Fan Filter Sensor?

[Revolvr]

Hi all,

I am thinking about making a sensor that will tell me if the intake fan filters are getting clogged. Seems like it would be easy by sensing the amount of light that passes through. An LED and a photo transistor to turn on a red LED if the filter is getting clogged, probably have a green LED when it’s OK, a trimmer pot to adjust.

Anyone done this, have advice, or can link me to schematics?

TIA

-- Rev

 

[sliperypete]

I wish I had the time to even think of an idea like that...

 

[Mohonri]

In industries where they have filters in pipes, typically they measure the pressure on either side of the filter. If there's a significant difference in the pressures, it means that the filter's clogged.

However, I don't think you'll want to spend $3000 for a couple of pressure gauges that are each at least the size of a soda can, so that option's out. I guess you *could* try an optical sensor, but both the light source and the sensor would have to be really close to the filter to get any sort of accurate reading. And if they're really close to the filter, then that part of the filter won't gather as much dust. And besides, by the time you got enough dust to see a difference, the filter would be long ago clogged.

My suggestion would be to simply remove and clean the filters once a week. You can even make a scheduled task to remind you...

 

[DirtyLude]

Actually there are plenty of small pressure sensors which would be useful for that, but they still aren't that cheap at ~$20 to $30 USD each. That's a good idea.

The only way I can think of doing it would be with a microcontroller and you would need to measure on and off light levels to counter any ambient light. Like Mohonri said, the LED and sensor would have to be far enough away from the filter to not block it.

Still it would not be difficult with any microcontroller with an ADC.

 

[Revolvr]

I guess one requirement is that this be so cheap and simple I won't kick myself if it turns out to be a complete waste of time.

I figure a light sensor, probably infrared, would be the best way to go. If the sensor and detector were in tubes, say half inch long, it may be fairly immune to ambient lighting.

A pressure or airflow sensor would be too expensive I expect and subject to changes in fan speed.

Here's a simple circuit. Don't know the values of the resistors, one would be a trimmer pot for calibration. The transistors are probably general purpose PNP and NPN's. This should be able to toggle a green and red LED. I'm not an ace at circuit design; if you see a glaring error let me know.

-- Rev

 

[Frank4d]

Here's a simple circuit. Don't know the values of the resistors, one would be a trimmer pot for calibration. The transistors are probably general purpose PNP and NPN's. This should be able to toggle a green and red LED. I'm not an ace at circuit design; if you see a glaring error let me know.

The problem I see is that the circuit would be very difficult to calibrate. While the filter goes from clean to dirty, the green LED is going to dim while the red LED gradually turns on. There is probably going to be a long period where both LEDs are on.

It would work better if you replace the NPN and PNP transistors with half of a LM339 type quad comparator IC and a couple of resistors and trimmer potentiometers to calibrate it.

Edit: I see that would have problems too. Would need some hysteresis or a latching circuit so when the red LED first turns on, it stays on until the filter is cleaned.

 

[BrainEater]

Really good idea !

 

[Mohonri]

I just had another idea. One thing you could do is mount a pressure sensor of some sort (could be as simple as a lightly spring-loaded switch) so that as the filter clogs (and the fan is still sucking air through it), it pushes harder and harder against the switch until it makes the contact. you can then do whatever you need to with an electrical signal from there--for example switching those circuits you drew up there.

(by the way, there's an even easier way to switch the LEDs if you're using a double-throw switch)

I have a switch around from a disassembled CD player that might be suitable. Again, though, the problem is with the calibration, and would definitely be harder than using a uController, as the adjustments would need to be mechanical instead of software-based.

 

[Frank4d]

This might also be a good use for one of those LM3914 LED bar/dot display chips http://www.national.com/pf/LM/LM3914.html. Hook the phototransistor up to the LM3914 and LEDs and calibrate it so the LEDs move from one side to the other as the filter gets dirty.

 

[DirtyLude]

Those are great ideas.

Like mentioned before, though, you wouldn't be able to use a fan controller with the fan as the pressure would change with fan speed.

 

[gee]

how about current sensing on the fan? a fan with a blocked input should draw more current.

 

[upriverpaddler]

how about current sensing on the fan? a fan with a blocked input should draw more current.

Attach a fanblade to a small radioshack dc motor (or just an extra fan) and place it in front of of your fan. The dc motor will output voltage that you can use to power an led. This shouldn't block your airflow too bad. When the led goes out, time to clean the filters.

Probably the fan won't put out enuff to actually light the led directly, but it can be done.

 

[Mohonri]

The fan blade idea gave me another idea--use the RPM sensor for the fan. When airflow gets reduced, the load on the fan will actually decrease, and the fan will increase speed. You could use the RPM sensor to detect an increase in speed, and once it hits a certain level above baseline, it turns on an LED. You'd have to use some sort of microcontroller for that, unless there's a product out there (and there probably is) I'm unaware of.

 

[Rondu]

Mohonri's idea is a good one, as the air flow is reduced the fan will speed up. If the fan is connected to the motherboard fan header there are several monitoring programs out there that can sound an alarm at a set speed so this will just have to be calibrated.

 

[DirtyLude]

Again, it won't work with a fan controller, and I have my doubts that fan speed will decrease enough for repeatably measurable results.

 

[Revolvr]

I decided to breadboard the circuit. The original idea probably works in theory but I could never get it balanced well. So I tried a simple logic inverter. Got it breadboarded and testsd. Seems to work!

Below is a schematic of what I tested and some shots using a typical filter, with 2,3 and 4 layers simulating various levels of clogation. It does transition fairly slowly from green to red, but is clearly either green or red with a clean vs. dirty filter.

I used a 2N2222 instead of a 3904 because the 3904 turned on too quickly resulting in red on even for a clean filter.

If you need a simple low cost project this holiday weekend, maybe you need to get away from your mother-in-law, give it a try.

Not sure I'll take this any further than a breadboard, mostly a mental exercise. Total cost would be less than $20 US probably, depending on what stuff you have in your parts bins.

Cheers,

-- Rev

The IR LED emitter is on a separate board on the right, pointing to the detector on the breadboard.

One filter between emitter and detector

Two filter layers, still mostly green but the red LED is slightly on.

Three filter layers, mostly red, a little green.

Four filter layers, most all red