Hi Peter, Most CMOS logic families are buffered, which means that there are additional stages of extra inverters inside to increase the gain without changing the logic function. For example, and actual 74HC04 is really three inverters per gate. This makes it very difficult to get a reliable negative feedback path by connecting a resistor from input to output because the gain is so high and the gain and/or phase margins are negative. This is similar to an uncompensated op-amp. You can get specifically "unbuffered" versions - like the 74HCU04 or the 4009UB. These usually work reliably in a negative feedback connection. I looked up the CD4060 you mentioned but I think the part number must be wrong because the 4060 appears to be a ripple counter. I think also that you and Tom may be mixing up different types of oscillators. For the typical Pierce crystal oscillator circuit, you do NOT want Schmitt trigger behavior. ST would be for relaxation-type oscillators using an RC circuit. Sean On Sat, Apr 20, 2013 at 12:53 PM, Peter P. wrote: > Problems with 74HC(T)xxxx and CMOS4xxx IC 'linear' circuits. > > The following reproducible problems were found in several > popular 'analog' applications of CMOS gate based circuits. > The circuits are tried-and-true 'old' circuits from the > 1980s and 1970s, as published in datasheets and application > notes from various name silicon makers. > > 1 - inverter-with-feedback-resistor does not work or oscillates > uncontrollably. Occurs with 74HC(T)xxx circuits, any. Using any > feedback resistor over a simple inverter in a 74HC(T) circuit > causes the gate to oscillate at over 10 MHz. The resistor value > does not matter. The oscillation also occurs if an external > precise bias source is used to bias the input, without using a > feedback resistor. Source impedances tried ranged from 1 ohm to > 10Megohms. Consequences: > > - It is not possible to build low frequecy Xtal oscillators > using 74HC(T)xxxx gates. Example: 32767kHz osc on 74HC(T)4060 > will not work, it oscillates at several MHz. Manipulating the > usual Pierce oscillator part values around the xtal does not > help. This does not appear anywhere in the datasheet. > > 2 - CMOS oscillators built as above with gate-and-feedback > resistor stop working when the supply voltage is taken up > above 5V. Example: CD4060 with 32768kHz Xtal using datasheet > Pierce oscillator values (2.2kohms, 2x33pF, 2M2 feedback), > works fine at 5V but stops working at about 7V supply and is > impossible to start at 12V. My assumption is, that there is > some 'new' 'feature' caused by die shrinks which increases > leakage at the input and/or lowers output impedance somehow, > or that gain decreases with voltage somehow in the 'deadband' > dc bias region. > > In all cases I tried several Xtals (3 makes), and several ICs > (2 or 3 makes each). All ICs were new (datecode >2007). I did > not try old old parts (CD4060 etc) beucause I know they work > fine. > > Any comments would be welcome here. I am a little stumped by > these new features. I tried the above several times (incredu- > lously, after having used cmos parts since the early 1980s > with no problems). > > What IS going on, in your opinion? > > Comments, etc., are very welcome. Tia, > > -- Peter > -- > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .