Kraken x62 water temps on Ryzen high...

[LordGabeNisBae]

Hello peeps,

 

I have a ryzen 7 1700 oc'ed to 3.8Ghz with 1.3125V vcore ans 1.2V SOC and running 2933 mhz memory. I have seen peeps run higher voltages for higher clocks on the same chip online (ryzen is supposed to run cool) but I am getting 70 C load temps running Aida 64 stress test. This makes no sense to me. People are getting this kind of temps with the included wraith cooler... not a $200 high end AIO like the Kraken x62. 

Specs as currently setup:

 

Ryzen 7 1700 @ 3.8Ghz 1.3125 Vcore

Asrock x370 taichi

64gb G-skill trident Z 3200 @ 2933 

Corsair HX850i psu
GTX 1050ti exoc @2ghz/8ghz

Case: NZXT S340 white with tempered glass.
Fans: All fans are Noctua NF-A14 + NF-F12 with all fan slots populated

I have done 2 repastes and 5 remounts of my cooler with no major change. I am using Arctic cooling MX-4 thermal compound. I know I live in Australia and the ambiant temps are around 35 C but a delta T over ambient of 35 C with liquid cooling seems off. My highly overclocked 1050ti runs around 66C which is aircooled. Pls halp. 

 

[BubblyCharizard]

70C is not an issue especially on a chip with a TJmax of~95C

[LordGabeNisBae]

2 minutes ago, BubblyCharizard said:

70C is not an issue especially on a chip with a TJmax of~95C

I am aware of that. But the liquid cooled temps are off. they are similar to the Wraith cooler i tested it with before install.
Isnt liquid cooling ,albeit with a lowly aio, supposed to make it a lot better?

[spat55]

Are you running the fans at max speed?

[LordGabeNisBae]

42 minutes ago, spat55 said:

Are you running the fans at max speed?

fans are on performance setting in the bios which pegs them to around 60% and up as temps rise. Bottom line: the fans are audible at my chair which is about 2-3 meters away from the system

[For Science!]

1 hour ago, LordGabeNisBae said:

fans are on performance setting in the bios which pegs them to around 60% and up as temps rise. Bottom line: the fans are audible at my chair which is about 2-3 meters away from the system

I think your expectations are just off the chart. 35 degree delta is exactly the performance of the x62.

 

[SpaceGhostC2C]

17 minutes ago, For Science! said:

I think your expectations are just off the chart. 35 degree delta is exactly the performance of the x62.

 

The performance of the x62 for which CPU, at which voltage? You cannot characterize the performance of a CPU cooler by one delta T. For low TDPs, idle CPUs, etc, all those delta Ts would converge to some low number. For a 7980x overclocked to its limits, all those delta Ts would go through the roof, many even fail to complete the test.

Maybe the original source was using a Ryzen 1700, but that's missing from the chart you pasted.

 

OP: your title speaks of water temps, but you never reported any. What are the water temps in the x62?

[SpaceGhostC2C]

Also:

3 hours ago, LordGabeNisBae said:

 but a delta T over ambient of 35 C with liquid cooling seems off.

There's no reason why it shouldn't be 35, nor lower, nor higher. Being air or water cooled doesn't restrict the delta T to some magic number.

 

3 hours ago, LordGabeNisBae said:

My highly overclocked 1050ti runs around 66C which is aircooled. Pls halp. 

That's not informative. GPUs have different die sizes than CPUs, plus direct on-die heatsinks. The temperatures on your GPU don't tell you anything about what to expect on the CPU (other than in extreme cases of terrible case air flow, and still they'll be just loosely correlated).

 

3 hours ago, LordGabeNisBae said:

People are getting this kind of temps with the included wraith cooler...

Problem is, who know what settings "people" are using, how they are testing, etc.

Much simpler test: what temps do you get if you keep your settings but use the stock cooler instead?

[LordGabeNisBae]

35 minutes ago, SpaceGhostC2C said:

The performance of the x62 for which CPU, at which voltage? You cannot characterize the performance of a CPU cooler by one delta T. For low TDPs, idle CPUs, etc, all those delta Ts would converge to some low number. For a 7980x overclocked to its limits, all those delta Ts would go through the roof, many even fail to complete the test.

Maybe the original source was using a Ryzen 1700, but that's missing from the chart you pasted.

 

OP: your title speaks of water temps, but you never reported any. What are the water temps in the x62?

I agree with the Delta T argument. For limited rads like the ones on most off the shelf AIOs, you will eventually be subjected to heatsoak when using obscenely hot CPUs like the 7980XE... Ryzen is not supposed to be anywhere near that hot. HWinfo indicates a max power draw of 120W by the overclocked chip with 8 cores @ 3.8Ghz (SMT: on). Thats nowhere near the power consumption of my old Haswell-E 5820K whick used to hit 160-180W easily and stayed in the 50s with a corsair H100... a clearly inferior cooler to the x62.

Gamers nexus said that the standard water temps for the Asetec based kraken coolers should be around 60 C operating. the pipes coming off the CPU block were ridiculously hot though... I am not taking the thing apart as I value my warranty and the thing cost and eyewatering 240 bux   perhaps I have a pump failure? of maybe this is what it is with AIOs... Coming from custom loops, the temps of the outlet tube of the cpu block was toasty. Maybe I was expecting too much of this cooler and should go back to custom loops. All these reviewers saying the performance between custom loops and AIOs have converged...  

 

 

39 minutes ago, SpaceGhostC2C said:

Also:

There's no reason why it shouldn't be 35, nor lower, nor higher. Being air or water cooled doesn't restrict the delta T to some magic number.

 

That's not informative. GPUs have different die sizes than CPUs, plus direct on-die heatsinks. The temperatures on your GPU don't tell you anything about what to expect on the CPU (other than in extreme cases of terrible case air flow, and still they'll be just loosely correlated).

 

Problem is, who know what settings "people" are using, how they are testing, etc.

Much simpler test: what temps do you get if you keep your settings but use the stock cooler instead?

1) There is a reason. The reason is surface area. The wraith cooler included with the chip has a much smaller surface area than the 2x140mm radiator of the x62. I have measured a max power draw of 120W which is peanuts for a rad this size... or so I beleive. So I beleive that the delta T should be lower for an objectively better cooler?

2) The point I was making there is that the case flow is not an issue so peeps dont assume the rad is getting heatsoaked due to having overly low spinning fans. 

3) I use the Aida64 stress test stresing the CPU, Memory and cache. Maybe the memory stress test is stressing out the SOC and generating extra heat? 


Either way, I was just intrigued by the ridiculous temperatures that the cooler was acheiving on what could be described as 'conservative' voltages... I realised that my motherboard bios was ancient and flashed it to the latest 3.20 iteration by asrock. As a result, I have been able to drop the voltages down a bit for what seems to be stable overclocks of 3.8Ghz @1.225V Vcore + 1.125V VSOC and tightened the timings of my memory to 14-14-14-34 @3200Mhz. Temps are down 10C but it is still far from ideal... 
 

[SpaceGhostC2C]

2 minutes ago, LordGabeNisBae said:

Gamers nexus said that the standard water temps for the Asetec based kraken coolers should be around 60 C operating. the pipes coming off the CPU block were ridiculously hot though... I am not taking the thing apart as I value my warranty and the thing cost and eyewatering 240 bux   perhaps I have a pump failure? of maybe this is what it is with AIOs... Coming from custom loops, the temps of the outlet tube of the cpu block was toasty.

I think it woudl be worse with a dead pump, but in any case you can check readings for pump rpm and, perhaps, temperature: the x60 had a USB header connector, so you could monitor liquid temps and pump rpm on NZXT software. That's whay I was asking about.

 

2 minutes ago, LordGabeNisBae said:

 

1) There is a reason. The reason is surface area. The wraith cooler included with the chip has a much smaller surface area than the 2x140mm radiator of the x62. I have measured a max power draw of 120W which is peanuts for a rad this size... or so I beleive. So I beleive that the delta T should be lower for an objectively better cooler?

But that's not what you said. No, there is no reason to expect the delta T (that is, the difference between ambient temps and CPU temps) to be any specific temperature. I was not talking abotu the difference between the stock cooler and the radiator, but the difference between ambient and CPU. It can be anything, case-specific.

Yes, you would expect better cooling from the AIO, but you haven't told us what your temps are if you change nothing but the cooler.

 

 

2 minutes ago, LordGabeNisBae said:

3) I use the Aida64 stress test stresing the CPU, Memory and cache. Maybe the memory stress test is stressing out the SOC and generating extra heat? 
 

It could be, I mean, stress tests are supposed to be stressful  At 70C your cooler seems to be handling your OC just fine, though, so your concern seems to be only in relative terms to other solutions, but for that we need to know what temps would look like under other solutions.

 

2 minutes ago, LordGabeNisBae said:

Either way, I was just intrigued by the ridiculous temperatures that the cooler was acheiving on what could be described as 'conservative' voltages...

I don't think 20C away from the CPU's max temps is ridiculous, though.

 

2 minutes ago, LordGabeNisBae said:

I realised that my motherboard bios was ancient and flashed it to the latest 3.20 iteration by asrock. As a result, I have been able to drop the voltages down a bit for what seems to be stable overclocks of 3.8Ghz @1.225V Vcore + 1.125V VSOC and tightened the timings of my memory to 14-14-14-34 @3200Mhz. Temps are down 10C but it is still far from ideal... 
 

If running an overclocked 8-core at 60C with ambien temps north of 30C is far from ideal, then maybe it does all boil down to unrealistic expectations...  

[LordGabeNisBae]

2 hours ago, SpaceGhostC2C said:

I think it woudl be worse with a dead pump, but in any case you can check readings for pump rpm and, perhaps, temperature: the x60 had a USB header connector, so you could monitor liquid temps and pump rpm on NZXT software. That's whay I was asking about.

 

But that's not what you said. No, there is no reason to expect the delta T (that is, the difference between ambient temps and CPU temps) to be any specific temperature. I was not talking abotu the difference between the stock cooler and the radiator, but the difference between ambient and CPU. It can be anything, case-specific.

Yes, you would expect better cooling from the AIO, but you haven't told us what your temps are if you change nothing but the cooler.

 

 

It could be, I mean, stress tests are supposed to be stressful  At 70C your cooler seems to be handling your OC just fine, though, so your concern seems to be only in relative terms to other solutions, but for that we need to know what temps would look like under other solutions.

 

I don't think 20C away from the CPU's max temps is ridiculous, though.

 

If running an overclocked 8-core at 60C with ambien temps north of 30C is far from ideal, then maybe it does all boil down to unrealistic expectations...  

pump header says 0 rpm lol. but its running I checked haha :P.

AFIK the specific heat capacity of water flowing at that speed is fat better at removing heat from the block than the evporative liquid used in vapor chambers/heatpipes. so delta t should be lower because the heat is being carried away and radiated faster than it would with an aircooler. Isnt this the reason why watercooling temps are so stable. 
You are right though  coming down from a 12x120mm radiator setup does bring about unrealistic expectations from a 280, a thin one as well. 

The chip passed a 3 hour Aida64 test with an average temperature of 61C. Doesnt seem like anything is broken so it should all be good. I might try some more serious undervolting later on to drop the temps even more.

Water cooling meant sub 50C temps to me but I guess you have to part with a few of those for the convenience of an AIO.

Edit: I have my 5960X oc to 4.5Ghz running sub 50 in my proper loop. That is a far hotter chip than the ryzen 7. But I digress... 280 doesnt compare to a dedicated quad rad