. The offical spec is for 500mA to be available from the port, but this varies between different manufacturers: some will only go up to 500mA, others on light load will go to 1-1.5A. But, as always, it's better to err on the side of caution when it comes to power issues.

The switching supply jereon suggests is one way round this, and if you want to have your DAC and amp powered by USB it seems sensible. Miss out the railsplitter in the MINT, and wire directly from -12V, GND, +12V. In the end, you can choose between jamonts and jereons implementations depending on cost and size requirements.

Regarding the DAC being line out, we decided on this ages ago as it was felt most who would want to build a DAC would have an amp of some variety as well.

g
 
Oct 29, 2004 at 11:16 PM Post #15 of 21
Quote:

Originally Posted by guzzler
We discussed USB power limitations a while back. According to the official spec, the USB controller has to enumerate everything that wants to draw over 100mA (1 unit load) from a single port. However, in practice this is rarely implemented. Indeed, Maxim even published an engineering guide about it and said they had never any limited ports. The offical spec is for 500mA to be available from the port, but this varies between different manufacturers: some will only go up to 500mA, others on light load will go to 1-1.5A. But, as always, it's better to err on the side of caution when it comes to power issues.

The switching supply jereon suggests is one way round this, and if you want to have your DAC and amp powered by USB it seems sensible. Miss out the railsplitter in the MINT, and wire directly from -12V, GND, +12V. In the end, you can choose between jamonts and jereons implementations depending on cost and size requirements.

Regarding the DAC being line out, we decided on this ages ago as it was felt most who would want to build a DAC would have an amp of some variety as well.

g



So PRR's design seems to be designed for a 12V source - what would I need to change to get it to work at 5V?

It sounds, though, like having only 5V available is going to be quite the limiting factor, so without adding an additional power source (batteries or wallwart) - maybe a chip like the MAX743 might be a good idea? Is 24V a good supply voltage to work with?

And so if I do end up using something like the 743 - any reccomendations for a small amp that doesn't draw *too* much power?

edit: if I'm reading correctly - jamont's solution was to get an external power supply right? I'd really prefer not to do that...
 

Users who are viewing this thread

Back
Top