(Trying to know how to do it without 555 IC) How do we generate oscillating signals today?

Thread Starter

blink13

Joined May 26, 2022
27
It might be a bit broad but I'd like it to confine on the astable operation of the 555 timer. That would be, generating a signal with a certain frequency and pulse width. Although we can also output sawtooth/triangle signal by using the capacitor connected to the 555, so there's also that.

But, how do we generate such signals in today's much more modern ICs?

I would also like to explore other methods because I know that in 555, the accuracy of my timing is highly dependent on the use of discrete components, resistors and capacitors, I get around at almost 10% and more difference on the frequency I am aiming for than what I have measured. I've also thought about using a digital potentiometer with the 555, but if there are other methods of signal generation without much dependence on discrete components and their actual values, I'd like to know more.
 

Thread Starter

blink13

Joined May 26, 2022
27
Have you even tried to peek about using Google, which serves well for such a generic query?
Not really, I've only known of things up to using the CMOS version of 555 as well as using ICs with Schmitt triggers on them. As for the use of microcontrollers, I've got an idea on Attiny.
 

Papabravo

Joined Feb 24, 2006
21,157
There are any number of oscillator circuits. Some are named after the inventors, and some have generic names.

For sinewave oscillators a non-exhaustive list might include:
Wien Bridge, Phase Shift, Bubba, Quadrature, Twin Tee, Colpitts, and Pierce
For square wave oscillators that list might include:
Relaxation and Multivibrator
 

Ian0

Joined Aug 7, 2020
9,667
We still use the 555, or probably a 74hc14 which is even less accurate. What makes you think that we need frequencies to be more accurate then they used to be? Most small microcontrollers are run from a relaxation oscillator with no great accuracy.
If we need accurate, then it’s a crystal in a Pierce oscillator, but with a CMOS inverter rather than a thermionic valve, and we’ve been doing it that way since the days of the MOS Technology 6502.

There are some laser-trimmed oscillators these days https://www.analog.com/en/product-category/timerblox.html
which achieve greater accuracy, but only because a Laser has been used to adjust the resistor value rather than having to do it manually on test with a preset.
 

BobTPH

Joined Jun 5, 2013
8,804
Most products today contain one ore more microcontrollers. These can supply nearly any pulsed signal you want while performing their other functions as well. If a crystal is used, the frequencies will be very accurate, if not, about 1%.

Bob
 

Reloadron

Joined Jan 15, 2015
7,501
Sounds like what you want is commonly called a "waveform generator IC" so you may want to Google that term. The link is to an AD9833 but there are several available depending on exactly what you want. Waveform and function generator ICs are nothing new. The XR2206 goes back maybe 30 years. You may want to Google XR2206. Many chips like this can be bought on a module as a turn key solution which you can modify or hack to suit your needs. Less knowing your specific needs this is about as good as it gets.

Ron
 

Thread Starter

blink13

Joined May 26, 2022
27
We still use the 555
That's a reassuring thing right there. At least there still is a reason for me on learning it. I thought, sure, I can do things with it, but if it's not the way how people do it now, then I'd be left behind.

What makes you think that we need frequencies to be more accurate then they used to be?
No specific reason, I just thought the more things improve the more they demand accurate parameters.

Laser has been used to adjust the resistor value
In my simpleton mind, I guess like a light dependent resistor but with lasers?

Most products today contain one ore more microcontrollers.
So almost or everything's going microcontroller, I should really start at least learning one.

If a crystal is used, the frequencies will be very accurate, if not, about 1%.
If we need accurate, then it’s a crystal in a Pierce oscillator, but with a CMOS inverter rather than a thermionic valve, and we’ve been doing it that way since the days of the MOS Technology 6502.
Ok so accuracy or at least, most of it, will depend on the crystal.

So, reading the other replies too, I take it that perhaps not much changed on how such signals are produced but it is already leaning towards the use of microcontrollers. Thanks a lot! you replies already get what I am trying to think of.
 

MrSalts

Joined Apr 2, 2020
2,767
In my simpleton mind, I guess like a light dependent resistor but with lasers?
No, laser trimming of resistors or crystals involves burning away small amounts of a resistor until the resistance matches the target value. This burning is done while the resistance or frequency is being measured and trimming continues until the value is within the specified range.
So almost or everything's going microcontroller, I should really start at least learning one.
Yes
Ok so accuracy or at least, most of it, will depend on the crystal.
mostly true for stable frequencies - even to the level that heaters are applied to some crystals to maintain a constant temp to maintain a constant frequency.
So, reading the other replies too, I take it that perhaps not much changed on how such signals are produced but it is already leaning towards the use of microcontrollers. Thanks a lot! you replies already get what I am trying to think of.
You're welcome.
 

DickCappels

Joined Aug 21, 2008
10,153
For one project I resorted to Direct Digital Synthesis. Here a stable reference frequency (generally a crystal oscillator) followed by frequency dividers and then a "DDS engine" and a look-up table and high speed DAC are combined to call up a specific frequency and waverform, often with resolution of a small fraction of a Hz. See the article below.

https://www.analog.com/en/analog-dialogue/articles/all-about-direct-digital-synthesis.html

https://www.analog.com/media/en/tec...4741144165304775709740692131461831AN823_0.pdf
 

MisterBill2

Joined Jan 23, 2018
18,167
Very few of my designs would use or even require a microcontroller or computer. Mostly they have a specific task to acomplish in a consistent manner, and do it accurately for years at a time. So why in the world utilize a micro and software when reliable and stable and totally immune to malware, hardware can do the task better and more reliably with less power required.

If the need is for a stable frequency with not much phase noise then a tuned circuit or a crystal oscillator is a good choice. The Hartley and Collpits oscillators with High "Q" tuned circuits can produce a quite good sine wave without much distortion, if one is able to avoid excessive feedback.
For much lower, audio, frequencies the inductive resonant circuits get big and heavy and so resistive-capacitive circuits are more commonly used.

So it gets down to a question of what frequency do you want, how stable and accurate do you want it to be, and how much distortion is acceptable. And none of those circuits demands any complex integrated circuits.

The very newest technology does not always mean an improvement, most often it means smaller and cheaper and almost good enough.
 

Papabravo

Joined Feb 24, 2006
21,157
Very few of my designs would use or even require a microcontroller or computer. Mostly they have a specific task to acomplish in a consistent manner, and do it accurately for years at a time. So why in the world utilize a micro and software when reliable and stable and totally immune to malware, hardware can do the task better and more reliably with less power required.

If the need is for a stable frequency with not much phase noise then a tuned circuit or a crystal oscillator is a good choice. The Hartley and Collpits oscillators with High "Q" tuned circuits can produce a quite good sine wave without much distortion, if one is able to avoid excessive feedback.
For much lower, audio, frequencies the inductive resonant circuits get big and heavy and so resistive-capacitive circuits are more commonly used.

So it gets down to a question of what frequency do you want, how stable and accurate do you want it to be, and how much distortion is acceptable. And none of those circuits demands any complex integrated circuits.

The very newest technology does not always mean an improvement, most often it means smaller and cheaper and almost good enough.
Contrary to popular belief, this is not new technology It has been around for quite a while.
 

djsfantasi

Joined Apr 11, 2010
9,156
Contrary to popular belief, this is not new technology It has been around for quite a while.
I believe this technology was used in latter Data General minicomputers. The problems it resolved and the problems it created is documented in “Soul of a New Machine”.
 

Papabravo

Joined Feb 24, 2006
21,157
I believe this technology was used in latter Data General minicomputers. The problems it resolved and the problems it created is documented in “Soul of a New Machine”.
I remember the use of PALs, but that was before the time of DDS chips as I remember them. The first semiconductor ROMs appeared in 1971.
 

MisterBill2

Joined Jan 23, 2018
18,167
Still we have no mention at all as to what frequency or range of frequencies the TS is seeking. That certainly matters a lot. Ther only hint was the desire to set "a certain frequency and pulse width." so probably the TS is not looking to generate RF . That does leave a rather broad range of frequencies. I am not aware of the 555 being good for over a megahertz, but maybe it is.

And the interesting things about using the very latest technology is finding all of the bugs that still need to be discovered and fixed, and discovering that there are production yield issues and so quantities are not available yet.
 
Top