I would try to reinstall the system, reset the BIOS settings and try everything stock. Don't plug anything in and don't install any bloatware. Then check the internal motherboard audio for clicks and only then add the external audio card.
Messing us with interrupt priorities and c-states obviously didn't help.
Changing MB from ASUS to Gigabyte - what can I expect in reinstall etc?
-
- KVRAF
- Topic Starter
- 6426 posts since 22 Jan, 2005 from Sweden
It might come to that, but still have some new discoveries.
Thanks for suggestions though.
Also some test suggested by Pictus I will have a closer look on, I already do backups with Macrium Reflect to restore after a reinstall if no difference.
Any usb ports on atx panel have been the noisiest.
Chassis ports with a 40 cm usb header cable at back of mb clearly less noise than atx panel.
Now I also tried a 25 cm usb header on short end of card, and this is the best so far.
- 4 usb 2 ports out on a card slot on back of case
Remember talking to RME rather early on if drivers limit transfers to just channels used on interface.
- no need they said, usb 2.0 handles this with no problem
This is 2 x 32 channels 48k 24 bits which would mean all the time sending 46 Mbps in two directions over usb all the time.
Cable shipped is 1.5m length, so ordered two high speed certified usb 2.0 cable just 0.5m and 1m and will see if any difference.
What happend when using short end usb header 25 cm was that not single thing heard for over 10 minutes. I maybe thought this is fixed now, but then small faint ticks started to appear. Like mb is warming up. Computer have been runnings all day, at least 10 hours so everything else should be luke warm.
Interface draw 250 mA 5V, and this 2x46 Mbps per second is maybe heating mb a bit or similar. Don't know why the delay until ticks appeared otherwise. This side of mb was pretty much no activity until I connected interface there. Just 3 midi interfaces, with no gear running mostly.
I did check once again on atx panel ports and those are real noisy. Network card is not connected, and just mouse and keyboard and not much activity. But a lot of connections on confined space on mb around this atx panel. The closer pcb copper lines over a distance, the higher capacitance.
As far as I could find out there is only one Intel usb 3 controller that runs if it is 16 ports or so, whether usb 3 or 2. 10 ports on atx panel, 6 through header at end of mb.
Only card slot used is graphics card, and nothing on atx panel for graphics used.
It was a comment from Asus support a couple of months ago, they contacted Taiwan on my case, and they listened to ticks then much worse than those above, and just said this is no inductance problem and said to switch cables or output.
Summing up the various tests I will see what better and shorter cables may do. It's so close to extinct now that is may work. It's already one notch better than the other day.
Just theories of course and trying to understand why some things make a difference altering this and that.
And still not all sure it's not a missing ground somewhere that is needed. I think how it acts up coming and going is a bit like that. But touched metal on various places to hear if any small ticks from that like ground issue might create. But nothing. External gear is all to a preamp rack which is isolated with optical cable from computer itself.
Anyway, thanks for input so far.
Thanks for suggestions though.
Also some test suggested by Pictus I will have a closer look on, I already do backups with Macrium Reflect to restore after a reinstall if no difference.
Any usb ports on atx panel have been the noisiest.
Chassis ports with a 40 cm usb header cable at back of mb clearly less noise than atx panel.
Now I also tried a 25 cm usb header on short end of card, and this is the best so far.
- 4 usb 2 ports out on a card slot on back of case
Remember talking to RME rather early on if drivers limit transfers to just channels used on interface.
- no need they said, usb 2.0 handles this with no problem
This is 2 x 32 channels 48k 24 bits which would mean all the time sending 46 Mbps in two directions over usb all the time.
Cable shipped is 1.5m length, so ordered two high speed certified usb 2.0 cable just 0.5m and 1m and will see if any difference.
What happend when using short end usb header 25 cm was that not single thing heard for over 10 minutes. I maybe thought this is fixed now, but then small faint ticks started to appear. Like mb is warming up. Computer have been runnings all day, at least 10 hours so everything else should be luke warm.
Interface draw 250 mA 5V, and this 2x46 Mbps per second is maybe heating mb a bit or similar. Don't know why the delay until ticks appeared otherwise. This side of mb was pretty much no activity until I connected interface there. Just 3 midi interfaces, with no gear running mostly.
I did check once again on atx panel ports and those are real noisy. Network card is not connected, and just mouse and keyboard and not much activity. But a lot of connections on confined space on mb around this atx panel. The closer pcb copper lines over a distance, the higher capacitance.
As far as I could find out there is only one Intel usb 3 controller that runs if it is 16 ports or so, whether usb 3 or 2. 10 ports on atx panel, 6 through header at end of mb.
Only card slot used is graphics card, and nothing on atx panel for graphics used.
It was a comment from Asus support a couple of months ago, they contacted Taiwan on my case, and they listened to ticks then much worse than those above, and just said this is no inductance problem and said to switch cables or output.
Summing up the various tests I will see what better and shorter cables may do. It's so close to extinct now that is may work. It's already one notch better than the other day.
Just theories of course and trying to understand why some things make a difference altering this and that.
And still not all sure it's not a missing ground somewhere that is needed. I think how it acts up coming and going is a bit like that. But touched metal on various places to hear if any small ticks from that like ground issue might create. But nothing. External gear is all to a preamp rack which is isolated with optical cable from computer itself.
Anyway, thanks for input so far.
-
- KVRAF
- Topic Starter
- 6426 posts since 22 Jan, 2005 from Sweden
So this is finally solved, the ticks thing.
I came up with the idea as I was having a usb 3 pci express card in my former daw to move that to new computer and see what happends.
This was the solution, audio ticks went away using that for audio interface, those that were trashing audio and you could see on my pics.
This card is based on a NEC chip, and creates a usb host controller and a usb root hub. Renasas Electronics drivers.
So windows runs both these controllers, and figure something needs update on Intel usb controller or the chipset or something. System is all stable, and usb sense that audio device and can handle that differently. Don't know, but seems this way to me.
Then the final mystery with the very faint ticks seems to be solved by turning off my air cleaner which as I discovered pushed out ionized air into the room. These ticks were strange in that I could not record them, they just appeared on phones and not in audio stream itself. Some static thing as said by posts in this thread as well.
But some windows and bios settings helped also
- trying to get a stable frequency on cpu no jumping up and down
- turning off E-cores help with that
- turning off C-states helped with that
- multicore enhancement in bios also helped doing all cores follow same frequency
So taskmanager say constantly 4.5 GHz running now.
Seen nothing alarming with temperatures, cpu 41 C, PCH 46 C, VRM MOS 46 C.
My largest audio project make about 16% utilization on 14 cores. Those in progress about 6 cores and 9% utilization.
With audio perfect I will see how main reason for new computer with multi camera editing works. But there is plenty headroom for that I think. My old computer benchmark at 5000 and new on 29 000 with E-cores off(cpubenchmark.net running their test).
I came up with the idea as I was having a usb 3 pci express card in my former daw to move that to new computer and see what happends.
This was the solution, audio ticks went away using that for audio interface, those that were trashing audio and you could see on my pics.
This card is based on a NEC chip, and creates a usb host controller and a usb root hub. Renasas Electronics drivers.
So windows runs both these controllers, and figure something needs update on Intel usb controller or the chipset or something. System is all stable, and usb sense that audio device and can handle that differently. Don't know, but seems this way to me.
Then the final mystery with the very faint ticks seems to be solved by turning off my air cleaner which as I discovered pushed out ionized air into the room. These ticks were strange in that I could not record them, they just appeared on phones and not in audio stream itself. Some static thing as said by posts in this thread as well.
But some windows and bios settings helped also
- trying to get a stable frequency on cpu no jumping up and down
- turning off E-cores help with that
- turning off C-states helped with that
- multicore enhancement in bios also helped doing all cores follow same frequency
So taskmanager say constantly 4.5 GHz running now.
