CPU temp 99C when connecting USB 2.0

So i just notice that my max temp in cinebench doing de CPU test is 76C with the USB 2.0 unplugged from the AIO H100IV2... when i connect it my temps go up to 92c on the package and even getting 99C... idk what is going wrong with it...

 

i just change the thermalpaste and the issue keeps going... i really dont feel the radiator hot to be on "99c"... only the tubes are warm... help pls

 

my cpu is a 7820x on a tuf mark 2, when i just get the water cooling everything was cool with the usb 2.0...

 

can someone explain me what does the usb2.0 connector on the pump does apart from the rgb lights?, thx

The USB cable is the communication line between the cooler -> motherboard -> software. It is necessary for you to have desktop control over the pump and fans, when they are connected to the H100i v2's internal fan controller.

 

Best guess is the H100i v2 is defaulting to a some very relaxed profile when you connect. Do you have Link or iCUE installed? Even if the above is true, a +16C swing is a bit much. It would be good to see the coolant temperature from Link/iCUE.

The USB cable is the communication line between the cooler -> motherboard -> software. It is necessary for you to have desktop control over the pump and fans, when they are connected to the H100i v2's internal fan controller.

 

Best guess is the H100i v2 is defaulting to a some very relaxed profile when you connect. Do you have Link or iCUE installed? Even if the above is true, a +16C swing is a bit much. It would be good to see the coolant temperature from Link/iCUE.

 

i have Link installed since i bought the AIO in january, and even with a agressive fan curve and in performance mode the 99c is still present... without the 2.0 usb the fans are at lower speeds than the maximum 2700rpm that i chose when reaching 99C

 

about the water temp its around 47-49 C at the "99c" temps... in theory without the usb 2.0 the AIO wont be recognize and will be just like a basic cooler right?

 

something i havent try is connecting the pump to the AIO PUMP, and the two fans to the CPU and CPU OPT, but i really dont think its a performance issue (considering a 76max temp good) i think that its a reading temp problem or something like that

 

btw i have not updated my bios, but as far as i know that shouldnt matter...

 

 

thx for replying, and sorry for my bad english

Your coolant temp is pretty high, although theoretically possible in a warm environment/case temp plus 8 cores at full load. On a H100i v2, I would expect a +8C or so coolant rise at full load. When you are at cool idle, what is the H100i v2 Temp? Normally, most people are about +4-7C over their room temp, although it can't be lower than the interior case temperature.

 

Without the USB cable, it will run its saved setting or some default profile in the firmware. I am not sure we really know what that is, but small changes in fan speed should not have dramatic effects on end temperatures. I think we should investigate this as a cooler issue before trying to learn the hidden default behavior.

 

What is the reported pump speed in Link? The Quiet setting should be just below 2000 rpm. The Performance closer to 3000 rpm. Make sure the motherboard header powering the H100i v2 is 'disabled' or set to 100% in some way. If not, you are cutting voltage to the unit and that affects all aspects of performance. It can be easy to forget to reset it after a BIOS update.

Your coolant temp is pretty high, although theoretically possible in a warm environment/case temp plus 8 cores at full load. On a H100i v2, I would expect a +8C or so coolant rise at full load. When you are at cool idle, what is the H100i v2 Temp? Normally, most people are about +4-7C over their room temp, although it can't be lower than the interior case temperature.

 

Without the USB cable, it will run its saved setting or some default profile in the firmware. I am not sure we really know what that is, but small changes in fan speed should not have dramatic effects on end temperatures. I think we should investigate this as a cooler issue before trying to learn the hidden default behavior.

 

What is the reported pump speed in Link? The Quiet setting should be just below 2000 rpm. The Performance closer to 3000 rpm. Make sure the motherboard header powering the H100i v2 is 'disabled' or set to 100% in some way. If not, you are cutting voltage to the unit and that affects all aspects of performance. It can be easy to forget to reset it after a BIOS update.

 

 

hi, the think you tell me about the header powering i did it a couple of hours ago setting the CPU FAN to DC at Max Speeds and it didnt work so i disable it... here ist a Screenshot of the corsair link at load, as you can see the temps are around 87C, here is the funny think, this is and older version of corsair link that dont even recognize my RGB ram but the cpu appears to be a bit more cooler...

 

 

about my water temps on idle when i just turn the PC are around 33C, my case its carbide 400c and just have 4 fans (2 cha fan and the fans that came with the AIO)

 

thx again for replying

OK, so 33 to 43C on load. That is about where I was expecting it to be. The other thing to watch is does the coolant then come back down after you quit. Normally it will drop about 1C/minute until you get down to the last few degrees above the starting point. Those take a bit longer, sometimes 20 minutes plus to come to a full baseline again. If it sticks at 40C plus, then the cooler has an issue or the entire case is now the same temperature. See next paragraph.

 

The one thing I notice in your screen shot is the GPU and drives are also about the same temperature as the coolant. This could be mere coincidence, but obviously raising the internal temperature of the case has an add-on effect to every component inside, including coolant and CPU temp. Just so I understand the layout, front intake in the H100i v2 and its fans, then one exhaust fans in the top and rear of the case? Or two up top?

 

The H100i v2 power supply is fine and "disabled" is the best setting when available. Regardless, I can see the pump and fans are getting full power so I know it is set up correctly.

OK, so 33 to 43C on load. That is about where I was expecting it to be. The other thing to watch is does the coolant then come back down after you quit. Normally it will drop about 1C/minute until you get down to the last few degrees above the starting point. Those take a bit longer, sometimes 20 minutes plus to come to a full baseline again. If it sticks at 40C plus, then the cooler has an issue or the entire case is now the same temperature. See next paragraph.

