Prototyping AK4395 based DAC

Status
Not open for further replies.
Jump to Latest
M

macgyver

Member
Joined 2002
#1
I have almost finished my latest project. It's based on AK4114 receiver, AK4395 DAC, OPA1632 filter (still testing some filters) and PGA2310 volume chip. I'm controlling the AK4114, AK4395 and PGA2310 with ATmega8535 controller. Display is 2*16 VFD and I use rotary encoder for reading volume. I also have button and spdif boards (no pics yet sorry).

Here are some pics (I update when I get new data/pics):

Pics

Some of the pics ain't good 🙄
 
#10
maxlorenz said:
Congratulations Macgyver, very nice looking DAC.

I wish I could build something like that 🙁

Have you considered using (instead of AK4114 receiver?) one of the attractive Sample Rate Converters from AKM like AK4125 or AK4127.

Owning a modified M-Audio superDAC (AK4393) fed by an upsampler I can tell that it does make a difference. 🙂

Good luck,
M
Click to expand...

Thanks 🙂 I haven't thought about it. The AK412X would come after the AK4114.
 
#14
macgyver,

I'm liking the project. That board looks awesome!

Just thought I'd add my two cents about that Genelec paper, since it has some glaring omissions.

First off, I think they're cooking their numbers when they say their amplifiers have 130dB of dynamic range. 130dB is an awful lot for a power amplifier.
Second, they neglect to mention that the best DACs have as much as 120dB of dynamic range and so would be a good match for their systems.
Third, and very importantly, they've stated that digital attenuation is "a rather gross approximation". Their description of the effects of digital quantization is only true for undithered quantization. If you dither your quantization you end up with perfect linearity right down to the noise floor which is, as I've said, at -120dB. It is essentially like an analogue gain stage, but with much better performance than a real analogue gain stage.

If you were to use a good DAC, with it's analogue output normalized so that full scale was equal to 120dBSPL after the speakers, then its noise floor would be at 0dBSPL after the speakers. Totally inaudible. Properly-implemented digital attenuation would add NO noise or distortion to this performance (the only noise you'd have would be the 24-bit dither which is some 20dB below the DAC's output noise.

Don't mean to threadjack or anything! But papers like that one presented as "fact" rile me right up.

That said, that TI volume chip's supposed to be pretty darned good and I don't think you're going to be losing out by using it 😉
 
#15
I just came across this great thread. This is very interesting because Behringer DCX 2496 uses AK4393 chips. As far as I understand AK4393 and AK4396 are pin compatible, so exchange on the existing board is possible. Receiver chip in Behringer is CS 8420 and that should be exchanged for CS8416 in order to use full potential of new chips.
Macgyver could you please confirm if my thinking is OK? Any news on your project?
Thank you
AR2
 
#16
AR2 said:
I just came across this great thread. This is very interesting because Behringer DCX 2496 uses AK4393 chips. As far as I understand AK4393 and AK4396 are pin compatible, so exchange on the existing board is possible. Receiver chip in Behringer is CS 8420 and that should be exchanged for CS8416 in order to use full potential of new chips.
Macgyver could you please confirm if my thinking is OK? Any news on your project?
Thank you
AR2
Click to expand...

Yes, AK4393 and AK4396 are pin compatible. I'm not familiar with the receiver chip, but I think you will find some info in the datasheets etc.

My project is not ready yet. I have a cabinet for it but right now I'm finishing my master's thesis, so I have been busy last 6 months.
 
#19

Attachments

  • ak4396-in-place.jpg
    ak4396-in-place.jpg
    99.5 KB · Views: 746
Status
Not open for further replies.
Top Bottom