Seen nothing alarming with temperatures, cpu 41 C, PCH 46 C, VRM MOS 46 C.
My largest audio project make about 16% utilization on 14 cores. Those in progress about 6 cores and 9% utilization.
With audio perfect I will see how main reason for new computer with multi camera editing works. But there is plenty headroom for that I think. My old computer benchmark at 5000 and new on 29 000 with E-cores off(cpubenchmark.net running their test).
-
- KVRAF
- 7400 posts since 17 Feb, 2005
You copied settings and things from the old computer with this USB card into the new? Also, have you tested the card with other USB periphials connected as well as sound interface?
-
- KVRAF
- Topic Starter
- 6426 posts since 22 Jan, 2005 from Sweden
No, no copying anything software, just move card and just install Renasas electronics driver from a mini cd that came with card. Driver is from 2013 I think, so still latest what I could find. I think card brand was CSL and their site had drivers too, but seems no newer.
This card only give 2 usb ports, so am not to bother audio with anything else.
But everything else is connected to Intel controller as it was before. Just Digiface USB connector to this new card.
Just so happy to be home with this solution.
I let three midi interfaces and the rest input devices be on Intel usb controller on mb.
Nothing from old computer os disk is transferred to new computer.
This card only give 2 usb ports, so am not to bother audio with anything else.
But everything else is connected to Intel controller as it was before. Just Digiface USB connector to this new card.
Just so happy to be home with this solution.
I let three midi interfaces and the rest input devices be on Intel usb controller on mb.
Nothing from old computer os disk is transferred to new computer.
-
- KVRian
- 937 posts since 21 Aug, 2017 from Brasil
Firmware Backdoor Discovered in Gigabyte Motherboards, 250+ Models Affected
https://www.tomshardware.com/news/gigab ... e-backdoor
Supply Chain Risk from Gigabyte App Center Backdoor
https://eclypsium.com/blog/supply-chain ... -backdoor/
-
- KVRAF
- Topic Starter
- 6426 posts since 22 Jan, 2005 from Sweden
Thanks for heads up.
I actually uninstalled all of that AppCenter, turned off that automatic download in bios and that stuff. Had issue for a good while with some GbtMaxCloud.exe that got messages in event log not finding it, since uninstalled of course, and I did not see where it was referenced. So probably this bios thing and it went away.
I was impressed at first, but not consistent how software interacted with fans vs bios etc. I could suddenly high speed on fans for no reason.
Intel had som crap too, that never worked "something went wrong". So turned off Intel services I thought had this function for updates.
Bios settings were actually quite neat. I made a manual curve, and picked which sensor to use for temp and which type of control voltage/pwm etc.
I have three system fans chained on one control, so maybe that disturbed how it worked from start. Three fans react to same voltage. And not sure max rpm on my fans was on par with what settings in software were etc. All temps seems fine(entering bios).
I have not run CineBench and that to super stress everything, just glad all is fine with audio.
Won't even breath on this computer unless windows update do something not so good. It's not online anyway on daily basis.
I actually uninstalled all of that AppCenter, turned off that automatic download in bios and that stuff. Had issue for a good while with some GbtMaxCloud.exe that got messages in event log not finding it, since uninstalled of course, and I did not see where it was referenced. So probably this bios thing and it went away.
I was impressed at first, but not consistent how software interacted with fans vs bios etc. I could suddenly high speed on fans for no reason.
Intel had som crap too, that never worked "something went wrong". So turned off Intel services I thought had this function for updates.
Bios settings were actually quite neat. I made a manual curve, and picked which sensor to use for temp and which type of control voltage/pwm etc.
I have three system fans chained on one control, so maybe that disturbed how it worked from start. Three fans react to same voltage. And not sure max rpm on my fans was on par with what settings in software were etc. All temps seems fine(entering bios).
I have not run CineBench and that to super stress everything, just glad all is fine with audio.
Won't even breath on this computer unless windows update do something not so good. It's not online anyway on daily basis.
