I have an SMSL SU-9n DAC that I've been enjoying for more than an year. A few weeks ago it stared having interruptions while playing music. Initially I thought it was something network related. So I stared resetting switches and routers but with no success. Here is my main setup: the SMSL DAC is connected via USB to a RPI3 Roon endpoint. The RPI is connected with cable to my network. I have a TrueNAS NAS that has a Linux virtual machine where the Roon core is installed. Then the DAC is connected to the amp via XLR cables. I also reset the virtual machine but with no luck. I then connected the laptop directly to the DAC via USB and I noticed that the issue was still there. This clearly means that it is something within the DAC. I contacted the support from Shenzenaudio (I bought the DAC from them) and they replied that I can send it back to them for a repair. Obviously I do not want to wait around 10 weeks to get the DAC back so I asked if they saw this issue and if they have a fix for it. Working in R&D (electronics) and having access to nice tools I can fix it myself. Or at least I can try 
. They replied that I can try to fix it and if I do not succeed, I can send it but they didn't say anything about the issue I mentioned.
I then started looking around the DAC. First I noticed that the AC/DC converter that should have an output of +/-12Vdc and 1A, only had the +12Vdc and it had 0Vdc on the negative rail. But I noticed that that it was jumping around as it was some kind of AC. I measured around the DC-DC converter that are on the board and I noticed that the negative rail was all shorted out. I removed the AC-DC converter and it looked like the short was on the AC-DC module itself. I thought that this was defective. I was wrong. As it turns out, there is no negative DC output at all. It is actually the voltage taken directly from the secondary of the transformer (high frequency AC in the shape of flyback converter). This AC is then rectified with a schottky diode and filtered and then a DC-DC converter generates the negative rail. So bottom line, the AC-DC converter is fine. The short I was measuring is actually the resistance of the secondary winding.
At this stage, the DAC board was powered from a lab power supply with +/-12Vdc and it seemed like the interruptions were not present anymore. But after a while, these stared appearing again. I started looking around the DAC and I noticed that when I touched one IC, the interruptions got better. This is the IC in the image below (with the part number erased):
I thought that it might be the case of bad soldering and I reordered the IC. Again, it worked just fine for the entire weekend but it did have a few random interruptions but definitely not as bad. I then asked the shenzenaudio support if they can sell this IC as I do not know the part number, therefore I can't order it myself. They said no problem that asked me to contact them after the 1st of March as they now have bank holidays. Well I said ok but I didn't really fancy waiting that long.
I then went back and removed the IC from the PCB and cleaned the PCB and the IC. I put it back on and the behavior was pretty much the same. What I noticed, is that when I put the top cover back on, the interruptions are a lot worse than when I run the DAC with the lid open. This pointed toward a thermal issue and when I touched that IC, I was actually cooling it down. I went ahead and teste this by touching the plastic part of the IC with a tip of the soldering station set at 70C. It instantly stared interrupting and then complete silence. I then touched the IC with my finger and the music was back on. So it is definitely a problem with the IC itself. Yes I can run the DAC with the top cover off but this is not ok as I have kids and there is AC mains inside the enclosure. So the only solution is to try and find out what IC this is.
I then stared looking around the IC with the scope and I see that when the IC warms up, the clock on pin 10 (49MHz) starts to jiggle and it actually interrupts as seen below. The 24MHz clock coming from the TCXO goes to pin 2 of the IC.
Good clock:
Bad clock:
So it is clear that the IC starts malfunctioning due to some temperature issue. The IC temperature is around 51C (measured with a FLIR camera) so it is not that hot. I think it is probably broken and I think that when they removed the markings from the package (probably by laser) it might have damaged it. Or it is just a bad sample (lucky me ).
Next I tried to figure out what IC this was and after a few minutes on the webs, I think I found it. I think it is a clock generator Si5351A. Below you can see the comparison from the datasheet of the Si5351A and the PCB of the DAC. I can tell for sure that on pin 5 I have something resembling data and this would indicate an I2C SDA pin.
I will order this IC and I will test it. I will come back with the results.
. They replied that I can try to fix it and if I do not succeed, I can send it but they didn't say anything about the issue I mentioned.I then started looking around the DAC. First I noticed that the AC/DC converter that should have an output of +/-12Vdc and 1A, only had the +12Vdc and it had 0Vdc on the negative rail. But I noticed that that it was jumping around as it was some kind of AC. I measured around the DC-DC converter that are on the board and I noticed that the negative rail was all shorted out. I removed the AC-DC converter and it looked like the short was on the AC-DC module itself. I thought that this was defective. I was wrong. As it turns out, there is no negative DC output at all. It is actually the voltage taken directly from the secondary of the transformer (high frequency AC in the shape of flyback converter). This AC is then rectified with a schottky diode and filtered and then a DC-DC converter generates the negative rail. So bottom line, the AC-DC converter is fine. The short I was measuring is actually the resistance of the secondary winding.
At this stage, the DAC board was powered from a lab power supply with +/-12Vdc and it seemed like the interruptions were not present anymore. But after a while, these stared appearing again. I started looking around the DAC and I noticed that when I touched one IC, the interruptions got better. This is the IC in the image below (with the part number erased):
I thought that it might be the case of bad soldering and I reordered the IC. Again, it worked just fine for the entire weekend but it did have a few random interruptions but definitely not as bad. I then asked the shenzenaudio support if they can sell this IC as I do not know the part number, therefore I can't order it myself. They said no problem that asked me to contact them after the 1st of March as they now have bank holidays. Well I said ok but I didn't really fancy waiting that long.
I then went back and removed the IC from the PCB and cleaned the PCB and the IC. I put it back on and the behavior was pretty much the same. What I noticed, is that when I put the top cover back on, the interruptions are a lot worse than when I run the DAC with the lid open. This pointed toward a thermal issue and when I touched that IC, I was actually cooling it down. I went ahead and teste this by touching the plastic part of the IC with a tip of the soldering station set at 70C. It instantly stared interrupting and then complete silence. I then touched the IC with my finger and the music was back on. So it is definitely a problem with the IC itself. Yes I can run the DAC with the top cover off but this is not ok as I have kids and there is AC mains inside the enclosure. So the only solution is to try and find out what IC this is.
I then stared looking around the IC with the scope and I see that when the IC warms up, the clock on pin 10 (49MHz) starts to jiggle and it actually interrupts as seen below. The 24MHz clock coming from the TCXO goes to pin 2 of the IC.
Good clock:
Bad clock:
So it is clear that the IC starts malfunctioning due to some temperature issue. The IC temperature is around 51C (measured with a FLIR camera) so it is not that hot. I think it is probably broken and I think that when they removed the markings from the package (probably by laser) it might have damaged it. Or it is just a bad sample (lucky me
).Next I tried to figure out what IC this was and after a few minutes on the webs, I think I found it. I think it is a clock generator Si5351A. Below you can see the comparison from the datasheet of the Si5351A and the PCB of the DAC. I can tell for sure that on pin 5 I have something resembling data and this would indicate an I2C SDA pin.
I will order this IC and I will test it. I will come back with the results.
I know. But I think this might help others in solving this issue. There are many people who will not be able to perform this fix themselves and will go to repair shops/guys. I do not have any repair business and I would not have the time t fix DACsYou did a very, very bad thing! You probably just told a million guys how to fix their own dac instead of charging them some money and fix the problem yourself
.I also want to add that I sent an email to SMSL directly to buy the part (kelly@smsl-audio.com). This was before I stared investigating what part this is. I had no reply for 4 days now. So I decided to go ahead and try to figure it out myself.
I also want to add that I really like the DAC and how it sounds. I think it's a nice product and it seems to be pretty well made. I also think that the solution to use a clock generator to generate any clock for the DAC chip is pretty nice and clever. So kudos to SMSL. I think it's just that I was a bit unlucky.
Just a short update. I ordered the IC and I will replace it once it arrives. I did another test and indeed the optical and coax input work just fine when the USB doesn't. SO this IC is only used for USB playback. I guess that the clock is generated differently for COAX and optical. I just want to add that I also measured the Vcc voltage of this IC and it is 3.28V. So this is definitely ok. Not an issue of over-stressing the IC. It does seem though that the problem is getting worse as time goes by. Now even if I cool it down, it won't work.
I received the IC and I managed to replace it today. BINGO!!! It works just fine. It plays without any interruptions. I looked with the scope at the 49MHz clock that is used by the ES9038PRO and it looks a bit better than with the original IC. I also uses the hot air de-soldering station to heat the IC up to 70C and I did not get any interruptions. With the previous IC, when I did this, it almost instantly stopped playing. Now it works just fine.
