CHANGSTAR: Audiophile Headphone Reviews and Early 90s Style BBS
Lobby => Head Amps, DACs, Sources, Portable Equipment Discussion => Topic started by: lachlanlikesathing on May 23, 2015, 12:19:01 AM
-
I found this last night and I'm just a little curious:
(http://www.apogeedigital.com/wp-content/uploads/2015/04/Groove-3-Quarters-Right-10001-103x300.png)
http://www.apogeedigital.com/products/groove (http://www.apogeedigital.com/products/groove)
It's a USB stick style DAC/AMP with 4 ESS Sabre chips per channel, done by Apogee who have a pretty good pedigree. $295 USD. All very fancy, but it's this bit that has got me interested:
The Groove uses Apogee’s trademark Constant Current Drive output stage, whose output voltage is dynamically altered by impedance variations of the connected headphones, which has the effect of smoothing-out the frequency response non-linearities that occur with traditional headphone outputs. ‘Apogee’s Constant Current Drive is a truly revolutionary approach to dealing with the wide impedance variations of headphones available on the market today,’ says Betty Bennett, Apogee Co-Founder and CEO. ‘Because of this, the Groove unlocks the highest possible quality of any headphones you have, allowing you to listen to the music you love with incredible life-like clarity.’
http://www.avhub.com.au/news/hi-fi/apogee-groove-portable-usb-dac-399726 (http://www.avhub.com.au/news/hi-fi/apogee-groove-portable-usb-dac-399726)
Any idea how this would work? It seems to be claiming that it will supply a constant voltage to a headphone at all frequencies regardless of impedance curve. I know next to nothing about electronics. The only thing I can think of is a reverse version of the resistor network on the Sony MDR-MA900 / MDR-F1 (http://www.changstar.com/index.php/topic,1725.msg46398.html#msg46398).
-
It's basically just an amp with a super high output impedance. So high, that in theory, the impedance of the transducer doesn't matter. It also meas essentially zero damping factor, which may matter for some headphones. I'm pretty skeptical about making that work properly off of just USB power though.
AFIK, there's not much benefit to using that method unless you're trying to get extra bass extension out over damped full range drivers in giant horns (http://www.firstwatt.com/pdf/art_cs_amps.pdf) or something similarly weird.
Maybe someone else has some more insight...
-
From the ad copy it is not clear that "constant current drive" is a current-source output, but apparently it is. This post on gearslutz (https://www.gearslutz.com/board/new-products-coming-soon/981900-apogee-unveil-groove-portable-usb-dac-headphone-amp-mac-pc-2.html#35) by Don Spacht of Apogee references this earlier post (https://www.gearslutz.com/board/new-products-coming-soon/981900-apogee-unveil-groove-portable-usb-dac-headphone-amp-mac-pc-1.html#5) in the thread that says:
Groove's output circuit is not a regular voltage out. It is current out. Which is a difficult concept to explain, but the net result is that a higher impedance will imply a higher output voltage.
-
But now you wind up with a frequency response that follows the impedance curve.
Can a current output amp wind up producing a voltage output across the load higher than it's internal rails?
-
It is basically a little Bakoon.
Probably good match with high-impedance headphones, but for low-impedance headphones and IEMs with complex crossovers... forget about it.
-
So I sent an enquiry to Apogee and got the following response:
Thanks for your email! The Groove’s Constant Current Drive reverses the relationship between voltage and current from how headphone amplifiers are traditionally built.
99% of all headphone outputs in the world use a voltage-drive design, where the current varies in order to compensate for load fluctuations and maintain a constant voltage. The disadvantage to this is that it is current that actually moves headphone drivers, and under certain conditions where the headphone itself isn’t perfectly matched to the headphone amp the actual diaphragm produces different audio levels at different frequencies - in effect, the frequency response becomes non-linear.
The Groove instead varies voltage to compensate for load fluctuations and maintain a constant current. By doing this, the circuit is built to compensate for frequency response non-linearity caused by impedance fluctuations. The headphone drivers get the consistent current they need to output a linear audio signal. In practice, the Groove’s circuit is compensating for both electromechanical and acoustic non-linearity, offering extremely consistent performance with any type of headphones.
