Hooked up the Amanero board correctly now and it plays 44.1 native usb from my old Android tv box! The Amanero board doesn't have any ON leds at all! I found out that the best confirmation that the Dac works is seeing the big numerals appear on the screen when it finds a source.
Is the output of this dac a bit low btw? Even with a step up 1:4 it doesn't appear to play as loud as my Arcam rDac. So I guess next step will be to build the opamp circuit discussed earlier in this thread.
Is the output of this dac a bit low btw? Even with a step up 1:4 it doesn't appear to play as loud as my Arcam rDac. So I guess next step will be to build the opamp circuit discussed earlier in this thread.
Any one compared 9038Q2M to AK4493? very similar board, and price
AK4493 EQ DAC DSD Decoder Board Supports IIS 384KHz DSD512 + OLED display | eBay
AK4493 EQ DAC DSD Decoder Board Supports IIS 384KHz DSD512 + OLED display | eBay
Sound quality will increase for sure, even more if you do AVCC with something decent. Doesn't have to be film caps or LTC6655, but should be a dual opamp regulator with its own voltage reference source. LT3042 can serve as a low noise reference, but not so good without the dual opamp buffers.
With a proper output stage and proper AVCC supplies the dac should start to sound pretty decent. You will probably like the sound a lot more then, that's what everyone says when they hear it.
With or without modifications? The 9038Q2M board sounds pretty bad IMHO until significant modding is completed.
Don't know about the AK4493 board, but it ought to depend a lot on the quality of USB-to-I2S board clocks since they are used to provide the dac MCLK.
In the attached picture you can see the AVCC driven with LiFe batteries. One LT3042 is trickle charging a low impedance LiFe battery capable of 35A.. So a dual mono psu, and stereo separation is really good. At the moment there is a 9 ohm resistor between the regulator and the battery, and the battery is connected directly with a few mm of copper to the pcb. Avcc is bypassed by the smd film capacitor very close to the chip(the two flat silver caps.) The dac plays well at this point but the test output transformers is not linear, and its too much undertones. Its very musical but not high end.
I saw in another thread that someone recommended the THS3491 chip for a virtual ground since it has a slew rate of 8000 V/µs! Unfortunately the noise figures isn't as good, but it would be fun to try out if one could get it stable.
Has anyone found a inexpensive virtual ground preamp at this point? I'm still hoping for something pre built to mod
I saw in another thread that someone recommended the THS3491 chip for a virtual ground since it has a slew rate of 8000 V/µs! Unfortunately the noise figures isn't as good, but it would be fun to try out if one could get it stable.
Has anyone found a inexpensive virtual ground preamp at this point? I'm still hoping for something pre built to mod
Attachments
Last edited:
Doh, I forgot to the attach the picture. I edited the post now. Too bad the picture quality got lost somewhere. Which filter do you think sounds best when using the upsampler? The filter L_slow sounds the most natural on decaying cymbals through the transformers at the moment.
Last edited:
Best to crop and reduce JPG quality yourself to keep attached pictures a little below 1MB in size. That way the forum software won't do it for you in a way you can't control.
Which upsampler, AK4137, HQ Player, etc.?
Also, hard to tell what you are doing with some of the wiring and components shown in the picture. Do you have a schematic or maybe a block diagram to help make sense of it? Hand drawn would be fine, if legible
EDIT: Digital audio filters described or named as 'Slow' usually cut into the top the audio band. As a result they are not the most accurate, but they may sound best if there are other problems with the system somewhere.
Last edited:
I was thinking of the AK4137. Did cymbals and the top end improve much for you with the AK4137?
The AVCC right is driven with a battery and regulator in parallel. The Regulator has a resistor to limit the charging current if the battery becomes very discharged. Excact same for AVCC left. This was quickly put together so it might have some unforseen drawbacks, but it seems good enough for going further to the IV stage. Thanks for the audio filter info, there is a treble problem but I think is has to do with the transformers. Does the filter change the phase at the top end? I'm thinking the additive phase change from the transformers + filter might cause phase change in the signal.
The AVCC right is driven with a battery and regulator in parallel. The Regulator has a resistor to limit the charging current if the battery becomes very discharged. Excact same for AVCC left. This was quickly put together so it might have some unforseen drawbacks, but it seems good enough for going further to the IV stage. Thanks for the audio filter info, there is a treble problem but I think is has to do with the transformers. Does the filter change the phase at the top end? I'm thinking the additive phase change from the transformers + filter might cause phase change in the signal.
Last edited:
The digital filters might be linear phase, don't recall how much the datasheet says about the them. Seems to me the first (non-slow) one is the most accurate. It should sound best if all problems with the dac have been fixed.