So I think it's safe to say that the problem came from this IC. It was either damaged during production when its markings were removed, or it wasn't a genuine part. I replaced it with a SI5351A-B-GT that I got from DigiKey (https://www.digikey.ro/ro/products/detail/skyworks-solutions-inc/SI5351A-B-GT/3679847) but it can be found at Mouser as well (https://eu.mouser.com/ProductDetail...nc/SI5351A-B-GT?qs=p9T7GgSe1IEO%2BMfC/KPFfA==). I will keep listening to the DAC for the entire weekend and I'll report back if I see any issues.
This is the new IC soldered on the board.
This is the 49MHz clock that is used by the ES9038PRO DAC IC when playing 192kHz sampled audio.
So I think it's safe to say that the problem came from this IC. It was either damaged during production when its markings were removed, or it wasn't a genuine part. I replaced it with a SI5351A-B-GT that I got from DigiKey (https://www.digikey.ro/ro/products/detail/skyworks-solutions-inc/SI5351A-B-GT/3679847) but it can be found at Mouser as well (https://eu.mouser.com/ProductDetail...nc/SI5351A-B-GT?qs=p9T7GgSe1IEO%2BMfC/KPFfA==). I will keep listening to the DAC for the entire weekend and I'll report back if I see any issues.
This is the new IC soldered on the board.
This is the 49MHz clock that is used by the ES9038PRO DAC IC when playing 192kHz sampled audio.
After almost two weeks everything is fine. No more interruptions and other issue. So the issues seems to be fixed.
SMSL replied to my email after almost two weeks (Chinese new year I suppose) and that advised me to get in contact with the seller (Shenzenaudio). Of course I won't be doing this but nevertheless. It's good to see that they reply and that the issues would have been solved this way as well. A lot longer but still ok.
SMSL replied to my email after almost two weeks (Chinese new year I suppose) and that advised me to get in contact with the seller (Shenzenaudio). Of course I won't be doing this but nevertheless. It's good to see that they reply and that the issues would have been solved this way as well. A lot longer but still ok.
I have been using my SU-9N for 9 months and all was well until a week ago the music got static occasionally then worse and then distorted like the pitch was wandering. Then loud static noise so bad it was like full peak to peak swings. Turning it off and letting it cool, works for a few minutes but comes back. Definitely a thermal issue as it warms up it goes bad. I have heard others on ASR having similar issues. Did a search on DIYA and here we are.
Thank you horias2000 for posting this. I am thinking this could be a common design issue and will try fixing it the same way. I have tools and experience to fix SMT parts like this. Super detective work!
Perhaps bonding a small heatsink onto the clock generator chip will help longevity?
Thank you horias2000 for posting this. I am thinking this could be a common design issue and will try fixing it the same way. I have tools and experience to fix SMT parts like this. Super detective work!
Perhaps bonding a small heatsink onto the clock generator chip will help longevity?
Glad I could help. This is exactly why I posted my findings here and on ASR. I hope others will be able to fix their DACs as well. The DAC itself is really good (to my ears) and I lile the sound. Addind a small heatsinks will improve the situation. As it does if you take the top lid off and it cools the IC down. But it will get worse over time. So the real fix is to replace the IC. It should not be to difficult to replace if you have access to decent soldering equipment and the part itself is cheap and available .
.
I think it uses the 24MHz VCXO as an input. I did not see any reference that pin 3 should be left floating and not grounded. The strange thing is that the part is not a "B or C" one, it's the A type. And this one does not have any input other than a crystal. It must be that even an "A" part can work with an external clock input.
I just replaced the bad clock IC as show. In this thread. It’s working again so far. Will report back with long term testing in a few days. Hope it holds up. The replacement is easy if you have a hot air SMT soldering pencil, paste, tweezers and some good young eyes, or magnifying binoculars or microscope.
Here is original chip with label rubbed out:
Under microscope:
Here is replacement under microscope:
Thanks to the OP for showing us how to fix this issue!
Here is original chip with label rubbed out:
Under microscope:
Here is replacement under microscope:
Thanks to the OP for showing us how to fix this issue!
Last edited:
