Post by thread:[OT]: Magic CMOS chips

Hi Everyone, Just thought I'd share a funny experience from the lab. We needed a 15-bit counter from parts that we have available to us in the lab, so we took two CD4024 7-bit counters and a dual JK Flip flop to do the job. Well, we laid them all out on a breadboard, conencted all the power to the right places, fired it up and saw that the output from the last of the CD4024's was pretty bad. It was counting correctly and everything like that, but only after we put some fairly large capacitance across power. Anyway, after a bit of playing around we realized that this was one of those breadboards with the split power rails, so we weren't actually powering the chip up! Needless to say, once we gave it power, the output seems to be a lot nicer. The strange part is that the thing was still counting properly, power or not! If only every chip ran like that! -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

The counter chip was being powered by one (or more) of the other counter chip outputs connected to it's inputs. The CMOS input protection diode connected from the input to VDD will conduct when forward biased, providing a postive rail one diode drop below the driving logic level. PIC's will do the same thing when an external voltage is applied to a port pin while no power is applied to the PIC. If the external source is strong enough, it will power the PIC (or in your case, the counter)... Veteran PIClister's have heard the question "The power is off -- WHY is my PIC still running?" many, many times! ;-) Matt Pobursky Maximum Performance Systems On Mon, 8 Oct 2001 12:29:51 -0700, Wade Carpenter wrote: {Quote hidden}>Hi Everyone, > >Just thought I'd share a funny experience from the lab. > >We needed a 15-bit counter from parts that we have available to us in the lab, >so we took two CD4024 7-bit counters and a dual JK Flip flop to do the job. > >Well, we laid them all out on a breadboard, conencted all the power to the >right places, fired it up and saw that the output from the last of the >CD4024's was pretty bad. It was counting correctly and everything like that, >but only after we put some fairly large capacitance across power. > >Anyway, after a bit of playing around we realized that this was one of those >breadboards with the split power rails, so we weren't actually powering the >chip up! Needless to say, once we gave it power, the output seems to be a lot >nicer. > >The strange part is that the thing was still counting properly, power or not! >If only every chip ran like that! > >-- -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam@spam@mitvma.mit.edu

<grin>. I try to favor CMOS 4xxx series logic to anything else when I can. There are several commercial designs of interfaces etc that use what you have discovered. The chips run on power from the H levels on the clock or data input, it being rectified into the supply rail through the upper (towards Vdd) input protection diodes. Since at low clocks these chips draw microwatts this works pretty well, as long as there is enough decoupling to supply during transients. Also a CMOS series flipflop with a simple 0.1uF decoupling and a diode to separate the supply would make a 1-minute persistent memory iff the output is not loaded (driving current) and the reset state is high, not low. Both can be arranged easily in a circuit (this is relevant to the one bit memory thread). Peter -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

Hi Wade, You can actually power a PICmicro in exactly the same way - leave the Vdd pin open on a PIC16F84 and drive a current through one or two I/O pins set to "input". You'll find that the oscillator will start up and the chip will run quite well (although logic output levels will be lower than the input pin voltages). myke {Original Message removed}

>The strange part is that the thing was still counting properly, power or not! If only every chip ran like that! You were powering it through the protection diodes. As long as one input is high, and you have ground, they will run. However, you were in danger of CMOS latchup, and as you noted, the outputs are pretty bad. -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads