Reserator Modding Results **NOW WITH PICS**

Bosk · Post by **Bosk** » Fri Jun 18, 2004 10:46 am

::: Introduction :::

Greetings everyone,

I am the proud owner of a Zalman Reserator 1 and have so far been reasonably satisfied with it.
Noise levels are amazingly low, but it's cooling ability wasn't quite as strong as I'd hoped, so I "took the plunge" and tried modding it

Below are results and images detailing the various mods I made to the system and their effects on performance.

If you have any questions about the Reserator, any of these mods or where to buy parts, please post here or PM me

::: Testing Parameters :::

- System is an Intel P4 2.8C, Abit IC7-G motherboard, Antec SLK3700BQE case, 2x 120mm Papst 4412FGL case fans @ 5V, Zalman 400W PSU.
- CPU was run at 1.525V and stock speed for all tests.
- Only the CPU was watercooled in these tests.
- Arctic Silver 5 was used in all tests.
- The IC7-G is well known for reporting CPU temps 10C+ higher than other motherboards.
- Prime 95 was used to put the CPU under maximum load.

::: Results :::

---------------------------------------------
The original Zalman Reserator system:

- Ambient Room Temp: 22C
- Case Temp: 27C
- CPU Idle Temp: 49C
- CPU Load Temp: 54C
---------------------------------------------

---------------------------------------------
Reserator with the following mods:

* Replaced the blue 3/8" silicon tubing included with the Reserator with clear Swiftech 3/8" tubing.
* Removed the Zalman Flow Indicator.

- Ambient Room Temp: 22C
- Case Temp: 27C
- CPU Idle Temp: 46C
- CPU Load Temp: 52C
---------------------------------------------

---------------------------------------------
Reserator with these later mods:

* Replaced Reserator's built-in pump with Eheim 1048.
* Replaced Reserator's Zalman waterblock with Silverprop Cyclone Evolution SE.
* Replaced the Swiftech 3/8" tubing with ClearFLEX 1/2" tubing.

- Ambient Room Temp: 22C
- Case Temp: 25C
- CPU Idle Temp: 43C
- CPU Load Temp: 48C

During this test, the only way I could connect the 1/2" tubing to the Reserator itself was to connect a short length of 3/8" tubing to the default fittings, and then squeeze the 1/2" hose over the 3/8" tubing and secure with a hose clamp.
Because the 3/8" Swiftech tubing had an OD of 1/2" it was a snug fit, but this makeshift arrangement definately reduced the overall flow rate slightly.
---------------------------------------------

---------------------------------------------
Identical to the above, but with a final mod:

* Replaced pump & Reserator fittings (see below).

- Ambient Room Temp: 22C
- Case Temp: 26C
- CPU Idle Temp: 43C
- CPU Load Temp: 46C

I found a local supplier of quality 1/2" fittings, Hydrautech. I replaced both plastic fittings of the Eheim 1048, and both 3/8" plastic fittings of the Reserator itself with professional-grade 1/2" brass fittings. These consist of a 1/2" barb and a 1/2" thread, except the pump inlet one which required a larger thread (3/4" I think? I'm not sure I let the sales guy fit them for me).
These have definately made a difference, I think that the previous makeshit 1/2" arrangement was killing my flow rate, hence slightly better temps I'm getting now.
---------------------------------------------

::: Images :::

Please excuse the poor quality of these pics, I'm using a rather cheap webcam which doesn't react well to low-light conditions. (or any conditions I suspect)

This shows the Reserator tower fitted with the 1/2" brass fittings, and 1/2" ClearFLEX tubing.

The Eheim 1048 pump resting on a folded t-shirt, connected to 1/2" ClearFLEX with 1/2" brass fittings.

The Silverprop Cyclone SE waterblock. Note the Swiftech MCX159-R chipset cooler I'm using to cool my northbridge (you can just make it out at the bottom edge of the pic).

