Review of Analog Devices ADAU1592 vs. Tripath TA2024

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
For the review I am comparing the Tripath TA2024 board from Sure Electronics (AA-AB32155) to the Analog Devices ADAU1592 evaluation board (EVAL-ADAU1592). Pictures of the two boards are below (assuming the attachment feature works.). Both chips will be run off the same power supply. It is a 12v regulated supply, though both if these board can take a little bit more voltage ( 18V max. and 14V max. respectively). I will only be trying one pair of full range speakers. The speakers are made by Tekton (Tekton Design - Tekton Lore / Mini Lore) they use the Japanese Fostex FE166 which is a 6” paper driver in an 18.5”x12”x8.5” mdf front vented enclosure, this model is now discontinued as far as I can tell. There is no cross-over. One day I hope to upgrade the box with a pair of 6” Lowthers, I can think of no better way to protest Japanese whaling policy. The speaker stands are Core Audio Designs (Core Audio Designs Furniture for Audiophiles) model B's and actually cost more than the speakers themself. The source is the Twisted Pear Audio (Twisted Pear Audio) USB DAC/Receiver using the Texas Instruments PCM2707 USB DAC chip (16-bit/48-kHz). All cabling is unshielded.


Price (quoted at 01/30/11):

Tripath TA2024 broad from Sure Electronics ----- $16.99 USD
Analog Devices ADAU1592 eventuation board ----- $213.00 USD*

* Evaluation Boards are always more expensive then 3rd party boards. I never understood this it seems counter intuitive to make it costly to get your product into developers hands.



Solder-ability (note: this bit is subjective):

It seems more and more the industry is moving towards smaller foot prints and using less plastic and closer lead spacing (0.5mm in the case of AD). Personally I'll take the DIP format any day. Both of these chips are absolutely dreadful when it comes to hand soldering, however relatively the Analog Devices chip is far worse. You can't even buy the chip in any format that even has leads coming out of the body any more. The Tripath chip at least has legs. The thermal pads are on the bottom for both of these chips. In all fairness both chips give off very little heat. But if I were working with these chips I would still much prefer the thermal slug on the top like the NXP's *TDA8950, that way I could just glue a heat-sink on the top. *Because of the really poor solder-ability by hand I purchased them already soldered on boards. The Tripath board is even cheaper than doing it myself. That said, some of the wizards on this site have managed to make their own PCB for these chips and solder them successfully, no doubt one will chime in on this thread in a moment and say “it's not so bad”.



Sound(note: this bit is subjective):

Let's start with the usual warning this bit is subjective and your mileage will vary with different speakers, source material, taste, and so on. The two chips have different characters. The TA2024 sounds more like a tube amp and the ADAU1592 sounds more like solid state class AB well controlled with feedback. The ADAU1592 is clearer it sounds best with rock because it preserves very fine detail. Some people will call the ADAU1592 too clinical it will not do as well with Jazz and vocal. If that’s your main diet I would go with the TA2024. The TA2024 has far more air and sound staging and imaging. There is one peace of music, on which the TA2024 seems to blow everything else out of the water, it is the song “Tu Vuo' Fa L'Americano” from “The Talented Mr. Ripley” sound track. I have even tried this peace on my 300B se amps, and still the TA2024 is better (again this only applies for this song on these speakers). With this song the bass gets nice and weighty images are just pinpoint, the sound stage just seems to wrap around and go back into the wall. The ADAU1592 and my 300B and some other tube amps I have laying around play it, but they really just seem to be going through the motions, just two boxes producing sound, no magic, I just don’t understand it.
Now moving away from that song, if I were playing “Pink Floyd’s ~ The Final Cut ”, I would use the ADAU1592. Sometimes I like to play music to help me fall asleep, so my mind stops churning over the goings on of the day. I don’t feel guilty about leaving these amps on overnight because they use so little power. The ADAU1592 has a little lightness in the bass compared with the TA2024. Some may see this as a bad thing, but there are many circumstances I can think of where this may be a positive. I have always found “The Final Cut ” to be a bit too dynamic and the ADAU159 damps the bottom down a little. And again the ADAU1592 has crystal clarity, not to generalize too much but it is the type of sound you get with the classical solid state arrangement that everyone uses and is also used in opamps as well. You know the one with the diff-pair followed by the voltage gain stage, followed by the push-pull follower stage wrapped in a global feedback loop.
I had stated that the TA2024 seems to have more bass then the ADAU1592. Of course remember I am playing on 6” full range drivers. If you want deep base on full range drivers the nearest you’ll come is with a transconductance amplifier. My 8 watt 300B amps have far more base then these class-d offerings, but this is because it knows when to push the current through the voice coil.
Both of these amps have decent high end. However the TA2024 seems to do that thing with the high end that tubes do where every little detail is flushed out. In some cases even embellished with flex of gold highlights. Tubes are amazing when it comes to top end, though I exclude my own 300B here, because it has a dark top end, I am told this is because of the grid capacitance of the 300B output tubes themselves.
Noise:

All class-d amps seem to have a problem with noise. Unlike the more classical / conventional classes of operation where most of the noise comes in from the PS and can be filtered to inaudible levels by ever increasing the filtering of the PS, in class-d amps noise is generated by the amplifier it's self so building a better PS can only help you up to a point getting the 60Hz (50Hz Europe) buzz out. Class-d amps have a hiss. It's actually more annoying than the 60Hz buzz because the spectrum is wider. My speakers are in the 93-94dB /Watt range, so power is usually not a problem but small amounts of noise can be. I also believe all audiophile grade gear (whatever that means) should be significantly quite so that the operator cannot tell if the device is on or off even if they hold their ear to the speaker. I should also state that if I am sitting at my computer desk, on the side of the room, my ear is only one foot away from the speaker, and I am in this position a lot, but don’t worry I turn the volume way down.

So how do they compare, well the TA2024 has considerably more noise than the ADAU1592. I suspect this has to do with the TA2024 being a 2nd order delta-sigma modulator and the ADAU1592 being a 5th order delta-sigma modulator. I have tried shielding the TA2024 with aluminum foil and using a better PS none of these things made any difference. On the ADAU1592 the gain was set to zero dB. Moving it up to 6dB resulted in slightly more noise but still not as much as the TA2024. Taking the gain to 12dB had no effect on the noise. The TA2024 has fixed gain (fix by the pcb not chip its self) and I did not measure what it is set to. I can't tell if some of the noise comes from the signal source (it's most likely a combination of both) and I have tried grounding the outputs, but these are really subtle differences and no one standing more than 5 feet away would perceive any noise from either module. I watch a lot of British mysteries and some of the older DVD’s (now ripped to my HD) have noise actually recorded in the audio, Inspector Morse comes to mind. I may get out my soldering iron and try converting the TA2024 to zero gain, to see if it lowers the hiss, if it didn’t have the hiss it would be the perfect amp.

Conclusion:

Well you really can’t go wrong with any of these amps. If you like rock and you want clarity go with the ADAU1592. If you prefer Jazz and vocals, sound stage and warmth tone and tubey high lights go with the TA2024. The TA2024 makes more of an emotional connection. If you need the lowest noise go with the ADAU1592 set to 0db gain. If cost and prefabricated boards are an issue go with the TA2024, it costs 10 times less. Though in future availability of the TA2024 may become an issue, right now stocks are plentiful, as far as I can tell. No amp does everything well, so I like to keep a few around.

B.T.W.
I am selling five UcD400’s ($100) and one UcD supply HG ($200). Looking to sell it all at once. If anyone is interested, shoot me an e-mail (my user name)@rcf.rhic.bnl.gov . Only 2 out of the five boxes are open.
 

Attachments

  • IMGP2732.JPG
    IMGP2732.JPG
    99.9 KB · Views: 1,051
  • IMGP2734.JPG
    IMGP2734.JPG
    858 KB · Views: 944
  • IMGP2735.JPG
    IMGP2735.JPG
    680.5 KB · Views: 945
  • IMGP2736_copy.JPG
    IMGP2736_copy.JPG
    267.1 KB · Views: 882
  • IMGP2740.JPG
    IMGP2740.JPG
    178.8 KB · Views: 905
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.