By the way, for anyone interested it looks like AKM released a 117-page data sheet for AK4499: https://www.akm.com/akm/en/product/detail/0054/
Hm, soxr does conversion to DSD and is open source available on github. I believe Volumio uses soxr for their upsampling to DSD as well.
The version of soxr on github does not appear to support DSD, not that I saw anyway. However, it looks like there is a modified version hosted elsewhere that does.
At least at this point, it still kind of looks a like there my be some licensing issue, otherwise hard to understand why there would not be wider support than there is.
Not that I know of. However, Twisted Pear makes a little ADM7150 AVCC board they use for Buffalo dacs. It's called Trident-SR. Trident-SR
So far as I hear through the grapevine, only some major manufacturers have the parts under NDA so far. Evaluation boards for everyone else are expected to hit the shelves at some of the usual electronics supply places we use, maybe sometime in June or July.
Right now I only have the dac chip data sheet, and the eval board manual. From looking at the eval board info, as is usual these days, the eval software won't run on Windows 10, or any 64-bit OS. They recommend WinXP. Time to learn how to use Arduino or other MCU, at least for those of us with a newer OS.
Hello to all! After a long wait for the components, they finally came! I'm already starting to work on powering VCAA, DVCC, Clock. So far made Clock. The ADM7154 transducers are back-mounted to the board with glue and soldered with thin wires to the components. Capacitors are used recommended by the manufacturer. I just added 0.1uf С0G in parallel to 10uf at the entrance and exit. There is still a lot of work ahead.
Recently, I got a question via PM about sharing one 5v supply for both the dac board and the AK4137 board. Since my response may be of interest to more than one person in the group I will post it here:
I use one LT3042 (with pass transistor) 5v, 1Amp, supply for both the dac board and for the AK4137 board. Both boards have additional 3.3v regulators, so for the most part sharing one 5v power supply is okay. However, not quite fully okay. Since the AK4137 I2S output is connected by ribbon cable to the dac board I2S input, it means there is a ground loop that includes the I2S ground wires and the 5v power supply ground wires. That caused what sounded like RF from the I2S ground to circulate around the ground loop and hurt sound quality a little. I put two clip-on ferrites on the 5v ground return line for one of the boards so that RF wouldn't flow through the power supply ground. That fixed the SQ problem so far as I can tell, but there are many different kinds of ferrite material and it comes in different sizes and shapes, so one might have to try some to see if they can help sound quality when using a shared 5v supply for AK4137 and the dac board.
I use one LT3042 (with pass transistor) 5v, 1Amp, supply for both the dac board and for the AK4137 board. Both boards have additional 3.3v regulators, so for the most part sharing one 5v power supply is okay. However, not quite fully okay. Since the AK4137 I2S output is connected by ribbon cable to the dac board I2S input, it means there is a ground loop that includes the I2S ground wires and the 5v power supply ground wires. That caused what sounded like RF from the I2S ground to circulate around the ground loop and hurt sound quality a little. I put two clip-on ferrites on the 5v ground return line for one of the boards so that RF wouldn't flow through the power supply ground. That fixed the SQ problem so far as I can tell, but there are many different kinds of ferrite material and it comes in different sizes and shapes, so one might have to try some to see if they can help sound quality when using a shared 5v supply for AK4137 and the dac board.
Best power config and amanero questions
*For an lt3042 board with 5v output what is the best AC voltage to supply to the board?
*I have an r-core that lists its output as 12v@.7A x2 and 9v@.7a x2, so four pairs of wires coming out of the r-core trafo. I also have a few e and l trafos of 12 and 24v @1amp
*Out of the Amanero, the DAC, and the akm4137 board, what are the most important to use clean power? I do hear a difference when my leaf audio amanero with upgraded caps uses an external supply
*The i2s inputs on the cheaper ak4137 board are numbered 1-6
*Is there a post that explains all the pins on the cheaper ak4137 board in this thread?
*My amanero has these 5 pins I use Gnd, MCK, LRCK, BCK, DATA, and it has the DSD pin which I understand outputs a stream of 1s when the amanero is outputting dsd.
*It also has FS0-FS3 Gnd, and 3v3 pins which I have never used. (Luckily it came with holes where the pins go and I scavenged many good quality pins from a digi002 that was scrapped.)
*(For the Amanero) I only have soldered pins into the Gnd, MCK, LRCK, BCK, & DATA holes.
*I still do not know about the fs0-3, and assume the second gnd and 3v3 are voltage outputs because it has an adapter input I assume is 5v and works @5v
What are the FS0,1,2,&3 pins for?
*For an lt3042 board with 5v output what is the best AC voltage to supply to the board?
*I have an r-core that lists its output as 12v@.7A x2 and 9v@.7a x2, so four pairs of wires coming out of the r-core trafo. I also have a few e and l trafos of 12 and 24v @1amp
*Out of the Amanero, the DAC, and the akm4137 board, what are the most important to use clean power? I do hear a difference when my leaf audio amanero with upgraded caps uses an external supply
*The i2s inputs on the cheaper ak4137 board are numbered 1-6
*Is there a post that explains all the pins on the cheaper ak4137 board in this thread?
*My amanero has these 5 pins I use Gnd, MCK, LRCK, BCK, DATA, and it has the DSD pin which I understand outputs a stream of 1s when the amanero is outputting dsd.
*It also has FS0-FS3 Gnd, and 3v3 pins which I have never used. (Luckily it came with holes where the pins go and I scavenged many good quality pins from a digi002 that was scrapped.)
*(For the Amanero) I only have soldered pins into the Gnd, MCK, LRCK, BCK, & DATA holes.
*I still do not know about the fs0-3, and assume the second gnd and 3v3 are voltage outputs because it has an adapter input I assume is 5v and works @5v
What are the FS0,1,2,&3 pins for?
Last edited:
I have heard a couple AKM DACs and they remind me of high end DENON cd players, Plasticky bass, too much of it, and harsh high ends. ESS wins in my book from what I have heard. I used the same preamp chip for all the boards I tested and the AKM (DAC) board seemed more well made. I prefer ESS and Cirrus and good ol' philips r to r and not too impressed by AKM.
BTW: what is this DAC-3 people were mentioning earlier? is it an upgraded signalyst dac?
RE
revious post: I just learned that fs0-3 are sampling rate indicators and 3v3 is indeed output, correct?
BTW: what is this DAC-3 people were mentioning earlier? is it an upgraded signalyst dac?
RE
revious post: I just learned that fs0-3 are sampling rate indicators and 3v3 is indeed output, correct?
Last edited:
Benchmark Media DAC-3:
Benchmark DAC3 HGC - Digital to Analog Audio Converter - Benchmark Media Systems
Benchmark DAC3 HGC D/A preamplifier-headphone amplifier | Stereophile.com
Benchmark DAC3 HGC review | What Hi-Fi?
Regarding AKM dacs, AK4499 will likely sound a bit different from earlier offerings. It has lower distortion than any Sabre dac (although only by a bit), and is current output like Sabre dacs. Its a change for AKM. We will have to see how it turns out.
Last edited:
The best AC voltage is the least voltage that allows the regulator to work, probably with a little headroom. 6VAC would be fine. Higher voltages are fine too so long as they don't result in exceeding the data sheet input voltage limit for LT3042. 9VAC is often okay too. Higher voltages mean it will dissipate more heat. So long as it doen't get too hot, its not a problem.
One 9VAC 700mA winding is probably enough.
Clean power is needed for everything in a good dac.
Amanero gets its power from USB, and has its own voltage regulator on-board.
Regarding Amanero, if you use it as the I2S master, its clocks will be used to clock your dac's I2S input so it can affect sound quality.
An ES9038Q2M needs three I2S signals. They are: DATA, BCLK, and LRCK. It also needs a ground for I2S, preferably three grounds, one for each of the I2S signals.
Depending on Amanero firmware, F!, F2, F3 indicate the incoming USB audio sample rate. The AK4137 board would like to see two of those, otherwise you might have to set the input sample rate manually.
Depending on Amanero firmware, the DSD pin is high when DSD is playing. The AK4137 board can use that signal too. However, a jumper needs to be removed on the low cost AK4137 board if using the Amanero DSD signal. Otherwise, you can set it manually.
Attached below is the dac chip pinout.
For the low cost AK4137 board, I have posted pictures of how to wire it up and a user manual I put together from all the different online ads for the board.
Pics are here: https://www.diyaudio.com/forums/digital-line-level/314935-es9038q2m-board-418.html#post5746444
To see the pics at full resolution, first open them by clicking, then find the white X in the lower left corner and click on that to open all the way. Pics can also be downloaded at full resolution by right clicking and selecting to save the image.
The manual is attached two posts down from where the pics are, in post #4175.
I suggest replacing the output pin header on the AK4137 board and the input pin header on the dac board with gold plated pins for lowest jitter. The nickel pin headers are rather poor when it comes to jitter. Otherwise, ribbon cable could be soldered between the boards instead of using connectors. Or, u.fl connectors could be added to both boards. Your choice. I also added extra I2S header ground pins for the I2S input and output on the AK4137 board, and on the dac board. That allows the ribbon cable to work properly for carrying RF signals with lower loss, lower noise pickup, and lower jitter. The extra ground pins on the AK4137 board are visible in one of the pics where they can be seen by the yellow color plastic at their base.
Attachments
Last edited:
Regarding my last post above I recommended using gold pins for one bit of wiring, but forgot to say that tin contacts in ribbon cable connectors should be avoided when using gold pins. Proper ribbon cable pin header female connectors should also use gold plated contacts. If it helps to save some expense for gold pin header connectors, ribbon cable could be soldered at one or both ends for the I2S connections going from AK4137 outputs to the dac board I2S inputs. Those are the most jitter sensitive connections, other pin header connectors don't really benefit as much from using gold and can use standard pins and ribbon cable connectors.
Last edited:
Hello to all. It's me again. I completed the revision of the Clock, VCCA, DVCC lines using the ADM7154. Now they eat separately.
One of these days I will receive the oscillator CCHD-575X-25-100.000M, I will use it.
Ahead still revision AVCC and the output part.
Mark, this is really very interesting and exciting. Now I can not stop doing better this DAC
BTW, I forgot to connect 2 power points on the VCCA line on both sides of the chip. But they are still connected inside. I think it's not so scary, but I will do it for sure.
One of these days I will receive the oscillator CCHD-575X-25-100.000M, I will use it.
Ahead still revision AVCC and the output part.
Mark, this is really very interesting and exciting. Now I can not stop doing better this DAC
BTW, I forgot to connect 2 power points on the VCCA line on both sides of the chip. But they are still connected inside. I think it's not so scary, but I will do it for sure.
Last edited:
Very nice and neat looking dead bug construction.
Regarding VCCA, yes, best to connect both sides, IMHO.
You could also replace the DVDD bypass cap at pin 15 of the dac chip. The recommended part is 4.7uf, 20%, X5R, 6.3v, 0402 size. The reason for so small a cap is to help minimize ESL (much of which is actually caused by the ground pad via).
I also replaced the power filter caps on top of the board for the clock, DVCC, AVCC, and DVDD. The old ones (probably mostly the ceramics) were rather piezoelectric (all high-k ceramics are to some extent) and pushing on the board with a finger could cause the clock frequency to drift. Better caps helped stabilize that quite a bit.
With a new clock, proper AVCC for each channel, and a good output stage, you will have the basis for a very nice sounding dac.
Do you think you will be getting into I2C bus and the dac registers at some point? If so, I better start working faster on a new clock divider for AK4137
The finishing touch on the dac to make it sound its best is to feed it synchronous DSD, however that is accomplished. There are other ways besides the way I did it, but it sounds very nice (possibly putting the modded dac just a bit below DAC-3 in sound quality, but it takes very careful attention to everything to get it to that point).
I probably paid bad attention to DVDD, but I knew that a 4.7uf X7R capacitor was recommended. I removed the old capacitors and put a new 4.7uf X7R Murata 1206. I did not know that we needed such a small size) Soon, then I will put the size 0402 (It is extremely small
)Did you notice a difference when changing capacitors to a DVDD or did you just put the recommended one?
About the rest of the ceramic capacitors, it is better to replace them too? Tantalum capacitors better to leave?
I did not understand a bit about the I2C bus. Is it the one that is connected to the microcontroller on the board? With the help of registers, we can change the settings of the DAC itself? I have not yet delved into this topic, but really want to understand all this. My board was with the display module and for the time being it suits me, I can switch the signal source, volume and filters.
But in the future I would like to work with the bus myself, is it possible to get more opportunities?
AK4137 allows upsampling and PCM conversion -> DSD. But is it necessary? How does this affect sound quality? So far, I'm new to this and I have a lot to learn)
I probably paid bad attention to DVDD, but I knew that a 4.7uf X7R capacitor was recommended. I removed the old capacitors and put a new 4.7uf X7R Murata 1206. I did not know that we needed such a small size) Soon, then I will put the size 0402 (It is extremely small
The smallest size you can reliably solder is probably fine. I did not notice a difference is SQ, but if replacing the other caps, may as well fix this one too.
I replaced both at once, so I don't know what would happen if only replacing ceramics. After the change, I could not shift the clock frequency to affect jitter by pressing on the board with my finger anywhere around the dac. Since all the caps are the same, if the clock caps were poor quality then the other must be poor too.
Yes. To get the best sound out of the dac it is essential to know how to program a few of the registers. It's easy to wire up if you have a board with 1.07 firmware since installed J1 and J2 at the same time will halt the board's MCU from accessing I2C bus. In that case you can control the dac chip yourself using Arduino or other MCU.
If all the other mods and done and if AK4137 board is modified slightly, then it can really improve the sound of the dac. The low cost AK4137 boards cost around $30 and are definitely worth having. What they do is upsample and convert PCM to DSD256. You could do the same thing using computer software, but then you would need to clock the dac differently and spend money on other parts that would end up costing more. However, AK4137 won't work to produce best sound quality without programming dac registers too. You need to program registers to (1)export MCLK on dac GPIO pin, (2)adjust DPLL bandwidth, and (3)use harmonic distortion compensation (that's all!
)EDIT: If you would like to know how to upsample and convert to DSD the software way, we could talk about that too, but like I said, it costs more.
Last edited: