These are the DC mV readings on my SMSL Sanskrit MK II DAC output while music is playing. Are these readings, which seem high to me, normal?
Its kind of unusual to post a link to a video of measurements here. Can you just post the results, or do you maybe want us to watch the video to help see if you are doing anything wrong?
It shows an unstable wandering offset that climbs up to 40 mV. I thought that perhaps the nature of the movement could provide some insight, hence the video.
Okay. That's not unusual when trying to measure DC in the presence of AC. To get an accurate picture of DC offset requires a LP filter in the DVM with a sharp cuttoff and a very low corner frequency, otherwise any AC also present can to some extent the influence the reading. It might take a longer integration time than your DVM offers to separate average DC from LF AC.
Help is still needed.
I have been investigating an offset due to a suspected problem, which has now been confirmed through repetition. When playing the Sanskrit 10 Mk II directly through my temporarily borrowed integrated Audiolab 8000SE amplifier, a slight but audible distortion develops in the upper range of my JBL 590 speakers after some time. This creates channel imbalance within a couple of dozen seconds. Sometimes the problem takes longer to manifest (over 30 minutes). Occasionally, switching the amp off and on resolves the issue. Although there have been listening sessions when it did not happen, it's a persistent problem that occurs more often than not. I suspect that the DAC is causing this issue because feeding the amplifier directly from the laptop's headphone output does not recreate the problem. Additionally, in retrospect, my previous setup of the DCG3 preamp and TSSA Class A/B amp (currently being serviced) showed signs of the same behavior with the DAC.
What kind of problem could I be facing? Is it a DAC-induced oscillation?
I have been investigating an offset due to a suspected problem, which has now been confirmed through repetition. When playing the Sanskrit 10 Mk II directly through my temporarily borrowed integrated Audiolab 8000SE amplifier, a slight but audible distortion develops in the upper range of my JBL 590 speakers after some time. This creates channel imbalance within a couple of dozen seconds. Sometimes the problem takes longer to manifest (over 30 minutes). Occasionally, switching the amp off and on resolves the issue. Although there have been listening sessions when it did not happen, it's a persistent problem that occurs more often than not. I suspect that the DAC is causing this issue because feeding the amplifier directly from the laptop's headphone output does not recreate the problem. Additionally, in retrospect, my previous setup of the DCG3 preamp and TSSA Class A/B amp (currently being serviced) showed signs of the same behavior with the DAC.
What kind of problem could I be facing? Is it a DAC-induced oscillation?
If you stop playing music does a DC offset at the output of the dac appear to persist, slowly decay, etc.? If you use a scope and suddenly pause play, do you see any DC offset at the output of the dac, even for a brief time?
When I stop a track, the offset turns quickly to a sub 1mV and reamains stable. Unfortunately, I don't have a scope to perform your test, but I am transitioning to a phase of researching scopes. Yesterday evening, I played the setup for a good 4 hours and the problem did not reoccur. I am quite puzzled by this mysterious issue.
It happened while listening to radio programs, watching TV streams, and listening to music. The time difference between occurrences since turning on the setup varied from approximately 10-15 minutes to 3 hours. I don't have anything more specific that could point to the source of the issue. It seems that the program material did not have an influence on it, but I could be wrong. I did accidentally switch the filter setting on the remote yesterday. Could that have influenced the behavior? Only the time will tell, I guess.
Here is my thoughts:
Dc offset can be embeded into the audio files. It happens all the time when producing/recording/mixing. If for any reason ( justified or not) it's not took into account ( corrected) then it ends up in the chain and it'll be present.
Everything broadcasted ( radio/tv) have a chance to be plagued by it because of the 'final treatments' used to gives a 'character/signature' which will amplify the issue if present.
Streaming probably sees same issue, i can't really tell as i don't know how it's done in this case.
Your better chance to identify it is to find a digital physical media which have the issue, makes a bit perfect copy to an harddrive and then analyse it through a soundeditor to define offset level.
From there you can see if the dac then behave the way it should or not.
If there was a filter engaged and then no issues i think you can reasonably conclude it's embeded in the source signal as it is very probably done ( filtering) in digital ( pre dac).
I use a Sanskrit 10th and have not noticed anything wrong with dc offset on this unit, but being mass produced goods...
Dc offset can be embeded into the audio files. It happens all the time when producing/recording/mixing. If for any reason ( justified or not) it's not took into account ( corrected) then it ends up in the chain and it'll be present.
Everything broadcasted ( radio/tv) have a chance to be plagued by it because of the 'final treatments' used to gives a 'character/signature' which will amplify the issue if present.
Streaming probably sees same issue, i can't really tell as i don't know how it's done in this case.
