PDA

View Full Version : SF800 Review and FAQ


Xilikon
12-11-2008, 09:56 AM
Since I will be getting mine very soon, ladderman is toying with one currently and few others already own one (either from a former group buy or ordered directly from Swissflow), I think it's time for us to have a official thread about the SF800 flowmeter, especially when I will send mine to Martinm210 for some pressure drop testing and possible review. I already feel that this will bring more interest on the SF800 :)

First of all, here is the specifications of the SF800 as taken from Swissflow (http://www.swissflow.com/) :

Technical Specifications (http://www.swissflow.com/en/SF800/Flow_Meter_Specifications)
Data based on water at 20°C (68°F)

Range
Flow velocity*: 0.5 – 20 litre per minute
Temperature: -20 – +90 °C
Operating pressure: 16 bar
Max. pressure: 40 bar

Technical information SF-800
Process connections: 3/8” hose barb; 3/8”BSP Male
Exitation: 5 to 24 VDC, 12 to 24 mA
Power consumption: 12 – 36 mA/s
Material: PVDF, Vectra (rotor), Viton or EPDM
Output frequency: 100 to 2000 Hz (depending on the flow velocity)
K-factor: +/- 6100 pulse/litre
Length cable: 15 cm standard (different length on request - Dirk confirmed that you can order with the needed wire length, just put a note)
Connector electronics: 3-wire flat cable sealed in housing (jack-plug or molex connector on request)

Reliability
Interchangeability**: +/- 2.25 %
Accuracy: +/- 1.00 %
Reproducibility: +/- 0.30 %

Medium
Medium type: clear or translucent liquids capable of transmitting infrared light
Viscosity: 1-1000 Cst
Examples: water, chemicals, oil, beer, syrup

How the SF800 work (Working Principle) ?

The geometry of the rotorhousing (causing a radial flow in) in combination with the light-weight Vectra rotor (0,04 grams) makes the rotor float in the medium. Because the rotor floats, the measurement process is nearly resistance-free, guaranteeing high accuracy, durability and linearity .

An integrated PCB emits an infrared signal, which is interrupted by the three blades of the rotor. SMD components interrupted transform the signal into a pulse. Because of the construction of the rotorhousing, the Swissflow flow sensors are practically insensible to high pressure.

The materials used (PVDF and Vectra) are to a great extent chemical-resistant and temperature-durable . Contrary to many other fluid flow measurement systems, the Swissflow sensors do not have to be calibrated .

There is also a pps presentation made by Swissflow to introduce the SF800 available here (http://pages.videotron.com/xilikon/LC/SF800.pps).

Here is the internal assembly diagrams :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1275.jpg
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1276.jpg

Why it's a great flowmeter ?

So far, it looks like it's one of the least restrictive flowmeter in the market, which is perfect for the current watercooling loops. ladderman did some rough testing and it looks like the same restriction as a EK FC8800 block (with 1/2" tubing, possibly almost nil with 3/8" tubing if further testing confirm this).

Is there any reviews about the SF800 ?

Here is a few we found so far :

effizienzgurus.de article on flowmeters (translated) (http://www.google.com/translate?u=http%3A%2F%2Fwww.effizienzgurus.de%2Fm ain%2Findex.php%3Fcontent%3Darticle%26action%3Dvie w_spec_article%26article_id%3D134&langpair=de%7Cen&hl=en&ie=UTF8)
ladderman's crude review (http://www.xtremesystems.org/forums/showthread.php?t=169956)

How to connect the SF800 to the current loop tubing ?

First of all, you need to read the guidelines in the leaflet back for some precautions (it come with a 10 year warranty so it's wise to follow them) :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1273.jpg

The SF800 come with a male 3/8" BSPP on both ends so we need a coupler to be able to use with a G1/4" set of barbs. Here is some possible sources for those couplers :

UK : http://www.airlines-pneumatics.com/webcat/Detprod.asp?ProductCode=s010104 (order code 103031)

US : http://www.mcmaster.com/ (item code 4860K675, not to be confused with 4860K657 which is a G3/8" to G1/4" adapter for Thermochill radiators)

Canada : <Currently attempting to source a online store, possibly http://www.adaptall.com/ to be confirmed by a phone call>

Alternatively, we can also order a 3/8" to 3/8" coupler and use a pair of Thermochill barbs sold by NCIX.

How can I monitor the SF800 signal ?

Currently, there is 3 possible ways to obtain the signal feed from the SF800 :

-Plugged on the motherboard fan header.
-Plugged on the T-Balancer Sensorhub.
-Plugged on the Aqua Computer PowerAdjust (radiical_53 commented about this possibility, to be confirmed).

When do I use the resistors as asked in the schematics sheet included with the SF800 ?

If we want to connect the SF800 on the motherboard fan header, you need to solder 2 resistors to make it work and the diagram is found on the SF800 installation guide included with each bag. The 50ohms/V mean you need to calculate 50 ohms for each volt you want to drop and the difference between the fan header and the 5V feed of the SF800 is 7 volts so you need a 350 ohms resistor.

Here's the picture of the schematics included in the package :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1271.jpg

With a M-Cubed Sensorhub, the resistors is not needed but you need to make sure the sensor plug wires is ordered like that : GND - 5V - SENSOR when viewed from above (In other words, the ground wire should be on your left and the sensor on your right). If you are unsure, just look at a fan lead and you will notice the black (GND) is on the left, red (power) on the middle and yellow (RPM) on the right.

For others, I'll add more informations as more people confirm the workings.

How can I get the proper LPM/GPM reading ?

<To be filled>

Is there any worklogs of users with a SF800 in use ?

Yes, there is at least one with a CryztalFontz LCD : http://www.crystalfontz.com/forum/showthread.php?s=&threadid=4452
Custom mod for readingthe flowrate on a custom monitor (electronic soldering skills required) : http://www.turbokeu.com/myprojects/flowmeter.htm
<more to be added - post or PM me if you want to add yours>

Be noted this FAQ is a work in progress and will be expanded with more informations coming from me and other SF800 users. If you have extra informations, comments or useful links about the SF800, please post and I'll add in the thread :)

Here (http://www.google.com/translate?u=http%3A%2F%2Fwww.effizienzgurus.de%2Fm ain%2Findex.php%3Fcontent%3Darticle%26action%3Dvie w_spec_article%26article_id%3D134&langpair=de%7Cen&hl=en&ie=UTF8) is a translation of the german efficiencyguru's article on flowmeters, including the SF800.

Working Principle

The geometry of the rotorhousing (causing a radial flow in) in combination with the light-weight Vectra rotor (0,04 grams) makes the rotor float in the medium. Because the rotor floats, the measurement process is nearly resistance-free, guaranteeing high accuracy, durability and linearity .

An integrated PCB emits an infrared signal, which is interrupted by the three blades of the rotor. SMD components interrupted transform the signal into a pulse. Because of the construction of the rotorhousing, the Swissflow flow sensors are practically insensible to high pressure.

The materials used (PVDF and Vectra) are to a great extent chemical-resistant and temperature-durable . Contrary to many other fluid flow measurement systems, the Swissflow sensors do not have to be calibrated .

A bit of impulse info translated from german in the M-cube SDK -

How can I interprete the values from the T-Balancer?

- FanMaxSpeed: add the low- and high byte to a 16bit value (256*HB+LB), multiply it with 10.5 to receive the value for turns per minute.

- Fan RPM and Volt are only interpolated values. RPM is linear interpolated from maximum, and voltage is quadratic interpolated (results from root mean square from ideal rectangular signals)

- The flowmeters are a little more complicated. You receive a simple impulse count value. But this value is dependent from the measuring windows (1 to 4 seconds) – 235 TWI_SHWerte[10]. The 232 and 233 then means the measured pulses per second, 2 seconds, 3 seconds or 4 seconds. Divided through the window width you get the pulses per second, multiplied to 60 seconds you get the pulses per minute. To get the liter per minute or hour you have to respect the adjusted flowmeter dependent pulses/liter value. Our standard flowmeter have 509pulses/liter. This value can be found in the TBAN database on the table “Einstellungen” with the keys “Durchfluss1” and “Durchfluss2”.

-Getting the pulses is similar to pumps and fans, but not exactly. It is only usable without damaging something, if the impulse function works with an open drain circuit. Standard fans uses the 12V for reference, the flowmeters are not powered their self and receive only 5V from the sensor hub. The signal input on the sensor hub is only build for 5V input. You only have to connect ground and pulse with a pushbutton to get an impulse.

Xilikon
12-11-2008, 09:57 AM
I got the SF800 package from Dirk last friday so here's the goodies while waiting for some tests :

Whole package contents (notice 2 SF800 given as samples :eek: ) :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1270.jpg

Package leaflet with the schematics and the resistors to add if you are plugging in a header feeding more than 5V :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1271.jpg

Back of the same leaflet :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1273.jpg

A shot thru the hole to see how big :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1272.jpg

Here's mine with the adapters fully assembled using 1/2" fittings (in reality, it looks more like 3/8", which is a bit restrictive) :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1274.jpg

A pic of the flowmeter header (the wires order is not correct and should be GND - 5V - Sensor and not 5V - Sensor - GND, to be fixed later) :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1277.jpg

Shots of the internals drawings in the included booklet :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1275.jpg
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1276.jpg

I have packaged the 2nd one to be sent to Martinm210 for pressure drop and flowmeter accuracy testing so expect the results in a few weeks (he has all the time since I have my own SF800 so no hurry).

Look also in the first post with the same pics to be used as reference.

Xilikon
12-11-2008, 09:57 AM
I finally got around to buy the needed resistors so here's my small HowTo with pics on building a adapter :

Here's what you need :

-2200 ohms resistor (the one with 3 red lines)
-350 ohms resistor (if the store doesn't carry them since it seems to be non-standard, get a 360 ohms since it's only a few millivolts difference)
-A fan extender wire that you can buy at any store. Mine is coming from my now defunct VLSys L.I.S.2 LCD display and it's very long.
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1313.jpg

First of all, since the 2 resistors will meet at one end, just twist them together :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1314.jpg

After this, splice both red and yellow wires (red is +5V and yellow is sensor) then connect one end of the 2200 ohms resistor to the yellow wire. Connect one end of the 350 ohms resistor to the red wire and connect the twisted end to the other end of the red wire. Make sure the yellow wire is continous. Here is a picture about how it all come together :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1315.jpg

Since this will be part of my own project, I also sleeved them for a neat look :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1316.jpg

That's all. I did a preliminary by connecting the SF800 to 2 bits of tubing, put a funnel at one end. I connected the SF800 on a fan header and opened SpeedFan. Then I poured water thru the sensor and in the bucket while looking at the fan sensor. Depending on the flow, the speed is fluctuating meaning it works. Remain to put it in a closed loop for a real-life testing and that's what Martinm210 will do as I won't disassemble my loop before receiving 10 feets of R-3400 tubing.

Martinm210 is currently preparing the 2nd one with very nice adapters in delrin (a real gem IMHO) and we are talking about the resistors. The HowTo I posted will also help him setup the same thing.

Xilikon
12-11-2008, 09:59 AM
Pressure drop testing done:

Testing in progress:

http://i30.photobucket.com/albums/c338/Xilikon/LC/sf8008ug0.jpg

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

http://i30.photobucket.com/albums/c338/Xilikon/LC/sf800pdresults1gx8.png

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

http://i30.photobucket.com/albums/c338/Xilikon/LC/sf800pdresults2ps4.png


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.

Xilikon
12-11-2008, 10:00 AM
I now plugged the flowmeter in my loop :

http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1317.jpg

Here's the SpeedFan reading I'm getting :

http://i30.photobucket.com/albums/c338/Xilikon/LC/SF800-Speedfan.jpg

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 :

http://i30.photobucket.com/albums/c338/Xilikon/LC/SF800-Samurize.jpg

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 :
http://i30.photobucket.com/albums/c338/Xilikon/th_DSCF1318.jpg


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

http://i30.photobucket.com/albums/c338/Xilikon/LC/SF800-Log.jpg

Here's a alternate method of connection using a separate molex to receive the proper 5V feed :

http://i30.photobucket.com/albums/c338/Xilikon/LC/SF800-Connection2.jpg

Ok, I created a new adapter based on a new idea of tapping 5V from a 4 pin or floppy molex. For this, I scavenged the cut off excess wire from the Yate Loon that I did for the radiator fan mod. Here is the cruddy result (I didn't bother sleeving it for a neat look right now for a simple test, I'll do it later) :
http://i30.photobucket.com/albums/c338/Xilikon/LC/DSCF1323.jpg

Installed in my loop in place of the other adapter :
http://i30.photobucket.com/albums/c338/Xilikon/Modding/DSCF1324.jpg

It works fine and I get exactly the same reading range as the old one so either method works without problems. Martinm210, I think you should try this way in case your 680i is giving funky 12V feeding (maybe due to PWM, which can cause erratical behavior with non-PWM fans from what I heard around) with only the sensor wire going in your motherboard header.

P.S. : Disregard the gross tubing, will be replaced by black Tygon shortly.

bigslappy
12-11-2008, 10:01 AM
Great work keep it coming !

RedRaider
12-11-2008, 10:02 AM
This is good shit...:coolup:

Xilikon
12-11-2008, 10:03 AM
Reserved for more details