ok for those who have not read this i wanted to build one but it seems there is no MIC502 available locally. id hate to spend for shipment just for 1 chip. is there another article i can follow that has more common parts used? is there a replacement for the MIC502 that i cud use which wud produce same results
Microchip do a series, TC642, TC646, TC647, etc that are similar, (but not identical). If you can get the chips, datasheets are here
Well, the chip doesn't ever drive the fan directly, it just produces a waveform with a varying duty cycle, that you then use to switch a transistor which drives the load. So with appropriate switching transistors, you could drive incandescent light bulbs from a MIC502 or any other PWM chip you care to use. Match the power transistor to the load.
i sure wish i understood what you guys just said. what do mean 555 timers? and linear, m not good at circuits. that is where i am "ignorant".
http://www.bit-tech.net/images/article/51/1.jpg The waveform that the chip produces (on pin 7) is fed to the base of a transistor. The transistor switches on and off the load (a fan) in response to the waveform, but the voltage and current are set independently, in this case 12V is supplied to the fan, less a couple tenths of a volt for the transistor. So the fan isn't directly driven by the IC. A 555 timer is a common, inexpensive, general purpose timing IC. It has a long history of being used in a wide variety of circuits and it's pretty well-known to most electronics hobbyists. (Google for "555 timer" for eight zillion pages of info.) You can screw around with pulse width modulation very cheaply using the 555, Uller has a tutorial here that explains theory and practice pretty well, and he develops a circuit that works around most of the goofy limitations. But by the time you do that, it was way easier to use a MIC502. Still interesting to think about though.
That article... Shouldn't there be a reverse diode in paralell with the fan to handle the inductive dumps? Also... I highly question the use of a MOSFET in a circuit for newbies... Newbies just don't have the knowhow to handle MOSFET transistors w/o killing them with ESD. A NPN transistor would be more suited.
Which article? No mosfet in the Bit-Tech one. Brushless fans don't let back-emf out, there are blocking rectifier junctions in the fan electronics, so a reverse diode is superfluos. Try spinning your fan with a voltmeter connected. And most people here have heard the handling warnings for memory, cards, cmos, etc.
You mean Uller's, right? He's got clamping diodes on the ones that are the full circuit, the one on this page seems to be just an oscillator without the driver. Agree on the FET, he loves them I hate them. I tried to talk him out of it, but he applied zero of my suggested edits (I proofread it before he published it).
HTF does that work with the mosfet gate grounded? The bonus from a mosfet is low voltage drop when full on, less than a bipolar. And most aren't that sensitive.
*lol* I didn't even see that ground connection. There is a reverse diode i paralell to the MOSFET, but that only handle negative currents... The positive inductive currents (when the MOSFET switches from on to off) will hit the MOSFET and it's reverse diode again, and again, and again... Most MOSFET's are sensitive enough for newbies to kill.
my teacher just told me that fans run under a regulated voltage meaning however low the voltage is, the fan rotation remains constant. and stalls when its too low. also the PWM may shorten a fans lifespan due to its way of adjusting fan speed.
You could try the TL598 by TI. It's one of the most advanced PWM chip I've ever seen, but I guarantee it would work. (Unfortunately I can't find any good docs on how to use it.) What's with the 741? (There are *much* better op amps out there, and cheaper I would presume.) Also, the anti-paralleled diode with the FET is redundant. All MOSFETs have a built in diode between the source and drain (go look at the datasheet). (It's inherent because the body is tied to the source with the body being p type and the drain being n+ type which yields a diode....)
I think it's hard to beat the 741 when it comes to cost... It is produced in vast quantities. Besides... there's no need for a better OP than the 741. It's always a good thing to use standard components as much as possible, as long as it doesn't affect performance or price in a significant way.
Uller said if he didn't use Radio S*ack components, no one would build it. Probably no one built it anyhow judging by the grounded gate.
How about the LMC662 or the LM324? The 741 doesn't do rail-to-rail outputs, has a high minimum voltage (10 volts), high input bias, and has a high voltage offset. The 741 costs $0.10/each while the 324 is $0.12/each and the 662 is $0.69/each (all prices on nsc.com). There are better parts yet, but the 662 and 324 are common from what I've seen.