The effective output impedance ends up being relatively high, actually, but that’s not a negative because it’s not a factor in how it’s moving the headphone driver like on a standard voltage-drive design. It’s kind of like asking for the MPG on an electric car - it wouldn’t be measured in the traditional sense.
Sorry if that was a little confusing, but I hope it helps!
So that seems to confirm that the Groove is a current source amplifier. I'm pretty intrigued by what this will mean given that the electrical damping factor will be very low. Again I have very little idea about this sparky stuff but I think I may order one of these just to see if it actually sounds different / better / worse with different headphones.
-
Okay, I got the Groove. This is very interesting. It sounds nice and clean, and I'm not sure if it's my imagination, but the low damping factor seems to make things sound a bit smoother/sweeter on the various gear I'm trying it with. Maybe this is the supposed euphonic sound that high output impedance tube amplifiers are famed for (I don't know, I don't own any.) But I'm enjoying the sound on the headphones that I regard as a little leaner or hot, like say the Audio Technica ATH-M70x, AKG K550 or Beyerdynamic DT770.
I was reading through this article on Innerfidelity: http://www.innerfidelity.com/content/musings-headphone-amplifier-output-impedance (http://www.innerfidelity.com/content/musings-headphone-amplifier-output-impedance)
Jan Meier seems to suggest that a current source headphone amp with a higher output impedance should actually help arrest oscillations in low mass, high efficiency drivers, which is the same argument as made in the article Maverickronin posted above.
This line is interesting:
The membranes of magneplanar drivers have a large surface and very low mass. Therefore they are already highly dampened by air. Proper electrical damping is of less importance. ...
Changing the effective output impedance only had very little effect on the current. It is the reason that current-amplifiers (instead of voltage amplifiers), like the Bakoon, which by nature have an extremely high effective output impedance, do work well with magneplanar headphones.
So if the Groove really is like a little Bakoon, there should be some headphones for which the Groove should be a great match...
Amusingly from Jan's article it appears Sennheiser headphones are not a good match despite Apogee's partnership with the company.
Axel Grell, the chief-designer at Sennheiser, once told me that their top-of-the-range headphones are always designed to sound best at a low output impedance.
-
Interesting. That's what I would expect for full size dynamic headphones. I think my K601s tend to sound better with a little extra Zout. I should probably so some measurements sometime and see if it really shows up on the CSDs or if the extra bass is just covering it up.
Now the big question is, how does it sound with BA IEMs? In theory, if the Zout is is high enough then the FR won't vary but since you'd also need really high PSU rails to go with it wouldn't work all that well in a USB powered device.
Maybe I should try and build my own current source amp too....
-
but since you'd also need really high PSU rails to go with it wouldn't work all that well in a USB powered device.
This is my chief concern. But then consider the bakoons: max power listed is 1W at 50ohm... that means ~2.23V and ~447mA, which is within USB 2.0 spec, right?
-
This is my chief concern. But then consider the bakoons: max power listed is 1W at 50ohm... that means ~2.23V and ~447mA, which is within USB 2.0 spec, right?
Most of the amps power gets "eaten" by the the output impedance though.
-
This is my chief concern. But then consider the bakoons: max power listed is 1W at 50ohm... that means ~2.23V and ~447mA, which is within USB 2.0 spec, right?
1W @ 50 Ohms = 7.07V (140mA).
For this you need 24V power supply rails (+/-12V) and a power supply that must be able to supply 0.4A.
The Groove can provide:
225mW into 30 Ohm = 2.6V
40mW into 600 Ohm = 5V
So most likely the internal supply voltage is around 16V (+/- 8V) with a current limit of 85mA/channel.
As it is powered by USB and the device draws around 150mA max from a USB bus which is well within spec.
-
1W @ 50 Ohms = 7.07V (140mA).
For this you need 24V power supply rails (+/-12V) and a power supply that must be able to supply 0.4A.
The Groove can provide:
225mW into 30 Ohm = 2.6V
40mW into 600 Ohm = 5V
So most likely the internal supply voltage around is 16V (+/- 8V) with a current limit of 85mA/channel.
As it is powered by USB and the device draws around 150mA max from a USB bus which is well within spec.