 

The one thing I notice in your screen shot is the GPU and drives are also about the same temperature as the coolant. This could be mere coincidence, but obviously raising the internal temperature of the case has an add-on effect to every component inside, including coolant and CPU temp. Just so I understand the layout, front intake in the H100i v2 and its fans, then one exhaust fans in the top and rear of the case? Or two up top?

 

The H100i v2 power supply is fine and "disabled" is the best setting when available. Regardless, I can see the pump and fans are getting full power so I know it is set up correctly.

 

yeap its one rear and one top, and the radiator in the front. right now im not getting near to 99 as i use to get with the newest version of Link, the package is around 87C, fans at 1380, i enable again the DC on the CPU Fan on the BIOS and still dont see any difference...

 

Really looks for me like a wierd compability issue with newest version Link doing an Offset of temperature... Should i update BIOS? or just unplug the USB 2.0 and control de fans using Aisuite or something like that? (btw i dont have Aisuite installed because i know there problems btw link and Aisuite)

 

the 76 max without the 2.0 connector should be the closest to my real CPU Temp cause i have touch my components, the left panel of the case, and its warm but not extremely hot to have a 87C CPU... my old CPU (4770 with a water 3.0) used to get to 70s-80s and i still remember how hot it was the case.. any other thing that i can try?

 

 

 

thx

 

 

UPDATE:

 

i just put the performance mode again and the temps have decrease to 69-71C (pump at 2900 and fans at 2700, water temp 42) look like the DC actually work but just when i use the performance mode... btw its really loud xd

UPDATE:

 

i just put the performance mode again and the temps have decrease to 69-71C (pump at 2900 and fans at 2700, water temp 42) look like the DC actually work but just when i use the performance mode... btw its really loud xd

 

 

OK, just getting to that. The thing is you should not see a 16C swing between a low and high fan speed on any cooler. That makes me wonder about pump speed, something that should not matter unless there is a restriction inside the cooler.

 

Try this: Let the coolant temp come back to a steady baseline, then run the test again with Link running, USB in. Set the fans on your normal test speed (1300 or whatever that is). However, click on the pump speed and change that to Performance. It should increase to ~2900 RPM. Run the same test. How does that compare? Another thing you can do it toggle the low/high speed setting on the pump in the middle of the test. Does the coolant temp jump one way or the other? There should be no change and it would take several minutes to show a measurable difference, if any.

 

You can also run this test on the desktop at idle with no load. Set the fans to a medium speed (running them super low can show a false positive), then turn up the pump speed. There should be no change. Maybe 0.5C under certain idle conditions with low speed. Again, if you see the temp drop a huge chunk in one moment, something is wrong.

OK, just getting to that. The thing is you should not see a 16C swing between a low and high fan speed on any cooler. That makes me wonder about pump speed, something that should not matter unless there is a restriction inside the cooler.

 

Try this: Let the coolant temp come back to a steady baseline, then run the test again with Link running, USB in. Set the fans on your normal test speed (1300 or whatever that is). However, click on the pump speed and change that to Performance. It should increase to ~2900 RPM. Run the same test. How does that compare? Another thing you can do it toggle the low/high speed setting on the pump in the middle of the test. Does the coolant temp jump one way or the other? There should be no change and it would take several minutes to show a measurable difference, if any.

 

You can also run this test on the desktop at idle with no load. Set the fans to a medium speed (running them super low can show a false positive), then turn up the pump speed. There should be no change. Maybe 0.5C under certain idle conditions with low speed. Again, if you see the temp drop a huge chunk in one moment, something is wrong.

 

hi, so the temps on idle just decrease around 1-2C when i set the pump to its maximum

 

if start the render with the pump on 1900, fans 2000rpm and change it in the middle to performance(the water temp, fans still 2000) the temps decrease to froms 82 to 70 even getting 66C sometimes...

 

about the water temps, when having the pump at maximum the water its around 37C and going up but really slow.. update: now its around 38.4C

 

thx for replying

It's starting to look like you have a partial blockage developing that is inhibiting your performance. With the USB unplugged and the cooler seeing a high coolant temp value, it was kicking into the highest pump speed mode. Normally, the pump speed should make little difference. When you see a big drop by kicking it into high, that probably means a restriction is really hampering the flow rate at the lower speed. However, if this is true, you really should see it as a coolant temperature change as well. You have to be careful using CPU package and core temp data only, as those change change so fast and for so many reasons. Regardless, based on the evidence I think you should contact Corsair Tech Support to see about replacing the cooler under the warranty terms. You are probably going to have to do some explaining. Focus on the fact the CPU temp takes a big drop when you toggle it on during load. There are possible alternate explanations for most of the other things, but that one really is the telling mark.

It's starting to look like you have a partial blockage developing that is inhibiting your performance. With the USB unplugged and the cooler seeing a high coolant temp value, it was kicking into the highest pump speed mode. Normally, the pump speed should make little difference. When you see a big drop by kicking it into high, that probably means a restriction is really hampering the flow rate at the lower speed. However, if this is true, you really should see it as a coolant temperature change as well. You have to be careful using CPU package and core temp data only, as those change change so fast and for so many reasons. Regardless, based on the evidence I think you should contact Corsair Tech Support to see about replacing the cooler under the warranty terms. You are probably going to have to do some explaining. Focus on the fact the CPU temp takes a big drop when you toggle it on during load. There are possible alternate explanations for most of the other things, but that one really is the telling mark.

 

ok... i will try yo do a RMA, thanks for all your help.