Interesting. I don't use a lot of CD4000 series, but I do know all the AC ACT HC etc that I regularly use have quite reliable switching points and hysteresis.
I also have doubts about your first transfer chart for the 4011B, where is it from? I've NEVER seen CMOS gate inputs that had zero hysteresis as shown. Hook one up with a pot on the input and you should see a very obvious hysteresis right around the expected voltages.
From my databook;
BU4011B 1.5 3.5
CD4011B 1.5 3.5
CD4011AE 1.5 3.5
HCC4011BK 1.5 3.5
out of 40 brands and specs listed fo the 4011 there are only about 6 that list 1.0 and 4.0v, the rest are 2.5 and 3.5 which are about the values you always see when testing the real chips.
Now the schmidt version;
BU4093B 1.5 3.5
MC14093BAL 1.5 3.5
again almost all are listed as 1.5v and 3.5v respectively.
It's been a lot of years since I tested a CMOS chip that did not have a large input hysteresis and quite relaible LO and HI threshold voltages. But maybe there are still some crap manufacturers that can't design a CMOS input pin circuit. Can you suggest a brand or chip I can test??
You can test any reliable brand of CD4000 parts. I believe that Fairchild, TI, and NXP (HEF series) are still in production. You might find new old stock of National, Harris, Intersil, Motorola, and ON Semi. There are probably others.
While you're at it, test some 74HC/HCT gates. As I said earlier, 74HC14 and 74HC132 are the only SSI parts in that series with Schmitt trigger inputs.
A simple test is to enable the (inverting) gate by connecting all but one input to the appropriate rail, then feed a 1 Meg resistor from the output to the remaining input, and a 100nF (or whatever) from the input to GND If the input has hysteresis, you will get a square wave on the output and a triangle wave on the input. If the input does not have hysteresis, the input will be biased at the input logic threshold (due to the negative feedback), as will the output, except the output may have a lot of noise on it.
I ran this test on 3 different gates. Vcc was 5.00V in all three cases. Results were as follows:
1. An HCF4011 quad NAND made by ST Microelectronics. Vin=2.63V, Vout=2.63V (with about 4.5V volts p-p of noise)
2. A 74HCT00 quad NAND made by RCA. Vin=1.40V, Vout=1.40V (with about 800mV p-p of noise)
3. A 74HC132 quad NAND with Schmitt trigger inputs, made by Philips. Vin was a triangle wave swinging between 1.8V and 2.5V at about 18Hz. Vout was a square wave at the same frequency.
You can't run this test on noninverting gates (OR, AND, noninverting buffer), because the resulting positive feedback will put both input and output at one of the rails.
Think about this: If all gates had Schmitt trigger inputs, why would only 2 in each family explicitly mention this?
Roman, you are misinterpreting the input specs in the datasheets.