Your better chance to identify it is to find a digital physical media which have the issue, makes a bit perfect copy to an harddrive and then analyse it through a soundeditor to define offset level.
From there you can see if the dac then behave the way it should or not.
If there was a filter engaged and then no issues i think you can reasonably conclude it's embeded in the source signal as it is very probably done ( filtering) in digital ( pre dac).
I use a Sanskrit 10th and have not noticed anything wrong with dc offset on this unit, but being mass produced goods...
Last edited:
There is a thread on modding Sanskrit 10th mk II. I received a defective one with one channel distorting (not always). The volume control of the DAC had influence on the distortion. I suspected the opamp being defective maybe because of the grinding off of the type number. It turned out to be a defective DAC chip.
Please note that the DAC chip is connected partly directly to the 5V supply line so without a voltage regulator in between. Overvoltage will kill the AKM chip.
However you could be lucky with just a defective opamp. An OPA1642 performs quite good in the Sanskrit and it is not expensive.
Please note that the DAC chip is connected partly directly to the 5V supply line so without a voltage regulator in between. Overvoltage will kill the AKM chip.
However you could be lucky with just a defective opamp. An OPA1642 performs quite good in the Sanskrit and it is not expensive.
Last edited:
Hmmm.
I run mine off a smartphone psu (regulated) without issue ( input through toslink).
I've got a LT3045 psu i use in one of my tube microphone, maybe i should try it? It already crossed my mind but i was lazy to dismantle the psu...
Mine run fully open ( level), good to know it can influence output distortion and opa1642 could be used.
Thank you Jean- Paul.
I run mine off a smartphone psu (regulated) without issue ( input through toslink).
I've got a LT3045 psu i use in one of my tube microphone, maybe i should try it? It already crossed my mind but i was lazy to dismantle the psu...
Mine run fully open ( level), good to know it can influence output distortion and opa1642 could be used.
Thank you Jean- Paul.
Yours is not defective, it is Siberia that possibly has a defective MKII one and like I said I had one too (I also have a good one so I could compare). Those that don't have a certain issue often wonder why others have issues with the device but such things happen 🙂 Just suppose devices with a series error and only yours still working after 1 year. Also never compare device X with issue 1 with device Y. MkII has a completely different DAC chip than the MkI has.
Smart phone SMPS are for smart phones. We do not know the brand and type so no one can tell if it has a stable 5V. This device is a DAC for audio, it benefits from a clean linear PSU. But make sure the PSU used has no leakage/overvoltage whatsoever as the DAC chip is connected to either AUX power or USB power via a best battery diode system directly. A burst of let's say 15V will put 15V on the AKM chip.
No the volume control had influence on the distortion, the distortion became less when volume was not fully open. Still it distorted so it was a fast and short life for that DAC. The next MKII will receive OPA1656 as a test.
Smart phone SMPS are for smart phones. We do not know the brand and type so no one can tell if it has a stable 5V. This device is a DAC for audio, it benefits from a clean linear PSU. But make sure the PSU used has no leakage/overvoltage whatsoever as the DAC chip is connected to either AUX power or USB power via a best battery diode system directly. A burst of let's say 15V will put 15V on the AKM chip.
No the volume control had influence on the distortion, the distortion became less when volume was not fully open. Still it distorted so it was a fast and short life for that DAC. The next MKII will receive OPA1656 as a test.
Last edited:
Thank you for your comments and insights. The distortion mostly appeared on one channel, although I'm not 100% certain, and at least once it shifted to the other speaker after turning the power amp off and on again within the same listening session. In any case, it's such a ghostly phenomenon that I need it to replicate itself again so I can narrow down the possible causes.
It is sad but it’s probably a goner. I judged the defective one as irreparable. DAC Chip desoldering requires special tools, finding an affordable replacement chip is both not easy and not cheap. Then soldering the new chip again…
The device is harder to disassemble than usual too (by design).
The device is harder to disassemble than usual too (by design).
Last edited:
Replaced a QFP dac chip with bottom solder pad before. It was complicated by the fact that it was a double sided PCB with components on both sides, with some THT components sticking out on the bottom near the underside of the dac. So, preheat was pretty much not going to work.
Got the chip out using hot air and Chip Quik from the top. Had to get it very hot, but the chip was no good anyway. Cleaned up the board, prepped the bottom solder pad with some 63/37 and a little paste flux. Hand soldered the exposed side pins on the new dac chip, then used hot air from the top to melt the soft solder under the dac chip. It worked, despite being a little nerve wracking at times.
Got the chip out using hot air and Chip Quik from the top. Had to get it very hot, but the chip was no good anyway. Cleaned up the board, prepped the bottom solder pad with some 63/37 and a little paste flux. Hand soldered the exposed side pins on the new dac chip, then used hot air from the top to melt the soft solder under the dac chip. It worked, despite being a little nerve wracking at times.
Hi Jean-Paul,
Mine is a mk2 too ( bought it in 2020 from audiophonics).
One thing i wonder: Siberia how is it connected to your source? Do you use usb for dac power?
Mine is in my secondary ( analog) chain, mainly dedicated to cd player and tv/radio. As such i didn't wanted to have to strugle with groundloops so toslink connected to tv/radio and (short) spdif to cd player ( which act as a transport). And quality isn't as critical as in my mains.
I stopped using usb powered gear when possible since an issue with a semi pro soundcard ( Digidesign Mbox): changing computer changed S/N ratio so it wasn't usb power agnostic at all and i have little confidence on computer psu anyway (at least when i didn't choosen it by myself as with notebooks for example).
Jean- Paul is right in saying it is best practice to use a dedicated supply. Linear maybe better but i'm not afraid of smps either ( my main converters use smps and are not bad at all, as my digital desks or other digital gear i use in my home studio).
The psu i use is of this kind:
https://fr.rs-online.com/web/p/adaptateurs-ac-dc/1228377
Mine is a mk2 too ( bought it in 2020 from audiophonics).
One thing i wonder: Siberia how is it connected to your source? Do you use usb for dac power?
Mine is in my secondary ( analog) chain, mainly dedicated to cd player and tv/radio. As such i didn't wanted to have to strugle with groundloops so toslink connected to tv/radio and (short) spdif to cd player ( which act as a transport). And quality isn't as critical as in my mains.
I stopped using usb powered gear when possible since an issue with a semi pro soundcard ( Digidesign Mbox): changing computer changed S/N ratio so it wasn't usb power agnostic at all and i have little confidence on computer psu anyway (at least when i didn't choosen it by myself as with notebooks for example).
Jean- Paul is right in saying it is best practice to use a dedicated supply. Linear maybe better but i'm not afraid of smps either ( my main converters use smps and are not bad at all, as my digital desks or other digital gear i use in my home studio).
The psu i use is of this kind:
https://fr.rs-online.com/web/p/adaptateurs-ac-dc/1228377
Last edited:
My suspicions why the DAC chip goes bad are a few possible factors that all need to be investigated:
- either leakage current translating to a too high voltage on the power rail (the one I had here had such a PSU). This occurs in situations with mixed other devices having either PE or non PE so potential differences.
- the 4700 µF caps discharging through the DAC chip at power off/disconnect. The value is extremely large compared to what AKM recommends.
- the odd combination of having 4.7V because of the series diode and then a step up converter making 5.6V that is made 5V again?!? Maybe a power on/off phenomenon in this part of the circuit? I already modified mine to see if is solved. No series diode and no more microUSB.
- AK4493EQ possibly is a lousy sensitive chip breaking down for no external reason after x hours (don't exclude this possibility)
Don't be afraid of SMPS in audio, just replace them for something better. Even a basic 7805 based linear PSU is better than the average extremely noisy SMPS with only a few exceptions. There is no valid reason to choose SMPS over linear PSUs for sub 25VA applications in audio except low cost which is a non relevant parameter. Please don't copy stupid behavior "because everybody does it" or "brand X does it". SMPS are a designers wet dream but less so for the average audio person. Make sure to choose or build one with a true mains switch too.
I suggest to go to the other thread as I won't maintain 2 threads regarding MKII. Building and modifying costs valuable time and so does reporting here. Make sure not to only get but also give information.
- either leakage current translating to a too high voltage on the power rail (the one I had here had such a PSU). This occurs in situations with mixed other devices having either PE or non PE so potential differences.
- the 4700 µF caps discharging through the DAC chip at power off/disconnect. The value is extremely large compared to what AKM recommends.
- the odd combination of having 4.7V because of the series diode and then a step up converter making 5.6V that is made 5V again?!? Maybe a power on/off phenomenon in this part of the circuit? I already modified mine to see if is solved. No series diode and no more microUSB.
- AK4493EQ possibly is a lousy sensitive chip breaking down for no external reason after x hours (don't exclude this possibility)
Don't be afraid of SMPS in audio, just replace them for something better. Even a basic 7805 based linear PSU is better than the average extremely noisy SMPS with only a few exceptions. There is no valid reason to choose SMPS over linear PSUs for sub 25VA applications in audio except low cost which is a non relevant parameter. Please don't copy stupid behavior "because everybody does it" or "brand X does it". SMPS are a designers wet dream but less so for the average audio person. Make sure to choose or build one with a true mains switch too.
I suggest to go to the other thread as I won't maintain 2 threads regarding MKII. Building and modifying costs valuable time and so does reporting here. Make sure not to only get but also give information.
Last edited: