@mara

Thanks for the solid recommendation! Good isolation is a big plus. I'll look into these before my final purchase.


@speculatrix and Benson

Thank you kindly for your insight. The trouble is, I have only a very faint clue as to what you both are talking about! If it's not too much, do you mind giving me a frank explanation? It would help out tremendously, as it really does sound interesting.


}:^)~
YARR!

Capt'n Corrupt

Well, I wasn't so much offering insight, as offering to take a couple measurements on my setup. Oh, and I was offering EE-jargon, too.

But if you think it's insightful, here's more:

Overview:
The sound hardware in the N800 has some stuff which generates a digital output.
This digital output gets run through a DAC (Digital -> Analog Converter). The DAC outputs a voltage corresponding to the numeric input.
This analog signal then goes through some filters, out the headphone jack, and presumably into a load.

So parameters of the system include:
Maximum output sample rate: This is how often the digital output gets updated. Probably 48kHz, but IDK. This is the most important factor in maximum frequency. (The highest output frequency is one-half the sample rate, since you need a positive and a negative peak in each cycle.)
DAC resolution: This is how many levels of voltage the DAC can output.
Probably 16 bits, maybe less. Highly unlikely to be more, though PC sound cards are readily available higher. This is the limiting factor on dynamic range.
Analog filtration stuff:
Since the DAC outputs a constant voltage for a given input, the output has stair-steps. This results in HF noise, at frequencies at and above the sample rate. Those are well above the range of human hearing in this case, but are still probably eliminated with a low-pass filter, which attenuates frequencies above about 24 kHz (sample rate / 2).
There could be more analog filters here; probably volume control is here (it definitely should be), and there may be circuits to compensate for the frequency-dependence of the load.

Finally, the output goes to the headphone jack. Now, if the outputs are directly connected, it'll work fine for headphones. But you could have trouble when connecting it to an amp, as the sound wave is riding on top of a DC level, and the amp may have its inputs referenced to ground. To avoid such trouble (which could cause anything from distortion to frying one of the components), the output is connected to the jack with a capacitor.

Brief discursion into capacitors and inductors:
Resistors have a given resistance value. It's constant for all frequencies of voltage/current you might care to apply. Capacitors and inductors (the ideal ones) have no resistance, but have impedance, which is like resistance. The distinction is that resistance doesn't change the phase; the voltage across it is in phase with the current through it. Impedance shows the same current-voltage relation (high impedance = low current for a given voltage, or high voltage for a given current), but also contains a phase difference between voltage and current. The interesting thing, for our purposes, is that the capacitor has an infinite impedance at DC, and a low impedance at HF. Inductors are the opposite.

So, our capacitor looks like an open circuit for DC -- no DC gets sent through, and if the load has a DC bias, it doesn't send any DC back to fry the N800. For high frequency, it looks like the jack is wired directly.
The only question is, at what frequency does the capacitor "start" conducting?

For the moment, we'll treat your load (earphones, external amp, speakers, whatever) as a resistor. It's not, of course, though the assumption is probably respectable for the amp. The others have strong inductance, but we'll ignore that in the interest of simplicity.

The output power will be one-half the pass-band power (-3dB) when the impedance of the capacitor is equivalent to the resistance of the load. This frequency (called the corner frequency) can be shown to be characterized as 1/(R*C), where R and C are the resistance and capacitance. This formula gives the ouput in radians/second, but we want Hz, so it's 1/(2*pi*R*C) Hz. Above this frequency, we can treat the capacitor as conducting, since the resistance of the load rapidly dominates. Below this, the capacitance dominates, so we can treat the response as falling off at 20 dB/decade. A decade is a power of 10, so if the corner frequency is 100 Hz, a 10 Hz signal would be attenuated by 20 dB, or have one-tenth the voltage, or one-hundredth the power, of a higher-frequency signal with the same amplitude. A graph of the output on a dB scale versus the frequency on a log scale looks like a straight line rising at 45 degrees up to the corner frequency, then level the rest of the way, hence the term corner frequency.

The corner frequency, then, is what we need to find. But it depends on the resistance of the load. So the only thing to do is hook the load (your earphones) up, and measure the voltage output while playing a bunch of sound files containing sine waves with the same amplitude (but a range of frequencies). Look for the point where the output really starts falling off as the frequency drops, and you know the corner frequency. (To best determine it, plot all points on a dB-log plot, giving you the entire frequency response, but if you just want the corner frequency, look for where the output amplitude is drops to 70.7%)
That is easiest with a scope, though you could do it with a good AC voltmeter.

HTH!

Benson, that was a tremendous post! Thank you kindly for your insight.

I think I understand most of it. The following confuses me a bit:
I'm not sure what this is. Are these assumptions made prior to the explanation? I think I understand the idea of corner frequency, even if my inchoate mental model is somewhat wobbly.

One more question: What is it that electrical engineers do exactly? I've always been interested in circuits, but have never taken the plunge to play around with some home brew electronics. I'm guessing that EEs do this, at a much higher level. Am I right in assuming that Computer Engineers, are like EEs only with a specialty involving the organization of circuits for computation (logic/arithmetic/etc)?

Thank you so much. Could you recommend a good book/author? I need to know more, and can't continue asking you questions directly!


}:^)~
YARR!

Capt'n Corrupt

We sell burgers for a living, lol. Just joking, I hope.

Assuming your interest is (or will be) electronics hooked up to a computer/microprocessor/system-on(-a)-chip, then I would recommend dev kits for the MSP430 or the very interesting Parallax "Propeller".

On subject (sort of): if anyone has the equipment I would be interested to know the frequency response of the inbuilt speakers for 770/800/810. Why? I have an idea, here's a hint: anyone remember principle of operation of early tv remotes?

Nothing wrong with burgers! They're delicious, and fill a great human need.


Amazing! This has turned out to be the best thread I've ever had the pleasure of contributing to. I'll check out the dev kit. I'm now seriously rethinking my desired Christmas gift; cologne was my first option.


Here's a guess: Using high pitched, nearly inaudible sound, one could control devices remotely. This post triggered a very old childhood memory I had, when I noticed that a very quiet high pitched noise came from the tip of the remote when I pushed buttons. I noticed that the pitch was different for each button pressed. I *never* would have remembered that, had it not been for this post!


}:^)~
YARR!

Capt'n Corrupt

For those interested in learning more, I've located a great resource of scientific texts online (by searching the ITT), SpringerLink. Together with the classical literature of sites like Manybooks.net, there's more free content than anyone could ever hope to read! Here's to hoping that I can additionally purchase the occasional book (DRM'd or not).

I very excited to get a N810 to call my own! It's going to be my e-book solution until e-ink based readers get better (colour e-ink is around the corner, and pixel density and refresh speed is bound to improve).


}:^)~
YARR!

Capt'n Corrupt

@Capt'n Corrupt: you are giving away your age if you remember those types of remotes, lol. But spot on, I'm envisaging a matchbox sized device with a highly selective audio filter that converts audio pulse train to IR. It would have the advantage of being cheap and not requiring additional hw (such as some BT/networked device). All assuming the RC codes can be captured or dl'ed somewhere. I am not an audio engineer though and suspect some specialist equipment is needed to test if this idea would be feasible.

Ah, early remotes.. I remember those. They used ultrasound. That didn't last long, because the ultrasound sometimes went through the walls and when you had just sat down to watch "Flipper" (which was popular back then.. your TV would switch channels because the neighbour operated his own ultrasound remote.

As for the bit that confused you, perhaps that wasn't the best way to put it. The -3dB point is the conventional place to define the "start" of the pass-band. In this sense, it is an assumption, of where we conventionally draw the line.
But the bit about the impedances being equivalent (by which I meant, have the same magnitude, but different phase), is not an assumption. It's simply that the output falls off to one-half the power at the point where the impedances are equivalent.

I should have thought of Wikipedia, and it's propensity for useful pictures: Corner frequency

As for what EEs do, yes, that's more or less it. You can go into power, where you get to spend your life bored to death specing overhead transmission lines and the like, or you can go to typical electronics stuff (most EEs), which consists of designing all manner of circuits. Or, you could go to various minor specialties, like IC design and such. Wikipedia is helpful again.
CpEs are really more like EEs who know how to program better. (And, on the flip side, aren't as good with heavy-duty analog stuff, like super-heterodyne radios.) But both EEs and CpEs have basic skills in analog and digital electronics.
I have more of a CpE bent, but took the EE course due to difficulties scheduling some of the CpE classes. Sometimes CpE is considered a specialty within the EE field, sometimes as a separate, though closely related, field.

Well, I expect the design specs weren't for "high-pitched, nearly inaudible sound", but rather for ultrasound, i.e. high-pitched, totally inaudible sound. But kids tend to have higher range of hearing, so you could hear it anyway.
I have even heard of some cell-phone ringtones at the top end of human hearing, where most schoolkids can hear them, but their teachers can't!

... which would explain why mention of infinite impedance made me nervous at first glance... :P

This would be an interesting project and a fun one. Your device has the advantage of being small and portable, and with bright IR LEDs, could easily drive remote devices.

A spin on this, would use a computer with an IR transmitter and a wireless network interface. Not as portable, but relatively easy to design.

Another neat alternative, would be to use the microUSB, a homebrew circuit (and code), and a small IR LED to drive devices remotely. It requires additional equipment used in conjunction with the N810, but it should work effectively.


Heh heh... The perfect platform for a few childish pranks...


}:^)~
YARR!

Capt'n Corrupt