Jump to content
Corsair Community

Hydro X Experience after 6 Months


Daknarian
 Share

Recommended Posts

Hey Everyone.

 

Wanted to drop a line here and give my experience with the Hydro X Cooling system.

 

Here is my Setup.

 

Ryzen 3950x - Running a Per CCD OC @1.3v

4.425ghz/4.3Ghz/4.3Ghz/4.425Ghz

(This was the only setup to get my CPU temps below 50c Idle, Running in PBO has my CPU Idling at 55c to 60c and gaming at 80c+)

 

ASUS Strix 2080ti OC - Running at 125% Power Limit & +100 on the Core with +650 on the Memory Clock

 

ASUS Strix Helios Case

 

Running Full Corsair Hydro X Cooling with Thermal Grizzly Kryonaut Paste

420mm Intake Rad with LL140 Corsair fans

360mm Exhaust Rad with LL120 Push/Pull fans

LL140 Exhaust fan at the Rear

All Fans set to Max (total 10 fans)

Pump set to 2500RPM

Set to Ryzen High Performance Power Plan

 

Idle Temps

Coolant Temp 34c

CPU temps at around 46c to 50c

GPU temps at around 37c to 39c

 

During Gaming

Coolant Temp 40c to 42c

CPU temps peak at around 75c

GPU temps peak at around 58c (1080p Max FPS)

GPU temps peak at around 64c (4k Max Graphics)

 

My Room Ambient Temps hit almost 29c during gaming, with a Ceiling Fan and House AC running.

 

Needless to say i'm not to happy with the current temps with all this cooling. Any tweaks you guys have would be appreciated. Been running this setup for 6 months. I love the Control that iCue gives me with the color schemes, but i'm lacking in performance.

 

Any feedback is appreciated.

Link to comment
Share on other sites

Well.. where to start...

Corsair used what they worked with for the longest of times : AIO coolers..

So they made custom loop waterblocks with microfins made for very low flow pumps.. and coupled that with a high flow D5 or DDC.. which doesn't work. needless to say it's far from efficient.

 

And on top of that, the XD5 pump is capped at half the flow rate of any other D5 on the market, probably because of bad pump top design..

 

The radiators are awesome since they rebranded the best, hardware labs, but Corsair also has the weakest fans when it comes to static pressure.. LL and QL on radiators give the results you see.

 

What i'd do is replace both waterblocks with any brand that knows custom loop : Watercool, EK, Alphacool.. plenty.. Blocks with thick fins that work best with high flow pumps.

 

Install a D5 pump that actually outputs the full 1500L/min, and not 800..

 

Install high static pressure fans so you dont blast your ears to have half decent performance.

 

With that amount of radiator surface, even at 30 ambient, your 2080TI could sit happily in the 45°C when pushed.

Your idle coolant temp is what i have under heavy load on OCed 2080 and 10900K with power limits removed.

 

In the meantime, i can only advise you to set your pump speed to max to try to lower the temps under load, if it's indeed an XD5 that you have.

Maybe if you have spare tubing, try a parallel loop. You may gain a few degrees from not sending slow flowing warm water from one block to the other.

 

Edit : i forgot ^^' the Strix Helios has terrible airflow too.. soooo it's kind of a lot of factors coming together.

Edited by LeDoyen
adding a note
Link to comment
Share on other sites

420mm Intake Rad with LL140 Corsair fans

360mm Exhaust Rad with LL120 Push/Pull fans

 

That's an issue right there.

You have a large intake radiator blowing warm air into the case. That warm air is then exhausted, at least partially, through your smaller exhaust radiator. You are pushing waste heat through one of your radiators. This will reduce your cooling efficiency pretty significantly. Setting the fans to max isn't going to help. Out of curiosity, do you have any idea of the internal case temp? What is the temperature of the air being blown through that exhaust radiator? Is it greater than or close to the coolant temp?

 

CPU temps are impacted by the heat spreader and TIM used by Intel. This has always been the case and it limits the amount of cooling available to the cores, increasing temps. Corsair did not start with AIO coolers (as previously implied) but designed HydroX from the ground up with a completely separate team focused exclusively on custom cooling.

 

And your cooling will increase the room ambient temperature. That's the cooling doing what it is supposed to do - getting the heat out of your case. And you are generating a LOT of heat with that CPU and GPU. It is going to heat up the room. Are you expecting otherwise? I'm not sure if the 29C in the room is a complaint that you somehow expect the cooling to not heat up the room? The laws of thermodynamics are against you there and they are pesky things.

Link to comment
Share on other sites

Coolant temp rise from 34 to 42C during gaming is just about right. Typically your would see CPU+GPU combined loads of 400-500W in that range. I have about the same combined wattage and it has taken me months to whittle it down from +10C to +7C with 3x360mm. The cooling system is doing its job and single degree coolant changes are what you are chasing with different fan types, different layouts, pulling the dust filters, etc. I thought your GPU temp looked warm at first, but at 42C coolant base, that's a +16C differential to GPU temp and exactly the same as my 2080 Ti.

 

And that brings us around to what really is the heavy hand in most end temps -- your baseline temperature. You can double of on rad thickness, add in 3000 rpm fans, throw on an extra radiator, and in the end, all those gains are swallowed up when the room temp goes up +4C. There's probably not much you can do about the 29C room temp or you would already be doing it. Such is Summer life and there is a reason no one is complaining about their hardware temps in Winter. If the room was 19C, everything in the case would be 10C lower and we probably wouldn't be talking about this. At 29C room temp, a 34C coolant baseline is right where most people will be. I generally quote +4-7C above room temp as the normal coolant temperature. That is mostly going to be down to radiator placement (intake/exhaust), case design, dust filters, etc. Pop the side panel off and you likely drop -2C and that is really the scope of it. However, there isn't much reason to chase this value because attempting to minimize it does not automatically carry over to load performance. Usually you wind up in the same place, absent some clear restriction.

 

Easy things to try:

 

1) Bump the pump speed up to 3500-4000 during gaming. It may have no impact whatsoever, which would be a good thing and means everything is flowing well. If your loop is somewhat restrictive, the water can slow down a bit too much in the CPU/GPU block. Regardless, you might get -2C or so on the GPU in very heavy loads like the 4K data.

 

2) Pop off the door or front panel on the case. Not aesthetically ideal, but a case is always a restriction. See what effect this has and you can determine if it is worthwhile. The largest gain would likely be from letting the top rad pull in some outside air instead of having to use the front radiator waste heat. This is likely something you only have to do in Summer. The side glass on my 011 XL comes off for certain games that really tax the system. For -4C, it's worth the effort so my side panel fans can breathe.

 

 

Harder stuff:

 

1) Normally we say flow path between pump/res, radiator, and components doesn't matter. On a lab table, it doesn't. However, when running a front/top radiator arrangement where the front exhausts into the top, there is a preference. Ideally, you would like the combined GPU/CPU heat to go to the top radiator and shed 1-2C before it goes to the front radiator. This hopefully keeps the front radiator exhaust temp that 1-2C cooler than the liquid temp in the top radiator and you still have some cooling effect. If the intake air temp and the coolant temp are the same, then you are mostly just passing the liquid through with no change. After quitting load, you can even see how it heats up the coolant as it passes through the warmer top. You may already have configured the loop this way and thus no change. If not, understand you are chasing maybe 1C in coolant reduction. These are small gains.

 

2) Another way to attack the above with far less effort would be to flip the front fans around to exhaust, along with changing the rear fan to intake. This is a reverse flow set-up. Now both rads can dump their heat directly out of the case and loop order is no longer important. This WILL NOT reduce your idle coolant temp. You might even be +1-2 more. However, the loads temps may be better. Reverse flow can work well if there is enough mesh/space for air to enter the back of the case. If you are solely reliant on a single 120mm with a solid back, it probably does not improve things. Room position may factor as well. Tucked into a corner, the air back there is probably warmer than the rest of the room and the PC does have a heat aura. On the other hand, if there is enough space or even an AC vent, you can make this work. When I run this configuration in Summer, I usually take a fan to blow air across me and then to disrupt the air in the back of the case, helping to make sure there isn't a hot spot around the back.

Link to comment
Share on other sites

CPU temps are impacted by the heat spreader and TIM used by Intel. This has always been the case and it limits the amount of cooling available to the cores, increasing temps.

 

He has a R9 3950x, not an 8th gen Intel core or older, with TIM issues :)

 

Corsair did not start with AIO coolers (as previously implied) but designed HydroX from the ground up with a completely separate team focused exclusively on custom cooling.

 

What i meant to say is they took AIOs as inspiration for their blocks, which is what makes them perform poorly compared to the competition.

 

Fan orientation on rads (one intake, one exhaust) won't affect much the coolant temp if there is sufficient airflow. Here, with LL fans who already perform poorly on rads and the very closed case front + mesh filter, the front rad does virtually nothing.

 

Setting both rads to exhaust will make the fans struggle even more, and there's still that big glass panel in front.

 

Removing the front panel as C suggested would be a good test to know if the case is causing most of the overheating.

Link to comment
Share on other sites

Thanks for the reply's. I have tried almost every idea you guys have shot out.

 

Today i ran the house AC almost all day, the room ambient temp is at 24.5c and after a good amount of gaming, the coolant temp didn't go over 37c.

 

I am running the XD5 pump and I did bump it up to 4000rpm, so that could also be a factor.

 

But moving forward, the LL fans are pretty, but not doing what I want. I'll be looking to change those out during the next Upgrade session. When the new RTX3000 series cards and the AMD Zen 3 come out.

 

As for the Reservoir and Blocks. I've been spoiled with the pretty lights and all around uniformity i'm getting with everything being Corsair. AND I MEAN EVERYTHING. But the performance just inst there. With this much cooling I shouldn't be running as hot as I am.

 

But what will I do for RGB? Currently I have everything Corsair.

LED strips

LS100

LT100

ST100

Hard Mousepad

K95 Keyboard

Virtuoso Headset (complete Junk.. stopped charging after 3 months)

CPU Block

GPU Block

XD54 Reservoir

Dominator RAM

10 LL Fan

 

Multiple RGB Apps running the background...

 

The only items in this setup that aren't Corsair is the Mouse and Motherboard. Since Corsair doesn't make a decent Ambidextrous Mouse or Left Handed Mouse.

 

Can I Haz Sponsorship now?

Edited by Daknarian
Link to comment
Share on other sites

  • Corsair Employees
Well.. where to start...

Corsair used what they worked with for the longest of times : AIO coolers..

So they made custom loop waterblocks with microfins made for very low flow pumps.. and coupled that with a high flow D5 or DDC.. which doesn't work. needless to say it's far from efficient.

 

And on top of that, the XD5 pump is capped at half the flow rate of any other D5 on the market, probably because of bad pump top design..

 

The radiators are awesome since they rebranded the best, hardware labs, but Corsair also has the weakest fans when it comes to static pressure.. LL and QL on radiators give the results you see.

 

What i'd do is replace both waterblocks with any brand that knows custom loop : Watercool, EK, Alphacool.. plenty.. Blocks with thick fins that work best with high flow pumps.

 

Install a D5 pump that actually outputs the full 1500L/min, and not 800..

 

Install high static pressure fans so you dont blast your ears to have half decent performance.

 

With that amount of radiator surface, even at 30 ambient, your 2080TI could sit happily in the 45°C when pushed.

Your idle coolant temp is what i have under heavy load on OCed 2080 and 10900K with power limits removed.

 

In the meantime, i can only advise you to set your pump speed to max to try to lower the temps under load, if it's indeed an XD5 that you have.

Maybe if you have spare tubing, try a parallel loop. You may gain a few degrees from not sending slow flowing warm water from one block to the other.

 

Edit : i forgot ^^' the Strix Helios has terrible airflow too.. soooo it's kind of a lot of factors coming together.

 

Before I focus the on actual topic of this thread, I have to reply to this post that has a lot of misinformation.

 

About the correlation with AIO coolers.

No, the cold plate design of the Hydro X CPU water blocks has nothing to do with AIO coolers.

The cold plates for the XC7 and XC9 CPU water block models were designed from scratch and actually have an impressively low hydraulic restriction given the density of fins. See here for a comparison with some of the other CPU water block models that are on the market. Both in line with competitor models.

 

Hydraulic restriction comparison.

Thermal performance.

 

+ As a consequence of high-density fin configuration, these cold plates perform very well even with low flow rates in the custom cooling system. Unfortunately, there are no CPU water block reviews that would go so deep when evaluating the CPU water block performance and I can't share any of internal evaluation data but at the end of the day, this is why the iCUE Hydro X pump profile is made in such way it is and it allows for lower pump speeds (lower noise operation).

 

XD5 pump 'cap'

The XD5 pump is not capped in any way. Like many other D5 pump motor based pump/reservoir units that are on the market, it's using a full power D5 PWM pump motor.

It's just a different approach on how to represent performance data of the pump. Some of the competitors' models like to use 'pumped up' numbers which are theoretical max flow rate and max head pressure. These would occur in an ideal scenario when there is no restriction (max flow rate) or no flow rate (max head pressure). The assembled custom loop does not present an ideal scenario.

The XD5 is showing a typical working point in a custom cooling system.

Yes, there are differences in how the pump volute is made. These will affect the P-Q curves of the pump but, with modern pump units, you will need laboratory equipment to measure differences once put in a real custom cooling system.

 

More information about P-Q curves:

Pump performance is presented as a P-Q curve. P (pressure), Q (volumetric flow). P-Q curve presents a relation between the flow and pressure in the system.

For presentation purposes, data from

https://www.hvacquick.com/catalog_files/Laing_D5_Vario_Catalog.pdf

On the image below, see the line marked with ‘5’. It’s the one which can serve as a guideline for a D5 PWM pump motor running at full speed.

hndlEVc.png

The values on the X and Y axis present the theoretical max numbers – 'pumped up numbers'

The black point on the curve '5' represents a typical working point and represents the XD5 flow and head performance figures.

Everything is on the same curve, thus the same performance.

Link to comment
Share on other sites

  • Corsair Employees
Hey Everyone.

 

Wanted to drop a line here and give my experience with the Hydro X Cooling system.

 

Here is my Setup.

 

Ryzen 3950x - Running a Per CCD OC @1.3v

4.425ghz/4.3Ghz/4.3Ghz/4.425Ghz

(This was the only setup to get my CPU temps below 50c Idle, Running in PBO has my CPU Idling at 55c to 60c and gaming at 80c+)

 

ASUS Strix 2080ti OC - Running at 125% Power Limit & +100 on the Core with +650 on the Memory Clock

 

ASUS Strix Helios Case

 

Running Full Corsair Hydro X Cooling with Thermal Grizzly Kryonaut Paste

420mm Intake Rad with LL140 Corsair fans

360mm Exhaust Rad with LL120 Push/Pull fans

LL140 Exhaust fan at the Rear

All Fans set to Max (total 10 fans)

Pump set to 2500RPM

Set to Ryzen High Performance Power Plan

 

Idle Temps

Coolant Temp 34c

CPU temps at around 46c to 50c

GPU temps at around 37c to 39c

 

During Gaming

Coolant Temp 40c to 42c

CPU temps peak at around 75c

GPU temps peak at around 58c (1080p Max FPS)

GPU temps peak at around 64c (4k Max Graphics)

 

My Room Ambient Temps hit almost 29c during gaming, with a Ceiling Fan and House AC running.

 

Needless to say i'm not to happy with the current temps with all this cooling. Any tweaks you guys have would be appreciated. Been running this setup for 6 months. I love the Control that iCue gives me with the color schemes, but i'm lacking in performance.

 

Any feedback is appreciated.

 

Hi!

 

The load (gaming) delta temperatures for the CPU and GPU are in line our internal testing.

 

For your specific setup, I believe some adjustments could be done to lower the coolant temperature.

I would approach in this way:

1. If not already, remove dust filters (top, front, bottom).

2. Convert to full exhaust fan configuration.

Link to comment
Share on other sites

I wouldn't call this impressively low flow restriction. it's in the average, but the major issue is mixing dense fins with what you marketed as a lower flow pump.

Why misrepresent the pump specs by using average values? average values in what kind of setup? Can i expect 800l/min with my 3 rads and 2 blocks? what size rads? angled fittings? what about parallel loops?

 

The typical working point isn't useful at all. At least with the actual pump specs we know what we start with and we then add losses as we build the rig. It's a bit like if Bugatti sold its Veyron, advertised with a speed of 120km/h because that's the average speed people will use it at. It just doesn't tell anything.

 

I still believe there's something off about the XD5 pump top, even if it works on the vast majority of cases. There's been a few examples in the forum about people struggling with low flow, some more than others.

Please have a look at this thread for example (and the video linked in it). https://forum.corsair.com/v3/showthread.php?t=198485

If the XD5 has the same specs as other brands, then there must be a logical explanation for what this user has experienced (with the same D5 pump, just different top), and for the flow performances in the video (that is far from using perfect scientific methods, i give you that ^^).

The only information i was passing is what i gathered on this forum, and on the XD5 page on the Corsair site.

Link to comment
Share on other sites

Why misrepresent the pump specs by using average values?

 

Why misrepresent the maximum specifications when they are not attainable in any natural circumstances? That is what most of them are doing with the 1500 L/hr values. It's not a reproducible result with actual equipment on the end, even minimal 1 block, 1 radiator on a table. To use your example, it's like running the Bugatti on a friction-less treadmill and calling it "road speed". You've removed the physical nature of the environment in which it must operate (friction losses, air resistance, etc). If you look at those companies that are nice enough to provide a P-Q curve, it doesn't always match their own specs. A good example of this is the new 'modified' version of a DDC pump I have from another brand. It's listed at 1100 L/hr. Their P-Q curve shows about half that. Actual measured flow at 100% for a 1 block, 1 radiator straight loop on a test bench was around 200 L/hr. I expected it to absolutely gush at maximum. It has less flow the XD3 it replaced at 100% in my actual loop (which is restrictive), despite the XD3's lower spec ratings on both pressure and max flow. Marketing is marketing, but that doesn't tell you what happens when faced with resistance and real hardware.

Link to comment
Share on other sites

Bah, neither the max flow plugged only to a flowmeter, and the average 800 figure tell anything since not 2 rigs are identical. but the max theoretical is the value with the less variables. That's what i personally would expect for a pump specification. what is the maximum pressure it's capable of delivering, and the maximum flow.

Marketing is marketing (...)

unfortunately ^^'

Link to comment
Share on other sites

Hi!

 

The load (gaming) delta temperatures for the CPU and GPU are in line our internal testing.

 

For your specific setup, I believe some adjustments could be done to lower the coolant temperature.

I would approach in this way:

1. If not already, remove dust filters (top, front, bottom).

2. Convert to full exhaust fan configuration.

 

Unfortunately I won't be able to test this until I drain the loop and flip the fans.

 

Also, you recommend me still using the LL series fans?

Change the rear exhaust fan to intake? Or keep it exhaust?

Changing out the fans to ML series would give me more flow?

 

Currently my flow pattern is XD5 -> GPU -> 360mm exhaust -> CPU -> 420mm Intake -> XD5...

 

Don't get me wrong, I love Corsair and their family of Products, and I would dump Logitech GHub in an instant If corsair made a decent mouse for me. But im starting to value performance over RGB. And i would love to keep the themes i have going on here. But running those temps in Gaming is a bit too much.

 

I attached some pics.

20200625_194705.thumb.jpg.c4e64c3602a56d3f9e67f17ff930a572.jpg

20200809_234115.thumb.jpg.e3f1a9e286e74765da842eae64ccd70b.jpg

Link to comment
Share on other sites

Currently my flow pattern is XD5 -> GPU -> 360mm exhaust -> CPU -> 420mm Intake -> XD5...

 

Well, that's not bad since 70-75% of the system wattage is GPU and it goes to the exhaust radiator first. I would not starting ripping tubing out to change it for this reason alone and other measures may be more productive.

 

A ML fan will move more air through a radiator than a LL at the same speed. However, don't confuse +15% more airflow with +15% better cooling. The relationship is not equivalent. +-2C to coolant temp would be a sizable for a fan change of any type. The larger your radiator surface area, the less it matters. So when I run my 2080 Ti (330W) on single 280mm radiator with 4 LL140 in push-pull, I can get about 3C reduction @1000 rpm by switching to ML. However, when I run that same GPU in 2x360mm set-up, along with the CPU in the same loop (now a 450W load), the difference was about 1.2C for switching from LL to ML120. The more radiator length you have, the less single fan airflow is going to matter. Whether or not you change fans or are just curious is totally up to you. Just don't expect miracles from this. It's small change. I've been spending the most of the last several months testing a variety of fans on a new 360x54mm radiator as part of a larger 3x360mm set up. The differences from QL to 50mm thick industrial models are measured in tenths of a degree at the in/out measurement point.

 

I think flipping the radiator fans to exhaust is the easiest, least expensive thing to try and Summer is a good time to do it. That would include using the rear fan as an intake to set up the reverse flow. However, the most dramatic changes will be anything you can do to less the environmental conditions, whether its more AC, a floor fan to move the waste heat coming off the case away, or other room management features.

Link to comment
Share on other sites

 Share

×
×
  • Create New...