Hello AIM, just for inform you, I wait the YJ double AK4495, since 21 days, I hope to recept it before Xmas...
I come back to you when i have it.
Charon
Hi, I've got the board pictured in post 49 with the xmos usb on the main board. Thought I'd share some of my observations. When I first fired up my board it had a DC offset of ~ 190mv on both channels. Problem was R8 in AIM's schematic was 3.2k instead of 3.6k. Good thing I checked since I plan to use this in a DC coupled headphone amp. Replacing it with 3.6k brought it down to ~17mv. Using an opamp calculator, (too lazy to do my own calcs!) I see that it is within the tolerance of the resistors but I would like to get it even further towards 0. The 150k resistor of r20 is used to bring it close to 3.4k which disregarding the small value resistors in the circuit balances the - gain with the + gain. I don't know why the the eval board circuit uses 3.6 but notice that they are using a much more complicated scheme for the VREFs including a separate power supply. I also found wrong values for some of the filter caps, even different values for the same cap on the other channel. Replaced all the filter caps with good Wima ones. I also had a opamp supply voltage difference of close to 1V. Replaced the zeners with the same ones from mouser and now have a close match. I am looking at every part on this board. I will replace the nover caps as soon as I decide if I want to keep the board. My board has 8 Nichicon fine gold caps around the DAC but with this board I'm afraid they might be fakes, gonna replace them also with the real thing....
Hi, that's interresting Nerd, I'll check tomorrow my DC offset. For now, on my board, I've no objective reasons to suspect fakes usage, so I'm not pursuing this direction, but I may be wrong....I didn't notice the same issues than you (component value, differents voltages, or unusual case or marking on components).
About the harsh sound in the mids, removing SSLOW enhance the thing, but I forgot to mention that at the same time I've done another mod and I agree with Spartacus post#117, so my mid enhancement may be the result of the mod and the digital filter change.
This mod is based on a post on a forum, post related to the LME49720. Idea behind the mod that is that most of the opamp we use (LME49720, OPA627, 603, MUSES,...etc) are very good ones and none of them may introduce such defect in the mid. So the issue is not the opamp but the way the opamp is used (some are very stable, other less...). Since have done the mod I remove the OPA627s and went back to the original LME49860 shipped with the board. (LM4562, LME49720, LME49860 are the same chip according to many source, between 49720 and 49860 there's a max power supply value difference)
Here's the link to the post :
http://www.head-fi.org/t/536589/new-opamp-anyone-tried-jrc-muses-8820-in-opa-rolling , this is post #55 at the bottom of the page.
It precisely describe the issue I've met. So I removed the two polyester cap close the two chemical ones (this is C02 and the symmetrical for other channel on the pictures of my previous post #99) and installed high quality MLCC X7R 100nf on V+ and V- and a 10nf NP0 between V+ and V- all directly soldered on the DIP socket under the board.
About the harsh sound in the mids, removing SSLOW enhance the thing, but I forgot to mention that at the same time I've done another mod and I agree with Spartacus post#117, so my mid enhancement may be the result of the mod and the digital filter change.
This mod is based on a post on a forum, post related to the LME49720. Idea behind the mod that is that most of the opamp we use (LME49720, OPA627, 603, MUSES,...etc) are very good ones and none of them may introduce such defect in the mid. So the issue is not the opamp but the way the opamp is used (some are very stable, other less...). Since have done the mod I remove the OPA627s and went back to the original LME49860 shipped with the board. (LM4562, LME49720, LME49860 are the same chip according to many source, between 49720 and 49860 there's a max power supply value difference)
Here's the link to the post :
http://www.head-fi.org/t/536589/new-opamp-anyone-tried-jrc-muses-8820-in-opa-rolling , this is post #55 at the bottom of the page.
It precisely describe the issue I've met. So I removed the two polyester cap close the two chemical ones (this is C02 and the symmetrical for other channel on the pictures of my previous post #99) and installed high quality MLCC X7R 100nf on V+ and V- and a 10nf NP0 between V+ and V- all directly soldered on the DIP socket under the board.
Nerd,
I’ve made the measurement, I’m fine with 5,4mV and 4,2mV. This output stage is unusual, in the literature and in lot of implementation we have R20=R6, R8=R7; R4=R2 and C5=C9 which simplify calculation and design (other resistor and cap are not involved in the mains characteristics of the filter). AKM guys in their reference design may have choose this option for fine tuning reasons.
About Vref, my board also have a separate power supply, provided by the adjustable LT1963, the other LT1963 is a fixed one delivering 3,3V for analog and digital parts of 4118 and 4495. My Vref is near 7V, this allow the 4495 to reach a S/N ratio of 123dB rather than 120 with Vref at 5V (see datasheet page 10) and a THD from -101dB to -105. Not sure my ears make the difference!…
As Vref has to be tied to VDDL/R, I believe this is the power supply of the output stage of the D/A converter inside the 4495 and the ref of the D/A.
Your board looks like mine, you may have the same Vref and the same LT1963. I believe this Vref supply is important for sound quality, I’m already done one mod but I want to check if my LT1963 is not a fake and may replace it by a low noise Ldo regulator bought in an authorized shop.
Could you share picture of your board and component in order to see if components, on their physical aspect, may be fake?, I can do same. My board is full of what should be good quality components: Sanyo Oscon cap, LT regulator, WIMA and ERO film cap, Nichicon gold cap; are they real ?!
Skrstic,
Mods are identified by yellow arrows in the two pictures.
On Above view :
• Restored control over SD and SLOW filter select switches : changed misplaced Dip Switch by jumpers
• Installed resistor attenuator on the output in order to reduce output level (-9,5dB)
• Changed 220uF on VDDL/R and VRefL/R on the output of the 6,7V regulator (LT1963) by 1500uF
• Removed film cap for high frequency decoupling of OpAmp
On Underneath view :
• Turned Off SSLOW filter :
o cut DVDD (+3,3V for Digital domain) track before and after the via feeding pin5 of the 4495.
o Connected the via to ground.
o Reestablished the connection to DVDD (brown wire)
• Resistor attenuatoro Connected the via to ground.
o Reestablished the connection to DVDD (brown wire)
• installed OpAmp decoupling as closest as possible to the chip and with rights caps
• Installed discharge resistor for the 1500uF cap (see above)
That’s it, the DAC sounds OK for now. Next mods are planned in order to replace the MCU by an Arduino and take full control of the AK4495 and 4118. I also have to find a way to check if some critical component (LT1963, Oscon Cap on the Xmos board, 22uF cap around the 4495,…) are not fake.
Attachments
Hi Aim , thanks for the info.
Just for clarity , where do the resistors for signal attenuation go and what values did you use?
Hi Ozcal,
this is fairly simple, as we say in my country :'a good sketch is better than a long discourse'.
I used low values in order to keep low output impedance (event if ly NAD amplifier as a 100k on on input); but not two low in order to not overload the opamp which will have to handle the feedback in high frequencies. Why 110 Ohms with two resistors : my shop sold me 220 Ohms ones by pack of 10...
Bests
Attachments
I've just done this mod. It's very worthwhile, makes the sound more natural and transparent. Thanks to both Bernie and AIM65.
vs. AK4495SEQ DAC I2S ?
Hi All,
Trying to decide between the AK4495+AK4118 DAC and the AK4495SEQ DAC I2S.
Has anyone tried the AK4495SEQ DAC I2S ?
Using the DAC with a Raspberry Pi.
Thanks !
Hi All,
Trying to decide between the AK4495+AK4118 DAC and the AK4495SEQ DAC I2S.
Has anyone tried the AK4495SEQ DAC I2S ?
Using the DAC with a Raspberry Pi.
Thanks !
Hi
Avoid boards with the two dip switchs, implementation is bad. You may have a look to this one Assembled AK4495SEQ AK4118 NE5534 Double Parallel Soft Control Board DOP DSD | eBay which provide serial control over the 4495 and a real dsd support.
Some DIYer reading this should give you some feedback about it.
regards
Guidance selecting AK4495 or AK4495SEQ
Thank you AIM65,
The Assembled AK4495SEQ AK4118 NE5534 Double Parallel Soft Control Board DOP DSD | eBay looks interesting.
Look forward to feedback about selecting a AK4495 or AK4495SEQ
All the best
Thank you AIM65,
The Assembled AK4495SEQ AK4118 NE5534 Double Parallel Soft Control Board DOP DSD | eBay looks interesting.
Look forward to feedback about selecting a AK4495 or AK4495SEQ
All the best
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.