Hi Joro,
Is the digital powersupply also reliant on the external ps? Is it not supposed to be 3.3V?
Ok 5v is also possible.
Gr GJ
Last edited:
According to docs:
AK4396: +5.2v (-0.2 +0.2) 200 mA
USB part: +4.1v (-0.2 +0.9) 400 mA
According to my measurements:
AK4396 around 335 mA
USB part min 92 mA max 111 mA depending of bitrate (min dsd 64 ; max 352800) I didn't measured dsd 128 and 384000
download xlsx
Sorry I wasn't clear enough. I mean in DSD mode, and not in PCM.
From what I know the "filter button" only works in PCM mode?
I didnot mean the "filter button"
I meant this "Reconstruction filter is made of MELF Vishay resistors and polyester capacitors."
But if it is the PCM or DSD optimized filter I do not know.
http://www.akm.com/akm/en/file/datasheet/AK4396VF.pdf
But Joro could answer that.
Gr GJ
Hi,
I assume that you have a technical mistake when you write the post 🙂 . You possibly meant:
AK4396 min 92 mA max 111 mA
USB part around 335 mA
Yes, the filter is only for PCM mode, AKM did that. I only able this feature on the board.
The filter is optimized for PCM signal and it's synthesized by me (if the filter was optimized for DSD, the filter would unnecessarily reject the PCM signals).
I attached filter characteristics (AK4396 DAC plus reconstruction filter) for 44.1kHz and 96kHz. The input signal is white noise. I don't measure the characteristics for 176.4kHz and up, because I use EMU0202 (moded) and this EMU can't measure this frequencies. For up 176.4kHz frequencies the characteristics have to be similar with the right proportion 🙂
Regards,
Joro
I assume that you have a technical mistake when you write the post 🙂 . You possibly meant:
AK4396 min 92 mA max 111 mA
USB part around 335 mA
Yes, the filter is only for PCM mode, AKM did that. I only able this feature on the board.
The filter is optimized for PCM signal and it's synthesized by me (if the filter was optimized for DSD, the filter would unnecessarily reject the PCM signals).
I attached filter characteristics (AK4396 DAC plus reconstruction filter) for 44.1kHz and 96kHz. The input signal is white noise. I don't measure the characteristics for 176.4kHz and up, because I use EMU0202 (moded) and this EMU can't measure this frequencies. For up 176.4kHz frequencies the characteristics have to be similar with the right proportion 🙂
Regards,
Joro
Attachments
In the AK4396 datasheet, Figure 16 on page 34 shows an example of an external LPF circuit with single op-amp - for PCM.
Figure 18 on page 36 is a similar circuit - for DSD.
The component values for the two circuits are significantly different.
There appears to be no easy solution to having a DAC which can switch between PCM and DSD sources.
I think I read that the ES9018 requires no additional external filtering for DSD, because its internal filters will deal with the extra HF noise (software configurable).
So ES9018 DAC's are a good choice, but not cheap of course.
One possible solution might be a transformer output stage - I understand that transformers roll off HF quite effectively, possibly adequate for DSD?
Figure 18 on page 36 is a similar circuit - for DSD.
The component values for the two circuits are significantly different.
There appears to be no easy solution to having a DAC which can switch between PCM and DSD sources.
I think I read that the ES9018 requires no additional external filtering for DSD, because its internal filters will deal with the extra HF noise (software configurable).
So ES9018 DAC's are a good choice, but not cheap of course.
One possible solution might be a transformer output stage - I understand that transformers roll off HF quite effectively, possibly adequate for DSD?
I have to say I'm really enjoying you product Joro. Thanks for the info on the filters.
Will you come out with different "dac" boards in the future?
Also, is it possible to get DSD256 and DSD512 support on the XMOS chip?
Will you come out with different "dac" boards in the future?
Also, is it possible to get DSD256 and DSD512 support on the XMOS chip?
Hi, Joro.
Thanks for USB converter. It is really excellent product.
I test with 2xpcm1794 in mono, i/u 100Om resistor || 78nF styrene capacitor without any output stage. Sounds very fine.
One question about 24-bit RJ (pin4 to pin1) - Left and right channels reversed. Stadart RJ has LRCK high for left and low for right ch. Which levels output you board? In configuration for pcm1704,63,61 etc (pin3 to 1) all is ok.
Thanks for USB converter. It is really excellent product.
I test with 2xpcm1794 in mono, i/u 100Om resistor || 78nF styrene capacitor without any output stage. Sounds very fine.
One question about 24-bit RJ (pin4 to pin1) - Left and right channels reversed. Stadart RJ has LRCK high for left and low for right ch. Which levels output you board? In configuration for pcm1704,63,61 etc (pin3 to 1) all is ok.
Guilty
. Yes, you are right. You are the first to notice that! Thank you for sharing.
When I was writing the firmware for RJ and LJ, in my mind is left by default I2S and..... One can easy deal with this issue with an invertor. Sorry about that. I updated the firmware - now the channels are OK.
Regards,
Joro
Last edited:
Got around to testing the power supplies for an XMOS/AK build. When I look at the supplies next to the XMOS/AK stack and the JG Filter... Well... seems a bit extreme 🙂
An externally hosted image should be here but it was not working when we last tested it.
Wired up the kit in the image above - and wow. My earlier build had a pedestrian 317 PS powering the JG filter. Switching to the Super Regulator shown took my platform up another level. It's not going to fit in a slimline case anymore, but who cares 🙂 Listening to it right now feeding a Crack with Speedball and then a pair of T90's. Very nice.
Sorry for the late reply, i tested my friend's XMOS+PCM5102 combination just this morning in my hifi system. The test DAC was totally USB-powered.
I have to say i don't agree with dhbiker: to my ears, my Xonar STX with 49990 opamps sound is clearly more convincing.
Joro's DAC sounds very good, anyway, in a very different way than my Xonar: the sound is a bit clearer and more open, basses are more controlled, detailed and precise. But Xonar sound is MUCH warmer and sweet and instrument timbres are much more realistic and pleasant, they have an "alive" feel that keeps listening fatigue near to zero. Sound is VERY pleasant in comparison...
XMOS+PCM5102 sound instead feels a bit "boxy" and cold, it finds it hard to come out from my speakers, like it was someway "filtered" or "dampened", with a lower volume. One can say it could be due to the lack of external power supply, but that can be said about my Xonar too, being it internally powered by my PC PSU as well.
BUT i must admit that Joro's DAC could improve substantially after a 50/100 hours break-in (it had just a couple of hours work until now): i don't know about this. What can you say?
Anyway, do you think AK4396 with external PSU could seriously improve in those fields or sound characteristics are similar to PCM5102 about the abovementioned factors?
Last edited:
You are soo wrong about external power supply. There is a huge difference between even a simple power supply and USB powered AK4396. Don't wait for break-in just go for one external power supply.
The comparison between internally powered Xonar and USB AKM is not making any sense.
Yes, i know, but we were talking about PCM5102 that is said to be less sensible to an external power supply.
For me, i'm looking forward to an AK4396 using 2 good external PSUs...
Sorry I missed PCM part 🙂.
I have them both (both with external power supplies), I played more with AKM but imo PCM5102 have a big improvement too from power supply.
Due to the fact that BiB GB was not finished when i received first AKM and PCM boards I used a simple power supply bought from a local electronics shop and it made a difference. Velleman nv - Item - Details - K7203
Now, thanks to MisterRogers advices, I have a working 2 Bib Powered AKM 🙂 and I moving forward to include my wandboard quad in the same setup, one ethernet in 2 RCAs out 🙂