You have probably connected all unused inputs to somewhere, like ground or VDD? Floating inputs do tricks with current consumption. There are or at least was 74HCu04 which were non buffered inverters, they were supposed to take less current as oscillators and amplifiers.
I have only made oscillators only from HC and 40 inverters and smitt triggers. And found nothing special. Those new LVC and so on series ICs are so fast that maybe this the reason for you current consumtion.
The reason for current consumption was already shown.
If you overlooked post 28 in this thread then see Fig 9 and 12 in:
https://pl.mouser.com/datasheet/2/916/74LVC2G14-1597848.pdf
Yes the current for the dual inverter is high by a margin more than I expected. Also…keep in mind that these graphs are usually typical.
There are no unused inputs on a single inverter in a 5 pin package. Allowing unused inputs to float is something of a beginners error, unless it is recommended in the DS. I may be fairly accused of many things but I am not a beginner.
Yes it would make sense that speed would correlate with the current flowing through the input stage (when the input voltage is in the threshold region.)
BTW in the 1970s I made a resonant LC oscillator with something like a 74S04 single schottky inverter. This was only experimentation; I would not expect the inverter to last long when it was running in this mode.
Current drawn in the linear region is not even specified for most gates, so it is a lottery what an oscillator will draw. I remember that we were advised to use the unbuffered “U” parts in these amplifier bodges as found in many CD player read amplifiers
Kind of reminds me of the original prototypes for the 3 custom chips in the prototype Amiga (known as the Lorraine). They were fanned-out books (my descriptive term) of wire-wrapped discrete circuitry. Ok, so not discrete transistors, but still impressive. Here is a picture that I found:
(picture source)
Just imagine the cajones on the original development team bringing those to a trade show to try to convince someone to buy into the project so they could take it to market.
1 Like
I would have been quite impressed. Both by the required effort to get a prototype of that complexity working, and by all that stuff fitting in three pieces of baked sand.
Some more pictures of Lorraine:
https://duckduckgo.com/?q=amiga+%2B+lorraine&t=hy&va=g&iar=images&iax=images&ia=images
But for schmitt input gate should be as such gate is simply expected to be used with slowly changing input voltage.
I think there is a question of how much is “slow”? I guess that a slew rate of 0.1V per microsecond would usually be considered slow these days. But if it is triggered by a pushbutton then the repetition rate should be less than a few Hz. Then the % of time that the input is in the threshold region will be less than maybe 20 microseconds/200 milliseconds = .01%. And the average current draw will be (very roughly) 2 x 0.01% x (the 20 mA level that I measured for example). That “2” comes from having both up and down edges for every pulse.
But for the oscillator, the input is in the transition region 100% of the time. I think it is a worst case.
Replace the worst with wurst, add some mustard and a bun and you’ve got something to eat. 
I was thinking how to do 10Hz oscillator taking less current and simplest solution is to replace LVC with HC. Don’t know how much current C-MOS versions of NE555 consume. The less will be loading capacitor with 10M resistor and discharging with something like single junction transistor or its replacement circuit with 2 transistors.
I have very little use for the NE555. A comparator (or even an op amp) can make a very stable 10 KHz oscillator. I think the frequency is more stable and predictable than that of an NE555.
At the time I drew this I was working for ADI but almost any comparator can work if it is fast enough for the frequency. Generally:
R24=R25 and each about 2*R26
R28 = R24//R25//R26 to balance input bias current effects.
If you use a SPICE simulator, go ahead and try this.
I don’t want to argue but NE555 have a potential to be stable in times when there were no Rai-Rail OpAms. NE555 use OC output to drive RC time setting circuit and it is not based on the stability of the push-pull output.
Here you use output assuming it is Rail-Rail. Both on and off offsets from supply lines cares (their dependency on temperature and supply voltage).
Input offsets problems look for me the same in 555 and this circuit.
But… you are thinking toward different goal then me. I considered 10Hz (not kHz) with stability not important, but total supply current in a range of few uA or better less. But it is needed for something I will be doing in a few years if at all.
Hi, Piotr
I learned about this oscillator design in about 1975 before we knew about rail to rail. In fact if you can insert an output offset into a calculation or simulation, you find that such offset affects both + and - op amp inputs in a manner which mostly cancels. That is one reason why this is a good design. I think you could build 4 good low frequency oscillators using an LM324 or using another amplifier with lower supply current.
Anyway if you like the CMOS 555; go right ahead and use it. I am not saying it is bad. But I am not yet aware of a 555 timer application which cannot be done as well or better with an op amp or comparator. (Or perhaps one 8 pin dual amplifier/comparator IC. The 555 is one 8 pin chip.) Does anybody know of one?
In our first product (Eprom Programmer “Piccolo”) designed in 1988 I used 555 as step-up (5V->25V) DCDC controller. We had very limited access to ICs and their data - it was the cheapest solution.
I know two ways of controlling output voltage with feedback (to pin 4 (Reset) or 5 (influencing the internal voltage divider). Reset just pauses generator if voltage is too high. Changing voltage at voltage divider influences t/T factor (PWM).
I don’t remember which one I used and why.
You might have come up with something which would be difficult to closely replicate with 1-2 comparators. But for many years I designed mostly AC-DC power supplies which operated in boundary mode (also called critical conduction.) It runs at variable frequency; lower frequency at full load and maybe 10X at light load. This used no power supply controller chip; only a TL431 (reference really also includes an error amplifier) and optocoupler and a few discrete parts.
I think that important part of 555 was voltage divider made of 3 resistors of the same value. I believe their values were consistent with each other with high accuracy and had the same temperature characteristics (even their actual value was not precisely defined). Thanks to that switching thresholds were precisely defined to be 1/3 and 2/3 of supply voltage.
When 555 was invented I suppose typical tolerance of resistors were 20% and there were no small 1% resistors so 3 such resistors could occupy more PCB place than whole IC.
NE555 was designed in a time that putting 40 or so transistors on a single piece of silicon was still impressive. (I can’t even make one for that matter) NE555 was not very accurate and control & trigger voltage levels did not have particularly tight tolerances, unless 10% was exceptional in that age.
Sidenote, “Zeptobars” makes pretty pictures of IC’s, including the 555, and has been doing so for almost 10 years and he’s build up quite a collection.
https://zeptobars.com/en/read/Ti-555-NE555-real-vs-face-china-chinese
Thanks, paulvdh
I remembered that the original 555 had poor timing accuracy as an oscillator. Just now I checked out
and I do not see that specified the same way. But I expect this newer one to be better in almost all regards. One possible exception:
The original 555 had fairly strong output drive but that came with some output totem pole shoot-through. The 555 might be able to drive some MOSFETs without a gate driver; more so than most comparators.
Page 9 of the datasheet specifies operating conditions for a 15V supply:
And indeed, the NE555 can be used just as a gate driver. It can source or sink 200mA which is much more then a typical uC pin, and at the same time have a higher output voltage swing so you have a much higher margin for FET’s with a high Ugs.
It’s not a great gate driver, but it’s “good enough” for a lot of situations, and combined with it’s wide availability makes it an attractive option for hobbyists.
I agree. But the hobbyist must be careful to bypass the supply rails right at the +Vcc pin…Bypass cap should be the next component to place after the chip… This is good practice anyway and was learned from the days of TTL logic.
Well, this brings back memories. One of the first 555 timing circuits I built was a little bit unstable on keeping frequency, and the first capacitor that I used for timing was EXTREMELY temperature sensitive. It was not the 555 causing the frequency instability; it was the first type of capacitor that I used (I think an electrolytic (because of price)).