IME, AK4137 sounded best with minimum stable DPLL bandwidth and or if synced with the dac clock. DSD sounded better than PCM (as often seems to be the case with modern dacs). IMHO, no point going to too much bother with clock syncing unless one is able to adjust DPLL bandwidth and HD compensation in the dac registers first.
Regarding the dac board, to summarize all the things that make it sound incrementally better they are: 3-opamp output stage (possibly with film caps on +-15v rails), opamp AVCC supplies, clock upgrade, Arduino or other MCU providing full control of dac registers, and separate local LDO voltage regulators for DVCC, VCCA, and clock.
Once all that is done, the dac will sound far, far better. At that point it will be ready to make best use of external devices such as upsamplers (possibly including DSD conversion) and or external interpolation filter for PCM playback.
Regarding dac register access, that should be a pretty high priority. Its no harder than some of the other mods and it makes a big difference. For some reason it seems to be something most diy'ers modding dacs appear least comfortable with. Don't know why.
We haven't heard from Kayata for awhile but he was getting very close to getting Arduino to work reliably. I sent him an I2C level converter, but don't know if he ever received it and or if he got it working.
IME, AK4137 sounded best with minimum stable DPLL bandwidth and or if synced with the dac clock. DSD sounded better than PCM (as often seems to be the case with modern dacs). IMHO, no point going to too much bother with clock syncing unless one is able to adjust DPLL bandwidth and HD compensation in the dac registers first.
Regarding the dac board, to summarize all the things that make it sound incrementally better they are: 3-opamp output stage (possibly with film caps on +-15v rails), opamp AVCC supplies, clock upgrade, Arduino or other MCU providing full control of dac registers, and separate local LDO voltage regulators for DVCC, VCCA, and clock.
Once all that is done, the dac will sound far, far better. At that point it will be ready to make best use of external devices such as upsamplers (possibly including DSD conversion) and or external interpolation filter for PCM playback.
Regarding dac register access, that should be a pretty high priority. Its no harder than some of the other mods and it makes a big difference. For some reason it seems to be something most diy'ers modding dacs appear least comfortable with. Don't know why.
We haven't heard from Kayata for awhile but he was getting very close to getting Arduino to work reliably. I sent him an I2C level converter, but don't know if he ever received it and or if he got it working.
Last edited:
Just to reaffirm, that experience was just subjective listening without level matching or bias controlling, right?
Hello. Sorry for the basic questions, but I'm hoping that someone will be able to help me get my board working. I have a 1.7 board (see post#822), that was bought some time ago, without the display.
My intention was to set the un-modded board up with a basic 317/337 type PS to get an idea of baseline performance. However, I couldn't even get to first base.
I powered the board up with J1 and J2 open, as I believed that was the correct configuration for I2S input from an XMOS board. Both the Lock and DSD leds came on immediately and remained solid, without even flickering, even when the USB was disconnected from the computer. All I could hear in the headphones was a solid hiss (white noise) with no sign of music.
I then tried shorting J1 and J2. This time both leds remained off, and never even flickered. However, the same hiss came through the headphones.
I initially thought that the XMOS board was faulty. However, it works without any problems, with one of the ES9028K2M boards discussed in an earlier thread.
Is my board faulty, or am I doing something wrong? Any help would be appreciated.
My intention was to set the un-modded board up with a basic 317/337 type PS to get an idea of baseline performance. However, I couldn't even get to first base.
I powered the board up with J1 and J2 open, as I believed that was the correct configuration for I2S input from an XMOS board. Both the Lock and DSD leds came on immediately and remained solid, without even flickering, even when the USB was disconnected from the computer. All I could hear in the headphones was a solid hiss (white noise) with no sign of music.
I then tried shorting J1 and J2. This time both leds remained off, and never even flickered. However, the same hiss came through the headphones.
I initially thought that the XMOS board was faulty. However, it works without any problems, with one of the ES9028K2M boards discussed in an earlier thread.
Is my board faulty, or am I doing something wrong? Any help would be appreciated.
Last edited:
Hi Naaling,
Doesn't sound too good from the way you describe it. If possible you might try checking power supply voltages, and then maybe try installing only J2 or J1 and use a SPDIF or TOSLINK digital audio source.
If either of those inputs work, then there could be some problem with I2S wiring or something. A scope could be very helpful to make sure.
The Lock and DSD LEDs are controlled by the MCU. With 1.07 firmware installing both J1 and J2 at the same time may halt MCU activity. Why the MCU is turning on both lights when it seemingly shouldn't be is a bit puzzling. Seems to me I may have seen that before when the MCU was locked up or malfunctioning for some reason. Often recycling power would fix it. Do both lights come on without any I2S connection hooked up?
Doesn't sound too good from the way you describe it. If possible you might try checking power supply voltages, and then maybe try installing only J2 or J1 and use a SPDIF or TOSLINK digital audio source.
If either of those inputs work, then there could be some problem with I2S wiring or something. A scope could be very helpful to make sure.
