PWM to Analog converter circuit

Tealc, forgot to add my thanks. Have been on holiday.

It works wonders, fans at 600 rpm on idle (down from 1000+ rpm), and still hit 100% when required (1250rpm - only when benching).

Thanks again,
Chris.
 
Tealc

Is their a reason why the PWM speed seems to drop to zero quite a bit ? It drops to Zero then it goes back to normal and gives a speed reading.

temps-pwm.jpg
 
The Matrix said:
Is their a reason why the PWM speed seems to drop to zero quite a bit ? It drops to Zero then it goes back to normal and gives a speed reading.
Click to expand...

I get this at very low rpms - the reading becomes unreliable or bursts to something crazy like 10000 rpm :D

I think this is down to the motherboard not being sensitive enough to read the correct speed. It's not just an issue with the pwm converter but also with fans connected directly. My motherboard just can't record revs under 700 rpm - I have an SP120PE connected directly with a LNR on it.... as the rpms drop under 700 the readings go haywire but above that all the way to 1700rpm it's fully reliable
 
I used to notice this all the time in CPUID's hardware monitor. Sometimes it'd drop to zero and then spike in the 10s of thousands then read normally.

However I also have my BIOS set with fan speed warnings/shutdown and they've never triggered with these zero RPM events.

I would have to say that since it's only happening in CPUID HWM and that the enormous RPM readings happen just after the zero-out reading that there's some bug in the code that cannot correctly process low-speed signals, then in a fit of delayed data, reads the signal feed plus all the delayed data at once and over-calculates the actual RPM.

I've been using Open Hardware monitor for a while now and I've not noticed the same issue.
 
Mine goes down to around 550 and then either shows zero or sometimes it picks up the PWM frequency of 20-25kHz. It's just one of those things.

As the voltage of a dc fan decreases the signal current it sends back to the tach circuit on the motherboard decreases with it. Eventually the motherboard will not be sensitive enough to read it, or it will start seeing the PWM pulses as a stronger signal.

Not sure if it happens with a true PWM fan so could just be a feature of the way the circuit works.
 
Finally got some of these puppies in.

P1060056_zps190269eb.jpg


They fit reasonably well to the board. Haven't tried one yet on the transistor but hopefully it'll work out well.

P1060056_zps190269eb.jpg


P1060058_zps067c968c.jpg



On another subject I've been contacted by two hotmail users via trust but have received no reply back. If you are one of those check your email.
 
Last edited:
I've been thinking about LEDs again recently and how I can get them onto the converter to make it more interesting, at least for me anyway as my converter is semi visible at the top front of my case.

Thinking along the lines of a spindicator...


...but using some junkbox parts rather than expensive micros.

Along the same lines as my spinning LED fan mod from a couple of years ago but without the persistence and running off the fan RPM tach.

Might have a quick fiddle about with this over the weekend on my breadboard.



In other news I'm looking to etch my own PCBs. I've already drafted out the circuit ready for toner transfer and SMD resistors, if I decide to go down that route rather than through hole. I should probably make one suitable for through hole as well as I have loads of the things :) If I etch my own they will be more professional, easier to build and make them less prone to circuit mistakes as you get with veroboard. I could even make kits for folk who fancy a bit of soldering themselves.

pcb1_zps92597fb1.png




I'm getting an old oscilloscope later in the month so I will be able to actually visualise the output of the converter properly at last rather than guessing.
 
Last edited:
Just thought I'd throw in an update as I just got my first oscilloscope.

gouldos3000_zpse8df78ab.jpg


A 40MHz old analog Advance Gould oscilloscope.

I recorded a short-ish video of how the voltage looks in real time as it reacts to the PWM converter and basically what the PWM signal looks like. I also threw a resistor on there and tried a couple of caps out to see what difference it makes.


I always thought the circuit was robust and safe for use and this confirms it. No nasty spikes, no back emf, just a nice smooth voltage running to the fan.
 
Thanks guys.

Here's an interesting one. Developed from this blueprint.

untitled_zps7895bc67.png


A GPU controlled converter with separate PWM output for the converter fan, pass through for GPU fan and a free tach wire for future tach hookup.

Looks a bit like this.

Lil-Diabo_zps27d1d28f.jpg


Works great on the oscilloscope too. 4 pin PWm connector fitted to the yellow tach wire as I just don't have any 3 pin connectors. Probably should get some :/

This one is for Lil-Diabo. :)

I wonder how many of these things I've made now.
 
Last edited:
Tealc, i'm really made up with the PWM adapter. The increase in performance from my fans since i fitted the adapter has been impressive to say the least, they used to be 1200rpm fans, now i'm getting 675000rpm .... Not bad for around a tenner ;)

 
 
I call it a feature :)

It happens when you drive the fans a touch on the slow side. PWM fans are generally limited to around 600-700RPM as a minimum speed and this is electronically prevented from going lower. There is no such device in a 3 pin fan. I'm not even sure what the device or sub-circuit would be as I've never really looked into it.
 
Tealc, I'm after cables for my 2 GELID Rev. 2 ICY VISION coolers, how bulky is the circuitry in your cables can they easily be hidden out of site, will it act just like the PWM fan from the manufactures cooler I removed from my 7950 card or is tweaking needed on the actual circuitry?
 
You must log in or register to reply here.
Back
Top Bottom