Low cost flow meter?

[zittwaredotcom]

Originally Posted by nexxo

I think the flow restriction on these is not nearly as great as people think (I mean, think about it...).

I also suspect that flow is, well, a bit overrated. I have 1.7 l/min (which I know for a fact). Yet this happily cools two Opteron 250s (2x85W) plus a Radeon 9800 (50W) on a single BIX, fanned by a low-flow Panaflo.

Agreed. But I understand their concern.

My Flow is measured at ~0.75gpm and I'm running 2 Xeon w/ 2MB cache at 3.2GHz. Each one dispating about 120W each [240W] (I run seti at home and cancer research all full time). I'm also cooling my onboard VRM with a custom water block.

My temperatures aren't great right now because of a reservior restriction issue; but at full load my cpu DIODE is measureing 54C with 100% utilization.
As this is a koolance system; I have my fans set to lowest RPM setting and the outside edge of the cpublock temperature is hovering around 43C.

The loudest thing in my case is the 550W Antec EPS-12V PSU.

[Etacovda]

Originally Posted by nexxo

I think the flow restriction on these is not nearly as great as people think (I mean, think about it...).

People often think that they are achieving the flowrates that their pump is capable of, but the reality is that by the time you add blocks and tubing this has gone down to 25%. Most people's system does not flow as fast as they think it does.

I also suspect that flow is, well, a bit overrated. I have 1.7 l/min (which I know for a fact). Yet this happily cools two Opteron 250s (2x85W) plus a Radeon 9800 (50W) on a single BIX, fanned by a low-flow Panaflo. The temps you see a few posts up were at an ambient room temp of 25C. Currently (at 21C) I get 45C and 43C for the CPUs and 28C for the system.

Some people are in it more for the cooler temperatures and are unwilling to compromise their loop with unneccessary restrictions...

I'm sure that you'll realise that around here at least 90% of people realise that their pumps arent pumping that 1200lpm at 0 head... we're all well aware of pq curves and flow restriction.

I know that my system will be pushing upwards of 4-5lpm without issue - hell, my 1048 based system was running around 1gpm, and this 50z has double the head... Id rather add another block than add a flow meter that will cause the same restriction. FWIW, im running a heatercore, 50z + mcw6002, 20 ambient, 31 load w/ 2500+ @ 2.4ghz, 1.8vcore.

Different systems, different aims.

[billbartuska]

Here are some photos of a DigiFlow 8000T installed. I removed the electronics, display, battery, and buzzer form the original case and lo and behold they fit perfectly into a 5-1/2" bay.
I put it on a bypass because it is only about 3/8" and appears to have alotta restriction. But the "cool factor" is high! I plan on using it for determining "relative" flow rates as I change parts in my watercooling system.
I found it at:
Fresh Water Systems
85 Commerce Center
Greenville, SC 29615
877 335 3339

[nexxo]

Hey, cool factor matters.

On that note, if you are able to feed your flow data to Speedfan or MBM, then you can use Samurize to display it on your desktop, or on a 5" LCD screen on your PC...

[MadHacker]

Originally Posted by nexxo

Hey, cool factor matters.

On that note, if you are able to feed your flow data to Speedfan or MBM, then you can use Samurize to display it on your desktop, or on a 5" LCD screen on your PC...

I realy Like that 5"LCD... can you give me a few details about it?

[nexxo]

It's a standard S-Video LCD TFT screen. Parts Express used to sell them but I think that nowadays you can't get anything under 6.4" widescreen (although that is good too. ).

If you're lucky you can get one on eBay, or else a PSOne 5" LCD screen. That one has the advantage that with a little modding you can hook up the VGA port.

[|kbn|]

My 1/2" flow meter idea:


What would happen if you:

Take a wide 40mm fan - although a 10mm one might do. Take it apart and disconnect the coils so it wont turn by the power when it is connected. Waterproof the inside electonics.

Cut/Drill two holes in the side of the fan casing - allowing water in and out in the right places as with normal flow meters.

Cut out a 40mm fan-sized hole in a peice of plastic and sandwitch it between more plastic, bolted togeter, with oring seals. The three wires (ground, Vdd, sense) coming out a small hole in back, sealed with some sealant..

Might need to bend fins and remove unneeded fins, so that the fan turns properly and is not restrictive.

This should all be easy, but one problem: I expect some of the parts of the fan are not water freindly - the bearings? the permenant magnets, etc..
I think this can be worked around though.

I will try making one sometime, I should be able to make a leak-proof fan housing to test if this would work.




Also http://www.sixdesign.de/sflow.html is the first site I had seen that shows a non-commercial flow meter using led/photo transisor. Not sure if they sell them but some of the other things shown on that site are sold online.

[Risky]

Drifting back on-topic....

Dave, what's the latest on this flow meter?

[Dave]

^ No CNC time, no injection time at either company

AVT and C-Systems have been overcome with OEM requests.

AVT is moving alot of there production equipment from US offices to Canada, and expanding Canadian operations, this is part AVT International seperations of companies back in January.

C-Systems has finished there expandsion, however this was planned only for our parent companies internal needs, and has already proven too little.

AVT Canada is expected to have the required capacity mid / end of Aug, and I hope to have machine time again then.

We have two additional machining centers (one new), and a new additional injection machine.

Yes, good news for the companies, not so good news for my project.

[RaptorRaider]

Originally Posted by billbartuska

Here are some photos of a DigiFlow 8000T installed. [...] I put it on a bypass because it is only about 3/8" and appears to have alotta restriction.

What are you basing this on?


Anyway, I've got one too, and since I recently acquired a decent manometer I guess I could try to make a P/Q curve if anyone is interested.


Also good luck with your project Dave.

[Mathelo]

I've been following this thread with interest as I'd like to include a flow meter or switch in my WC setup for protection but without hurting waterflow or incurring great expense. I've looked at getting the Swissflow and not bad at 50 euros but they only take payment by wire transfer. Wiretransfer will end up costing almost as much as the flowmeter itself.

Isn't the simple answer to install an inexpensive meter / switch in parallel? Certainly won't give accurate flow rate but I don't really care much about that. I'm after system protection.

Louis

[nexxo]

Then you can use a flow switch. Gentech makes some useful ones.

[CoolROD]

Just wanted to post a link to Gems Sensors Paddlewheel Flow Sensors . I haven't seen it mentioned in this thread, but I have talked about it here before. They are available with pulse train output (Which I bought), Visual indication only, and mabye contact closure at some preset. They wouldn't serve for many scientific purposes, but suited my application.

I can dig up the invoice Monday, if necessary -I believe they charged a nasty handling fee.

[Mathelo]

Can someone help me understand this chart - http://www.gemssensors.com/pdf/380PFlowChart.PDF for this flow sensor - http://www.gemssensors.com/PDF/Catalog/FS380P.pdf?

My WC will include:
Swiftech Storm CPU Block
Laing D5 pump
Thermachill PA 160
1/2 tubing.

If I've properly interpreted other guidance I've read on this forum, I should expect to see 5-6 lpm of flow with this configuration (without the flow sensor). What kind of hit can I expect to see if I put this flow sensor in the loop in series? For the 3/8" NPT port, the chart indicates .5 psid (.035 bar???) at approximately 6 lpm. How does that translate? Do I have the psid to bar correct?

Putting the flow sensor in parellel remains an option but complicates the loop. Moreover, I'm still considering purchasing the Swissflow800 (even though it will be stupidly overpriced after adding in the wire transfer fees) but with my TBalancer I like the idea of being able to control the pump along with all the other noise generation devices in the box. To do that effectively it seems that I need to have a fairly accurate read on flow.

The Swissflow pressure drop is approximately .25 bar at 6 lpm. That seems significantly worse than the flow switch but may still be immaterial.

Thoughts?

Louis

[Mathelo]

Or maybe a low cost pressure switch like this would be even simpler?

http://www.gemssensors.com/PDF/Catalog/PS31_0321.pdf

$17.00.

[bobkoure]

Originally Posted by Mathelo

Or maybe a low cost pressure switch ...

Hmmm... 5 to 150 psi.
And the commonly-used-for-PC-watercooling pumps that generate at least 5psi?
mcp 600/650/350 and maybe the 50Z (same as the 600?) - and I don't think any of these is rated for more than about 5.5 psi
So maybe not this particular switch.

[Mathelo]

Originally Posted by bobkoure

Hmmm... 5 to 150 psi.
And the commonly-used-for-PC-watercooling pumps that generate at least 5psi?
mcp 600/650/350 and maybe the 50Z (same as the 600?) - and I don't think any of these is rated for more than about 5.5 psi
So maybe not this particular switch.

Hmmm, I'm confused. I'm looking at the Laing pump. It says Max Pressure: 50 PSI (3.5 BAR). Is that just a safety rating? Working pressure is actually 10% of that?

More research on the math ... foot of head x .433 = psi. Looking at the Laing website this pump has about 7.5 ft of head at 6 lpm so that should equal a working pressure of just over 3 psi. Did I do that right?

Assuming that is right, then I'll need something more like this:
http://www.gemssensors.com/PDF/Catalog/PS11.pdf. Goes down to .55psi. Unfortunately, no price is given on this one and it is a bit larger but should be workable.

Louis

[bigben2k]

The max pressure rating there refers to the max pressure of the pump housing; it has nothing to do with the pump's ability to move a fluid. i.e. if the pump increases the flow pressure by 1 psi, it can do so from 25 to 26 psi (relative to atmosphere).

[dacooltech]

Nexxo: I hooked one up to a fan header, told MBM to divide the signal by six, and presto: flow in millilitres right there on my desktop.

Nexxo if I'm not mistaking the motherboard fan header uses /2 for fans. So I think the Fan Type needs to be set to /3 instead of /6 in MBM to get the correct l/m value right?

Then you get 3.514 l/m instead of 1.757 l/m which equals to 0.92 gpm, for example... (nice LCD BTW )

Joe, pH, Rosco since you guys are the SF 800 experts, pls correct me if I'm wrong...

[CoolROD]

Quick search here for Dwyer Flow returned nothing similar to this...

One of my trade publications posted an ad for a Dwyer Instruments paddlewheel-style flow indicator with output sensor and visual sight indication. Solid brass body w/ tempered glass window and Red PBT paddlewheel. Supposedly made of UV-Stabilized materials for outdoor use.

Dwyer Sensor Line

SFI-100T (Advertised Sensor)

FYI --->These are also pricey.

[JSimmons]

Is this flowmeter project dead?

[RaptorRaider]

Probably.

The market is most likely too small anyway.

[zittwaredotcom]

I'm still planning on providing my sensors. I just haven't had the time to finish the shopping cart among other things.

I have been building one-off sensors for various modders / worklogs. If your interested in more information contact me by email at zittware at hotmail.com.

[stev]

The subject here was LOW COST FLOW METER.

I had posted this in the other forum thread.

This is a typical $10 dollar part.

http://www.kartell.it/uk/labware/sch...?IDProdotto=21



http://www.cynmar.com/product_info.p...49b83c0e13f255

http://www.highspeedpc.com/Merchant2...de=FanTempFlow

Here's what the Kartell Paddle Wheel Flow Indicator looks like next to a 120mm fan in the Corsair HydroCool 200.



Stev

[Brians256]

That's a nice design. Cheap and easy to mfg.

Now, if you could hook that up to a photosensor circuit that shuts power off if flow is impeded for more than <insert user preference> seconds, that'd be killer.