Light Harmonic Geek DAC

[Savage]

A good write-up on the daily swarm - The Daily Swarm - Sonic Satori: Tripping The Light Harmonic - Bringing Serious Sound to the Masses Via Kickstarter...

[lithium1085]

Last chance for the geek today!

[rdsu]

Last chance for the geek today!

Last 3 hours...

 

I hope we can be blown away with the final result, in January...

[barrows]

Looking at those measurements, there is some really good engineering going on in the Geek, as that is pretty impressive performance for a small, USB powered device. I ordered one… looks like a good match for simple closed back 'phone like the Senn Momentums.

[Ben-M]

I got a Super Duper Geek just in case the need arises for that extra power. I really don't get how they're squeezing 1000mW out of a USB stick, but who knows what load they're quoting that with. Overall I think it's going to make a great and simple computer speaker rig for my regular/gaming PC. Then again, maybe it'll knock my socks off and I'll have to find a more ~dignified~ role for it?

[Savage]

Looking at those measurements, there is some really good engineering going on in the Geek, as that is pretty impressive performance for a small, USB powered device. I ordered one… looks like a good match for simple closed back 'phone like the Senn Momentums.

 

They are evaluating a mobile dac chip to use instead of the TI 1795, hoping the measurements stay the same.

 

They also have a $30.00 off coupon code for pre-orders - iamageeksfriend

[Superdad]

Does anyone know what chip they are using for headphone amp? None of the popular TI headphone amp chips put out anywhere near 450mW, let alone 1,000mW. Just curious as to what they are using. It's not like they have room for any sort of discrete amps for the output on such a small device.

[barrows]

Does anyone know what chip they are using for headphone amp? None of the popular TI headphone amp chips put out anywhere near 450mW, let alone 1,000mW. Just curious as to what they are using. It's not like they have room for any sort of discrete amps for the output on such a small device.

 

Yeah, these are good questions. You can do a discrete buffer in a pretty small space these days with smd parts, but with the TI 1795 DAC they need I/V and buffer, so I would assume they are using ICs for both. They might be able to fit a discrete I/V, which really only requires a single transistor and a couple of resistors and then an IC for the buffer. Good thing is, whatever parts they are choosing, they are running them at high bias, claiming class A operation.

As to the power, I think it should be possible to do, as long as the USB port is up to the 5V and 500 mA spec, as that gives 2.5W.

[barrows]

They are evaluating a mobile dac chip to use instead of the TI 1795, hoping the measurements stay the same.

 

They also have a $30.00 off coupon code for pre-orders - iamageeksfriend

 

That mobile chip is the ESS Sabre "M", it has better specs than the TI 1795, so no worries for the measurements getting worse if they end up choosing the ESS.

[Miska]

Does anyone know what chip they are using for headphone amp? None of the popular TI headphone amp chips put out anywhere near 450mW, let alone 1,000mW. Just curious as to what they are using. It's not like they have room for any sort of discrete amps for the output on such a small device.

 

TPA6120A2 is very popular, I have it in at least two different devices. But not very suitable to be powered from USB.

 

Then there are quite a bunch of class-D options commonly used in mobile phones with output powers up to 3W (intended to drive 4/8 ohm speakers).

[barrows]

TPA6120A2 is very popular, I have it in at least two different devices. But not very suitable to be powered from USB.

 

Then there are quite a bunch of class-D options commonly used in mobile phones with output powers up to 3W (intended to drive 4/8 ohm speakers).

 

They are claiming class A bias for their output circuit.

[Miska]

They are claiming class A bias for their output circuit.

 

A rail-to-rail op-amp and class-A biased two-transistor push-pull output buffer would do. I designed quite a bunch of such headphone amps some 20 years ago.

 

Problem in a small device would be heat, 1W class-A would produce minimum 2W heat, more typically at least 4W.

[4est]

That might explain the aluminum housing.

A rail-to-rail op-amp and class-A biased two-transistor push-pull output buffer would do. I designed quite a bunch of such headphone amps some 20 years ago.

 

Problem in a small device would be heat, 1W class-A would produce minimum 2W heat, more typically at least 4W.

[Jud]

Problem in a small device would be heat, 1W class-A would produce minimum 2W heat, more typically at least 4W.

 

I wonder if "Class A" will turn out to be marketing-speak similar to "non-oversampling." Or, if it actually is Class A and the aluminum housing is truly supposed to act as a heat sink, how hot that might get over the course of a 5-hour plane or car trip (situations for which I currently use the Dragonfly). Guess we'll see in a few months.

[rdsu]

I wonder if "Class A" will turn out to be marketing-speak similar to "non-oversampling." Or, if it actually is Class A and the aluminum housing is truly supposed to act as a heat sink, how hot that might get over the course of a 5-hour plane or car trip (situations for which I currently use the Dragonfly). Guess we'll see in a few months.

At least they talked a lot about the "Class A", so it seems to be true...

 

The non-oversampling issue was more bad communication between Gavin Fish and Larry Ho...

[Superdad]

A rail-to-rail op-amp and class-A biased two-transistor push-pull output buffer would do. I designed quite a bunch of such headphone amps some 20 years ago.

 

Can that be done with just a single-ended PS? Only +5VDC on USB bus. And if it is just a chip, then I don't know of ANY charge-pump headphone amp devices that can put out more than about 140mW.

[Ben-M]

Can that be done with just a single-ended PS? Only +5VDC on USB bus. And if it is just a chip, then I don't know of ANY charge-pump headphone amp devices that can put out more than about 140mW.

Guess we'll see in a few months.

Hahaha, but that really is the answer to everything about this thing. Not to much real info on it. Really though, I am quite eager to find out how those 1000mW are being put out and if this thing is going to be an Easy Bake Oven running in Class A to do it. I also really hope that this is some kind of wonder stick and kicks the pants of stuff and of course puts a huge smile on my face for owning it.

[barrows]

Well, I think it is pretty clear that this thing will not run in true class A all the way to 1W output, that would not be possible on USB supply. 1W of true class A would require a 4W power supply, USB has 5*.5=2.5W available total. I am sure they are just high biasing the circuit, and it runs in class A for, perhaps, the first .25W or so. Still, that first .25W is what most are going to hear. While the higher power versions can probably drive some hungry headphones, I suspect this thing will sound best when running mostly in class A bias, with relatively efficient headphones. You cannot get blood out of a stone.

Class A can be very misleading the way many manufacturers talk about it. In power amps, very few manufacturers really tell the truth when they say an amp is class A, almost always (except Pass) they are referring to a class A/B circuit, which runs class A in the first handful of watts.

Note how large a true class A amp has to be, take a look at Pass Labs XA 100.5 monoblocks, these put out 100 watts in class A each (and then transition to A/B where they can produce more power for peaks) Kudos to Pass for being honest rating these as 100 watt class A amplifiers: they really produce much more power (before clipping) but they produce 100 watts of real class A.

[Ben-M]

Well, I think it is pretty clear that this thing will not run in true class A all the way to 1W output, that would not be possible on USB supply. 1W of true class A would require a 4W power supply, USB has 5*.5=2.5W available total. I am sure they are just high biasing the circuit, and it runs in class A for, perhaps, the first .25W or so. Still, that first .25W is what most are going to hear.

 

Hmm, but if this is a USB 3 device it may have more on tap. USB 2 is limited to 500mA, but USB 3 has 1A, I believe. At least if it's one of those fast charging circuits it has more on tap, but then again, they'd have to tell you that's part of the 1W output deal.

 

I've been wondering this since I bought in, is the Geek a USB 2 or USB 3 device? I really hope 3.

[filhodoassis]

Some real competition:

Resonessence Labs HERUS: an ultra-portable DSD DAC & headphone amplifier | Digital Audio Review by John Darko

Its DAC Wars! A new hope.

DSD Strikes Back!

And maybe later, Return of the PCM!

May the Sound be with you.

[labjr]

They are evaluating a mobile dac chip to use instead of the TI 1795, hoping the measurements stay the same.

 

I hadn't heard that. I thought the design and choice of the TI 1795 was pretty much a done deal.

 

Hmm, but if this is a USB 3 device it may have more on tap. USB 2 is limited to 500mA, but USB 3 has 1A, I believe. At least if it's one of those fast charging circuits it has more on tap, but then again, they'd have to tell you that's part of the 1W output deal.

 

I've been wondering this since I bought in, is the Geek a USB 2 or USB 3 device? I really hope 3.

 

I'm pretty sure they'll be using USB 2. USB 3 would add to the cost.

[barrows]

labjr: it is mentioned on the FaceBook page for the Geek that they are listeneing to other chips as well, including the ESS SabreM (M=mobile, lower power consumption model for mobile devices).

Well, if you could get 1A @ 5V from a USB3 port, I expect that 5 watts would drain a laptop battery pretty quickly. I am only interetsed in Geek for travel use with my laptop, so battery power run time is an issue for sure. All things considered I ordered the normal Geek, which is still more powerful than most competition, and will just pair it with appropriate 'phones... I figure the lower power model is also going to produce a greater percentage of its output in class A mode, so expect better sound from the Geek with relatively efficient headphones.

[Jud]

All things considered I ordered the normal Geek, which is still more powerful than most competition, and will just pair it with appropriate 'phones... I figure the lower power model is also going to produce a greater percentage of its output in class A mode, so expect better sound from the Geek with relatively efficient headphones.

 

Better sound with lower power is also the reason I stuck with the basic model.

 

I hoped to use this with XXHighEnd, but unfortunately the Windows driver doesn't support kernel streaming. So looks like it will be OS X and Audirvana Plus exclusively.

[rdsu]

I hoped to use this with XXHighEnd, but unfortunately the Windows driver doesn't support kernel streaming. So looks like it will be OS X and Audirvana Plus exclusively.

Why do you say that their drivers will not support KS?

[labjr]

Barrows, I think they're exploring ESS chips for future projects. I'd think, changing from TI to ESS DAC at this stage would be a rather complicated design change. They've worked on Geek Prototypes for a year. I imagine at this point they would be done with listening to prototypes and about to send the final version to manufacturing. I guess someone could ask Larry Ho to be sure.