[RESOLVED] How to make an artifical fan

[Bartleby]

I think all your "rpm replicator" will need to output is a quasi-digital signal; something with a regular square wave. The 555 timer would be fine, and here is an alternative method.

I am reminded of my highschool electronics project in which I built an LED oscillator.

I found this schematic here
copy the ascii text to a fixed-width editor for the schematic. I could scan a better copy, but you would have to wait a while

All you need are:
- 2 NPN transistors (2n3904 or 2n2222 will do fine)
- 4 10kohm resistors
- 2 ~3uF capacitors
- One solder board from radioshack
- some wire

-----+--------+--------+--------+------> +12V
| | | |
R1 R2 R3 R4
| | | |
| + | | + +------> SIGNAL
+---C1---+ +---C2---+
c| | | |
\ | | |c
Q1 |------|--------+ b /
/ b +---------------| Q2
e| \
| |e
+--------------------------+------> GND

(note the polarity on the capacitors if you use electrolytic capacitors - the round cylinder ones)

Connect the circuit as shown, using the V+ and GND from the fan header

The operating frequency is dependent on the values of R2/R3 and C1/C3, approximately by the following formula:
f=3/(RC) - assuming R2/3 and C1/2 are respectively the same values.
Using the values above, you get 100Hz, or 6000rpm. You most likely won't find 3uF capacitors, so use the nearest standard values - 3.3uF or 2.2uF.

Hope this helps!

[CALV]

Balderunner do you think you could make the circuit in the above pic that I posted? if so, I'll work out some approximate values for you for the other components, But I need to know what frequency you need (i.e. RPM that the "imitation fan" needs to run at), the circuit can easily be built on veroboard if you are familiar with putting stuff like this together, Idealy you would need a scope to measure the frequency, but you could replace Ra or Rb in the pic with a potentiometer which would give you some flexibility, does your mobo have a spare fan header that allows speed monitoring, if so, you could build the circuit and plug it into the spare header and use MBM5 or similar to check the speed before you attatched it to your vid card.

[cyril1]

For 6000rpm the freq would work out to 100hz but I would want to verify what the fan signal looks like before I built a circuit.

I suspect it is a square wave between gnd and 12V. Very easy to make this with the 555.

[Gasman]

Wow!
This is a lot of tech talk.

I think you should change the topic heading:

" Who can make me one of these? "

The only time I ever got to play with a CRO was at school.

Oh yeah, the reason for my post, my Asus Smart Doctor has two temp gauges. Doesnt yours? Check your new GF3 CD.

[cyril1]

After re-reading your post Blade I believe there may be no voltage supplied by the Fan signal wire. The NC would stand for No Connection in other words open circuit. That means the fan either supplys an open circuit or a ground depending on where it is in the rotation.
The actual voltage is probably supplied by a pull-up resistor on the board. If that is the case then the circuit design must be one that that supplies grounds or opens at a 100hz rate.
If you have a voltmeter measure that signal pin to ground with the fan disconnected to see if there is a voltage on it.
You would not want to supply 12V to the circuit if there is already 5V there. The current has to sink somewhere.

[CALV]

if the fan gives an RPM reading then it will prodice an oscillation, if however, as Cyril mentioned, it only detects weather the fan is either running or not running, then it will be even easier for you to fool the fan sensor.

[Wuzup101]

I just wanted to say I'm impressed by all the knowledgeable people here. Since I'm still in highschool I have no idea what the hell you are all talking about. But carry on!

[cyril1]

Slow here at work tonight so I breadboarded the 555 oscillator circuit that Cal sent you.

I believe it will work just fine for you. Thought you would be able to just plug it right into your fan header but that is not going to work.

It probably has something to do with the Frequency selected but at 12V the frequency is not stable and runs away until it hits around 800hz. The good news is if you select a +VCC of 5V the circuit works perfectly and gives a steady output of 100hz gnd pulses if you remove RL.
C=0.1uF RA=47Kohm RB=47Kohm and the capacitor that is marked 10 to 100uf will work with either. I tried both.
After you build your circuit I would try it out with a speaker first to make sure it works. Just connect it where it shows RL. The speaker should emit a low frequency hum just slightly higher than powerline hum.
Once the circuit is working remove RL(speaker) and the wire used to hook in the speaker. Connect O/P to the signal pin for the fan on the video card. Connect +VCC to a 5V wire and gnd to a gnd wire. There you have it.
If you do not feel capable of building the circuit yourself a budding Electronic Wizard at your local community College would probably do it for a few pints.

