A challenge - PA120.3 vs HE120.3 vs PA160.

A combination of unfortunate events recently resulted in an old familiar friend making an appearance on my doorstep... my old TEC rig had returned home...!

http://www.over-clock.com/marci/2003...s/DSC00106.jpg

The guy I sold this to (R.I.P. d00d) sadly couldn't get his head round the TEC side of things so ran it as a standard watercooling setup. Once it got back to me however, twas time to get the TECs back on the job.

http://marci.over-clock.com/tec/S4100255.jpg

226w TEC on CPU (A64 Clawhammer 3200+), 172w TEC on GPU (X800ProVIVO >> X800XT-PE). Figured whilst I was in there I'd strip out the previous 240v Eheim1250 pump... hell... all the watercooling got ripped out basically... and refit with more up-to-date equivalents....

http://marci.over-clock.com/tec/S4100260.jpg

The radiator in this rig is the first ever ThermoChill HE120.3 prototype from way-back-when. Still going strong...

http://marci.over-clock.com/tec/S4100261.jpg

So... what was I gonna do with the rig now? Well, with the new PA Series rads finally getting sorted, figured may as well do a comparison of HE120.3 vs PA120.3 with the same fans... to do so required some monitoring tools in the form of the mCubed T-Balancer and some handiwork with silicon... thus giving me liquid in and out temps...

http://marci.over-clock.com/tec/teclatest/liqtemps.jpg

And air in and out temps

http://marci.over-clock.com/tec/teclatest/airin.jpg

http://marci.over-clock.com/tec/teclatest/airout.jpg

Add in the handy feature of the tbalancer software being able to talk to LCDC and we get...

http://marci.over-clock.com/tec/teclatest/tbanlcd.jpg

MBM5 will collect CPU temps, and ATITool will handle GPU temps. So, that was the rig all ready... time to case it all up but leave rad external (just easier for swapping it over later)

http://marci.over-clock.com/tec/teclatest/casedup.jpg

Started off with unit under full load with as much heat as possible, whilst keeping speeds at stock.

CPU was set to 1.75v. AGP VDDQ to 1.6v. 100% load inflicted.

Temps were meaured every 20mins. Surveying them, they levelled out rapidly, within an hour, and hovered around the same points for the rest of the day (controlled ambient as best possible).

Max variance in ambient temp was 1.5 deg C. This was reflected across the board.

Note, temps shown are just a quick cross-section of readings recorded, not complete end-to-end results...

HE120.3 with 3x Panaflo M1A@7v

Air<: 23.0 | 24.0 | 23.5 | 23.0 | 23.5 | 24.5 | 24.0
Air>: 26.0 | 27.0 | 26.5 | 26.0 | 26.5 | 27.5 | 27.0
Liq<: 33.5 | 34.0 | 34.0 | 34.0 | 34.0 | 34.5 | 34.5
Liq>: 33.0 | 33.5 | 33.5 | 33.5 | 33.5 | 34.0 | 33.5

CPU: 17.0 | 17.0 | 18.0 | 18.0 | 18.0 | 19.0 | 18.0
GPU: 19.0 | 19.0 | 20.0 | 19.0 | 20.0 | 20.0 | 20.0

< = In
> = Out

So, averages are being used.

I'll summise testing of the HE120.3 with Panaflo M1A (fans at 7.04v by multimeter) as follows:

Avg Air <: 23.6
Avg Air >: 26.6
Avg Air Dt: 03.0
Avg Liq <: 33.9
Avg Liq >: 33.5
Avg Liq Dt: 0.40

Next step - remove HE120.3, install PA120.3 with same fans set to same voltage by Rheostat (measured with multimeter). This gave me the following recordset...

PA120.3 with 3x Panaflo M1A@7v

Air<: 22.5 | 24.0 | 22.5 | 23.0 | 23.0 | 23.5 | 22.5 | 23.0 | 22.5 | 22.5
Air>: 27.5 | 28.0 | 27.5 | 28.0 | 28.0 | 28.5 | 28.0 | 28.5 | 28.0 | 28.0
Liq<: 30.5 | 30.5 | 31.0 | 31.0 | 31.0 | 31.0 | 30.5 | 31.0 | 30.5 | 31.0
Liq>: 30.0 | 30.0 | 30.0 | 30.0 | 30.5 | 30.5 | 30.0 | 30.0 | 30.0 | 30.0
CPU: 15.0 | 15.0 | 15.0 | 14.0 | 15.0 | 15.0 | 15.0 | 15.0 | 15.0 | 15.0
GPU: 15.0 | 15.0 | 15.0 | 16.0 | 16.0 | 16.0 | 15.0 | 15.0 | 15.0 | 15.0

Avg Air <: 22.9
Avg Air >: 28.0
Avg Air Dt: 05.1
Avg Liq <: 30.8
Avg Liq >: 30.1
Avg Liq Dt: 00.7

Already, we see a substantial drop in CPU and GPU temps between the two rads, with the PA120.3 shaving 4 or 5 degrees off each over the old HE series.

Figured I'd go for some self-amusement next...

PA160 with 1x Panaflo M1A@7v
Well... I'm impressed...

Air<: 24.0 | 23.5 | 23.0 | 23.5 | 23.5 | 23.0 | 24.0 | 23.5
Air>: 31.5 | 31.5 | 31.5 | 31.5 | 31.5 | 31.5 | 32.0 | 31.5
Liq<: 37.0 | 37.5 | 37.5 | 37.0 | 37.0 | 37.5 | 38.0 | 38.0
Liq>: 35.5 | 36.5 | 36.0 | 36.0 | 35.5 | 36.0 | 36.5 | 36.5
CPU: 21.0 | 22.0 | 21.0 | 21.0 | 21.0 | 21.0 | 21.0 | 21.0
GPU: 21.0 | 22.0 | 22.0 | 22.0 | 22.0 | 22.0 | 22.0 | 22.0

Avg Air <: 23.5
Avg Air >: 31.5
Avg Air Dt: 08.0
Avg Liq <: 37.4
Avg Liq >: 36.0
Avg Liq Dt: 01.4

Only 1 or 2 degrees hotter than the HE120.3 manages on CPU & GPU temps, with 2 less fans = less noise. Always a bonus!


PA160 with 2x Panaflo M1A@7v (PushPull with dual shrouds)

Air<: 23.0 | 23.0 | 24.0 | 23.0 | 23.0 | 23.5 | 24.0 | 23.5 | 24.0 | 23.5
Air>: 27.0 | 28.0 | 27.5 | 27.5 | 27.5 | 27.5 | 27.5 | 28.5 | 28.0 | 29.5
Liq<: 34.5 | 35.5 | 36.5 | 36.5 | 36.5 | 36.5 | 37.0 | 37.5 | 37.5 | 38.0
Liq>: 34.0 | 35.0 | 35.0 | 35.0 | 35.0 | 35.0 | 35.5 | 36.0 | 36.5 | 36.5
CPU: 19.0 | 20.0 | 21.0 | 20.0 | 21.0 | 21.0 | 21.0 | 21.5 | 21.5 | 21.5
GPU: 19.0 | 21.0 | 21.0 | 21.0 | 21.0 | 21.0 | 21.0 | 22.0 | 22.0 | 22.0

Compare these figures to figures above - result? 2x fans in Push-Pull is of no major benefit over 1x fan in push.

PA120.2 with 2x Panaflo M1A@7v

Air<: 22.5 | 22.0 | 22.5 | 22.0 | 22.5 | 22.5 | 22.0
Air>: 32.0 | 31.5 | 32.0 | 32.0 | 32.0 | 32.0 | 32.0
Liq<: 33.0 | 33.0 | 33.0 | 33.0 | 33.5 | 33.5 | 33.0
Liq>: 32.5 | 32.0 | 32.5 | 32.5 | 32.5 | 32.5 | 32.5
CPU: 18.0 | 17.0 | 18.0 | 18.0 | 18.0 | 18.0 | 17.0
GPU: 18.0 | 18.0 | 18.0 | 18.0 | 18.0 | 18.0 | 18.0

Avg Air <: 22.3
Avg Air >: 31.9
Avg Air Dt: 09.6
Avg Liq <: 33.1
Avg Liq >: 32.4
Avg Liq Dt: 00.7

2x PA160 with 1x Panaflo M1A@7v on each

AirIn <: 23.5 | 24.0 | 23.5 | 24.0 | 23.5 | 23.5 | 23.5 | 23.5
AirRad1>: 29.0 | 28.5 | 29.0 | 29.0 | 29.0 | 29.0 | 29.0 | 29.0
AirRad2>: 29.0 | 29.0 | 29.0 | 29.0 | 29.0 | 29.0 | 29.0 | 29.0
Liq <: 31.0 | 30.5 | 30.5 | 31.0 | 31.0 | 31.0 | 31.0 | 31.0
LiqMid : 30.5 | 30.0 | 30.0 | 30.0 | 30.0 | 30.5 | 30.0 | 30.0
Liq >: 30.5 | 30.0 | 30.0 | 30.0 | 30.0 | 30.0 | 30.0 | 30.0
CPU : 16.0 | 16.0 | 15.0 | 16.0 | 16.0 | 16.0 | 16.0 | 16.0
GPU : 15.0 | 15.0 | 15.0 | 15.0 | 15.0 | 16.0 | 15.0 | 15.0

Avg Air Into Rads <: 23.6
Avg Air Out of Rads>: 29.0
Avg Air Dt : 05.4
Avg Liq Into Rads <: 30.8
Avg Liq Out of Rads : 30.0
Avg Liq Dt : 00.8

Break it down...

PA160 with 1x Panaflo M1A @ 7v: CPU - 21 GPU - 22
HE120.3 with 3x ^^ ^^ ^^ ^^ : CPU - 18 GPU - 19
PA120.2 with 2x ^^ ^^ ^^ ^^ : CPU - 17.7 GPU - 18
2x PA160 with 1x ^^ on each ^ : CPU - 15.8 GPU - 15.1
PA120.3 with 3x ^^ ^^ ^^ ^^ : CPU - 15 GPU - 15


ADDENDUM - JULY 2006 - Further independant comparative results vs other rads - Testing Performed by Radical_53 on behalf of HardwareLuxx Magazine - Germany

120.3 radiator formats tested. Results scale down perfectly for all other formats (120.2, 120.1) from same radiator family (as originally proven in BillA's testing of the HE Series MANY years ago)

1 - Pressure Drop - flow set at 4lpm at pump, pressure measured after radiator. % loss shown - smaller the loss, better the head pressure... listed best to worst.

Least Pressure Drop...
ThermoChill PA120.3 - 6%
Watercool HTFX3-x - 14%
HWLabs BlackIce GT360 - 17%
HWLabs BlackIce GT360 XFlow - 17%
HWLabs BlackIce Xtreme - 17%
HWLabs BlackIce Xtreme XFlow - 17%
Watercool HTSF - 17%
Cooltek MCR320 - 18%
NexXxos Xtreme - 23%
NexXxos Pro - 27%
...highest Pressure Drop

As you can see, you could theoretically run TWO Thermochill rads, and still have better head pressure than if you were using a SINGLE rad of any other brand.

2 - Performance order (best to worst) based on final coolant temp, cooled with YateLoon 120mm fan - lower the "k" figure (difference between coolant temp and air temp), cooler the coolant, therefore better the performance...

At 600rpm:
Thermochill PA120.3 - 7k
Watercool HTSF - 8k
Watercool HTSF3-x - 8.4k
Cooltek Maxistream / Swiftech MCR - 8.4k
Black Ice GT - 9.2k
Black Ice Extreme - 11k
Alphacool NexXxos Pro - 11.2k
Black Ice GT X-Flow - 12.1k
Black Ice Extreme X-Flow - 12.3k
Alphacool NexXxos Extreme - 14k


At 900rpm:
ThermoChill - 6.2k
Cooltek - 6.9k
BIGT - 7.2k
WC HTSF - 7.3k
BIGT-XFlow - 7.3k
WC HTSF3-x - 7.4k
NexXxos Pro - 7.5k
BIX - 8.5k
NexXxos Extreme - 8.5k
BIX-XFlow - 9.5k


At 1200rpm:
ThermoChill - 4.2k
Cooltek - 5.6k
BIGT - 5.7k
BIGT-XFlow - 5.9k
BIX - 5.9k
NexXxos Pro - 6.1k
NexXxos Extreme - 6.1k
HTF3-x - 6.3k
HTSF - 6.3k
BIX-XFlow - 6.6k

The rad with the lowest delta (k) can also (because it is returning the coolant to a lower temperature) handle more heat at the same temp as another brand rad... eg: You could shift 500w with Brand X at 1200rpm. With ThermoChill at 1200rpm you could shift MORE than 500w - more cooling power for the same noise level... or you could shift 500w on it's own at LOWER than 1200rpm, therefore same cooling power, but less noise.

Couple the fact that ThermoChill PA Series is FAR less restrictive to head pressure than any other rad, AS WELL as the fact that it outperforms all the rest at every fan speed tested, with it's lead increasing the quieter the fan... and the results are obvious. One family of radiators out there will always beat all the rest (and I'm glad someone finally tested all those out there to allow me to actually say this with some proof to back it up!) - the lower pressure drop alone will mean better performance out of EVERY block in your system, as you'll have better pressure thru-out the entire setup.

Note, these results are from independant testing by HardwareLuxx Magazine in Germany, and were published in May 2006 in their magazine in a head to head shootout of the top 10 Triple-120mm radiators available today. The original article in German is available to download from http://www.thermochill.com/pa1203.php, and since it's publication, we've seen ThermoChill sales to the German market soar, to the point that big name companies over there who currently rely on the chinese-clone rads as their main rad (ie: XSPC rads, AlphaCool rads) are beginning to switch over to using ThermoChill products instead...

These results also highlight the fact that singlepass rads on the market currently, which are all a dualpass with tanks knocked off and singlepass tanks slapped on, perform worse than their dualpass counterparts, and offer no benefits when it comes to pressure drop (the entire black ice range, whether dualpass or singlepass, still poses a 17% restriction). Other than the ThermoChill PA160 which was designed from the ground up as Singlepass...

So, if you're using silent low airflow fans, depending on your current rad, you could be seeing coolant temps upto 7 degrees cooler from a ThermoChill rad over another brand. The higher the airflow, the smaller the difference (eg: 2.4 degrees at 1200rpm rather than 7 degrees at 600rpm)... but in this day and age where everyone wants as quiet as possible... the PA series is the best choice.


ADDENDUM - OCTOBER 2006 - Further independant comparative results vs other rads - Testing Performed by Radical_53 - Germany

ThermoChill PA120.2 vs Swiftech MCR320
PA120.2 fitted with shroud and 2x YateLoon D12-SL fans, Swiftech MCR320 fitted with 3x of the same...

http://marci.over-clock.com/pa2-vs-mcr3.jpg

As you can see from the above figures, the PA120.2 with shroud manages to effectively match the Swiftech MCR320, with less noise due to fewer fans in play...


ADDENDUM - JANUARY 2007 - Further independant comparative results vs other rads, and complete curves for ENTIRE PA Series Published - Testing Performed by Bill Adams - (Formerly of Thermal-Management-Testing, Swiftech, and CoolingWorks)
See http://www.thermochill.com/PATesting/ for full test results. Test rig specified here: http://thermal-management-testing.co...pment_list.htm - Methodology specified here: http://thermal-management-testing.co..._procedure.htm

http://www.thermochill.com/PATesting...VSFlowrate.jpg


ADDENDUM - NOVEMBER 2007 - EffizienzGurus review 7 of the best 3x120 Format Radiators
Once again, the ThermoChill PA120.3 was found to win hands down vs the rest... and the PA120.3 was awarded the "Efficiency Award".
Full Review

Now those are some very conclusive results....:eek:

Good review Marci! Thanks for sharing the info :D

So the HE got kicked by the PA huh?

TY for the test results so fast marci :)

gonna help me out a lot on choosing which rad to get :up: :up: :up: :up:

And I though that the PA160 wasn't going to be enough for two OC'ed 7800GTX's.....guess I was wrong.

Nice results Marci, nice idea of the benefits of the PA series there, thanks for the info.

G

Wow very nice results there marci..

Seeems that PA160 is handing thigns better than i thought. In my brief testing it didnt do nearly as well

Hey marci I'd like to see another pa160 thrown into the comparrison ...like PA120.3 vs HE120.3 vs PA160 vs PA160x2. Won't cathar be doing some testing?

There's NO SUCH THING as a PA160.2, there's only the PA160. The theoretical PA160.2 equates to more or less same cooling area as a PA120.3, therefore the PA160.2 is pointless - just buy the PA120.3 - it's easier to mount for a start.

PA120.2 data added into the above.

marci he meant dual pa160's....

Added Dual PA160 test data.

Would it be too much to ask you for 2x120.2 numbers? :)

About 2 deg better than a PA120.3 (estimated). There's only so long I can spend doing this... other work is piling up and me sitting logging temps doesn't bring in the wages.

For me your estimate is just fine. Thank you.

Quote:

---

...other work is piling up and me sitting logging temps doesn't bring in the wages.

