Author Topic: Best choice oscillator for low powerdraw, low component count, ca 1-5kHz  (Read 1924 times)

0 Members and 1 Guest are viewing this topic.

Offline ALondaTopic starter

  • Contributor
  • Posts: 23
  • Country: se
    • Bizarre Electronics
Im a self-taught beginner level hobbyist trying to design some basic voltage doubler/booster circuits for other various projects.

My current choice of doubler is a boost circuit that requires an input from an oscillator to function. While I know how to make simple 555 and inverter oscillators, and have schematics for opamp oscillators, I dont know how to make RC oscillators or anything else.

As I am a bit of a perfectionist and minimalist, naturally im always looking to minimize component count and power consumption, as well as trying to make it as efficient as possible.

Ofcourse I could probably just order some cheap complete IC, but then I wouldnt learn anything so whats the point of that?

So I would like to put the question to all you knowledgable guys n gals, if you could maybe help me rank different kinds of oscillators in terms of component count (simplicity) and power consumption?

Also, is squarewave or sine better for a coil based booster? for a diode/transistor based doubler?

Ive heard something about a crystal based oscillator only drawing microamps, but the output would be in the 30kHz range, which I dont think would work with the 1mH coils I have for my voltage booster (my target voltage is 5V from about 3.5-4V, as well as 14V (only ~2uA) from 5V).
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15156
  • Country: de
For best efficiency one wants to switch the main current hard, so more like a square wave. The voltage can actually be more sine like in a resonant circuit.

There a several types of low power oscillator:
Lowest part count would be a small µC running of internal RC clock - this could be a few µA.
There are charge pump chips (e.g. ICL7660 like) that provide there own clock - but again kind of overkill.
From the NE555 there are similar CMOS versions (e.g. LMC555) with lower power than the old NE555, but usually still not very low power.
There is RC type with a low power comparator (e.g. MCP6541 or similar with supply current below 1 µA).
However on has to be careful with supply current specs - the current can go up, it the comparator / Schmidt trigger is close to switching.
This is a catch when considering a HC1G14 schmidt-trigger as an oscillator.
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23099
  • Country: gb
LTC6906. Much win that IC. 12uA. One resistor. Done!

Although you'd be better looking at the canned ICs that do this.
« Last Edit: July 09, 2019, 04:40:37 pm by bd139 »
 

Online Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Depends on how precise the frequency needs to be. If it's uncritical, a simple RC oscillator using a single-gate Schmitt trigger inverter is the simplest.
For better precision, an oscillator using a 32,768 kHz crystal plus divider is better.

Both can be built with consumption in the uA range.

You'll need to be a bit more specific.

 

Offline ALondaTopic starter

  • Contributor
  • Posts: 23
  • Country: se
    • Bizarre Electronics
bd139, its always disheartening to see replies like yours, especially since I specifically put

Quote
Ofcourse I could probably just order some cheap complete IC, but then I wouldnt learn anything so whats the point of that?

in my original post. I mean, whats the point of learning, eh? And you even mention a 10kHz+ IC when the topic says 1-5kHz.

Great job.
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23099
  • Country: gb
My apologies. That was an utter failure on my part.

By way of compensation, some advice and thoughts: I think your requirement of 1-5KHz is the wrong end of the problem you're looking at. Getting from the requirement back to the appropriate solution requires throwing existing approaches away and starting with a specification. Then gain understanding of the problem and research first before applying existing knowledge. From experience comes understanding and from understanding comes simplicity.

Art of electronics chapter 9, available free here: https://artofelectronics.net/wp-content/uploads/2016/02/AoE3_chapter9.pdf ... see section 9.6 on switching converters.

The trick is to learn the theory first, then look inside all the ICs via datasheets and see how they work. This is neatly packaged up in AoE but you need to go looking further. Then go back to the problem when educated.

You will then see why my assertion that the 1-5KHz and random 1mH inductors is the wrong end of the problem.

I built something very similar a couple of years back. I needed to derive four voltages from a 12V supply, one of which was 2KV. Applying knowledge attained, plus some external research, I built one that worked out of the box first time which had two transistors and a Royer architecture which is as minimal as it gets.

My design, now departed as it was a toy, is similar to the converter shown here: https://ludens.cl/Electron/scope/scope.png . These can be scaled down to tiny things running with a couple of 2n3904's using a few mA up to 2n3055 beasts delivering 10 amps.

But before all this you need to learn to walk so build some simple R-C oscillators and skip worrying about what they do and the efficiency. In fact if you want to design and build something discrete you need an intimate understanding of the parts. Learning the Art of Electronics, the companion book of the above is great and even covers building boost converters and how they work later on.
« Last Edit: July 09, 2019, 08:32:38 pm by bd139 »
 

Offline ALondaTopic starter

  • Contributor
  • Posts: 23
  • Country: se
    • Bizarre Electronics
thank you, bd139.

Ive never been one for theory. I think Ive learned pretty much everything by testing and failing until I eventually succeed. Programming, as well as electronics. I know very little about transistor theory, but as I experimented with them I think I know what results to expect from them.

Ive got a ton of breadboards with half-finished to finished circuits. But now im making the step to PCB and want to know if Im on the right (according to my own goals) path, before I commit to an order. Ive built a "voltage doubler" , working off a 1kHz 555 oscillator and testing a coil voltage booster which runs on a 2.5kHz oscillator feeding a 1mH coil getting over 14V on the output. Im not sure if I should just make a generic 555 breakout pcb to hook up to whatever else I make, or if I should include the 555 on the project board itself, or if I should go with something else (type oscillator) entirely. That is why I asked for advice.

I found an RC oscillator at https://www.electronics-tutorials.ws/oscillator/rc_oscillator.html but failed to get it working with a transistor, but got it working with opamp. I failed to get https://www.electronics-tutorials.ws/waveforms/astable.html working. Not sure why. Wrong values or low voltage. In both cases, all the theory explanations just goes in one ear and out the other -situation.

Ive got plenty of boost and buck converter breakout boards, so my goal isnt to get it done fast, my goal is to build it on a breadboard, figure out what makes it tick, and put it to use in my main projects. While I dont understand the voltage doubler, I atleast figured out that some of the diodes changed into schottky diodes made it perform better. If I had a proper 555 oscillator where I could alter frequency and duty cycle I could probably figure out more. But because of disability I work extremely slow. But Ive learnt to be patient.

I understand that theory can be important, but my brain is really too dim to retain that stuff anymore. But the practical stuff seems to stick around - Ive only burn up one arduino in a couple of years time.

I seem to have lost track of what to say so I'll stop here for now.
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23099
  • Country: gb
@ALonda. Same approach here for reference. My day job is software engineering. There isn't a lot of theory to be productive and most of that is simple rules of thumb and the Art of Electronics teaches those in an easy to understand model and gives you a mental model to work with.

555 is however fine for this type of thing if you must go down that route. Proof - here's a boost converter I built with one, an LM311 (to stop it oscillating once the voltage was right) and a MOSFET. This was used to generate high voltages to test leakage on 250V capacitors. This used a small radial inductor to do its thing, although I forget the value. This is based on the boost converter topology in Art of Electronics, literally hacked up with what I had on hand.



The RC oscillator you saw does work. It's called a phase-shift oscillator. They are difficult to get working sometimes. Thus you need to understand feedback and Berkhausen critera to debug one when it doesn't. The astable one is a relatively food proof one but it'll give you a headache, firstly because it will only source current and not sink it, so you need a driver stage after it before your circuit. Again referring to Art of Electronics, that's BJT amplifiers, feedback, oscillators that you need to understand first.

Working slow is fine. In fact it's best to absorb this slowly. Most people rush through and fail to develop an understanding. Plus if you go too fast you tend to miss some of the magical things that are hiding in these problems.
« Last Edit: July 10, 2019, 07:23:29 am by bd139 »
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20361
  • Country: gb
  • 0999
bd139, its always disheartening to see replies like yours, especially since I specifically put

Quote
Ofcourse I could probably just order some cheap complete IC, but then I wouldnt learn anything so whats the point of that?

in my original post. I mean, whats the point of learning, eh? And you even mention a 10kHz+ IC when the topic says 1-5kHz.

Great job.
The trouble is, you have conflicting requirements: low power draw/component count and educational.

If you want low power draw and component count, then a canned oscillator or an IC such as the LTC6906, are the way to go.

If you really want to learn, then start a discrete oscillator, such as the two BJT astable multi-vibrator, but it's not the lowest power or component count solution on offer.
https://en.wikipedia.org/wiki/Multivibrator#Astable_multivibrator

My advice is do both: play around with the discrete astable for fun and use an IC such as the LTC6906 or the real thing.
 

Offline Old Printer

  • Frequent Contributor
  • **
  • Posts: 796
  • Country: us
bd139, its always disheartening to see replies like yours, especially since I specifically put

Quote
Ofcourse I could probably just order some cheap complete IC, but then I wouldnt learn anything so whats the point of that?

in my original post. I mean, whats the point of learning, eh? And you even mention a 10kHz+ IC when the topic says 1-5kHz.

Great job.

This is a public forum, and while the OP's are usually just interested in their own problem being solved, the hundreds of others who read these posts can learn from everything posted as responses. A simple "thanks, but...." will serve you better in the long run. Don't bite the hand that feeds you, even if it's not the cookie you asked for.
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23099
  • Country: gb
I respect someone who calls me up on my incompetence ;)
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Does it have to be stable or precise?
If not, a simple RC oscillator around a single-gate inverter with a schmitt trigger input should not draw much at around 1kHz...
 

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1507
  • Country: ca
Why do you want a voltage doubler/booster circuit? I have designed and made thousands of circuits that use a supply voltage or parts that work properly together without doubling or boosting the voltage.

If you want to learn about the simple phase shift or mutivibrator oscillators you made that did not work then please post all the details (supply voltage, parts list and load resistance) of the ones you made.
Maybe you simply got the transistor pins connected backwards. The pins on an American little transistor are EBC but on a European transistor they are reversed at CBE, as shown on their datasheets.
 

Offline ALondaTopic starter

  • Contributor
  • Posts: 23
  • Country: se
    • Bizarre Electronics
Does it have to be stable or precise?
If not, a simple RC oscillator around a single-gate inverter with a schmitt trigger input should not draw much at around 1kHz...

I imagine not, for the 14V was to adjust it with a variable-zener circuit afterwards as it only has to supply 2uA (verified).

For the 3.5 to 4V -> 5V I have bought some cheapo 5V 200mA regulators.

Kleinstein mentioned that voltage booster circuits should run on square wave for best efficiency, so I think that rules out the RC oscillator in this application, even though Im still interested in learning to put together an oscillator with just resistors, caps and single transistor, or designs like that. The astable multivibrator would work I guess, but I havent made it work personally.
 

Offline ALondaTopic starter

  • Contributor
  • Posts: 23
  • Country: se
    • Bizarre Electronics
Why do you want a voltage doubler/booster circuit? I have designed and made thousands of circuits that use a supply voltage or parts that work properly together without doubling or boosting the voltage.

Your question made me question my own plans (for world domination). Ive designed a kind of data transmission over powerlines kind of circuit (I know there are ICs) and to protect against the "short circuit" of signals, voltage drops are involved. So my simple caveman brain just though to boost the voltage at the receiver side with some simple circuitry to get a stable nice 5V. But I might aswell run the powerline at 9V and straight up regulate at the receiver end, that'd even improve the power efficiency (the powerline will be upwards of 10 meters long, but <200mA expected)

But the 14V application is for a double random noise generator circuit that only needs ~2uA and Itd be too klunky to use a 14V+ powerbrick just for that when all the other circuitry is 5V.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15156
  • Country: de
A very simple oscillator with schmidt-trigger  (like 1/6 of 74HC14 or  74HC1G14) and resistor and capacitor. It is a little similar to the OP with RC circuit, just the positive FB part already included in the chip.  The supply current tends to be not extremely low as the chips are quite fast and need more current when near transition, but it's probably OK (e.g. some 20 µA).  So maybe a little less than the CMOS555 version.

Besides a separate oscillator and boost circuit, there is also the option to use an LC oscillator with inductor taps / combination with a transformer.
Such circuits are known as Joule-thieves - they are kind of fun for the beginning. It does not have to be rectangle. Hard switching in just usually the higher efficiency.

Having up to 200 mA available does not sound like one needs really low power. This is more like something for a normal SMPS chip for the 5 V and some kind of added charge pump/voltage doubler utilizing the SMPS clock for the 14 V.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf