[Nemo de Monet]

Quote:

Originally Posted by rds /img/forum/go_quote.gif I agree with joneeboi - don't waste that current. Keep it for headroom. Do you really need a LED to tell you that the power supply is working? Personally I'd be able to tell that something is wrong when the amp is turned on and no sound is coming out

Is an extra, oh, 2mA of current really going to help any? You wouldn't be losing voltage swing, just - literally - a few mA of current, where there is, under every circumstance I can think of, going to be plenty available.

Also, isn't that why we put blinkenlichten on these things in the first place - to indicate, in fact, that they're on? As the design stands, all the LED will tell you - because the switch is on the other side of the converter - is that the amp is plugged into a running computer. Even when you turn the amp off, the LED will stay lit. How useful is that?

Quote:

Besides that there's no reason to think the dc-dc converter will suddenly die. By this logic you'd want LEDs for every chip to indicate whether they are working or not.

Um, no. See my comments above about the relative uselessness of the current LED position. Also, have you read the datasheet for the converter? It'll shutdown if it gets too hot - thermal overload protection, natch - but I'm unclear what if anything it'll do if you try to pull more current than it can provide (due, for example, to oscillating opamps, or a huge capacitor inrush current). Sag in voltage? Hold voltage and current steady at the maximum limit? Shutdown? Couldn't tell you. There are also a couple fun circumstances where the chip will shutdown, just as one example, if the input voltage at the chip sags too low, however momentarily.

My feeling is that if someone puts the amp together for the first time, plugs it in, gets a lit LED, turns it on, and nothing happens, that lit LED should provide some more useful troubleshooting information than "your USB port is supplying at least some voltage"; putting it on the isolated side of the converter will at least say "the power supply circuit is working as intended", or will narrow the problem down to one at the power end of things.

 

[Juaquin]

Not sure what all the fuss is about. Apparently people just do not want to see USB powered anything. I mean, an LED? Keep it in, cut it out, whatever. It's just a DIY project (no offense to the project).

 

[DKJones96]

I'd love to see more USB powered amps, I'm just not sure why you need to boost the voltage. I've heard some nice op-amps that work at under 5v. The USB dac/amp I'm working on runs on 3.3v(+-1.65v) for the amp. I've tested the amp on headphones at that voltage and it gets more than plenty loud(80ma/Amp ftw). In fact, once I found the right caps for it, aside from being loud and clean instead of distorting at high volumes you can't even tell it is there. But my goal is always transparency.

 

[rds]

Quote:

Originally Posted by Nemo de Monet /img/forum/go_quote.gif Is an extra, oh, 2mA of current really going to help any? You wouldn't be losing voltage swing, just - literally - a few mA of current, where there is, under every circumstance I can think of, going to be plenty available. Also, isn't that why we put blinkenlichten on these things in the first place - to indicate, in fact, that they're on? As the design stands, all the LED will tell you - because the switch is on the other side of the converter - is that the amp is plugged into a running computer. Even when you turn the amp off, the LED will stay lit. How useful is that?

Just move the switch.

 

[joneeboi]

I really appreciate all the discussion that's going on. It's good.

Allow me to give some context. I'm boosting the voltage because I'm trying to replicate the Mini^3. In some of my feeble attempts at producing a board for a CMoy, I kept coming back to what the Mini^3 had and was. I talked to amb about it, so now I'm more worried about doing the name justice. I was thinking of trying a chip that accepts 3.3V to 4.75V, but might come after this project.

Boosting to +/- 12V doesn't give me enough current, plus the 5V-in, 12V-out chip that this paper says exists no longer is listed on the DCP02 product page. Having 2W would be a lot easier, but neither Mouser nor Digikey supply the product which seems to no longer exist. Plus, the OPA690 doesn't accept that high of a power supply voltage.

Regarding the LED, I put it in at the end as an afterthought because, as rds mentioned, I don't really care about the light indicator, plus I could easily move the switch. If it plays music, it's good enough for me. I merely put it in as an afterthought for people who might want to build this thing as well, but personally, I wouldn't put in an LED. I was thinking of just using the planes to power it, thereby removing any need to route anything, but I didn't want to waste current from the TLE. I put it in the back just to satisfy the mere fact of having an LED, but I clearly didn't think it through. That all being said, it's just an LED. Let's worry about something more important for now.

I'm quite tired of looking at the board editor in EAGLE, so let's discuss schematic and layout changes all at once to reduce my going back and forth to switch little bits here and there. I figured it would be good to put a capacitor between the TLE and TL750 as an input/output capacitor. They're so close I didn't think it would matter, but that seems to not really make sense. I'll throw C3 between the two and make it some ceramic/low ESR cap. I was also thinking of rotating the ground channel opamp just like how it is in the Mini^3. I don't know why I didn't start off like that. Anyway, that's something that will help with the parts placement, and it'll be more familiar and easier to build. Reducing the size of the test points will also help with the board space. Something else I was wondering about was the meaning of something in the TL750 datasheet. It says the output regulation voltage is 50mV to 100mV. Surely, that doesn't mean its output ripple is 50mV to 100mV, right?

Another big question is about the wires. If I'm connecting both ground planes to the case, will I even need to connect the audio or power supply ground wire from the BantamDAC to the board? The ground plane will lie underneath the board, and it will connect to the case. There should be lots of current through those exposed planes, so it will save a little room and wiring if I don't have to connect that. The USB connector will sit in the back and will presumably connect to the case. Either way, it might be easier to connect the wires anyway, but I just wanted to feel out the possibility.

Also, I was wondering what the full power supply draw will be from the AD8397. It says the supply current is maximum 12mA/Amplifier, but it also says the highest output current is 250mA at 3Vpp and Rload 12 Ohm. I was wondering what the power draw will be in total from the AD8397. Does 250mA * 12 Ohm ~= 0.75W make sense?

Also, could anyone who understands USB weigh in here? I've read a little about it, but there's still a lot I don't get. I get that there's a handshake that occurs at the beginning, but I don't think the PCM2702 would request for full power output. It could very well demand 100mA and be done with it since it does have such a relatively low current draw at about 55mA, at least with the voltages cetoole chose. I get that it limits inrush current at plug-in and then gives the full power after enumeration, but do I have to get that enumeration to take place? How would I do that?

In summation, my questions are:

1) What does output regulation voltage mean?
2) Would it be worthwhile connecting two ground wires between the boards?
3) What is the max power/current draw of the AD8397?
4) Do I need USB enumeration?

More out loud thinking to come later.

 

[joneeboi]

Quote:

Originally Posted by DKJones96 /img/forum/go_quote.gif I'd love to see more USB powered amps, I'm just not sure why you need to boost the voltage. I've heard some nice op-amps that work at under 5v. The USB dac/amp I'm working on runs on 3.3v(+-1.65v) for the amp. I've tested the amp on headphones at that voltage and it gets more than plenty loud(80ma/Amp ftw). In fact, once I found the right caps for it, aside from being loud and clean instead of distorting at high volumes you can't even tell it is there. But my goal is always transparency.

Just a quick mention on this. Even the iPod Wolfson DACs work on a low supply and can feed headphones directly, so maybe one can skip the amp altogether. Plus there are plenty of MP3 players that work off a mere AAA battery. Who really knows what's going on in there, but there is some merit to what you're saying though we won't get into the "audiophile" side of the discussion.

 

[error401]

Quote:

Originally Posted by joneeboi /img/forum/go_quote.gif 2) Would it be worthwhile connecting two ground wires between the boards?

Should work either way, though you should probably use a separate ground for each signal if possible, plus it's hard to get a really consistent connection without solder or compression.

Quote:

What is the max power/current draw of the AD8397?

(Iq (per amplifier) * number of amplifiers) + dynamic load current

Probably your average load current is going to be very small or you're going to be damaging your hearing. Except a couple mA at most.

Quote:

Do I need USB enumeration?

Nope, the PCM2702 handles all of that. All you need to do is wire it up as specified and it will work. The datasheet isn't clear as to what it specifies as its power requirement, but I bet it enumerates as a 100mA device. I know the larger chips (PCM2707) allow you to choose between 100mA and 500mA.

 

[joneeboi]

I've updated post 3 with the current layout, but I'll add it here for convenience. I'm going to stick with the two ground wires for audio and bus power. Maybe I'll add another one for kicks.

My head is going to explode if I look at this layout editor any more. I wonder how bad it would really be if I just sent what I have right now to a board house. I haven't picked one yet, so that research will come within the next few days. Anyone have any recommendations for the layout? One of the big differences from the Mini^3 board is that there is a V- plane on top of the board. I wonder if that will cool the chips any better, but I also wonder about the instantaneous power requirements of the opamps and whether the V- plane will be good for that. There are a few extra capacitors next to the opamps, so we'll see how well that plays out.

I also read on this thread that power will be readily available from the USB regardless of enumeration. I hope that that's true.
diyAudio Forums - USB powered HIFI headphone amp/DAC - Page 1

 

[amb]

You should carve a hole in the top plane around each of the opamps. Stray capacitance from the opamp pins to the plane is bad. On the Mini³ layout, I put the ground plane on the bottom side for this reason, and even there it has cutouts around the opamps.

 

[Maxvla]

Interesting project. Looking forward to seeing a working prototype with amp+DAC.

 

[joneeboi]

Quote:

Originally Posted by amb /img/forum/go_quote.gif You should carve a hole in the top plane around each of the opamps. Stray capacitance from the opamp pins to the plane is bad. On the Mini³ layout, I put the ground plane on the bottom side for this reason, and even there it has cutouts around the opamps.

Right, how abouts do I do that?

 

[MisterX]

Use the Polygon tool, choose layer 41 to draw a restricted area in the top layer or layer 42 for the bottom layer.

 

[amb]

I used rectangles, but the effect ought to be the same.

 

[error401]

Quote:

Originally Posted by joneeboi /img/forum/go_quote.gif I also read on this thread that power will be readily available from the USB regardless of enumeration. I hope that that's true. diyAudio Forums - USB powered HIFI headphone amp/DAC - Page 1

Yea, you can pretty much assume at least 1A will be available on a desktop machine even if you plug a 'dumb' device into the port, but it's a good idea to stick to the spec anyway. I don't think I've ever seen one that imposes any sort of limit based on the enumeration power requirement. I think it's mostly there so the computer can warn you if you're trying to draw too much power from one port. Laptops can be more finicky about power/port, but the enumeration still has nothing to do with it.

As far as board houses, BatchPCB usually works out cheapest for very small runs (a couple boards), but their turn time is very bad. Olimex is a bit better for quality and speed, but more expensive and your board has to work well with their panel sizes, and having to fax your details the first time is pretty irritating. If you're doing any sort of production run, Imagineering or Advanced Circuits are good choices. In Canada, APCircuits in Calgary is pretty good.

 

[joneeboi]

Thanks a bunch, fellas. I'll have to get even more creative with the layout. That's fine. I'll probably just cut out the V- plane near the opamps entirely. And thanks, error401, for suggesting APCircuits. I moved to Calgary in January, so that ought to save me a little on shipping.