At 07:11 PM 1/5/2004 -0500, you wrote: >----- Original Message ----- >From: "Peter L. Peres" >To: >Sent: Monday, January 05, 2004 4:27 PM >Subject: Re: [EE:] LM324 as schmidt trigger > > > > >>Sorry for being cryptic. The LM324 will not pull the output down to= GND > > >>(it cannot sink current when this happens). Therefore a hysteresis > > >>positive feedback resistor cannot be used when the input is only 65mV > > >>unless a pulldown is used at the output of the LM324. > > > > > > Really? > > > > National LM324 datasheet (ed. 1994) p.5 graph "Output Characteristics, > > Current Sinking", 5V supply line crosses the abscissa at about 10mV, 1uA > > sink. This is 10kOhms Ro without a load or a wire attached, and it gets > > worse with lower output voltage. So what? In the circuit I posted it's divided down by 350:1, so the output swing affects the threshold by 0.0mV/+10.2mV (~+/- 5mV). Even if the output only swung from 1.5V (rather than 10mV) to 3.5V you *could* still use +ve feedback with a 60mV signal, just it would require some parts to get logic levels at the output. The 10mV divided by 350:1 yields ~30uV. Try working the numbers through if you don't see how this works- ascii-matic reposted at end. The main limitation is that the LM324 is a bit marginal wrt Vos for this application without trimming (+/-7mV maximum at 25=B0C, with *no* maximum TCVos specified. There are much better performing (and much more expensive) choices. > > Same datasheet, p.8 "Driving TTL" shows what needs to be done to achieve > > Vil for a TTL chip input (which sources much less than a mA). A TTL > > compatible CMOS input will not need such a strong pulldown, but with a > > wire of some length attached, some noise, a scope probe, and pretensions > > to make the hysteresis work as expected while pushing the frequency up > > towards 1kHz a pulldown is required imho. Maybe 240 ohms is drastic. No pulldown at all is required (or desirable). Vout(low) will be around= 10mV, just as you say above. Normal TTL input level is 800mV for inputs and 400mV for outputs (400mV of noise immunity guaranteed). BTW, adding a very high resistance pulldown won't reduce the ~10mV output saturation voltage much, as it's coming from the *internal* base current (in the 50uA current sink), not the externally applied current. Adding a pulldown will reduce Vout(high). BTW, using an op-amp as a comparator is attractive in certain special applications, but there are many cases where it is not a good idea. The internally compensated op-amps have slow response times and slew rates, but high gain. If the input signal is larger than a few hundred mV, many parts have stuff connected to the inputs inside, not shown on the data sheets (since they tend not to show schematics anymore) that will cause current to flow in or out of the inputs when the differential voltage gets high. This is of little concern in an op-amp application, but can affect proper operation of circuits with comparators. In some cases, even the finest engineer can't determine this in advance as it's NOT indicated on the data sheet in ANY WAY. As previously posted, but with values indicated: In |\| o-------------|-\ Out | >--+-------o +-|+/ | | |/| | | | | 1M0 | | ___ | +-------+-|___|-+ | | .-. .-. 294K | | | | 2K79 | | | | '-' '-' | | 5.0VDC GND Best regards, Spehro Pefhany --"it's the network..." "The Journey is the= reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com -- http://www.piclist.com hint: To leave the PICList mailto:piclist-unsubscribe-request@mitvma.mit.edu