Official SF800 FAQ

[Martinm210]

Thanks!
I just noticed this after I soldered this one all togther...the only thing I noticed this:

Testing my fan header on my motherboard I get pins.

Black - Ground
Red - Power measured 12.41V
Yellow -Tach/Sensor measured 10.48V

I didn't realize the sensor wire would have power?

I also just wired a 100 + 100 + 150 resistor in series to get 350, but a single 360 would have been easier.


I also assumed the grey wire with lettering on the SF800 is the ground...I hope that's right?

Have you hook yours up and blown through it to see if it reads?

[Xilikon]

Blowing thru it didn't do anything but SpeedFan isn't registering changes instantly, there is a delay of 2-3 seconds. The funnel/bucket testing did register the changes.

Interesting to know the sensor pin is giving voltage and that's probably why Swissflow have this diagram with resistors.

On the grey wire with the SF800, the colored edge is ground, the middle is sensor and the opposite edge is 5V so you need to swap them so the 5V is on middle and sensor on the other edge.

[Martinm210]

Ok, I think I did everything right, but before this electronic greenhorn hooks it up can you look at it?

The resistors I bought at Radioshack:



The voltage reading I got on the sensor wire and power wires:


This is how I figured on coming up with 350 ohms, measures 345. A single 360 would have been easier if they had them I don't know.



How I wired everything together:


And the final product:


Should I go for it?
Hook it up and watch for smoke?..

[Xilikon]

Let's go This is exactly what I did. I never thought of running resistors in serie and it would have been better because my local electronics store also have 330 ohms resistors and 20 ohms resistors.

Great pics by the way, clearer than mine

[Martinm210]

Let's go This is exactly what I did. I never thought of running resistors in serie and it would have been better because my local electronics store also have 330 ohms resistors and 20 ohms resistors.

Great pics by the way, clearer than mine

Ok, going to pressure drop and run a few tests, then I'll be back and try hooking it up to the motherboard and run a test pump loop next to my computer with both the SF800 and my King Flow meter together in the same loop.

[Martinm210]

Pressure drop testing done:

Testing in progress:


Results, (Ignor the equation, it should have read y = 0.709x2 + 0.124x):


Comparison to other blocks, it's less restrictive than an apogee GTX, probably close to an Apogee GT or regular Apogee.



It's a little more restrictive than I thought, but I think most of the more accurate flow rate meters are a little restrictive, even my King Instruments flow rate meter is a bit restrictive.

[MrToad]

Thanks again Martin

Actually your curve is more benevolent than the one provided by the manufacturer...

I wonder under which circumstances they tested their meter.

This makes my decision even more difficult...

http://www.xtremesystems.org/forums/...d.php?t=174684

[Xilikon]

Thank you for the testing results. Remain to see the flowmeter accuracy test to do then we will have a clearer picture if the SF800 is worth it

On my side, I started toying with Samurize and I'm trying to make a config file available for everyone to simplify their live with the SF800.

[Martinm210]

Thank you for the testing results. Remain to see the flowmeter accuracy test to do then we will have a clearer picture if the SF800 is worth it

On my side, I started toying with Samurize and I'm trying to make a config file available for everyone to simplify their live with the SF800.

I'm thinking I might wait for my crystalfontz cables and try hooking it up to that. Not quite sure how the crystalfontz guy did his, but with a calc program it appeared he was able to convert the pulses into an actual flow rate.

I've got my crystalfontz sitting right here, just waiting for my order for the serial cable and drive bay bracket to come in. Should hopefully be here in a week or so

I guess I'm still leary of plugging it into my motherboard, something worries me about that 2.2K resistor and the 10V coming from the sensor wire....yeah I'm paranoid... But plugging it into the crystalfontz board is no big deal if I lost that board, but my 680i I don't want to just yet...

[Martinm210]

I went ahead and tried it on one of my 680i fan headers anyway.

Plug in a fan, I get a reading in a second or two, plug in the SF800 in a simple little DDC2 > SF800 > T line loop and nothing reads.

At least not in speedfan...might have to watch health in the bios and see if it reads there.

[Martinm210]

Nope, I'm not getting any RPM readings in either Speedfan or my Motherboard health monitor in the bios. Tried a bunch of different flow rates in the loop by changing the voltage of the pump and pinching the tubing to really slow flow rates down, but still nothing.

I do have a long fan header extension cable, maybe it doesn't like all the length. I'll keep looking at it, but no luck so far..

[KVL]

Martin,
Have you tried taking out the flowmeter and test it with your DC volt meter?

here how:
apply 12v to your flowmeter, 12v to red, gnd to black. Use your volt meter measure with the black(-) of DVM (digital volt meter) to black of power, and red of DVM meter to the signal wire(yellow) of flowmeter. try to spin the blade/propeller inside the flowmeter to see if your DVM toggle between 12v and 0v.

[Martinm210]

Did a little testing with the voltmeter:

IN
12.0V ->Red
Ground -> Black

After Resistors
After 350 Ohm 11.71V-> Red
After 350 Ohm and other side of 2.2K ohm 11.71V - > Sensor

Checked current running across the meter by feeding ground to black and then bridging the 12V power going to red and get 0 mA, nothing..


I'm wondering if the grey wire with writing on it is really supposed to be power not ground, and I have them backwards?

I also did some voltage testing on the sensor wire that comes out of my pump and it's only registering like 1.3V or so, so there is definately something different here.

Or do motherboards monitor ground pulses instead of power pulses..?

[Xilikon]

Did a little testing with the voltmeter:

IN
12.0V ->Red
Ground -> Black

After Resistors
After 350 Ohm 11.71V-> Red
After 350 Ohm and other side of 2.2K ohm 11.71V - > Sensor

My first reaction was (The resistors aren't working, but then I realized they don't lower voltage, that lower current) Electronic Noob I am..

I'm wondering if the grey wire with writing on it is really supposed to be power not ground, and I have them backwards?

I also did some voltage testing on the sensor wire that comes out of my pump and it's only registering like 1.3V or so, so there is definately something different here.

Or do motherboards monitor ground pulses instead of power pulses..?

That's weird. I did the same connections and I was able to register the changes but I had to try 1-2 headers before finding the one which can accept the signal. Maybe there are various types of headers or you may have to disable the PWM option because we don't want to control the SF800 voltage. If nothing works, then I think the CrystalFontz option would be easier (just do Sensor / 6000 to get the LPM reading or Sensor / (6000*3.78) for GPM).

If you have a hard time with it, I'll drain the loop partially by clamping the tubing where I'll place the SF800. After this, I'll do my own testing to find out how things work.

Another thing to try would be to invert the ground and sensor wires on the fan connector in case this one is different. Mine has a red edge on the wire but not yours.

[Martinm210]

That's weird. I did the same connections and I was able to register the changes but I had to try 1-2 headers before finding the one which can accept the signal. Maybe there are various types of headers or you may have to disable the PWM option because we don't want to control the SF800 voltage. If nothing works, then I think the CrystalFontz option would be easier (just do Sensor / 6000 to get the LPM reading or Sensor / (6000*3.78) for GPM).

If you have a hard time with it, I'll drain the loop partially by clamping the tubing where I'll place the SF800. After this, I'll do my own testing to find out how things work.

Another thing to try would be to invert the ground and sensor wires on the fan connector in case this one is different. Mine has a red edge on the wire but not yours.

I can try a different fan head, but before and after each test I plugged a fan into and it was reading the fan's RPM right away.

I guess I should take the flowmeter apart quickly and do a quick check to make sure it's spinning.

[Martinm210]

Flow meter is spinning fine.

I forgot that resistors just lower current and not voltage. The little directions say 50 ohm/Volt. Maybe try 600 ohms on the power line?

The operating specs say 12-24mA operating range. Maybe you can check what you're getting for current across?

[Xilikon]

I'll check this in a bit but tell me how to do this as my multimeter skills is not great I will also plug the flowmeter in my loop in a bit.

[Xilikon]

I now plugged the flowmeter in my loop :



Here's the SpeedFan reading I'm getting :



However, it vary between 2-3 values and not dynamically so I'm unsure if the motherboard reading is good, maybe due to the PWM option ?

Here's my rough experiment with Samurize :



When I can get a reliable reading, I will make this meter more prettier and more clearer then I'll make the config file available for anyone who want to use it with the SF800.

Here's also a video of the dynamic reading in the BIOS :

[Xilikon]

Yes, there is a fine mesh in the flowmeter to hold the impeller in place. It's possible some air get blocked there. Mine seems to read a bit low so I moved the flowmeter in all directions and slapped it a few times. Some air bubbles escaped. It's also possible the adapter provided by Swissflow is a bit restrictive as the inner diameter looks to be around 5mm while a 1/2" barb is over 10mm.

[Xilikon]

While doing some testing, I decided to try and clamp the tube partially and the flowrate dipped into the 1,5 LPM area. When I released the clamp, it came back in the normal range.

[Xilikon]

It seems to be accurate actually and not producing random values. Here is a screenshot of the log for ~30 mins :

[Martinm210]

Great work!

I may have had air in there, I just used a very simple T-line loop, next time I'll hook it to my pump test reservoir and bleed it out good.

I should also have my crystalfontz up and running pretty soon. My motherboard RPM monitor is fairly sparatic with PUMP RPMs. I hooked a DDC2 to it yesterday and it hopped all over the place, a fan however is very good.

I'll have to look into Kapt's info above, that might be just the trick.

Oh, for current, just set a multimeter to 20m and you need to basically splice the power wire. The power coming in goes to the red multimeter probe, and the black at the other end so any current litteraly has to pass through the meter..(Just like a flow rate meter... Only reason I know is because I needed to do it for my pump testing. Been learning alot about multimeters since then.

[Xilikon]

Thanks for the multimeter trick, i'll try this (since my adapter can be unplugged, I will basically hook a probe on a end and another on the SF800 end then check the current.

Kapt's theory is also very possible and only SpeedFan was able to get a reading. Everest doesn't even show the sensor and this happen when it read 0... If we really need a diode to work, I assume it should be plugged on the yellow wire ? As I said before, electronics isn't my strong points but I understand some basics. I'll email the questions to Dirk from Swissflow to see if there are some things to consider because he said it works with the M-Cubed Sensor Hub flowmeter port (might be possible this port has a diode to work with hall's effect of the flowmeters).

[MrToad]

As far as I understand it, you need a diode between the negative and the sensor wire and another one between the sensor wire and the positive.

Bear in mind diodes are polarized, and you don't want to wire them inverted.

It would go 0 VDC = 1N4148 >¦ Signal = 1N4148 >¦ +5 VDC

The ">¦" sign is the cathode, the "banded" part in the diode.

However, I'm only an "aficionado", so if KaptCrunch would give it a look before people start blowing up their equipment due to my ill advice, that would be much better

[Xilikon]

Thanks for the details and I'll wait for some firm confirmation by a expert before attempting this mod as well. I know diodes is polarized at least