Electronics Project- Fan Failiure Alarm

[Gundog48]

Now, I'm not sure if this is the right place to put this, but it is a project and it is related to PCs!

I will be calling on people's advice who know their electronics well or who work with PCB. For my GCSE Electronic Products course I am making a PC fan failure alarm, I know many computers have this built in but that doesn't really matter for what I'm doing!

Here's what I am looking to achieve:

• A circuit that sets of a buzzer when a PC fan stops rotating or falls below an acceptable RPM
• The alarm should stop if the fan restarts
• It should be mounted to a PCI bracket

The simplest example of a circuit I could find was the one below which I will be basing my own circuit on, where the 741 comparator compares the voltage from the fan tacho line to the adjustable one coming via the potentiometer. Fairly simple stuff, except the fan outputs pulses rather than straight voltage, so it must be converted.

Full info: http://pcbheaven.com/circuitpages/PC_Fan_Failure_Alarm/

So I'm going to be needing some advice on this project. Firstly, the circuit relies on the fact the fan emits 2 pulses per rev- is this the standard for most fans? Secondly, can you see any issues with this circuit? I am going to get it to run from a single 12V rail rather than all those inputs used, which I do not see as a problem. As I am not allowed to run the circuit from a PSU, I will have to run it from a 12v battery or even a few 9v batteries, do you see this causing problems? I will be putting a Molex connector for the power rails, and then I can modify the other end of it to connect to a 12v battery, so it can be used with a computer PSU without any modification. Lastly, I am going to need it to attach to a PCI bracket. The first issue is finding a blank PCI bracket, so suggestions on where to get them would be appreciated, and secondly, actually attaching the thing! I was planning the drill about 4 holes along the front of the PCB then get some kind of tiny, 90* bracket and then screw it into the PCB and either screw or rivet it to the bracket.

Suggestions on how to overcome the problems, possible issues or improving the circuit would be appreciated.

 

[thingemajib]

i did a fan controller for my AS level project, but i used chips that were programmed in BASIC (easier way of finding the flaws in my designs that taking the analogue route). Part of the project involved measuring the speed and displaying it on an LCD. Whilst i never actually confirmed this properly, i worked on the assumption that the fans emit one pulse per rev, and it seemed to do the trick. If i remember correctly, its a light gate of some description that actually makes the pulses (ill see if ive got a copy of my coursework floating around in a sec). It was basically (pun unintended) a case of hooking up the tach wire to an input pin, and telling the chip to listen for and count the pulses on that input for a certain amount of time (50ms IIRC). Then do some crazy maths hoodoo (you can tell ive finished 6th form...), and you have a figure for RPM. Due to deadlines and such, i only ever made it accurate to 100RPM, as i didnt have the time to finish it, let alone fine tune it (still got full marks though ). Now i dont know what components you have at your disposal, but i discovered that ADC is fairly common, but DAC is a little trickier to come by. Or at least in the world of chips. For example the chips i used did have ADC inputs, but only had digital outputs. This meant i couldn't vary the fan speed using voltage, and again i didnt really have the time (or PCB space) to work out the circuitry equivilent.

Like you, i also used a molex power connector. It was conveinient as the fans needed 12v, but the chips i used needed 5v. As for PCI brackets, there are ones that have a small 90 degree tab on and is screwed to the PCB. I did mine in using a knackered dvd drive case (admittedly did a pretty **** poor job of butchering it but it worked in concept). I guess that would be a tad on the large side for yours, so you could try an old floppy drive if PCI slots arent working out.

Im really missing D.T. in general, so feel free to send a PM and ill be glad to help (no promises that i can say the right thing though )

 

[egnshane]

Hi
Hopefully I will be able to give you a hand. I am currently doing electronic engineering at uni.
I have worked with the 741 a lot too in my first year so fingers crossed. You would be best setting it up with 2 9v batteries so that you can make a virtual earth otherwise your 741 wont work correctly. Also for your buzzer if your just using A speaker for this you will need to generate a signal, 555 timer would do the job.

Im going to have a think about the best way of converting the pulse to a constant voltage.

I cant think of any other way of doing it without using a micro controller and some programming.

Edit:

For the pulse to dc you might want to try using a small capacitor, im not sure this would work, I have no idea about the pulse out of a fan.

 

[egnshane]

It would be worth having a look at the datasheet for the 741 too.

http://www.national.com/ds/LM/LM741.pdf

 

[Gundog48]

Hi
Hopefully I will be able to give you a hand. I am currently doing electronic engineering at uni.
I have worked with the 741 a lot too in my first year so fingers crossed. You would be best setting it up with 2 9v batteries so that you can make a virtual earth otherwise your 741 wont work correctly. Also for your buzzer if your just using A speaker for this you will need to generate a signal, 555 timer would do the job.

Im going to have a think about the best way of converting the pulse to a constant voltage.

I cant think of any other way of doing it without using a micro controller and some programming.

Edit:

For the pulse to dc you might want to try using a small capacitor, im not sure this would work, I have no idea about the pulse out of a fan.

The fan sends out two pulses per rev. The problem I am having is simulating it in circuit wizard. There is a function generator that seems to do the job of creating waves/pulses but I cant work it for the life of me!

 

[egnshane]

I haven't used circuit wizard before, PSpice is a pretty good free software tool.

I mentioned I had no idea about the pulse of the fan meaning the pulse length and voltage etc. You would need to know those details although it would be easy enough to figure out with an oscilloscope.

 

[Akempster]

the pulse lengths vary between fans slightly but generally your looking at a signal which (at 100% high) will send a new signal every ms, the faster you have it the less jerky the fan will be. PWM works by turning the signal on and off with different amounts of on and off time, its hard to explain but this table should make it clearer:

speed %time on % time off
100 100 0
90 90 10
80 80 20
70 70 30
60 60 40
50 50 50
40 40 60
30 30 70
20 20 80
10 10 90
off 0 100

taking say 60% speed you would then, out of say 1ms if we use that as our data sampling speed need to have the signal on for 0.6ms and off for 0.4ms and keep repeating this over and over. To do this i would suggest using a microcontroller with an external frequency crystal it would require programming but it would be the better way of going round it