:)p13 :)p13 :)p13 Thanks solderdude ... I'm an idiot, I always fuck up the zeros :spank: My calculations reflect FIVE ohms resistance, not FIFTY.
-
"I think my K601s tend to sound better with a little extra Zout. I should probably so some measurements sometime and see if it really shows up on the CSDs or if the extra bass is just covering it up."
This, I would love to see. Actually, what I would love to see is the HD650 (my HP of choice) FR and CSD measured on the same system 1) from a voltage source 2) from a current source (decreased electrical damping), and 3) with removal of the rear damping foam a la Purrin (? decreased acoustical damping).
My experimentation with high Z out suggests that there are some real positive possibilities there sound-wise. These improvements remain even if I use a quality EQ plug-in to counteract the (calculable) frequency response changes. And I do end up using, for the HD650, a filter at 80-90Hz, q 0.4-0.6 to partially correct the response changes in the low frequencies, o/w it is a little too boomy, muddy (which is how I hear the rear-foam removal mod).
I am really starting to think that some of the older companies, with their older designs, measured and voiced their designs with amps with non-zero Z out (? the old 120 ohm "standard")
Bill Brown
-
I did some volume matched comparison against the Geek Pulse and the O2/ODAC. My usual method of volume matching is to plug the two amps into an A/B switch and then feed that into the line-in of my Zoom H2n, where I visually match the graphical audio meters on the recorder. This has worked in the past but completely failed to work with the Groove - I could volume match with the meter but end up with very different levels with the headphones plugged in. There's probably something that I'm not understanding here because of my limited understanding of the electrical stuff. I ended up volume matching the 1khz tone by ear with each headphone I tried, so this is far from scientific. In tests I can usually identify differences in levels of about 1-2dB.
Firstly I can say that the differences between all the sources was surprisingly small on most of the headphones. Just about the only obvious difference I could immediately pick out was that the Groove has a lower noise floor than the ODAC/O2, which itself has a lower noise floor than the Geek Pulse. That's actually pretty good because the O2 is the blackest sounding source I've heard.
I tried it with my M70x, GR07, ATH-IM02, Shure SRH-1540, AKG K712 and DT770.
I can't really say I can hear any strong advantage of the constant current drive's supposed ability to more accurately control the voice coil, nor can I hear any any real obvious harm from the supposedly low or negative damping factor the Groove has. Most of the times I thought the Groove maybe sounded a little sweeter or maybe a tiny bit cleaner, but it could well have been placebo and I wouldn't trust myself to pick it in a blind test.
The equipment that I could actually hear the most distinct difference with was the IM02, which ended up sounding more upper treble forward with the Groove. I'm assuming this is because of the way current drive amplifiers have output that follows the impedance curve - but on this point, I'm a little confused. If the Groove has something crazy like a 2.5 mega-ohm output impedance like the Bakoon, shouldn't the impedance curve of most headphones be almost irrelevant? I sent another email to Apogee asking what the actual output impedance of the Groove is.
From some of my (http://www.customanalogue.com/various/Current-Amplification.pdf) reading (http://www.customanalogue.com/tlcl_speaker.htm) it seems that the most appropriate headphones to pair with the Groove are highly efficient, low impedance headphones with minimal impedance variation across the frequency range. In my collection that would be the Audio Technica ATH-M50x and the V-Sonic GR07. I think they sound great with the Groove (again, Placebo?) but I just wanted to make a comment - as much as a lot of people hate the M50x, it seems like Audio Technica did try to prioritise making a headphone with a near ruler flat impedance and phase curve so that it could match with the widest variety of output sources.
Imma do some listening but at this point I am decently happy with the Groove. It's quite expensive for the power, but it's small and compact, doesn't look like some DIY science project on my desk like the O2/ODAC stack, has proper digital channel balance at low frequencies, doesn't make me nervous that it will eat my headphones like the Geek Pulse from power on/off pops or otherwise, doesn't need its own power brick - and has a cleaner output than both.
-
I got some additional helpful information from Apogee, apparently from the lead electrical engineer:
"It is a tricky concept to explain. I’ll try.
Ideal voltage outputs have zero output impedance. Ideal current outputs have infinite output impedance.
In the real world that is of course not possible, it is almost zero for voltage and very high impedance for current.
So the standard formula for damping will not apply with current out, as the electrical model is different. However the “comparable” damping factor is very high as the current source will always try to pump the same amount of current through the load, thus preventing the speaker from making motions other than those intended (as there is no series resistor to allow “free-wheeling”)"
I'm assuming this refers to the same idea of the current source amplifier ignoring electrical damping oscillations from various sources:
[A current source amplifier] ignores the impedances in series with the circuit, the resistance and inductance of the wire and voice coil and the back electromotive force (EMF) produced by the cone motion... the current through the voice coil can be made constant regardless of the variations in the acoustic environment. (http://www.firstwatt.com/pdf/art_cs_amps.pdf)Source: Nelson Pass
The trade off is that frequency response now follows the headphone's impedance curve?
I was also provided some more specification figures:
Crosstalk: -110dB at 1kHz
IMD: (Depends of method used). With -3dBfs 19k/20k method 1k artifact is -115 dB with 300 Ohm load
-
I'm assuming this refers to the same idea of the current source amplifier ignoring electrical damping oscillations from various sources:
The trade off is that frequency response now follows the headphone's impedance curve?
Yes, that is exactly the trade off.
Some MA drivers may sound quite wrong from this amplifier.
Most headphones will only react mildly as the impedance is quite flat (= do not need/expect/want any electrical damping) only some headphones will react more audibly.
Headphones are designed (voiced ?) with (close to) 0 Ohm in mind, perhaps some older headphones are excluded.
NONE of them are designed to be current driven which is something that may be worth mentioning.
The fact that some headphones sound 'better' probably has more to do with euphonic coloration of the FR/phase than with being closer to the 'intended' signal.
Most headphone drivers break-up at higher frequencies, the diaphragm is thus not making the same movements as the voicecoil does so that type of distortion isn't reduced.
Phase distortion (when caused by a voicecoil impedance rise) is lowered but at the expense of an increased SPL which may be beneficial to the sound ... or not.
In the end, the only thing that matters is if you like the sound, regardless if it has 'changed' somewhat.
Yes, the bass and treble might be different somewhat compared to 'normal' amps with some headphones but headphones themselves usually are 'all over the place' in the FR department all over the frequency range anyway. Most of them well over 5dB some even up to 20dB.
None of them accurately reproduce the applied electrical waveform, regargdless if they are voltage, current, or both (output R between 0 Ohm and a few hundred Ohm) driven anyway.
-
Has anyone here heard this with the HD600 or HD650? Any impression? It's currently on sale at Adoram w/ the HD650 for $499.
-
I can make a prediction for you.
+4dB in the bass (as if you pull up a 100Hz slider on a 5 band equalizer by 4db) + 1dB extra upper treble energy.
So more bassy/warmth than it is stock.
You could also search for experiences people have with the Bakoon + HD600/650 to get an idea.
The HD650 was NOT designed to be current driven.
-
I can make a prediction for you.
+4dB in the bass (as if you pull up a 100Hz slider on a 5 band equalizer by 4db) + 1dB extra upper treble energy.
So more bassy/warmth than it is stock.
You could also search for experiences people have with the Bakoon + HD600/650 to get an idea.
The HD650 was NOT designed to be current driven.
Wow. So I'm staying away. Thanks!
-
The HD650 was NOT designed to be current driven.
Neither was the HD800. I don't care how many people say they 'like' it. Poor HD800s.... :(
-
I heard it with some HD800s and it was just ... weird. Really messed with the FR.
-
The HD600/HD650 and HD800 have very similar impedance bumps so the effect will be the same.
Perhaps it mates well with DT880's when you find them lacking in the bass ?
These will become more 'bathtub shaped' with boosted bass and upper treble :-00.
Anyway when considering this device (or the bakoon) I would check innerfidelity's plots and have a look at the impedance plots of the headphone(s) you want to use with it.
http://www.innerfidelity.com/headphone-data-sheet-downloads (http://www.innerfidelity.com/headphone-data-sheet-downloads)
A massive 'bump' in any parts will emphasize that part of the frequency range.