P8Z68/GEN3 Deluxe Noisy USB Bus
I was trying to listen to some MP3s on my PC via iTunes, and suddenly
I was like "What the heck - this sounds like crap". I was getting random clicking and crackling noises. It basically sounded like I was playing back a badly worn LP. First I tried the same track several times in a row and confirmed that the clicks and pops were random - not in the same location in the track. I tried different tracks from other albums - same story. Then I turned to hardware troubleshooting - before I tell you about that let me describe the sound system on my computer: It's PC to USB DAC to tube preamp to amplified speakers. I am not using the onboard sound on the motherboard. The USB DAC (digital to analog converter) is a little gadget like a USB thumb drive that plugs into a USB port on the back of the PC and outputs an analog signal on a standard 1/8" phone jack. The tube preamp provides volume control (and that "good old tube sound"), and the speakers each have their own amplifiers built in. First thing I thought was I had a bad connection somewhere. I unplugged and reseated the DAC and all the analog connections. Didn't help. I ruled out a bad speaker because the noise was occurring in both channels. Next I pulled out an older DAC and swapped the DACs out. Didn't help. I moved the DAC to two other USB ports - didn't help. I moved the DAC to my add-in Inateck USB 3.0 board (I'm not using the onboard USB 3.0 on this motherboard). This seemed to make the noise even worse. To eliminate the tube preamp as a source of noise I disconnected the preamp and speakers, and plugged a set of headphones directly into the jack on the DAC. I Still heard the same noises via the headphones. So now I've isolated the noise all the way back to the PC itself - my troubleshooting has eliminated anything external to the PC as a possible source of noise. I suppose it might be a noisy power supply (and I do have a spare) but that's a lot of work to change out unless I truly have to. The only other thing I could think of trying at this point was to remove the DAC and activate the onboard sound chip on the motherboard, so that's what I did. Asus had a massive Windows 10 audio driver on their web site that came out two years ago - but I decided to let Windows try and install the sound device by itself, figuring I'd probably get a more up to date driver that way. It did find a driver for it - it's a pretty basic driver, but I don't need any fancy features. Pure unadulterated two channel sound is all I was looking for. So I hooked up my preamp to the onboard sound port and played an MP3. Sounds fine, zero noise. I'm not going to say it's as good as the USB DAC, but it sounds a heck of a lot better than I was expecting. This leaves questions though. If the power supply was noisy, shouldn't that noise have permeated the onboard sound system as well? What is the nature of this noise I'm hearing through the USB DAC? If there is random noise on my USB bus, I would think I'd be at risk for corrupted data on USB devices like my big backup drive or my USB thumb drives? Any advice is appreciated! Configuration: Asus P8Z68 Deluxe/GEN3 motherboard 4x4GB Corsair Vengeance LP DDR3 1600 RAM Intel i7-2700K CPU EVGA Supernova 850 T2 power suppy EVGA GTX 970 FTW+ 4GB video card SanDisk Extreme Pro 480GB SSD boot drive 2x1TB Western Digital Black data drives Western Digital 8TB My Book Duo USB backup drive (in Raid 1) Antec P280 case Noctua CPU cooler and 3 Noctua case fans Inateck KTU3FR-2O2I add-in USB 3.0 card Pioneer 16X Blu-ray burner Windows 10 Pro version 1803 Meridian Explorer 2 DAC ifi iTube preamp Paradigm Shift A2 speakers Thanks in advance. Jerry |
P8Z68/GEN3 Deluxe Noisy USB Bus
Tater wrote:
I was trying to listen to some MP3s on my PC via iTunes, and suddenly I was like "What the heck - this sounds like crap". I was getting random clicking and crackling noises. It basically sounded like I was playing back a badly worn LP. First I tried the same track several times in a row and confirmed that the clicks and pops were random - not in the same location in the track. I tried different tracks from other albums - same story. Then I turned to hardware troubleshooting - before I tell you about that let me describe the sound system on my computer: It's PC to USB DAC to tube preamp to amplified speakers. I am not using the onboard sound on the motherboard. The USB DAC (digital to analog converter) is a little gadget like a USB thumb drive that plugs into a USB port on the back of the PC and outputs an analog signal on a standard 1/8" phone jack. The tube preamp provides volume control (and that "good old tube sound"), and the speakers each have their own amplifiers built in. First thing I thought was I had a bad connection somewhere. I unplugged and reseated the DAC and all the analog connections. Didn't help. I ruled out a bad speaker because the noise was occurring in both channels. Next I pulled out an older DAC and swapped the DACs out. Didn't help. I moved the DAC to two other USB ports - didn't help. I moved the DAC to my add-in Inateck USB 3.0 board (I'm not using the onboard USB 3.0 on this motherboard). This seemed to make the noise even worse. To eliminate the tube preamp as a source of noise I disconnected the preamp and speakers, and plugged a set of headphones directly into the jack on the DAC. I Still heard the same noises via the headphones. So now I've isolated the noise all the way back to the PC itself - my troubleshooting has eliminated anything external to the PC as a possible source of noise. I suppose it might be a noisy power supply (and I do have a spare) but that's a lot of work to change out unless I truly have to. The only other thing I could think of trying at this point was to remove the DAC and activate the onboard sound chip on the motherboard, so that's what I did. Asus had a massive Windows 10 audio driver on their web site that came out two years ago - but I decided to let Windows try and install the sound device by itself, figuring I'd probably get a more up to date driver that way. It did find a driver for it - it's a pretty basic driver, but I don't need any fancy features. Pure unadulterated two channel sound is all I was looking for. So I hooked up my preamp to the onboard sound port and played an MP3. Sounds fine, zero noise. I'm not going to say it's as good as the USB DAC, but it sounds a heck of a lot better than I was expecting. This leaves questions though. If the power supply was noisy, shouldn't that noise have permeated the onboard sound system as well? What is the nature of this noise I'm hearing through the USB DAC? If there is random noise on my USB bus, I would think I'd be at risk for corrupted data on USB devices like my big backup drive or my USB thumb drives? Any advice is appreciated! Configuration: Asus P8Z68 Deluxe/GEN3 motherboard 4x4GB Corsair Vengeance LP DDR3 1600 RAM Intel i7-2700K CPU EVGA Supernova 850 T2 power suppy EVGA GTX 970 FTW+ 4GB video card SanDisk Extreme Pro 480GB SSD boot drive 2x1TB Western Digital Black data drives Western Digital 8TB My Book Duo USB backup drive (in Raid 1) Antec P280 case Noctua CPU cooler and 3 Noctua case fans Inateck KTU3FR-2O2I add-in USB 3.0 card Pioneer 16X Blu-ray burner Windows 10 Pro version 1803 Meridian Explorer 2 DAC ifi iTube preamp Paradigm Shift A2 speakers Thanks in advance. Jerry It sounds like you already have good debugging skills, so I won't need to add much. You can record the output of your DAC, using the onboard sound and the LineIn on it. That's one possibility. Using Audacity, you could record the output waveform, and compare it to your stimulus, and see how the noise modulates the waveform. Audacity doesn't support playback and recording at the same time. Which usually means I need two programs, one to playback a stimulus and a second program to record, to get the data I want. Sometimes clicks or pops, are actually flat spots caused on the output waveform, by "buffer runs dry". The hardware has to signal that the buffer is getting low, and the system coughs up a few packets of data to fill up some kind of queue. If the queue ever runs dry (because the buffer cannot be serviced in time), the last voltage level on the DAC "sits there" until a packet of new samples arrive. The flat spot makes an audible noise, although not as loud a sound as you would think. When digital samples are corrupted (like the MSBit changes state), that creates a huge pop. And you'd notice that. So while it's fun to pretend the source is analog, what you could be hearing is a "delivery system problem". If you play a sine wave on some LineOut, and then record on a LineIn, you may notice disturbances which are not part of a sine wave. And that will give you some idea what class of fault this is. ******* The onboard sound has a positive voltage for the digital portion of the chip. It also has a positive voltage for the analog portion. That analog power source, is a relatively small three terminal regulator. By using a separate regulated power source, the intention is to reduce the level of "digital noise" riding on the motherboard sound power rails. Or on any analog amp after that point. It's possible to get inductive coupling into the onboard sound. That happens when the NIC wires go too close to the LineOut wires. But that would not be a likely mechanism for an external USB DAC. And your motherboard sound is clean. So at this point, I recommend recording the defective output, so you can diagnose the fault type. ******* People who buy "audio workstation" motherboards, they usually run "DPCLat" or delay procedure call latency test. But it doesn't work on Windows 10. The time it takes to service a "delayed procedure call", is a function of the motherboard performance. If the BIOS uses the SMM (system management mode) interrupt mechanism, it usurps the OS completely. A USB DAC might be waiting for a millisecond for data, until the SMM routine exits and the OS is running again. When this happens, it ruins the motherboard with respect to running audio workstation functions. If you can find a replacement program for "DPCLat", one that works on Windows 10, you can check out your "responsiveness" and see whether real-time performance is compromised for some reason. When SMM fires off, the OS doesn't know what happened. All it could notice, is a high performance clock could be off by a millisecond. The OS has no other record that it's been swapped out. Programs like DPCLat, attempt to quantify this problem, by "indirect measurement". If the DPC Latency has spikes in it, that's generally a sign that SMM is active. (It's because we don't know of any other high-enough-priority activity that can do that to the OS. Intel made sure that SMM was king.) Note that there is one hardware event, which is absolutely huge. When your video card changes into 3D gaming mode, there is a 15 millisecond spike in the DPCLat trace. Which implies the video card driver is "stuck" and holding the kernel hostage for a frame time, while it switches modes. Normally, people doing audio workstation work, aren't screwing with 3D games in the middle of recording their rock group. So this isn't a problem. I've not done extensive testing of what happens to tunes, when the video card changes modes, so I don't know what that one sounds like. Paul |
P8Z68/GEN3 Deluxe Noisy USB Bus
On Sat, 01 Sep 2018 23:24:49 -0400, Paul
wrote: Tater wrote: I was trying to listen to some MP3s on my PC via iTunes, and suddenly I was like "What the heck - this sounds like crap". I was getting random clicking and crackling noises. It basically sounded like I was playing back a badly worn LP. First I tried the same track several times in a row and confirmed that the clicks and pops were random - not in the same location in the track. I tried different tracks from other albums - same story. Then I turned to hardware troubleshooting - before I tell you about that let me describe the sound system on my computer: It's PC to USB DAC to tube preamp to amplified speakers. I am not using the onboard sound on the motherboard. The USB DAC (digital to analog converter) is a little gadget like a USB thumb drive that plugs into a USB port on the back of the PC and outputs an analog signal on a standard 1/8" phone jack. The tube preamp provides volume control (and that "good old tube sound"), and the speakers each have their own amplifiers built in. First thing I thought was I had a bad connection somewhere. I unplugged and reseated the DAC and all the analog connections. Didn't help. I ruled out a bad speaker because the noise was occurring in both channels. Next I pulled out an older DAC and swapped the DACs out. Didn't help. I moved the DAC to two other USB ports - didn't help. I moved the DAC to my add-in Inateck USB 3.0 board (I'm not using the onboard USB 3.0 on this motherboard). This seemed to make the noise even worse. To eliminate the tube preamp as a source of noise I disconnected the preamp and speakers, and plugged a set of headphones directly into the jack on the DAC. I Still heard the same noises via the headphones. So now I've isolated the noise all the way back to the PC itself - my troubleshooting has eliminated anything external to the PC as a possible source of noise. I suppose it might be a noisy power supply (and I do have a spare) but that's a lot of work to change out unless I truly have to. The only other thing I could think of trying at this point was to remove the DAC and activate the onboard sound chip on the motherboard, so that's what I did. Asus had a massive Windows 10 audio driver on their web site that came out two years ago - but I decided to let Windows try and install the sound device by itself, figuring I'd probably get a more up to date driver that way. It did find a driver for it - it's a pretty basic driver, but I don't need any fancy features. Pure unadulterated two channel sound is all I was looking for. So I hooked up my preamp to the onboard sound port and played an MP3. Sounds fine, zero noise. I'm not going to say it's as good as the USB DAC, but it sounds a heck of a lot better than I was expecting. This leaves questions though. If the power supply was noisy, shouldn't that noise have permeated the onboard sound system as well? What is the nature of this noise I'm hearing through the USB DAC? If there is random noise on my USB bus, I would think I'd be at risk for corrupted data on USB devices like my big backup drive or my USB thumb drives? Any advice is appreciated! Configuration: Asus P8Z68 Deluxe/GEN3 motherboard 4x4GB Corsair Vengeance LP DDR3 1600 RAM Intel i7-2700K CPU EVGA Supernova 850 T2 power suppy EVGA GTX 970 FTW+ 4GB video card SanDisk Extreme Pro 480GB SSD boot drive 2x1TB Western Digital Black data drives Western Digital 8TB My Book Duo USB backup drive (in Raid 1) Antec P280 case Noctua CPU cooler and 3 Noctua case fans Inateck KTU3FR-2O2I add-in USB 3.0 card Pioneer 16X Blu-ray burner Windows 10 Pro version 1803 Meridian Explorer 2 DAC ifi iTube preamp Paradigm Shift A2 speakers Thanks in advance. Jerry It sounds like you already have good debugging skills, so I won't need to add much. You can record the output of your DAC, using the onboard sound and the LineIn on it. That's one possibility. Using Audacity, you could record the output waveform, and compare it to your stimulus, and see how the noise modulates the waveform. Audacity doesn't support playback and recording at the same time. Which usually means I need two programs, one to playback a stimulus and a second program to record, to get the data I want. Sometimes clicks or pops, are actually flat spots caused on the output waveform, by "buffer runs dry". The hardware has to signal that the buffer is getting low, and the system coughs up a few packets of data to fill up some kind of queue. If the queue ever runs dry (because the buffer cannot be serviced in time), the last voltage level on the DAC "sits there" until a packet of new samples arrive. The flat spot makes an audible noise, although not as loud a sound as you would think. When digital samples are corrupted (like the MSBit changes state), that creates a huge pop. And you'd notice that. So while it's fun to pretend the source is analog, what you could be hearing is a "delivery system problem". If you play a sine wave on some LineOut, and then record on a LineIn, you may notice disturbances which are not part of a sine wave. And that will give you some idea what class of fault this is. ******* The onboard sound has a positive voltage for the digital portion of the chip. It also has a positive voltage for the analog portion. That analog power source, is a relatively small three terminal regulator. By using a separate regulated power source, the intention is to reduce the level of "digital noise" riding on the motherboard sound power rails. Or on any analog amp after that point. It's possible to get inductive coupling into the onboard sound. That happens when the NIC wires go too close to the LineOut wires. But that would not be a likely mechanism for an external USB DAC. And your motherboard sound is clean. So at this point, I recommend recording the defective output, so you can diagnose the fault type. ******* People who buy "audio workstation" motherboards, they usually run "DPCLat" or delay procedure call latency test. But it doesn't work on Windows 10. The time it takes to service a "delayed procedure call", is a function of the motherboard performance. If the BIOS uses the SMM (system management mode) interrupt mechanism, it usurps the OS completely. A USB DAC might be waiting for a millisecond for data, until the SMM routine exits and the OS is running again. When this happens, it ruins the motherboard with respect to running audio workstation functions. If you can find a replacement program for "DPCLat", one that works on Windows 10, you can check out your "responsiveness" and see whether real-time performance is compromised for some reason. When SMM fires off, the OS doesn't know what happened. All it could notice, is a high performance clock could be off by a millisecond. The OS has no other record that it's been swapped out. Programs like DPCLat, attempt to quantify this problem, by "indirect measurement". If the DPC Latency has spikes in it, that's generally a sign that SMM is active. (It's because we don't know of any other high-enough-priority activity that can do that to the OS. Intel made sure that SMM was king.) Note that there is one hardware event, which is absolutely huge. When your video card changes into 3D gaming mode, there is a 15 millisecond spike in the DPCLat trace. Which implies the video card driver is "stuck" and holding the kernel hostage for a frame time, while it switches modes. Normally, people doing audio workstation work, aren't screwing with 3D games in the middle of recording their rock group. So this isn't a problem. I've not done extensive testing of what happens to tunes, when the video card changes modes, so I don't know what that one sounds like. Paul Paul, Thanks for the incredibly thoughtful and highly technical response. One possibly important clue I forgot to mention earlier is that there is no noise coming through the DAC when there is no signal. That in itself might mean that it's not the USB bus itself that's noisy, but that the noise is being generated by the SATA controller when the MP3 file is being fetched for playback. I'm goiong to try other sources as a means of troubleshooting this. If it is the SATA bus, then I should also get the noise when playing back a CD (because the SATA controller is fetching the WAV file from the CD drive). However I would not expect to get the noise when playing Pandora, as that is purely network traffic and doesn't involve the SATA bus. I'll do further troubleshooting and report back on Sunday. Jerry |
P8Z68/GEN3 Deluxe Noisy USB Bus
On Sun, 02 Sep 2018 00:33:43 -0700, Tater wrote:
On Sat, 01 Sep 2018 23:24:49 -0400, Paul wrote: Tater wrote: I was trying to listen to some MP3s on my PC via iTunes, and suddenly I was like "What the heck - this sounds like crap". I was getting random clicking and crackling noises. It basically sounded like I was playing back a badly worn LP. First I tried the same track several times in a row and confirmed that the clicks and pops were random - not in the same location in the track. I tried different tracks from other albums - same story. Then I turned to hardware troubleshooting - before I tell you about that let me describe the sound system on my computer: It's PC to USB DAC to tube preamp to amplified speakers. I am not using the onboard sound on the motherboard. The USB DAC (digital to analog converter) is a little gadget like a USB thumb drive that plugs into a USB port on the back of the PC and outputs an analog signal on a standard 1/8" phone jack. The tube preamp provides volume control (and that "good old tube sound"), and the speakers each have their own amplifiers built in. First thing I thought was I had a bad connection somewhere. I unplugged and reseated the DAC and all the analog connections. Didn't help. I ruled out a bad speaker because the noise was occurring in both channels. Next I pulled out an older DAC and swapped the DACs out. Didn't help. I moved the DAC to two other USB ports - didn't help. I moved the DAC to my add-in Inateck USB 3.0 board (I'm not using the onboard USB 3.0 on this motherboard). This seemed to make the noise even worse. To eliminate the tube preamp as a source of noise I disconnected the preamp and speakers, and plugged a set of headphones directly into the jack on the DAC. I Still heard the same noises via the headphones. So now I've isolated the noise all the way back to the PC itself - my troubleshooting has eliminated anything external to the PC as a possible source of noise. I suppose it might be a noisy power supply (and I do have a spare) but that's a lot of work to change out unless I truly have to. The only other thing I could think of trying at this point was to remove the DAC and activate the onboard sound chip on the motherboard, so that's what I did. Asus had a massive Windows 10 audio driver on their web site that came out two years ago - but I decided to let Windows try and install the sound device by itself, figuring I'd probably get a more up to date driver that way. It did find a driver for it - it's a pretty basic driver, but I don't need any fancy features. Pure unadulterated two channel sound is all I was looking for. So I hooked up my preamp to the onboard sound port and played an MP3. Sounds fine, zero noise. I'm not going to say it's as good as the USB DAC, but it sounds a heck of a lot better than I was expecting. This leaves questions though. If the power supply was noisy, shouldn't that noise have permeated the onboard sound system as well? What is the nature of this noise I'm hearing through the USB DAC? If there is random noise on my USB bus, I would think I'd be at risk for corrupted data on USB devices like my big backup drive or my USB thumb drives? Any advice is appreciated! Configuration: Asus P8Z68 Deluxe/GEN3 motherboard 4x4GB Corsair Vengeance LP DDR3 1600 RAM Intel i7-2700K CPU EVGA Supernova 850 T2 power suppy EVGA GTX 970 FTW+ 4GB video card SanDisk Extreme Pro 480GB SSD boot drive 2x1TB Western Digital Black data drives Western Digital 8TB My Book Duo USB backup drive (in Raid 1) Antec P280 case Noctua CPU cooler and 3 Noctua case fans Inateck KTU3FR-2O2I add-in USB 3.0 card Pioneer 16X Blu-ray burner Windows 10 Pro version 1803 Meridian Explorer 2 DAC ifi iTube preamp Paradigm Shift A2 speakers Thanks in advance. Jerry It sounds like you already have good debugging skills, so I won't need to add much. You can record the output of your DAC, using the onboard sound and the LineIn on it. That's one possibility. Using Audacity, you could record the output waveform, and compare it to your stimulus, and see how the noise modulates the waveform. Audacity doesn't support playback and recording at the same time. Which usually means I need two programs, one to playback a stimulus and a second program to record, to get the data I want. Sometimes clicks or pops, are actually flat spots caused on the output waveform, by "buffer runs dry". The hardware has to signal that the buffer is getting low, and the system coughs up a few packets of data to fill up some kind of queue. If the queue ever runs dry (because the buffer cannot be serviced in time), the last voltage level on the DAC "sits there" until a packet of new samples arrive. The flat spot makes an audible noise, although not as loud a sound as you would think. When digital samples are corrupted (like the MSBit changes state), that creates a huge pop. And you'd notice that. So while it's fun to pretend the source is analog, what you could be hearing is a "delivery system problem". If you play a sine wave on some LineOut, and then record on a LineIn, you may notice disturbances which are not part of a sine wave. And that will give you some idea what class of fault this is. ******* The onboard sound has a positive voltage for the digital portion of the chip. It also has a positive voltage for the analog portion. That analog power source, is a relatively small three terminal regulator. By using a separate regulated power source, the intention is to reduce the level of "digital noise" riding on the motherboard sound power rails. Or on any analog amp after that point. It's possible to get inductive coupling into the onboard sound. That happens when the NIC wires go too close to the LineOut wires. But that would not be a likely mechanism for an external USB DAC. And your motherboard sound is clean. So at this point, I recommend recording the defective output, so you can diagnose the fault type. ******* People who buy "audio workstation" motherboards, they usually run "DPCLat" or delay procedure call latency test. But it doesn't work on Windows 10. The time it takes to service a "delayed procedure call", is a function of the motherboard performance. If the BIOS uses the SMM (system management mode) interrupt mechanism, it usurps the OS completely. A USB DAC might be waiting for a millisecond for data, until the SMM routine exits and the OS is running again. When this happens, it ruins the motherboard with respect to running audio workstation functions. If you can find a replacement program for "DPCLat", one that works on Windows 10, you can check out your "responsiveness" and see whether real-time performance is compromised for some reason. When SMM fires off, the OS doesn't know what happened. All it could notice, is a high performance clock could be off by a millisecond. The OS has no other record that it's been swapped out. Programs like DPCLat, attempt to quantify this problem, by "indirect measurement". If the DPC Latency has spikes in it, that's generally a sign that SMM is active. (It's because we don't know of any other high-enough-priority activity that can do that to the OS. Intel made sure that SMM was king.) Note that there is one hardware event, which is absolutely huge. When your video card changes into 3D gaming mode, there is a 15 millisecond spike in the DPCLat trace. Which implies the video card driver is "stuck" and holding the kernel hostage for a frame time, while it switches modes. Normally, people doing audio workstation work, aren't screwing with 3D games in the middle of recording their rock group. So this isn't a problem. I've not done extensive testing of what happens to tunes, when the video card changes modes, so I don't know what that one sounds like. Paul Paul, Thanks for the incredibly thoughtful and highly technical response. One possibly important clue I forgot to mention earlier is that there is no noise coming through the DAC when there is no signal. That in itself might mean that it's not the USB bus itself that's noisy, but that the noise is being generated by the SATA controller when the MP3 file is being fetched for playback. I'm goiong to try other sources as a means of troubleshooting this. If it is the SATA bus, then I should also get the noise when playing back a CD (because the SATA controller is fetching the WAV file from the CD drive). However I would not expect to get the noise when playing Pandora, as that is purely network traffic and doesn't involve the SATA bus. I'll do further troubleshooting and report back on Sunday. Jerry I hate computers! I just now got around to messing with the PC again. I uninstalled the onboard sound via Device Manager, then rebooted and entered the BIOS to turn it off. I then re-attached my Meridian Explorer 2 USB DAC and installed the driver for it (even though the manufacturer claims it works under Windows 10 without drivers, I've never been able to make that work - I always have to load the Windows 7 driver). I checked Device Manager, and the Device Properties dialogs in the Sound section of Settings to make sure everything was looking normal and set up correctly. Then I figured I'd better confirm the noise again before trying different things to troubleshoot it. And naturally, the noise is now GONE. I want to pound my head against the wall. All that troubleshooting yesterday was a waste of time. I never did find the cause of the problem, and I have no idea why it went away. Guess I won't be buying a new PC after all. Paul, thank you again for taking the time to provide all that wonderful advice for me. I really appreciate it. I'm sorry that it proved unnecessary in the end. Jerry |
P8Z68/GEN3 Deluxe Noisy USB Bus
Tater wrote:
On Sun, 02 Sep 2018 00:33:43 -0700, Tater wrote: On Sat, 01 Sep 2018 23:24:49 -0400, Paul wrote: Tater wrote: I was trying to listen to some MP3s on my PC via iTunes, and suddenly I was like "What the heck - this sounds like crap". I was getting random clicking and crackling noises. It basically sounded like I was playing back a badly worn LP. First I tried the same track several times in a row and confirmed that the clicks and pops were random - not in the same location in the track. I tried different tracks from other albums - same story. Then I turned to hardware troubleshooting - before I tell you about that let me describe the sound system on my computer: It's PC to USB DAC to tube preamp to amplified speakers. I am not using the onboard sound on the motherboard. The USB DAC (digital to analog converter) is a little gadget like a USB thumb drive that plugs into a USB port on the back of the PC and outputs an analog signal on a standard 1/8" phone jack. The tube preamp provides volume control (and that "good old tube sound"), and the speakers each have their own amplifiers built in. First thing I thought was I had a bad connection somewhere. I unplugged and reseated the DAC and all the analog connections. Didn't help. I ruled out a bad speaker because the noise was occurring in both channels. Next I pulled out an older DAC and swapped the DACs out. Didn't help. I moved the DAC to two other USB ports - didn't help. I moved the DAC to my add-in Inateck USB 3.0 board (I'm not using the onboard USB 3.0 on this motherboard). This seemed to make the noise even worse. To eliminate the tube preamp as a source of noise I disconnected the preamp and speakers, and plugged a set of headphones directly into the jack on the DAC. I Still heard the same noises via the headphones. So now I've isolated the noise all the way back to the PC itself - my troubleshooting has eliminated anything external to the PC as a possible source of noise. I suppose it might be a noisy power supply (and I do have a spare) but that's a lot of work to change out unless I truly have to. The only other thing I could think of trying at this point was to remove the DAC and activate the onboard sound chip on the motherboard, so that's what I did. Asus had a massive Windows 10 audio driver on their web site that came out two years ago - but I decided to let Windows try and install the sound device by itself, figuring I'd probably get a more up to date driver that way. It did find a driver for it - it's a pretty basic driver, but I don't need any fancy features. Pure unadulterated two channel sound is all I was looking for. So I hooked up my preamp to the onboard sound port and played an MP3. Sounds fine, zero noise. I'm not going to say it's as good as the USB DAC, but it sounds a heck of a lot better than I was expecting. This leaves questions though. If the power supply was noisy, shouldn't that noise have permeated the onboard sound system as well? What is the nature of this noise I'm hearing through the USB DAC? If there is random noise on my USB bus, I would think I'd be at risk for corrupted data on USB devices like my big backup drive or my USB thumb drives? Any advice is appreciated! Configuration: Asus P8Z68 Deluxe/GEN3 motherboard 4x4GB Corsair Vengeance LP DDR3 1600 RAM Intel i7-2700K CPU EVGA Supernova 850 T2 power suppy EVGA GTX 970 FTW+ 4GB video card SanDisk Extreme Pro 480GB SSD boot drive 2x1TB Western Digital Black data drives Western Digital 8TB My Book Duo USB backup drive (in Raid 1) Antec P280 case Noctua CPU cooler and 3 Noctua case fans Inateck KTU3FR-2O2I add-in USB 3.0 card Pioneer 16X Blu-ray burner Windows 10 Pro version 1803 Meridian Explorer 2 DAC ifi iTube preamp Paradigm Shift A2 speakers Thanks in advance. Jerry It sounds like you already have good debugging skills, so I won't need to add much. You can record the output of your DAC, using the onboard sound and the LineIn on it. That's one possibility. Using Audacity, you could record the output waveform, and compare it to your stimulus, and see how the noise modulates the waveform. Audacity doesn't support playback and recording at the same time. Which usually means I need two programs, one to playback a stimulus and a second program to record, to get the data I want. Sometimes clicks or pops, are actually flat spots caused on the output waveform, by "buffer runs dry". The hardware has to signal that the buffer is getting low, and the system coughs up a few packets of data to fill up some kind of queue. If the queue ever runs dry (because the buffer cannot be serviced in time), the last voltage level on the DAC "sits there" until a packet of new samples arrive. The flat spot makes an audible noise, although not as loud a sound as you would think. When digital samples are corrupted (like the MSBit changes state), that creates a huge pop. And you'd notice that. So while it's fun to pretend the source is analog, what you could be hearing is a "delivery system problem". If you play a sine wave on some LineOut, and then record on a LineIn, you may notice disturbances which are not part of a sine wave. And that will give you some idea what class of fault this is. ******* The onboard sound has a positive voltage for the digital portion of the chip. It also has a positive voltage for the analog portion. That analog power source, is a relatively small three terminal regulator. By using a separate regulated power source, the intention is to reduce the level of "digital noise" riding on the motherboard sound power rails. Or on any analog amp after that point. It's possible to get inductive coupling into the onboard sound. That happens when the NIC wires go too close to the LineOut wires. But that would not be a likely mechanism for an external USB DAC. And your motherboard sound is clean. So at this point, I recommend recording the defective output, so you can diagnose the fault type. ******* People who buy "audio workstation" motherboards, they usually run "DPCLat" or delay procedure call latency test. But it doesn't work on Windows 10. The time it takes to service a "delayed procedure call", is a function of the motherboard performance. If the BIOS uses the SMM (system management mode) interrupt mechanism, it usurps the OS completely. A USB DAC might be waiting for a millisecond for data, until the SMM routine exits and the OS is running again. When this happens, it ruins the motherboard with respect to running audio workstation functions. If you can find a replacement program for "DPCLat", one that works on Windows 10, you can check out your "responsiveness" and see whether real-time performance is compromised for some reason. When SMM fires off, the OS doesn't know what happened. All it could notice, is a high performance clock could be off by a millisecond. The OS has no other record that it's been swapped out. Programs like DPCLat, attempt to quantify this problem, by "indirect measurement". If the DPC Latency has spikes in it, that's generally a sign that SMM is active. (It's because we don't know of any other high-enough-priority activity that can do that to the OS. Intel made sure that SMM was king.) Note that there is one hardware event, which is absolutely huge. When your video card changes into 3D gaming mode, there is a 15 millisecond spike in the DPCLat trace. Which implies the video card driver is "stuck" and holding the kernel hostage for a frame time, while it switches modes. Normally, people doing audio workstation work, aren't screwing with 3D games in the middle of recording their rock group. So this isn't a problem. I've not done extensive testing of what happens to tunes, when the video card changes modes, so I don't know what that one sounds like. Paul Paul, Thanks for the incredibly thoughtful and highly technical response. One possibly important clue I forgot to mention earlier is that there is no noise coming through the DAC when there is no signal. That in itself might mean that it's not the USB bus itself that's noisy, but that the noise is being generated by the SATA controller when the MP3 file is being fetched for playback. I'm goiong to try other sources as a means of troubleshooting this. If it is the SATA bus, then I should also get the noise when playing back a CD (because the SATA controller is fetching the WAV file from the CD drive). However I would not expect to get the noise when playing Pandora, as that is purely network traffic and doesn't involve the SATA bus. I'll do further troubleshooting and report back on Sunday. Jerry I hate computers! I just now got around to messing with the PC again. I uninstalled the onboard sound via Device Manager, then rebooted and entered the BIOS to turn it off. I then re-attached my Meridian Explorer 2 USB DAC and installed the driver for it (even though the manufacturer claims it works under Windows 10 without drivers, I've never been able to make that work - I always have to load the Windows 7 driver). I checked Device Manager, and the Device Properties dialogs in the Sound section of Settings to make sure everything was looking normal and set up correctly. Then I figured I'd better confirm the noise again before trying different things to troubleshoot it. And naturally, the noise is now GONE. I want to pound my head against the wall. All that troubleshooting yesterday was a waste of time. I never did find the cause of the problem, and I have no idea why it went away. Guess I won't be buying a new PC after all. Paul, thank you again for taking the time to provide all that wonderful advice for me. I really appreciate it. I'm sorry that it proved unnecessary in the end. Jerry So what you discovered was, the noise was indeed temporal servicing noise, not analog noise, because now the DAC is clear. And in future, you can test for "responsiveness" on the PC, as a root cause. Normally, this should never be a problem. The hardware interrupts from USB, should be serviced right away. 1) Stub service routine responds practically instantly. This is the driver in Ring0, down where the kernel lives. 2) Extended service (filling buffer for next time, preparing DMA chain or whatever), that service request is queued on the DPC queue. Delay procedure calls run in Ring3, like they were an application. 3) The OS monitors the DPC queue, and if the OS is falling behind, it's supposed to have an internal setting where more of the cores work on reducing the depth of the DPC queue. If all of that crap keeps your sound buffer full, there won't be any of "that kind" of noise. Paul |
P8Z68/GEN3 Deluxe Noisy USB Bus
On Sun, 02 Sep 2018 21:19:35 -0400, Paul
wrote: Tater wrote: On Sun, 02 Sep 2018 00:33:43 -0700, Tater wrote: On Sat, 01 Sep 2018 23:24:49 -0400, Paul wrote: Tater wrote: I was trying to listen to some MP3s on my PC via iTunes, and suddenly I was like "What the heck - this sounds like crap". I was getting random clicking and crackling noises. It basically sounded like I was playing back a badly worn LP. First I tried the same track several times in a row and confirmed that the clicks and pops were random - not in the same location in the track. I tried different tracks from other albums - same story. Then I turned to hardware troubleshooting - before I tell you about that let me describe the sound system on my computer: It's PC to USB DAC to tube preamp to amplified speakers. I am not using the onboard sound on the motherboard. The USB DAC (digital to analog converter) is a little gadget like a USB thumb drive that plugs into a USB port on the back of the PC and outputs an analog signal on a standard 1/8" phone jack. The tube preamp provides volume control (and that "good old tube sound"), and the speakers each have their own amplifiers built in. First thing I thought was I had a bad connection somewhere. I unplugged and reseated the DAC and all the analog connections. Didn't help. I ruled out a bad speaker because the noise was occurring in both channels. Next I pulled out an older DAC and swapped the DACs out. Didn't help. I moved the DAC to two other USB ports - didn't help. I moved the DAC to my add-in Inateck USB 3.0 board (I'm not using the onboard USB 3.0 on this motherboard). This seemed to make the noise even worse. To eliminate the tube preamp as a source of noise I disconnected the preamp and speakers, and plugged a set of headphones directly into the jack on the DAC. I Still heard the same noises via the headphones. So now I've isolated the noise all the way back to the PC itself - my troubleshooting has eliminated anything external to the PC as a possible source of noise. I suppose it might be a noisy power supply (and I do have a spare) but that's a lot of work to change out unless I truly have to. The only other thing I could think of trying at this point was to remove the DAC and activate the onboard sound chip on the motherboard, so that's what I did. Asus had a massive Windows 10 audio driver on their web site that came out two years ago - but I decided to let Windows try and install the sound device by itself, figuring I'd probably get a more up to date driver that way. It did find a driver for it - it's a pretty basic driver, but I don't need any fancy features. Pure unadulterated two channel sound is all I was looking for. So I hooked up my preamp to the onboard sound port and played an MP3. Sounds fine, zero noise. I'm not going to say it's as good as the USB DAC, but it sounds a heck of a lot better than I was expecting. This leaves questions though. If the power supply was noisy, shouldn't that noise have permeated the onboard sound system as well? What is the nature of this noise I'm hearing through the USB DAC? If there is random noise on my USB bus, I would think I'd be at risk for corrupted data on USB devices like my big backup drive or my USB thumb drives? Any advice is appreciated! Configuration: Asus P8Z68 Deluxe/GEN3 motherboard 4x4GB Corsair Vengeance LP DDR3 1600 RAM Intel i7-2700K CPU EVGA Supernova 850 T2 power suppy EVGA GTX 970 FTW+ 4GB video card SanDisk Extreme Pro 480GB SSD boot drive 2x1TB Western Digital Black data drives Western Digital 8TB My Book Duo USB backup drive (in Raid 1) Antec P280 case Noctua CPU cooler and 3 Noctua case fans Inateck KTU3FR-2O2I add-in USB 3.0 card Pioneer 16X Blu-ray burner Windows 10 Pro version 1803 Meridian Explorer 2 DAC ifi iTube preamp Paradigm Shift A2 speakers Thanks in advance. Jerry It sounds like you already have good debugging skills, so I won't need to add much. You can record the output of your DAC, using the onboard sound and the LineIn on it. That's one possibility. Using Audacity, you could record the output waveform, and compare it to your stimulus, and see how the noise modulates the waveform. Audacity doesn't support playback and recording at the same time. Which usually means I need two programs, one to playback a stimulus and a second program to record, to get the data I want. Sometimes clicks or pops, are actually flat spots caused on the output waveform, by "buffer runs dry". The hardware has to signal that the buffer is getting low, and the system coughs up a few packets of data to fill up some kind of queue. If the queue ever runs dry (because the buffer cannot be serviced in time), the last voltage level on the DAC "sits there" until a packet of new samples arrive. The flat spot makes an audible noise, although not as loud a sound as you would think. When digital samples are corrupted (like the MSBit changes state), that creates a huge pop. And you'd notice that. So while it's fun to pretend the source is analog, what you could be hearing is a "delivery system problem". If you play a sine wave on some LineOut, and then record on a LineIn, you may notice disturbances which are not part of a sine wave. And that will give you some idea what class of fault this is. ******* The onboard sound has a positive voltage for the digital portion of the chip. It also has a positive voltage for the analog portion. That analog power source, is a relatively small three terminal regulator. By using a separate regulated power source, the intention is to reduce the level of "digital noise" riding on the motherboard sound power rails. Or on any analog amp after that point. It's possible to get inductive coupling into the onboard sound. That happens when the NIC wires go too close to the LineOut wires. But that would not be a likely mechanism for an external USB DAC. And your motherboard sound is clean. So at this point, I recommend recording the defective output, so you can diagnose the fault type. ******* People who buy "audio workstation" motherboards, they usually run "DPCLat" or delay procedure call latency test. But it doesn't work on Windows 10. The time it takes to service a "delayed procedure call", is a function of the motherboard performance. If the BIOS uses the SMM (system management mode) interrupt mechanism, it usurps the OS completely. A USB DAC might be waiting for a millisecond for data, until the SMM routine exits and the OS is running again. When this happens, it ruins the motherboard with respect to running audio workstation functions. If you can find a replacement program for "DPCLat", one that works on Windows 10, you can check out your "responsiveness" and see whether real-time performance is compromised for some reason. When SMM fires off, the OS doesn't know what happened. All it could notice, is a high performance clock could be off by a millisecond. The OS has no other record that it's been swapped out. Programs like DPCLat, attempt to quantify this problem, by "indirect measurement". If the DPC Latency has spikes in it, that's generally a sign that SMM is active. (It's because we don't know of any other high-enough-priority activity that can do that to the OS. Intel made sure that SMM was king.) Note that there is one hardware event, which is absolutely huge. When your video card changes into 3D gaming mode, there is a 15 millisecond spike in the DPCLat trace. Which implies the video card driver is "stuck" and holding the kernel hostage for a frame time, while it switches modes. Normally, people doing audio workstation work, aren't screwing with 3D games in the middle of recording their rock group. So this isn't a problem. I've not done extensive testing of what happens to tunes, when