[This message has been edited by cyril1 (edited 06-21-2001).]

[BladeRunner]

I must say I'm impressed with all the input I thought it would be seen as a boring question and get ignored.

I will try to find out what signal the card is getting from the fan but I would assume it is just a normal fan, (as it is not made by Asus).

If I'm wrong on this let me know but I think it works in a similar way to a crankshaft or ABS sensor which gives and induced "pulse" or many "pulses" per revolution.

I'd also guess it does this with the negative side. So the fan has +12v and Ground supply going in to the motor. One of these will give a fluctuating AC voltage down the rotation wire. I could therefore check this with a voltmeter?? just to see if this is the case.

Gasman

That's the old v7700 Smart Doc app that didn't have ram temp monitoring. It was just already up on my site and I couldn't remember the file name of the V8200 image at the time. Laziness really

[cyril1]

[QUOTE]Originally posted by BladeRunner:
[b]Bartleby

I found this info:-

MOLEX (three-pin motherboard) connector

The Molex connector has three wires: + (usually 12V, usually red), GND (usually black), and SIGNAL (usually yellow or white, but there are no clear standards). Unless you are a board designer, only the + and GND wires are of interest to you. (For the curious: The SIGNAL switches between NC and GND twice per rotation, resulting in a square periodical signal if you connect a (low) voltage.

According to this statement the fan connection only supplies ground or open and not a voltage. This most likely means the voltage is supplied from the mainboard through a pull up resistor. If this is the case then the 555 circuit with the values I supplied should work perfectly. Just omit the load resistor.
Unfortunately a voltmeter will probably not work too well as it will average everything. You really need a scope to tell if anything is comming from the fan.
The best use for the voltmeter would be to disconnect the fan, power up the PC and check to see if there is a voltage on the pin that the signal wire would connect to. This voltage would be a steady DC voltage.
If the voltage is there then the circuit without a load resistor would work perfectly since the it would pull up to whatever voltage is there and twice per revolution it would go to ground at a 100hz rate. ie 6000rpm.



[This message has been edited by cyril1 (edited 06-21-2001).]

[SDS]

http://members.nbci.com/jimas/trnastbl.html

[SDS]

http://ece.skku.ac.kr/~biwon/electro/oscillator.html

------------------
I²C bus, TO3 case, P4 CPU.

[This message has been edited by SDS (edited 06-22-2001).]

[BladeRunner]

cyril1

I'm out of my depth here but I take the explanation to mean the fan returns a, (varying according to fan speed) pulsed negative voltage (A/C) with the ground supplied to the fan from the ground header pin on the card (lets call it -12v for discussions sake although it's not the correct term).

So what I would need is an oscillated -12v to the rotation pin header. what I'm not sure of is whether this ground "voltage" needs to be from the -12v pin on the card or could be from -12v out of a four pin plug?

Like I said it's over my head really and I'll struggle to build anything without a detailed circuit diagram & component description.

I'll go a see what I'm getting on the fan rotation pin

[cyril1]

Look at the fan connector plug. It should have a red wire(12V) a black wire(gnd) and a yellow or white wire (signal) at least according to the info you supplied in your post above.

Unplug the fan and measure between where the black wire would have been and where the white or yellow wire would have been. Is there a DC voltage there when the PC is on?

If there is what is the voltage? I suspect there is a voltage there that is provided by a resistor on the board. The purpose of it is to keep a high on the input to the fan detection cct. If the cct never detects a low or has a high or a low all the time it alarms.

What the fan most likely does is supply a gnd twice per revolution to pull that voltage to gnd. In essence it makes a square wave of positive and gnd pulses.

----- ----- ----- +V

----- ----- ----- GND

Just pretend that looks like a square wave above. That might look like AC to the layman but it is actually ocillating dc. It is either off or on but never changes direction.

If you measure a positive DC voltage there then the 555 circuit using the values I supplied should work.

That was supposed to look like a square wave above but for some reason after I posted, it got compressed.

[This message has been edited by cyril1 (edited 06-23-2001).]

[CALV]

hmm ive been thinking, when my watercooling stuff arrives (if it ever does) I will be faced with the same problem, I suppose I could cheat and use a small fan, but now theres a problem in hand... well I just CANT use a fan can I!!!. It probly wont be for a week or 2 but I'll have a go at sorting the pulse issue out and let you know what happens... if you hear a rather loud bang, it means I did it wrong