This pic shows the final modded Reserator setup, though it's a bit hard to make out unfortunately. Left edge of the pic is the Reserator, middle is the Eheim 1048 and in the right area you can see the case and waterblock.

Assorted fittings. Top are the 1/2" brass fittings I used, left are the default Reserator fittings, and right are some crappy plastic 1/2" fittings I bought from Mitre 10.

These are the fittings which I screwed into the base of the Reserator. they have a 1/2" barb and a 1/2" thread and fit the Reserator's inlet & outlet holes perfectly.

The Zalman waterblock with the aftermentioned fittings used to replace the default 3/8" fittings. They fit the block perfectly.

These are the standard Eheim 1048 fittings. The outlet has a 1/2" barb & 1/2" thread, and the inlet has a 1/2" barb & 3/4" thread, both with O rings.

A top down view of the various tubing I used. Left is the 1/2" ClearFLEX, middle is the 3/8" Swiftech and right is the 3/8" blue silicon supplied with the Reserator.

A front on view of the various tubing. Left is ClearFLEX, middle is Swiftech & right is blue silicon.

::: Conclusions :::

I'm now much more satisfied with the performance levels. The fact that my case is marginally cooler is an added bonus I didn't expect.

Noise from the Eheim 1048 is actually lower than the standard Reserator pump. I've sat mine on a folded t-shirt and though it vibrates a tad it's become virtually silent. I also removed the plastic cover from the inlet, and the plastic base connected to it (which is used to hard-mount the pump using screws) which also helped reduce rattling.

Performance levels have improved noticably, but weather the level of improvement warrants the cost of the project (which was significant) or the effort involved, I will leave up to the reader.

Personally I feel that an 8C drop in CPU load temps is pretty impressive for any passive cooling system, and the modded Reserator setup must surely rank as one of the most powerful passive CPU cooling systems money can buy.

Even so, my only complaint would be that the system is still not quite powerful enough to satisfy the demands of extreme overclocking, at least not compared to car heatercore based watercooling setups.

Given that most people reading this would be more concerned with low noise levels than massive overclocks, I suppose it might not be looked upon as too big a disadvantage. Nevertheless, I for one feel that to go to the trouble and expense of watercooling a computer makes me expect to get riduculously low temps and huge overclocks in return.

That aside, the system is practically silent and yet more effective than any heatsink/fan combo I've used.

Mods aside, the Reserator itself has taught me a lot about watercooling, and has allayed my apprehensions of it forever

Bosk.

Bosk · Post by **Bosk** » Sat Jun 19, 2004 8:06 am

Maxamus wrote:so the 3/8" ID - 1/2" OD are a snug/tight fit?

Any 3/8" Inside Diameter (1/2" Outside Diameter) tubing should fit the Reserator just fine.
The blue silicon tubing supplied with the Reserator has a slightly smaller OD than other kinds though, so using other types/brands of 3/8" tubing will probably result in a marginally tighter fit, but I guess that's a good thing anyway.

But maybe you were referring insted to my very ghetto method of wrapping the 1/2" tubing over a small length of 3/8" to connect it?

chylld · Post by **chylld** » Sat Jun 19, 2004 1:46 pm

i'm glad you got to see good results out of the mod AussieFellah!

going against the rules by comparing temps from different systems here: i think it might be safe to say that the reserator is the limitation in your system (no surprise here). i'm using the same pump, same waterblock, same tubing and getting much lower temps than what you've reported. in addition, my cpu voltage is at 2.06v and i have a gpu block as well.

of course, this is no surprise since yours is a passive solution and mine is an active solution, this is just an observation of what could be the limiting factors in our systems

Bosk · Post by **Bosk** » Sun Jun 20, 2004 8:38 am

chylld wrote:i'm glad you got to see good results out of the mod AussieFellah!

Thanks very much chylld

Yeah I must say I've been particularly impressed with the case temps of this modded Reserator setup.
Considering my SLK3700BQE is running with 2 Papst 120mms both @ 5V - virtually silent case cooling yet only 3C over ambient.. and thats under load! Pretty incredible stuff I reckon.

chylld wrote:going against the rules by comparing temps from different systems here: i think it might be safe to say that the reserator is the limitation in your system (no surprise here). i'm using the same pump, same waterblock, same tubing and getting much lower temps than what you've reported. in addition, my cpu voltage is at 2.06v and i have a gpu block as well.

of course, this is no surprise since yours is a passive solution and mine is an active solution, this is just an observation of what could be the limiting factors in our systems

Indeed, I think there's little doubt that even the best Reserator setup can't match a custom designed watercooling system like yours when it comes to pure performance - I've learnt to supress my natural urge to overclock the cr*p out of my system which helps me tolerate less performance I guess

If the Reserator has one advantage though, it's probably going to perform better than your Rad with passive cooling.
I'm convinced that a Reserator/medium-heat CPU/Passive PSU/9600 system could be made to run with no fans whatsoever.
A case with awesome ventilation would be needed, maybe a duct for the PSU or some kind of additional heatpipe system, and perhaps the Zalman hard-drive cooler for any drives.

In fact, I've become rather tempted to buy a proper Rad and a second pump (the standard Reserator pump hasn't impressed me) so I'll have everything I need to watercool *both* my systems, for active watercooling on my main rig, and passive Reserator watercooling on my 2nd box (P4 2.6B and SIS655 motherboard).
A fanless P4 system for my DivXs & MP3s would be so damn nice.

chylld · Post by **chylld** » Sun Jun 20, 2004 12:19 pm

AussieFellah wrote:In fact, I've become rather tempted to buy a proper Rad and a second pump (the standard Reserator pump hasn't impressed me) so I'll have everything I need to watercool *both* my systems, for active watercooling on my main rig, and passive Reserator watercooling on my 2nd box (P4 2.6B and SIS655 motherboard).
A fanless P4 system for my DivXs & MP3s would be so damn nice.

that's what i've been thinking

if you buy a proper rad and fans for it (or nick the papsts you have as case fans) then you have enough for two watercooling setups

it'd be the way to go imo; get the most out of your 1048/cyclone

halcyon · Post by **halcyon** » Sun Jun 20, 2004 11:27 pm

Excellent AussieFellah!

Thanks for the write up.

Could you comment a little bit more on the sound differences (based on your memory) between stock Reserator and the one with your last mods (Eheim 1048, not damped)?

I*m considering a similar setup for my next system, which will probably be dual + have a somewhat hot gpu.

So I _need_ a bigger pump and perhaps even dual radiators.

However, I'm not interested in going with fanned radiators (at least not yet).

Silence is paramount, but it needs to be achieved with good cooling

Bosk · Post by **Bosk** » Mon Jun 21, 2004 5:27 pm

halcyon wrote:Excellent AussieFellah!

Thanks for the write up.

My pleasure.

halcyon wrote:Could you comment a little bit more on the sound differences (based on your memory) between stock Reserator and the one with your last mods (Eheim 1048, not damped)?

Actually I'm glad you brought it up because I've noticed the opposite has happened since I've been running the modded setup for a few days.

Now that all the air has left my system and I've had a chance to address the 1048's vibration (it's resting on a folded t-shirt which is surprisingly effective) I've actually found that the 1048 is *quieter* than the pump supplied with the Reserator.
No-doubt this is mainly due to the standard Reserator pump being bolted to the inside of the tower, and there's literally nothing you can do to quiet it's vibrations.

I do think if I suspended the 1048 or encased it in a box lined with anti-sound material I could quieten it completely, but as it stands I can barely notice the noise over my two 120mm Papst case fans @ 5V.

halcyon wrote:I*m considering a similar setup for my next system, which will probably be dual + have a somewhat hot gpu.

So I _need_ a bigger pump and perhaps even dual radiators.

However, I'm not interested in going with fanned radiators (at least not yet).

Silence is paramount, but it needs to be achieved with good cooling

Two Reserators in a single loop would be pretty interesting. You could set one on either side of your rig like a pair of speakers, and maybe even mount blue cold cathodes on a wall behind each one so they'd glow... would look awesome with an expensive see-through glass desk.

Copper · Post by **Copper** » Mon Jun 21, 2004 5:46 pm

No-doubt this is mainly due to the standard Reserator pump being bolted to the inside of the tower, and there's literally nothing you can do to quiet it's vibrations.

Same T-shirt theory. I have hardwood floors and that Reserator was just a singing without a little help. I actually used two items. Folded up bubble wrap under a folded up hand towel. No vibrations!

I am, unfortunately, starting to get a little high pitched motor noise though. It's not bad but a step down form the total silence I was enjoying there for a while.

halcyon · Post by **halcyon** » Mon Jun 21, 2004 11:22 pm

Excellent!

Thanks for the update!

Copper · Post by **Copper** » Tue Jun 22, 2004 4:02 am

I don't know what's causing the problem, but it's definately coming from the pump, and it's not a hum. Maybe there is air in the pump? I've shut it of and restarted it and the sound is still there.

Bosk · Post by **Bosk** » Thu Jun 24, 2004 12:30 am

Copper wrote:I don't know what's causing the problem, but it's definately coming from the pump, and it's not a hum. Maybe there is air in the pump? I've shut it of and restarted it and the sound is still there.

Perhaps the pump's faulty?
You could try draining the system and re-filling to see if it makes any difference though I wouldn't expect so..
You also might want to check if the black base is connected to the Reserator properly, if it's loose it will cause a bit of a rattle.

Copper · Post by **Copper** » Fri Jun 25, 2004 6:04 pm

I'm wondering if I heat damaged it. I was doing some overclocking while cooling two cpu's and had cpu temps into the low 60's at times. I don't know what that would translate into water temp but I know the Reserator was hot to the touch. With the pump submerged in the water and producing a little heat of its own it must have been quite warm.

Bosk · Post by **Bosk** » Sat Jun 26, 2004 4:51 am

Ouch, Copper that doesn't sound too healthy.

I've got no idea what temperatures the Eheim pumps are rated to handle but it would be interesting to know. Somehow I couldn't imagine them struggling unless the water temp became really extreme, but who knows?

Seal · Post by **Seal** » Sat Jun 26, 2004 8:45 am

Hmmm thats 1/2 confirmed my long term doubts about the reserator, ive always been skeptical about its ability to cool well from what i have learned in my own watercooling experiences telling me that a setup like that is difficult to get good temperatures with despite other people telling me otherwise.

Congrats on the great results otherwise, i was quite interested to see the temp changes between changing different things such as the pump and the waterblock, are you sure that the ambient stayed the same as the ambient can greatly affect temps.

Bosk · Post by **Bosk** » Sat Jun 26, 2004 11:13 pm

I was always pleased with the Reserator's performance levels (compared to low-noise air cooling), but I guess when you have all the hassles of watercooling to deal with you tend to expect much better results as compensation, which was mainly why I modded it.

We're now in the middle of winter in Australia, but this room has a split system heater so I can control ambient temps pretty accurately, and I checked them many times during each test.
I would guess that all of those temps listed (not just ambients) would be accurate to one degree either way, though if any of them varied from what I listed it would more likely be the ambient temps.

I wouldn't advise taking the resulting temps to heart as being absolutely and 100% accurate, but I did try very hard to ensure each test was carried out under identical conditions, and I'm pretty satisfied with overall level of accuracy as I percieve it.

Bosk · Post by **Bosk** » Wed Jun 30, 2004 11:17 pm

Oh, in case anyone was getting confused I asked MikeC to change my board name from AussieFellah to Bosk.

Bosk is the name I use on most other tech boards so it makes things a bit simpler for me

Cheers!

Bosk.

Reserator Modding Results NOW WITH PICS

Reserator Modding Results NOW WITH PICS