The Lock and DSD LEDs are controlled by the MCU. With 1.07 firmware installing both J1 and J2 at the same time may halt MCU activity. Why the MCU is turning on both lights when it seemingly shouldn't be is a bit puzzling. Seems to me I may have seen that before when the MCU was locked up or malfunctioning for some reason. Often recycling power would fix it. Do both lights come on without any I2S connection hooked up?
Tried the dac this evening with a virtual ground solution with some 5532 op amps through the balanced filter (recommended earlier in this thread) and into the transformers. I use the transformer to combine the differential signal. Not surprising the cheap virtual ground made a huge improvement.
The new head amp with XLR connection isn't ready yet, so I'm running unbalanced into the old amp with some slight hum. So does the negative on the rca from the amp go to Vref or to dac ground? I don't think I had any hum before I added the op amps.. I also got less volume with the op amp installed with the filter. Is the filter eating up some volume? I see some op amps are now capable of driving 100 ohms load, could I lower the filter resistance with those? I'm sorry if these questions have been asked before, its been 6 months since I took a break from this dac project so I have forgotten a lot..
The new head amp with XLR connection isn't ready yet, so I'm running unbalanced into the old amp with some slight hum. So does the negative on the rca from the amp go to Vref or to dac ground? I don't think I had any hum before I added the op amps.. I also got less volume with the op amp installed with the filter. Is the filter eating up some volume? I see some op amps are now capable of driving 100 ohms load, could I lower the filter resistance with those? I'm sorry if these questions have been asked before, its been 6 months since I took a break from this dac project so I have forgotten a lot..
Hi,
I have the same display, but it has no effect.Display works and I can see it changes sources (at least on display) but nothing happens with real input selection. Do I have to short any of the jumpers for this to work?
Seller description does not have any info about this..
Hi Mark. Thanks for your quick reply.
Yes, both lights come on without the XMOS board connected. However, I only hear the hiss when it is in place.
I don't think there is any problem with the power supply. I checked the supply voltages at the power pins of the op amp and they were OK. The ES9028K2M board works with the 317/337 power supply (although its only connected to half of it) and I also shorted J6 on the DAC and connected a 12v Wall wart, and there was no change.
I don't have SPDIF or Toslink sources to check the other inputs, but I wouldn't be hopeful. Both leds still light up when the jumpers are set for either of those inputs.
I think I have to accept that the board is dead. Thank you for your help.
impuls60,
Please read carefully the document at http://www.esstech.com/files/4514/4095/4306/Application_Note_Component_Selection_and_PCB_Layout.pdf
...It is old and a bit dated in some ways, but mostly still applicable. Vref is a DC-offset sorta-virtual ground. It is DC the voltage that the non-inverting inputs of the I/V opamps are connected to. Originally, Vref was equal to AVCC/2, so a divide-by-two voltage divider from AVCC was used to produce Vref. The divider is made from two 10k resistors with a 10uf cap added to the output to act as a filter. In more recent times, Vref may be adjusted to some voltage between AVCC/2 and ground, but closer to AVCC. That can help lower a certain small, but measurable distortion effect.
After looking at the document in the context of what I said about Vref, please let me know if that explains it or if any questions.
Please read carefully the document at http://www.esstech.com/files/4514/4095/4306/Application_Note_Component_Selection_and_PCB_Layout.pdf
...It is old and a bit dated in some ways, but mostly still applicable. Vref is a DC-offset sorta-virtual ground. It is DC the voltage that the non-inverting inputs of the I/V opamps are connected to. Originally, Vref was equal to AVCC/2, so a divide-by-two voltage divider from AVCC was used to produce Vref. The divider is made from two 10k resistors with a 10uf cap added to the output to act as a filter. In more recent times, Vref may be adjusted to some voltage between AVCC/2 and ground, but closer to AVCC. That can help lower a certain small, but measurable distortion effect.
After looking at the document in the context of what I said about Vref, please let me know if that explains it or if any questions.
If using the display to control the dac with the rotary encoder, then encoder should work to select the inputs. The only exception might be if jumpers J1 and J2 are installed at the same time, which can cause the display MCU to stop controlling the dac chip.
Otherwise, the dac can be tested without the display. IIRC, no jumpers selects I2S input, J1 selects optical, and J2 selects SPDIF.
Best to check at the output of the voltage regulators that drop the +15v down to 5v, and then down to 3.3v. If you don't know where to check, if you post a pic of the board you have we could probably point out where to probe.
If you have a scope or know someone who does, you could try checking the I2S signals and the dac clock to make sure everything is okay there.
Its possible you have a bad dac board. In that case, if the dac was purchased recently you might try contacting the seller to return it for repairs or replacement. Otherwise, you could try fixing it yourself, or maybe give up on it.
Then again, maybe someone else has some ideas of something else to try.