Looking at the Newegg page for liquid cooling I can see where you've dropped the ball. The size quoted (28/360 is the effective length of the radiator surface, which excludes the end tanks and fittings. I would assume that it would be a simple matter of finding a set of suitable fans to replace those that are making rude noises, for your 280 radiator that would mean a pair of 140mm fans, and Newegg have plenty on offer (at least on their UK site). at a whole range of prices and performances (and a pair of the most expensive far less than the cost of a new radiator or a new case)

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

The failures of the Gen. 1 Liqtech AIO's was in the pump not the fans.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Ah - wrong end of the stick at my end of the stick :-(

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

I am running a stock-clocked Ryzen 7 1700X with 32GB of DDR4-3000, and cooled with my re-purposed Thermaltake Big Typhoon that cooled my previous system of a Phenom X4 9950 for 10 years before the motherboard finally gave up the ghost, and my two previous systems before that as well.

I have BOINC/SETI set to use 15 threads to allow 5% headroom to do typical web browsing and watching videos while my Ryzen 7 digests SETI units at an average of 1:11 to 1:13 per WU. I only crunch during off-peak electricity rates, which is after 7PM on weekdays and all weekend long and holidays. At 95% system load with a stock clock of 3.5GHZ (as indicated within my Linux OS) I'm seeing an average CPU temperature of 52 to 55 degrees Celsius. I feel I might be able to lower that a few degrees by using better thermal paste, as I couldn't find my trusty Arctic Silver 5 when assembling my system a few weeks ago and had to use something else on hand.

I also have an nVidia 1060 that I use to occasionally crunch Asteroids@home units, however my particular model of card exhausts the air from the heatsink back into the case rather than blowing it out of the case, and thus raises my CPU temperatures closer to 60 degrees Celsius when crunching, which I am not comfortable with. I may experiment with an expansion slot cooler over the GPU fan and reapply some better CPU thermal paste, and report back with my results in the future.

Overall I'm quite impressed with the performance of my 1700X so far. I was debating investing in a Threadripper when planning out my system build but I feel the power consumption between the architecture and cooling requirements wouldn't have suited me.

That's great temps for a air-cooled 1700X. If you can keep it out of the 60°C. range, you are doing great. I am usually around 72-76°C. on my 2700X running 8 cpu tasks and 3 gpu tasks on a Corsair H105 AIO. Running Auto on the Vcore at 4.025Ghz which ends up around 1.33V.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

It sounds like you got the generation 1 version. Which I warned you about. Lots of defective returns for that generation due to excessive corrosion in the bearings and impeller. I suggested the Gen. 2 360mm version as the best choice. No issues with that version.

Sigh, the 360 is radiator is 15.5" long according to the NewEgg spec page.
The 280 radiator is 12.3 " long. I don't think the length includes the fittings coming out of one end of either one

My current case is 14.5" in that location. :(

This is a perfect example of a preceding decision constricting the choices of the next decision. I had an incorrectly sized water cooler that we sized the chassis/case for before ordering the 280.
If I could do it over again I would get another case like the one my Intel is running in. If the next copy of the 280 craps out again very soon I probably will :(

The joys of DIY :)

Tom

I can technically run a 120 to a 480, though in an aio the biggest I could run is a 360, there is a 420 aio, but My last 240 from that manufacturer leaked cause the maker did not tighten the base screws tight enough, My case has no mounting holes for a 140-420 radiator, let alone a 280, I know the case could handle 140-420 spec equipment in theory, but that would require accurate drilling, drilling I can't do, besides I have a 360, several in fact. What I would like to do is mount a 360(cpu) and a 120(gpu) aio in the top of My case, probably nearly impossible to do.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

The replacement water cooler is due early next week.

After playing with my Intel systems while waiting on it I am more than half convinced that I should try the following and sit and watch it for maybe a couple of weeks or more.

1) No OverClocking.
2) Enable XFR (sp?)
3) Run the default CPU thingy that does CPU boost.
4) Disable Quiet&Cool.
5) Run BOINC/Seti at 90% of threads/cores.
6) Use -nobs on the GPUs.

Discussion: Previous testing seems to produce a turbo boost of about 3.3GHz which is about what my e5-2690 v2 is running at. So depending on how the distribution of the CPU processing task times sort out, this would give me somewhat comparable test of 36 (-2 cores for gpu) threads/cores vs. 57 (-2 cores for gpu) for a RAC score.

Since my latest experience with a working memory system seems to indicate the slowest CPU task times I can expect are in the range of 3 hours and the Intel is running most of its CPU task times at or under 1 hour and 10 minutes the question is 1) What will the 2990WX's "average" cpu processing time. 2) What will be the range of processing times (will I still get some 47 minute processing times or not?) and how many at or near 3 hour processing times will I get.

If I can muster enough patience to produce a solid baseline of tasks/RAC then I can start experimenting with manual OverClocking.

I am hoping to pick up some affordable prices for the ram upgrade during or after the Thanksgiving black Friday sales.

Tom

---

A proud member of the OFA (Old Farts Association).

"Best ram chips (new/used) for 2990WX. Prices for 4? sticks (probably 32GB) NewEgg then eBay

CMD32GX4M4C3200C14C, n/a, n/a
CMD32GX4M4C3200C14M,n/a , n/a
CMD32GX4M4C3200C14T, n/a,$404+ s/h
F4-3200C14Q-32GFX, n/a,auction $227.50
F4-3200C14Q-32GTZ, $679-promo code, $756 (from UK)
F4-3200C14Q-32GTZR,$678, $681 international
F4-3200C14Q-32GTZRX, $489, $489
F4-3200C14Q-32GTZSW, $639, $572 international
F4-3200C14Q-32GVK, $547, $324
F4-3200C14Q-32GVR, $429, $564 international
F4-3200C15Q-32GTZKY, n/a, n/a
F4-3200C15Q-32GTZSW, $393, $393
TXBD432G3200HC14BQC01, n/a, $562 international

I went shopping and bought the low price leader on this list.

Tom

---

A proud member of the OFA (Old Farts Association).

I have 1950x...

First thing you should do is D/L HWiNFO64 ... Then you can monitor CPU frequency, Tdie temps, and Core Voltages...

The thing you should know is that TR, max temp is 68 before it starts to thermal throttle. So.... Don't overclock or run any type of utility like AMD master utility ...

Fist step would be to just try to get a baseline with your cooler at stock frequencies ... with your cooler...

After you can achieve all cores 100% running for a few hours... you may just want to let it run this way for a few days... and record how fast you are processing work units. Once you have a stable CPU running at stock below 60C.... You are doing good!

Next....

Go into your bios (always do manual settings) and always take screenshots of what you set it to so you can go back....

Anyway, in bios... now is the time to bump up the frequences just a little bit... and again let the system run this way for a few hours.... You might see the temps climb, and don't worry.

What you will find, as the push the chip higher... the temps will go up, what you want to see is 59~65 ... you really don't want to go any higher than this (on Tdie) sensor... Because the chip will always fluctuate with the temperatures and you really don't want to run a TR over 68C.

Next... This is the tricky part... you want to adjust your voltages down to .8 .9 volt is a good number.... The lower the voltage, the cooler the chip will run, but the lower voltage, the more likely it will crash. So... you gotta balance this out... try to balance a -stable- chip running 100% load... with the lowest voltages and highest clocks without crashing, when you find a good balance. again... maybe let the chip run for a few days without crashing. If it crashes give it a bit more juice ... and if the temps get too hot back off on the frequences.

When you have a stable chip at its highest frequency and within 60C. I like to keep mine no higher than 62C .... Mine will fluctuate between 59 and 62C (depending on AC settings)... Then you will know how far you can push your chip that is mostly due to the cooling solution you got.

Note... I bought a Enermax Liqtech when it first came out for my TR.... But... After about 6 months of good cooling the AIO cooler started to fail. I tried to take it apart and apply Thermal Paste 6 times before I realized it was the pump that was probably going bad. So I RMA'ed it, and got a new one back, that lasted about 3 months and did the exact same thing. Everything was great, and temp would start going WACKO... chip would Thermal Throttle badly.

Water cooling is great, I've just had bad luck with Enermax. I think if your gonna run TR 100% 24/7, an AIO cooler may not be for you, you might want to install a custom loop. But... I am now using a Noctua air cooler with very good results for a 16 core CPU. Stable CPU for a long time and I never see the temps go up past 64C ... Altho, I just slightly OC my CPU. Where with water cooler you could OC more (when working properly).

Also, you should have a case with PLENTY of airflow. I have a thermaltake view with 6 140mm fans running, 3 in the front pushing air in, and 2 on top and 1 in the back... When running the enermax, I mounted the rad on the front so that the 3 140mm fans would push fresh air through it. then the stock fans on the back would pull the air ... so... it was getting plenty of cool air.

BTW, I run my TR at 3450 stable (small OC) all 32 threads on air. Not bad...not bad at all!

Good Luck to you!

[url]

I have 1950x...

First thing you should do is D/L HWiNFO64 ... Then you can monitor CPU frequency, Tdie temps, and Core Voltages...

Thank you. Just found the Linux listing for it. I remembered a similarly named utility for Windows and was afraid that was what you were referring to.

http://www.linuxintro.org/wiki/Hwinfo

The thing you should know is that TR, max temp is 68 before it starts to thermal throttle. So.... Don't overclock or run any type of utility like AMD master utility ...

Unless I have missed it, there isn't a Linux version of that "master utility" or is there?

Fist step would be to just try to get a baseline with your cooler at stock frequencies ... with your cooler...

After you can achieve all cores 100% running for a few hours... you may just want to let it run this way for a few days... and record how fast you are processing work units. Once you have a stable CPU running at stock below 60C.... You are doing good!

Next....

Go into your bios (always do manual settings) and always take screenshots of what you set it to so you can go back....

Anyway, in bios... now is the time to bump up the frequences just a little bit... and again let the system run this way for a few hours.... You might see the temps climb, and don't worry.

What you will find, as the push the chip higher... the temps will go up, what you want to see is 59~65 ... you really don't want to go any higher than this (on Tdie) sensor... Because the chip will always fluctuate with the temperatures and you really don't want to run a TR over 68C.

Next... This is the tricky part... you want to adjust your voltages down to .8 .9 volt is a good number.... The lower the voltage, the cooler the chip will run, but the lower voltage, the more likely it will crash. So... you gotta balance this out... try to balance a -stable- chip running 100% load... with the lowest voltages and highest clocks without crashing, when you find a good balance. again... maybe let the chip run for a few days without crashing. If it crashes give it a bit more juice ... and if the temps get too hot back off on the frequences.

When you have a stable chip at its highest frequency and within 60C. I like to keep mine no higher than 62C .... Mine will fluctuate between 59 and 62C (depending on AC settings)... Then you will know how far you can push your chip that is mostly due to the cooling solution you got.

Note... I bought a Enermax Liqtech when it first came out for my TR.... But... After about 6 months of good cooling the AIO cooler started to fail. I tried to take it apart and apply Thermal Paste 6 times before I realized it was the pump that was probably going bad. So I RMA'ed it, and got a new one back, that lasted about 3 months and did the exact same thing. Everything was great, and temp would start going WACKO... chip would Thermal Throttle badly.

Water cooling is great, I've just had bad luck with Enermax. I think if your gonna run TR 100% 24/7, an AIO cooler may not be for you, you might want to install a custom loop. But... I am now using a Noctua air cooler with very good results for a 16 core CPU. Stable CPU for a long time and I never see the temps go up past 64C ... Altho, I just slightly OC my CPU. Where with water cooler you could OC more (when working properly).

Also, you should have a case with PLENTY of airflow. I have a thermaltake view with 6 140mm fans running, 3 in the front pushing air in, and 2 on top and 1 in the back... When running the enermax, I mounted the rad on the front so that the 3 140mm fans would push fresh air through it. then the stock fans on the back would pull the air ... so... it was getting plenty of cool air.

BTW, I run my TR at 3450 stable (small OC) all 32 threads on air. Not bad...not bad at all!

Good Luck to you!

Thank you for your detailed "How to".

Tom

---

A proud member of the OFA (Old Farts Association).

No Hwinfo64 is a DOS or Windows only application. Don't know if it works under Wine or not. Nope. Just tried. Hangs trying to probe the PCI bus.

There really isn't any good program for Linux in the same venue as Hwinfo64. There is one called Hardinfo but it doesn't do much. Certainly can't tell you anything about temps or voltages.

If by 'master' you are referring to Ryzen Master utility, no that is Windows only too. The closest thing you can get to hardware monitoring is the builtin tools like lspci or lshw. Or use the kernel mode drivers for the SIO chip on the motherboard and see what it can report.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

No Hwinfo64 is a DOS or Windows only application. Don't know if it works under Wine or not. Nope. Just tried. Hangs trying to probe the PCI bus.

There really isn't any good program for Linux in the same venue as Hwinfo64. There is one called Hardinfo but it doesn't do much. Certainly can't tell you anything about temps or voltages.

If by 'master' you are referring to Ryzen Master utility, no that is Windows only too. The closest thing you can get to hardware monitoring is the builtin tools like lspci or lshw. Or use the kernel mode drivers for the SIO chip on the motherboard and see what it can report.

That was the one I was thinking of, apparently there is a near name alike in Linux: http://www.linuxintro.org/wiki/Hwinfo

Tom

---

A proud member of the OFA (Old Farts Association).

I am running a stock-clocked Ryzen 7 1700X with 32GB of DDR4-3000, and cooled with my re-purposed Thermaltake Big Typhoon that cooled my previous system of a Phenom X4 9950 for 10 years before the motherboard finally gave up the ghost, and my two previous systems before that as well.

I have BOINC/SETI set to use 15 threads to allow 5% headroom to do typical web browsing and watching videos while my Ryzen 7 digests SETI units at an average of 1:11 to 1:13 per WU. I only crunch during off-peak electricity rates, which is after 7PM on weekdays and all weekend long and holidays. At 95% system load with a stock clock of 3.5GHZ (as indicated within my Linux OS) I'm seeing an average CPU temperature of 52 to 55 degrees Celsius. I feel I might be able to lower that a few degrees by using better thermal paste, as I couldn't find my trusty Arctic Silver 5 when assembling my system a few weeks ago and had to use something else on hand.

I also have an nVidia 1060 that I use to occasionally crunch Asteroids@home units, however my particular model of card exhausts the air from the heatsink back into the case rather than blowing it out of the case, and thus raises my CPU temperatures closer to 60 degrees Celsius when crunching, which I am not comfortable with. I may experiment with an expansion slot cooler over the GPU fan and reapply some better CPU thermal paste, and report back with my results in the future.

Overall I'm quite impressed with the performance of my 1700X so far. I was debating investing in a Threadripper when planning out my system build but I feel the power consumption between the architecture and cooling requirements wouldn't have suited me.

(Sorry to be "late" on this...)

Arctic has a NEW Thermal Paste that they rate lasts 8 Years, while Arctic Silver 5 can last 4 Years!

"Arctic MX-4 Thermal Compound Paste, Carbon Based High Performance..."

Currently at $9.99 USD for one tube/syringe .


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

Arctic MX-4 isn't NEW. It's been around for years. Been using it forever. Think I last bought my 20 gram cartridge back in 2011 when I first built the modern crunchers. Need to reorder some more as my tube has about one application left.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Arctic MX-4 isn't NEW. It's been around for years...

There's a (in)famous test somewhere on the web whereby the various 'super-expensive' 'thermal' pastes were tested and found to be no better than using home made jam. The only advantage was that the inorganic pastes would perhaps rot more slowly than the organic jam which is food for lots of things ;-)

In brief:

Best results are gained by 'lapping' (using fine grinding paste) your CPU heat spreader to the heatsink base. Carefully maintaining constant but flexible pressure using a spring loading system (things move with changing temperature!) ensures consistent reliable results. Keep things clean. Use whatever material ensures greatest physical contact for minimum thickness.

Even without 'lapping', just ensuring the flat surfaces are clean is good enough.

Nothing 'special' needed... Just nicely polished flat surfaces, silicon grease, and no air bubbles. "Copper-slip" works very well also!


Happy cool crunchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Arctic MX-4 isn't NEW. It's been around for years. Been using it forever. Think I last bought my 20 gram cartridge back in 2011 when I first built the modern crunchers. Need to reorder some more as my tube has about one application left.

I hadn't heard of MX-4 until just this year; that's why I thought it was "New"... My local Computer Source typically just pushes Arctic Silver 5; BUT, when I mentioned the MX-4 to them a couple of months ago they said that they can get that too.


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

I've always used AS5 because it was locally available and appeared to be far superior to the generic 'paste' usually shipped with aftermarket coolers, and I usually sticks with what works for me.

With that being said, I never considered the age of the compound as I didn't think it ever went bad, and after finally finding my two tubes of AS5 tucked away in some dusty corner I realise they're well over 10 years old; the tube of OCZ synthetic silver compound I ended up using in my new system came out like peanut butter, and is probably just as old! So I'm probably going to invest in a new tube of AS5 and re-apply to my rig soon.

And coincidentally enough I did lap my old heatsink before installing it in my new system, as the stock finish on it looked like Freddy Kreuger left his mark on it right from the Thermaltake factory (oddly enough it didn't seem to hurt my Phenom X4 for well over 10 years), and I did it by using successively-finer grades of wet sandpaper on a sheet of 1/4" plate glass all the way down to 2000 grit until I could see my reflection in it like a mirror after 8 hours of polishing. I even cut and shaped my own AM4 mounting bracket out of a 1/8" sheet of stainless steel and polished that before installing - that took 12 hours!

As I mentioned before, once I re-apply some new paste on my cooler I'll report on any temperature drops. :)

Thermal Grizzly Kryonaut/Hydronaut is even better. but i do like MX-4. it's a solid choice for almost any application.

and then you have liquid metal TIMs, but there's really no good reason to use that on top of the CPU IHS.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

I don't know why . . . . but every time I try to use Thermal Grizzly Kryonaut, it sucks compared to my MX-4. I've posted about my issues a lot if you do a search on my posts. I still have the tube of it I tried earlier in the year in another attempted experiment. It got one use and thrown back on the shelf. I guess it is a backup in case I run out of my MX-4. I used my last of my Noctua HT-1 in an application earlier in the year. It was just as good as MX-4 but the little tube that came with the Noctua cooler is really only a one shot application.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

yeah weird. honestly i havent used it much. maybe once or twice, but nothing to complain about.

i do use MX-4 more, since it's cheaper and easier to get at the local Microcenter.

NT-H1 is also excellent. i can't say i've ever bought it, only used the sample tubes that come with their coolers. haha

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours