Thanks!
Any thoughts on the instructions that say 50 ohms/V on the posotive? So if using a fan header with 12V, would you want to use 300 ohms?
It also doesn't make sense to me what the 2.2Kohm resistor does.??
Thanks!
Any thoughts on the instructions that say 50 ohms/V on the posotive? So if using a fan header with 12V, would you want to use 300 ohms?
It also doesn't make sense to me what the 2.2Kohm resistor does.??
With all due respect, I disagree.Originally Posted by KaptCrunch
Semantically, anode = positive, cathode = negative.
Regarding diodes:
http://www.see.ed.ac.uk/~gaa/Digilab...nts/node5.html
If Vcc is 12 and your target is 5 then 12-5=7*50=350 Ohm
I vaguely remember the 2k2 resistor having something to do with excitation. But it's just a hazy memory, so KaptCrunch (or someone with equal knowledge) will have to step in on this one
Thanks for the small electronics lesson
So basically, the Swissflow way in the schematics is the correct one. However, I like the floppy connector idea so maybe it's much simpler to have a Y adapter with a end plugged on the SF800 and 1 lead to the floppy connector or 4 pin molex (IIRC, red wire = 5V and yellow wire = 12V) and 1 lead just for the sensor ? Is that way correct ?
What an educated way to tell me that I'm not only old but also obsolete.
I will make arrangements to be decommissioned...
Yeah, the manners of new think. How appropriate, that flows from POS to negative too.
See you in the grumpy old men forum, Toadie.
MrToad, you are quite correct. The cathode IS negative (compared to the positive, or anode). There is no "new science" bullsh*t.
From the very same Wiki page linked:
The current flow through the diode is indeed limited towards the banded end, meaning the unbanded end should be on the positive 5/12V side and the banded end to ground. I just built a working fan-control circuit with two of the damned things so I ought to know.
Last edited by IanB; 02-05-2008 at 03:47 PM.
Well, that was my feeling too, and the material I linked comes from the University of Edinburgh, School of Engineering and Electronics. And is dated in late 2000.
However, I'm too old to embark myself on discussions over whether angels have gender or not.
Besides, I feel rather out of my depth on this subject. I'm sure I can find easier ways to make myself look like a fool before being "retired" (with the same kindness shown to the Replicants)
Just wondering if anyone is working on a config file for Samurize to show LPM / GPM before I set about making another one ??
Xilikon is working on it, with Martin and KaptCrunch involved in... hmmm... ironing the corners so to speak.
Pity I've taken a completely different route. That case of yours with the LCD display looks like pure unadulterated pr0n
LOL, might of known Martin would have been involved
I’m interested now, whats the completely different route you’re taking MrToad??
Coolercases has taken on the UK sales side for Swissflow so I’ve an avid interest at present.
Congratulations on the official connection. I don't suppose they can be persuaded to consider releasing a version of this tailored directly for the water-cooling market, with an external housing using 1/2" fittings that match our usual standard and electronics already converted to plug straight into mainboard fan headers..? Seems the R&D would be minimal, and since there must be enough market potential for small companies to produce watercooling parts tailored only for us at a profit, a specially modified version of the SF800 shouldn't be a loss-maker...
Wow, 100 posts at last. As they say, w00t.
I'm working on the software, which shouldn't be a problem, being graduated into computer engineering (specialized into programming and analyzing). We are trying to complete together by merging our expertise together. When this is done (pending the accuracy checking, which I doubt a bit and the reason why there are talk about resistors, diodes and fan header connection), I'll release the software and will work out the North America side of the SF800 distribution if there is enough demand.
As for the connections, it's very easy to either make the adapters by Martinm210 if the demand is under 50-100, or source the lot from the UK (I saw a nice plastic adapter on a UK site for pennies each). If coolmiester is willing to work on the UK side, I bet we could arrange something like that. With the adapter figured out, remain for users to buy the barbs themselves and for me to make the adapter harness (I just bought a molex crimper from PPCS so it should be very easy). IanB, I know there will be a market for sure and that's why I'm willing to dive into it, like Iandh did for the 8800GT ramsinks
That's the main purpose of this thread to gather enough informations in a centralized place and for everyone to come here for reference. The SF800 exist since 2-3 years already and there isn't enough data beside a few users reports (I saw a great samurize screenshot with the flowmeter running but no config file available to the public unfortunately).
With respect, Xilicon, if the electronics are correct there's no need for software. Many of us just want an RPM meter equivalent to a CPU fan that can be used to trigger standard BIOS protection if the pump fails. We don't need conversion to real flow rates. If the unit came simply with a three-pin mainboard fan header connection pre-fitted, it would be a snap buy. All that is required is for their marketing people to realise there is a market for a specialised unit. Same with the adapters and stuff, for the least pressure loss and flow restriction surely a redesigned housing that was 1/2" straight through (or threaded G1/4" so users' own preference fittings were easily added) would be better. The internals wouldn't need changing.
Modding this particular unit is OK, but persuading them to produce a plug-in-and-go specialised version for our market would be better.
For your information, if you buy a DDC pump, it already come with a rpm output wire that you plug on the motherboard and you can initiate a shutdown if the rpm get to 0.
The real reason for a SF800 isn't for checking if there is no flow (it could but you get 2 notices instead of 1) but to see if the flow is normal or low. When it get low, it mean something is clogging the loop and it's time to fix this or to see if the blocks is too restrictive for your taste.
The SF800 as is can be used on the motherboard and it already provide a diagram for you to just solder the resistors as instructed. However, the real issue is to see if it works really as expected and Martinm210 failed to get a output even if he did things correctly. That's the purpose of the discussion and the research.
I'll try my wired plug on the crystalfontz when my stuff comes in. If that doesn't work, I'll just plan on sending you the unit back.
We know the unit works with a Big NG.
The really strong point I see the SF800 is the specified accuracy and digital readout. 1% accuracy is exceptionally good for a flow rate meter, even my 14" tall King Instruments flow rate meter is only good to 2%.
So the meter should also be a good meter for testing purposes.
I'll tinker around with this one some more in some better loops, then I'll send it back your way with the extra couplers. Probably just my picky 680i motherboard, I hooked a DDC2 to it the other day and was getting all sorts of sporatic RPM readings that I wouldn't trust for anything...wierd.
That's disturbing. My DDC2 rpm readings are very stable using Everest to read the evga680i.
It's not a problem if you didn't fully test it because the main reason I'm sending you a unit is mostly for a pressure drop testing, which is already done anyway. I hope it works fine with the Crystalfontz and if you say you can't even get a stable reading of the DDC2 rpm, I bet it's your motherboard which is very picky. Mine is very stable hovering around 4400-4500 rpm for a DDC 3.2.
The only thing I might be disappointed is the lack of accuracy testing versus the King Instruments flow meter but we'll see with the Crystalfontz
Great
With this in mind, do we need to put a 50 ohm resistor or we can get away with only a 2K2 resistor across the sensor and 5v wires ?
Great
With this in mind, do we need to put a 50 ohm resistor or we can get away with only a 2K2 resistor across the sensor and 5v wires ?
you can get away with the 50 ohm resistor but you need the 2k2 resistor. The 2k2 is essential. it is required for open collector output of the sensor.
Antec 900, Asus P5K Deluxe, E6600 2.4GHz OC @ 3.600GHz 1.472V, OCZ DDR2 PC2-8500 Reaper HPC Edition - 2GB, GeForce 8800 GTS 320M
2x Samsung SyncMaster 245B
H2O Cool: External Homemade H2O case, PA120.3, 3 x 14cm Aerocool fans, MCP655, D-Tek Fuzion, EK-FC8800, Gemstar Flowmeter RFO-2500P
Front Display: ThermalTake CL-W0033 and Matrix Orbital LK204-24-USB
Good, I'll try to build another adapter harness on this principle and do a new testing to see if it change anything. Here's the schematics :
I'm using the kit from www.keiang.de as fan/flow/temperature (and I'm looking into pressure) monitor/controller, and I've already sourced a nice 240 * 64 graphic LCD to go with itLOL, might of known Martin would have been involved
I’m interested now, whats the completely different route you’re taking MrToad??
Coolercases has taken on the UK sales side for Swissflow so I’ve an avid interest at present.
Regarding the flow meter, I'm going for this Digimesa solution.
Thanks to keiang (having access to the person that actually designs and produces the kit is great), who has changed the sample window to 8 seconds on his firmware, I can use this flow meter with reasonable accuracy (+-6.9 l/h) and with negligible pressure drop (0.020 bar / 0.29 psi).
Digimesa also produces a "family" of flow meters that would be IMHO great for our application, the FHKU(C) 70s. The one that really catches my eye is the one spotted by Marci a few months back, when effizienz gurus published their comparative of flow meters: FHKU 70 Schlauch Ryton Thermo
Not only has a built-in thermal probe, it has also the same negligible pressure drop (l@@k @ the curve ^^), smaller footprint than the one I'm using (which is, let's face it, huge), and a nice 330 ppl, which means a great accuracy with the default 2s sample window most controllers use. (1/330)*60*60/2=5.45 l/h deviation.
Unfortunately falmer, which seems to be the official distribution for the UK of Digimesa doesn't stock these.
First I tried to source one as "sample" directly from Digimesa but I was utterly ignored (which is surprising, even companies like Tyco, Meanwell, JST, Krone and Fluke didn't ignore me, but hey, Swiss people are different). Then I tried with falmer, asking politely if they could source one of each as samples and then resell them to me at a price they considered adequate for their troubles. No joy so far.
All in all, although I think the SF800 is a great flow meter, I think the Digimesa solutions are more suitable for us. They provide similar levels of accuracy and repeatability with a fraction of the pressure drop.
If Digimesa doesn't ignore the watercooling market, it would have been a great choice and I also like the FHKU 70 flowmeter. I looked at the distributors and there is hardly anyone in the flowmeter area, only in semiconductors.
I wonder how Efficienzgurus got a sample...
Contact inside or within their distribution channel, good PR, or even as simple as being able to speak natively in the same language.
Too bad. If you find out, let us know
I also realized my diagram is wrong. I should have put the 3-pin connectors drawings the other way but the description is right
Posting Permissions
Reply With Quote
Bookmarks