the video card changes modes, so I don't know what that one sounds like. Paul Paul, Thanks for the incredibly thoughtful and highly technical response. One possibly important clue I forgot to mention earlier is that there is no noise coming through the DAC when there is no signal. That in itself might mean that it's not the USB bus itself that's noisy, but that the noise is being generated by the SATA controller when the MP3 file is being fetched for playback. I'm goiong to try other sources as a means of troubleshooting this. If it is the SATA bus, then I should also get the noise when playing back a CD (because the SATA controller is fetching the WAV file from the CD drive). However I would not expect to get the noise when playing Pandora, as that is purely network traffic and doesn't involve the SATA bus. I'll do further troubleshooting and report back on Sunday. Jerry I hate computers! I just now got around to messing with the PC again. I uninstalled the onboard sound via Device Manager, then rebooted and entered the BIOS to turn it off. I then re-attached my Meridian Explorer 2 USB DAC and installed the driver for it (even though the manufacturer claims it works under Windows 10 without drivers, I've never been able to make that work - I always have to load the Windows 7 driver). I checked Device Manager, and the Device Properties dialogs in the Sound section of Settings to make sure everything was looking normal and set up correctly. Then I figured I'd better confirm the noise again before trying different things to troubleshoot it. And naturally, the noise is now GONE. I want to pound my head against the wall. All that troubleshooting yesterday was a waste of time. I never did find the cause of the problem, and I have no idea why it went away. Guess I won't be buying a new PC after all. Paul, thank you again for taking the time to provide all that wonderful advice for me. I really appreciate it. I'm sorry that it proved unnecessary in the end. Jerry So what you discovered was, the noise was indeed temporal servicing noise, not analog noise, because now the DAC is clear. And in future, you can test for "responsiveness" on the PC, as a root cause. Normally, this should never be a problem. The hardware interrupts from USB, should be serviced right away. 1) Stub service routine responds practically instantly. This is the driver in Ring0, down where the kernel lives. 2) Extended service (filling buffer for next time, preparing DMA chain or whatever), that service request is queued on the DPC queue. Delay procedure calls run in Ring3, like they were an application. 3) The OS monitors the DPC queue, and if the OS is falling behind, it's supposed to have an internal setting where more of the cores work on reducing the depth of the DPC queue. If all of that crap keeps your sound buffer full, there won't be any of "that kind" of noise. Paul Paul, the Meridian Explorer 2 DAC driver installs a small "control panel" that allows you to tweak some settings for the DAC. I have left these at the defaults since I never understood what they did, but in light of my recent experience, do you think I should change any of the settings? The control panel settings a USB Streaming mode: Minimum latency Low Latency Standard Relaxed Reliable -- Default setting Safe Extra Safe ASIO Buffer Size: Auto -- Default setting 64 128 256 512 1024 2048 4096 8192 16384 Jerry |
P8Z68/GEN3 Deluxe Noisy USB Bus
Tater wrote:
On Sun, 02 Sep 2018 21:19:35 -0400, Paul wrote: Tater wrote: On Sun, 02 Sep 2018 00:33:43 -0700, Tater wrote: On Sat, 01 Sep 2018 23:24:49 -0400, Paul wrote: Tater wrote: I was trying to listen to some MP3s on my PC via iTunes, and suddenly I was like "What the heck - this sounds like crap". I was getting random clicking and crackling noises. It basically sounded like I was playing back a badly worn LP. First I tried the same track several times in a row and confirmed that the clicks and pops were random - not in the same location in the track. I tried different tracks from other albums - same story. Then I turned to hardware troubleshooting - before I tell you about that let me describe the sound system on my computer: It's PC to USB DAC to tube preamp to amplified speakers. I am not using the onboard sound on the motherboard. The USB DAC (digital to analog converter) is a little gadget like a USB thumb drive that plugs into a USB port on the back of the PC and outputs an analog signal on a standard 1/8" phone jack. The tube preamp provides volume control (and that "good old tube sound"), and the speakers each have their own amplifiers built in. First thing I thought was I had a bad connection somewhere. I unplugged and reseated the DAC and all the analog connections. Didn't help. I ruled out a bad speaker because the noise was occurring in both channels. Next I pulled out an older DAC and swapped the DACs out. Didn't help. I moved the DAC to two other USB ports - didn't help. I moved the DAC to my add-in Inateck USB 3.0 board (I'm not using the onboard USB 3.0 on this motherboard). This seemed to make the noise even worse. To eliminate the tube preamp as a source of noise I disconnected the preamp and speakers, and plugged a set of headphones directly into the jack on the DAC. I Still heard the same noises via the headphones. So now I've isolated the noise all the way back to the PC itself - my troubleshooting has eliminated anything external to the PC as a possible source of noise. I suppose it might be a noisy power supply (and I do have a spare) but that's a lot of work to change out unless I truly have to. The only other thing I could think of trying at this point was to remove the DAC and activate the onboard sound chip on the motherboard, so that's what I did. Asus had a massive Windows 10 audio driver on their web site that came out two years ago - but I decided to let Windows try and install the sound device by itself, figuring I'd probably get a more up to date driver that way. It did find a driver for it - it's a pretty basic driver, but I don't need any fancy features. Pure unadulterated two channel sound is all I was looking for. So I hooked up my preamp to the onboard sound port and played an MP3. Sounds fine, zero noise. I'm not going to say it's as good as the USB DAC, but it sounds a heck of a lot better than I was expecting. This leaves questions though. If the power supply was noisy, shouldn't that noise have permeated the onboard sound system as well? What is the nature of this noise I'm hearing through the USB DAC? If there is random noise on my USB bus, I would think I'd be at risk for corrupted data on USB devices like my big backup drive or my USB thumb drives? Any advice is appreciated! Configuration: Asus P8Z68 Deluxe/GEN3 motherboard 4x4GB Corsair Vengeance LP DDR3 1600 RAM Intel i7-2700K CPU EVGA Supernova 850 T2 power suppy EVGA GTX 970 FTW+ 4GB video card SanDisk Extreme Pro 480GB SSD boot drive 2x1TB Western Digital Black data drives Western Digital 8TB My Book Duo USB backup drive (in Raid 1) Antec P280 case Noctua CPU cooler and 3 Noctua case fans Inateck KTU3FR-2O2I add-in USB 3.0 card Pioneer 16X Blu-ray burner Windows 10 Pro version 1803 Meridian Explorer 2 DAC ifi iTube preamp Paradigm Shift A2 speakers Thanks in advance. Jerry It sounds like you already have good debugging skills, so I won't need to add much. You can record the output of your DAC, using the onboard sound and the LineIn on it. That's one possibility. Using Audacity, you could record the output waveform, and compare it to your stimulus, and see how the noise modulates the waveform. Audacity doesn't support playback and recording at the same time. Which usually means I need two programs, one to playback a stimulus and a second program to record, to get the data I want. Sometimes clicks or pops, are actually flat spots caused on the output waveform, by "buffer runs dry". The hardware has to signal that the buffer is getting low, and the system coughs up a few packets of data to fill up some kind of queue. If the queue ever runs dry (because the buffer cannot be serviced in time), the last voltage level on the DAC "sits there" until a packet of new samples arrive. The flat spot makes an audible noise, although not as loud a sound as you would think. When digital samples are corrupted (like the MSBit changes state), that creates a huge pop. And you'd notice that. So while it's fun to pretend the source is analog, what you could be hearing is a "delivery system problem". If you play a sine wave on some LineOut, and then record on a LineIn, you may notice disturbances which are not part of a sine wave. And that will give you some idea what class of fault this is. ******* The onboard sound has a positive voltage for the digital portion of the chip. It also has a positive voltage for the analog portion. That analog power source, is a relatively small three terminal regulator. By using a separate regulated power source, the intention is to reduce the level of "digital noise" riding on the motherboard sound power rails. Or on any analog amp after that point. It's possible to get inductive coupling into the onboard sound. That happens when the NIC wires go too close to the LineOut wires. But that would not be a likely mechanism for an external USB DAC. And your motherboard sound is clean. So at this point, I recommend recording the defective output, so you can diagnose the fault type. ******* People who buy "audio workstation" motherboards, they usually run "DPCLat" or delay procedure call latency test. But it doesn't work on Windows 10. The time it takes to service a "delayed procedure call", is a function of the motherboard performance. If the BIOS uses the SMM (system management mode) interrupt mechanism, it usurps the OS completely. A USB DAC might be waiting for a millisecond for data, until the SMM routine exits and the OS is running again. When this happens, it ruins the motherboard with respect to running audio workstation functions. If you can find a replacement program for "DPCLat", one that works on Windows 10, you can check out your "responsiveness" and see whether real-time performance is compromised for some reason. When SMM fires off, the OS doesn't know what happened. All it could notice, is a high performance clock could be off by a millisecond. The OS has no other record that it's been swapped out. Programs like DPCLat, attempt to quantify this problem, by "indirect measurement". If the DPC Latency has spikes in it, that's generally a sign that SMM is active. (It's because we don't know of any other high-enough-priority activity that can do that to the OS. Intel made sure that SMM was king.) Note that there is one hardware event, which is absolutely huge. When your video card changes into 3D gaming mode, there is a 15 millisecond spike in the DPCLat trace. Which implies the video card driver is "stuck" and holding the kernel hostage for a frame time, while it switches modes. Normally, people doing audio workstation work, aren't screwing with 3D games in the middle of recording their rock group. So this isn't a problem. I've not done extensive testing of what happens to tunes, when the video card changes modes, so I don't know what that one sounds like. Paul Paul, Thanks for the incredibly thoughtful and highly technical response. One possibly important clue I forgot to mention earlier is that there is no noise coming through the DAC when there is no signal. That in itself might mean that it's not the USB bus itself that's noisy, but that the noise is being generated by the SATA controller when the MP3 file is being fetched for playback. I'm goiong to try other sources as a means of troubleshooting this. If it is the SATA bus, then I should also get the noise when playing back a CD (because the SATA controller is fetching the WAV file from the CD drive). However I would not expect to get the noise when playing Pandora, as that is purely network traffic and doesn't involve the SATA bus. I'll do further troubleshooting and report back on Sunday. Jerry I hate computers! I just now got around to messing with the PC again. I uninstalled the onboard sound via Device Manager, then rebooted and entered the BIOS to turn it off. I then re-attached my Meridian Explorer 2 USB DAC and installed the driver for it (even though the manufacturer claims it works under Windows 10 without drivers, I've never been able to make that work - I always have to load the Windows 7 driver). I checked Device Manager, and the Device Properties dialogs in the Sound section of Settings to make sure everything was looking normal and set up correctly. Then I figured I'd better confirm the noise again before trying different things to troubleshoot it. And naturally, the noise is now GONE. I want to pound my head against the wall. All that troubleshooting yesterday was a waste of time. I never did find the cause of the problem, and I have no idea why it went away. Guess I won't be buying a new PC after all. Paul, thank you again for taking the time to provide all that wonderful advice for me. I really appreciate it. I'm sorry that it proved unnecessary in the end. Jerry So what you discovered was, the noise was indeed temporal servicing noise, not analog noise, because now the DAC is clear. And in future, you can test for "responsiveness" on the PC, as a root cause. Normally, this should never be a problem. The hardware interrupts from USB, should be serviced right away. 1) Stub service routine responds practically instantly. This is the driver in Ring0, down where the kernel lives. 2) Extended service (filling buffer for next time, preparing DMA chain or whatever), that service request is queued on the DPC queue. Delay procedure calls run in Ring3, like they were an application. 3) The OS monitors the DPC queue, and if the OS is falling behind, it's supposed to have an internal setting where more of the cores work on reducing the depth of the DPC queue. If all of that crap keeps your sound buffer full, there won't be any of "that kind" of noise. Paul Paul, the Meridian Explorer 2 DAC driver installs a small "control panel" that allows you to tweak some settings for the DAC. I have left these at the defaults since I never understood what they did, but in light of my recent experience, do you think I should change any of the settings? The control panel settings a USB Streaming mode: Minimum latency Low Latency Standard Relaxed Reliable -- Default setting Safe Extra Safe ASIO Buffer Size: Auto -- Default setting 64 128 256 512 1024 2048 4096 8192 16384 Jerry Generally, you run an ASIO driver, for minimum latency. And that means selecting a smaller buffer size, such that timing-wise, we know the USB can refill the buffer before it is too late. If, for some reason, that wasn't important (the low latency) you could play with the buffer size. Perhaps the side effect, if it was malfunctioning, is the frequency of the noise events would change with buffer size. Sound doesn't need a lot of bandwidth. The OS can service at least 10,000 interrupts a second. The only way this can break, is if something manages to "hog" resources. For example, in the old days, I could use a "critical section" around some low level code I've written. Which locks out the OS for a period of time. A person using such a technique, asks the system architect, what the max time for a critical section is, without breaking stuff. And then that person works on their code, polishing it, so they stay well away from the breakage point. That doesn't explain how the video card gets away with blanking the system for 16 milliseconds when it switches to 3D mode. But then, we don't have high expectations on desktop performance while doing that anyway. I don't think I've heard anyone actually complain about that (other than to "notice" it in the DPCLat graph). Paul |
All times are GMT +1. The time now is 01:09 PM. |
Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
HardwareBanter.com