---

Well, it may be an insignificant wage, but most probably that answer will be decisive for me to buy two PA120.2 instead of only one :D

Bear in mind that that estimate only applies if you're heatload is more than the heat capacity of the 120.3

If you're not putting enough heat into the loop to over-stress a 120.3, then 2x120.2's won't particularly do any better... and it takes a HELL of a lot of heat to stress a 120.3

I know. Thing is I think that one PA120.2 may not be enough, I would have to make alot of compromises to fit a PA120.3 in my stacker although I came up with an idea to fit two PA120.2.

It may be overkill, but this is for my brother in Portugal and if I find it's not enough I will have to wait one year until coming back to portugal to finish the job.

But feel free to say that 1 PA120.2 is perfectly enough to cool down a X2 3800 and a D5 pump. Although I have no money concerns here, I don't like to waste it.

A single 120.2 will handle a 3800X2 and D5 very easily.. no issues what so ever..

Well, if you two guys say so, the noob here will just do what you suggested.

Now for noobish question of the day (so noob that I've been looking and couldn't find it here).

What CPU load temp drop (high clock, high volts) can I expect when going from a good aircooler (Big Typhoon, for instance) to such a watercooling setup?

Nice comparison Maci. I'm impressed with how well the single fan PA160 did, very nice.

Would have liked to have seen it with fans @ 12v as well.

Would not expect it to radically change rankings of the rads, but interested in seeing how it effects the temp spread between the rads, most of all with the PA160.

Thanks for this Marci !

Quote:

---

Originally Posted by amrgb

Well, if you two guys say so, the noob here will just do what you suggested.

Now for noobish question of the day (so noob that I've been looking and couldn't find it here).

What CPU load temp drop (high clock, high volts) can I expect when going from a good aircooler (Big Typhoon, for instance) to such a watercooling setup?

---

Its hard to tell ... every ones components runs a little differently. I thought id be able to atleast get 2650Mhz on my Newcastle. I'v seen some guys get decent clocks on those. Come to find out my L2 cache fails after about 2.65 sooo .. every ones chips are different. Only way to find out .. is to try it out your self.

Depends on c/w of cooler vs total c/w of watercooling system - s'the only way it can be predicted... and we don't have enough info for that.

Guess I'll have to find out by myself.

I hope I'm not spending $300-$400 to reduce load temps by only 4-5C (from 50's).

Need minimum your full current temps, what cooling you're using, what your room temperature is, what your internal case temperature is, and what hardware you're cooling.

Quote:

---

Originally Posted by amrgb

Guess I'll have to find out by myself.

I hope I'm not spending $300-$400 to reduce load temps by only 4-5C (from 50's).

---

I went from low 50s, to high 30s (under load) when I went to my modified koolance setup. A well designed system will def outperform that Im almost sure of it. So if you have plans for a $350 cooling system, you will def have some decent temps. What temps .. its hard to say rite now like Marci said. .. but as long as your room temps arent super high ... I would think you would notice a good gain.


Marci ... I have a question .. I asked a few times here with no real responce. Does a heatercore from a car perform as well as rads designed for watercooling. I love my heatercore and plan on running a second one when I get my new block and pump setup. I dont want to spend the money if the water cooling rads perform much better. They are basicly the same thing, just minor details which set them apart.

Quote:

---

Originally Posted by Marci

Need minimum your full current temps, what cooling you're using, what your room temperature is, what your internal case temperature is, and what hardware you're cooling.

---

Ok, as detailed as I can get:
- room temps about 20-22
- CM stacker with good airflow
- X2 3800 cooled with Big Typhoon, expecting load temps on the mid 50's (the cpu will arive today)
- pump Swiftech MCP655
- rad PA120.2 inside the Stacker (but I expect to get good airflow for the rad anyway)
- 2 delta 130 CFM on the rad (on a fan controller)
- Swiftech Storm cpu block
- anything more on the loop

I'm expecting mid 50's on this processor with this cooler based on what I've seen and on what my temperatures are on a Winchester @2650 @1.725v on a XP120 with 70 CFM fan on a stacker.

2 questions:

1. Is the PA120.2 enough? What kind of gains (if any) could I have on going to 2xPA120.2?

2. Roughly how many degrees C should I expect load temperatures to drop? Something around 5, 10?

I just need a rough estimate and I will take as that.

Thank you for your patience with me.