Sign in with
Sign up | Sign in
Your question

Innovatek Flow Meter Rev 2.0 : Restrictive?

Last response: in Overclocking
September 21, 2006 3:38:50 AM

You should look at the internal hole on this thing... Im going to run some tests... but I would bet you that this is the single reason why my watercooling isnt as good as it could be... Im so bumed... The slot is the width of a penny and the hight of the internal tube...

Its such a great way to test and see if the pump is running... Im bummed... anyone see any tests on it?

Im going to run some with my machine this weekend... I fear this thing is a complete waste of money!

Any Ideas?
September 21, 2006 5:12:26 AM

What is the tubing size that you use in your cooling loop?
September 21, 2006 7:06:19 PM


September 21, 2006 8:43:21 PM

Just so I will know for future reference, Why don't yoiu list your water cooling equipment...

If you are using a kit, let me know what kit it is. 8mm (1/4ID) is highly restrictive. Perhaps there is somehting that can be done to improve on things.
September 21, 2006 11:17:59 PM

YES, that thing is VERY restrictive, the only flow meters that don't restrict flow as much are high end industrian quality flow meters, and those run up to 1000-2000 dollars
September 22, 2006 1:50:14 AM

It was a custom case from Germany... it was real nice...

Its called an "Alphacool Ready 2 Go Coolermaster Stacker"! I like it... Just working out the kinks. You cant find it in the states... I got it on super discount, cause I live near frozencpu... they had a bunch down there... but they werent going to carry the box, cause of the poor markup... so he gave it to me for $400

The equipment isnt available from stores... seems kinda custom made for the've seen the flow-meter...

The Pump I cant find... its a submersive came with the coolermaster...
The Label Reads
Eheim Typ 05092
12v 50hz 11w
HMax M 1.3
l/std 600

I got a
Alphacool NexXxoS XP CPU Water Block - Socket 775

And I got the DangerDen
Danger Den Tyee ATI X1800 / 1900 Video Card Cooler

and a Radiator...

PrimoChill Ice Non-Conductive Liquid Cooling

Lastly a
...Which isnt restrictive...

... but I only bought it cause I got....

and wasnt told that it wouldnt fit my shit... turns out the temp gaudge on that last one is REAL restrictive... But the LED gaudge is cool! It fits right between my two switches...

Not much I can do... I dont have much more money... when I had it running the first time (before I swapped out MoBos) the temp wouldn't go past 110f

Now, It idles at 110f... I got to try reseating the waterblock and redoing the articsilver...

I get 10,000-11,000 RMP from my Innovatek when added with

...which translates to 1-1.1 L/min

...and I dont know if I will get any better then that... I was looking at frozencpu and they dont have many pumps that go beyond that... and they are all really expensive...

Im going to test that damn flow meter... but I think its reporting the pumps maximum flow... Though I cant find any info on the pump.

So I might not get anything more out of it... I have no Idea what size or restriction the radiator is... The machine is silent, until the powersupply comes on... Really pisses me off... I bought a wonderfull tower from them and then ask them what powersupply. The tower I buy is silent... well its not...

Im really not happy with frozencpu... They just sell allot of shit, and when you ask them for help, it seem like you come out with a bunch of shit...
September 22, 2006 6:03:24 AM

Yes, perhaps if you could find a more powerful pump. I would imagine that would make a difference - especially with the Tyee GPU block that you are using as it is a very restrictive water block in its own right - couple that with the CPU block that you are using and your water flow must be straining to make it through and maintain some sort of effectiveness.
September 22, 2006 12:59:13 PM

What do you recommend as a 8mm pump...

I probably can go 1/4" if I could find out what type of radiator I am using... It would be useless to go 1/4" if my radiator isnt 1/4" or more...

I think I got all of the 1/4" fittings... I would have to double check, but Im pretty sure... except for the radiator... I could ask the guys at frozen Cpu... since I would probably buy the pump from them, unless I can find it somewhere else for allot cheaper... Ill try to find the radiator for this box...

September 22, 2006 1:13:05 PM


Are you saying that this pump is rated for 600 l/hr?

thats 10 l/min

I am reporting 1 l/min on my flow meter... I am willing to bet that if I remove this flow meter and test it, that I will get a huge difference in flow rate...

All of the parts I bought are for 1/4"... the most restrictive part is my hosing. The pump I have, is only set up for 8mm, I dont think I can put anything bigger in it... Like I said, I dont know what the radiator is... so I cant say if raising the size of my tube would do me any good. I dont think any of the parts I am using are causing the restriction... I truely beleive that this flowmeter is the problem...

Maybe Im missreading it the flow meter... I asked a guy on this website to read the instructions (they are in german)..
he said that every 1000 rpm's reported with the gaudge are .1 l/min

that means Im getting 1/10th of the pump...

It is almost impossible to beleive that any other part would suck so much from this pump... if it is infact 600 l/hour...

That also means that my pump is under HUGE strain, and that it will die if I dont fix it... Im going to do the soda bottle test this weekend...

2/l and time it... with and without the flow meter...

I can squeze my hose quete a bit before I even get a drop in the reading from this device... If this device is causing these issues... then I am going to take it back to frozen CPU... if they give me shit, I am going to bad mouth them from now till I die...

September 22, 2006 2:03:05 PM

Maybe you could try plumbing in a bypass? Install a tee before and after the flow meter and have a length of pipe inline with the flow meter. The fluid will still flow through the flow meter but you'd also resolve the restriction problem. You will probably get bad/low results from it at that point, but I imagine you're going to find that the flow meter is restrictive with the soda bottle test.

Also, in response to how you were told to interpret the readings meaning like 1000rpms equals .1Lpm that could also be wrong. If the meter is maxed out then maybe more fluid is passing through then reported. It will only spin so fast. I don't imagine these instruments are engineered by someone who specializes in hydrodynamics.
September 22, 2006 2:11:09 PM

One thing about your equipment - specifically, your rad, CPU water block and GPU water block - do they use fittings based on, say, G1/4? Or, are the connectors soldered onto the components themselves?
September 22, 2006 4:17:10 PM

The cpu waterblock is a 1/4" waterblock, and had been fitted with a 8mm when I bought it... it has screw in fitting.

The Dangerden VideoCard cooling had 3/8" fittings, and I had to put 8mm replacements in them... they leaked a bit, but not enough that a little pipe tape wouldnt fix.

The radiator and pump both were 8mm. The system came with the radiator and fans installed, with a pump and the waterblock... all of it was 8 mm... thats why I stayed 8mm.

I have no Idea what size the Radiator channels are, and if I could find out what the radiator is, then I could find out if it has 1/4" defaults.
It would do little or no good to increase the size of the tubing if the radiator is small. I really have to look.

The pump is definately 8mm... though I will look at it, and see if the fittings are replaceable with 1/4... I have fittings for most of the items, which will support 1/4"...


the .1 l/meter = 1000rpm is in fact confirmed...

Luckily someone who lives in Germany helped me understand the instructions... and Fixed the errors that google left behind...

Turns out thoose strange germans use "," instead on "."

------- Water Compression and the wheel -------
if the wheel spins only so fast, then the water can only go though so fast... Liquid doesnt compress... It stays the same volume no mater how much pressure you put on it (well unless your talking thousands of tons, in which it will compress so little that it is barely noticable)

------- Basic elements water has to cool ---------
The only things water coolers have to work with is water volume, water velocity, surface area, and turbulance. Water Volume is a constant... cant change it..

--------- Water Velocity --------
Water velocity matters because it dictates how long the water stays ontop of the hot spots... The other side of it, is that the faster you move water, the less time it has to sit in the radiator. The perfect velocity would be one that would create a situation that makes water going into the radiator as close to the same temp as the water coming out. This is also saying that the water going into the heating surfaces is near the temp of the water leaving the heating surfaces.

as a side note:
The amount of time the water spends in the radiator and the heatsource stays constant unless the velocity of the water hits 0... As you increase velocity the time the water spends on the heatsource and the radiator is the same... it just cycles faster.

I read a good article explaining this... The example was this... an oval track with a race car, speeding around the track at 60 mph would spend the same time on a specific 100 foot spot area as a car traveling at any other speed. This is because a 60 mph car would go around the track slower then a 100 mph car...

-------- Cooling / Heating Properties -------------
If the water temp exiting the radiator is near the temp of the air, and the heating elements remain hot, then the only way to increase cooling compacity is to increase water <-> coper surface area, or to move the water through the system faster. In physics these two constants are the same.

If the temp of the water is higher then the temp of the air, then more water <-> air surface area is the only way to drop temps...

------------- Laws of Diminishing returns --------
Now without a waterfall size cooling system the ability for the water system to be near perfect (water temp entering is the same as water temp exiting and water temp is near air temp) is near impossible

... it follows the law of diminishing returns.

----------- My Conclusion ---------
What makes me confident that it is the velocity of the water that causing the problem is that the temp of the water will hit 90f after a few hours without the waterblock on the videocard even connected. And that it reads hot exiting the system even at idle... BTW, this temp is before it goes into the radiator! This tells me that the water entering the waterblock is sitting on the waterblock for too long without entering the radiator...


All of the equipment except for the radiator is confirmend to have flowrates that support above or at 1/4" tubes... The Videocards DangerDen channels are so big, that I would have to conclude that their is almost 0 resistance to it.

Now resistance is related to turbulance... and it must be noted that in some circumstances turbulance can increase the rate at which things increase or decrease the amount of heat absorbed... For example... if the water had 0 turbulance... then it would be like a car... only the bottom portion of the water would touch the hot street. Now if you shake things up, you will find that the water would mix and slosh, allowing more of the water to touch the surface.

All of this makes compairing water and radiator elements very dificult. If you tested the same water block in 3 different systems and compair it with another water block you may get mixed results on which ones the "best"... So benchmarks are not accurate unless they use multiple systems...

Note: The exchange of heat is measured in the loss and gain of calories.
September 22, 2006 6:37:11 PM

Maybe you could try plumbing in a bypass? Install a tee before and after the flow meter and have a length of pipe inline with the flow meter. The fluid will still flow through the flow meter but you'd also resolve the restriction problem. You will probably get bad/low results from it at that point, but I imagine you're going to find that the flow meter is restrictive with the soda bottle test..

The Bypass is a good idea, if the soda bottle test fails... The issue with the bypass is to present just enough back pressure on the alternateline to push water in through the more restrictive path... I feel this would be very hard to do, without a huge supply of very different sizes hoses, and the Y sections needed. I would assume that the parts are not availble for the sizes needed, and that this would take a large amount of test parts...

If you build too much of a restriction, you get low flow rates... too little and the flowmeter doesnt work...

It seems to me that 1/4" is the most popular... 1/2 inch is cool, and 1 inch is totally impracticle.. I looked at my radiator, and it probably would go to 1/2 inch, but to do that, I would have to buy a new waterblock and a new pump... It is an expensive proposition, since some of these high flow pumps are hundred plus dollars...

I also made a mistake, in my above assumptions... I thought 8mm was smaller then 1/4 inch...

1/4 inch = 6.35 millimeter
Im running 8 millimeter pipe... The 1/4 inch waterblock with 1/8th inch adapters may be a restriction...

1/2 inch = 12.75 milimeter

Im assuming this is the pipe internal diameter

Which means
1/2inch: 126.677mm area
8mm: 50.265mm = area
1/4inch: 31.669mm = area

You can see the differences a few mm makes...

This makes me very confident that the flow rate available for a this system should be able to cool the processor more then what it is...


as far as 1/2" the flows are amazing

$100 =
900gph = 3407 l/hour

Thats a 4x improvement if I could get the full potential out of my pump... or if we could assume that both systems would have the same % of restrictions on both sizes...

September 22, 2006 6:50:36 PM

Humor me... what is the best 1/2 inch cpu block for the 775?

September 22, 2006 6:58:20 PM

While this is just a generalized calculation, head pressure (basically, the resistance a pump encounters to its flow rate) is 1 foot per 20ft of horizontal tubing length and 1 foot per 90 degree turn. This will show why certain water blocks are more inclined to greater resistance to flow than others. A pumps head rating is, basically, the greatest height to which it can maintain a certain flow rate. Beyond that rating it begins to lose its effectiveness.

8mm = 1/4ID
10mm = 3/8ID
12/13mm= 1/2ID

In the U.S., ½ ID is the inner diameter of choice, in Europe, it’s ¼ ID.

That Hydor pump is powerul but, if you look at the specifications, it's roughly the size of a volkswagen...

5 inches long
5 inches tall
almost 3 1/2 inches wide

Not something that would be practical for a case and the 900gph is more than overkill, it's ridiculous.
September 22, 2006 7:02:07 PM


Well, for a system with a powerful pump (MCP655) the Swiftech Storm is ideal. It needs a stronger pump because it is restirctive in its design. For a cooling loop with a weaker pump (say, the MCP350), the Swiftech Apogee is ideal because of its low flow resistance characteristics.
September 23, 2006 1:06:27 AM

Soundslike with my 6 HD's, 3 DVD Readers/Burners, Sounblaster X-Fi front, and my Water temp reader + switches, I have very little room in front of the case... um like 0...

Anyways, the big pump WOULD be nice, but we would be talking a $200 increase in price... IF the radiator wouldnt restrict the hell out of it...

And that doesnt solve the issue of WHERE too put the damn pump and resevor... Its not like hanging it off the side is very attractive:( 

On top of my case I got the fans and the radiator... I was thinking of making a cover for it, to prevent spilled water from getting in my machine... It would be cool if that cover could be a resivor.

I could fit the VW in the bottom of my case... though it may be tight..

anyways, its a bit of dreaming... I always wondered if I could run a Peltier unit off one cycle... didnt think so.. but hey...

My CPU temps are just too high... they shouldnt go above 60 at how I am configured...

BTW, I misspoke above... Callories are the amount of heat it takes to warm a substance... 1 Callorie of energy will warm 1cc of water 1 degree.

When I talked to my father he talked in terms of BTU's, which is another measurement... I wonder if BTU's are considered metric...
September 23, 2006 6:59:35 AM

Peltiers - effective ones anyway (say 226 - 350 watt for a CPU) - would require it's own seperate cooling loop for certain. My own is a 226 watt TEC unit. Peltiers get extremely HOT on their hotside and for their coldside to be effective on the CPU, you must remove as much of that heat as possible. I have a ViperVenomIII GPU peltier water block that was around 245 or so watts. I gve it its own loop also but the resulting added heat to the coolant (I ran two seperate cooling loops, unified only at the reservoir) was very noticeable at the reservoir - this with running two external dual 120 rads with four 120mm fans on each). Due to a faulty videocard (XFX 7900GTX Extreme) I had to remove it though. Now, I am just watercooling the GPU and have only a TEC unit on my CPU.

Peltier units are supremely effective. Even under load playing Doom3 at max settings my CPU stays at 0 degrees (This on a Presler 955 core - I have yet to install my Conroe as I am waiting on some related parts).

Peltier cooling produces very nice results but, the water cooling involved for the TEC must be configured as efficient as possible. I wouldn't use anything less than 1/2 ID tubing (and related ID parts) for the loop. It is the next generational step up from watercooling (and is much more practical than phase change).

You could run a peltier in your cooling loop if you adapted to a larger ID configuration (1/2 ID) and dedicated the loop to only the CPU.
September 24, 2006 3:20:06 PM

As an Answer to my Topic... I found out some answers...

First the flow meter is reporting an accurate number... 1 l/min.

It is not restrictive at that flow (or very little).

Outside the loop, my pump pushes no where near the 600l/hour standard... but more like 2-3 l/min, and at that rate the Flow Meter is very restrictive cutting the flow by almost 1/2.

This is very sad news for me... However I am not sure of how effective the system is, because I havent been supplied with temp numbers associated with the X6800, and the settings it works with... It is also difficult to know if allot of my heat, when I overclock it, is because of the MotherBoards poor bios/design. Anyways... When I reset my bios and run the machine at default speeds (changing only the memory timings and the memory voltage) I get around a 100f idle, I got to test my max still...

Does anyone know of resources that give temps for X6800 on different cooling setups.

If the watercooling is less effective then the aftermarket aircooling, then I would increase the flow, and increase the fanspeeds on my radiator. Right now it is a waiting game to see how the next bios version works... And to gather data on other peoples temp results...


Any help finding temp numbers under prime95 would be great...