DC DC Step Up/Down - How it Works?

Ran across this little guy...anyone know what's going on? Is that a little transformer on this thing, or how does it work? Or is it a little joule-thief? Would it be hard to DIY?

Thanks

http://www.ebay.com/itm/192154856978

  SMPS (switching mode power supply)

  We are pretty familiar at this point with those, in fact today I have a test about those  :-\

  There are plenty of single chip solutions, you need to add a few external components and ready to go. They do need an inductor but nothing you can't wind yourself. They are (usually) much more efficient than the joule-thief you mention, they use an inductor and switch it between two positions making the conversion ideally lossless.

JS

Thanks Joaquin!

If I were to DIY one would you know of any preferred designs or good chips? Or should I just google it?

I want to take a AA and run an LED for a project.

TIA!

PS Good Luck on the test!

Ps found a great article on these. I'd still like to be able to DIY one so if anyone has a thought that'd be great.

https://www.maximintegrated.com/en/app-notes/index.mvp/id/4087

Phrazemaster said:

Ran across this little guy...anyone know what's going on? Is that a little transformer on this thing, or how does it work? Or is it a little joule-thief?

Click to expand...

First, the converter to which you linked is a step-up only.  While it claims a given input range and output range, the input must be less than the output, it cannot step down to an output lower than the input.

As for how a step-up (boost) converter works, start here.



You have two storage elements, an inductor and capacitor, and a switch. When the switch is closed, current builds up in the inductor  (remember that current through the inductor cannot change immediately). When the switch is opened, the path changes, and the current is dumped into the capacitor through the diode. Close the switch again, and the diode prevents the capacitor from discharging back into the inductor. Switch fast enough, and your output voltage builds up and remains stable.

The switching is the trick. You need something smart which controls the switching frequency and duty cycle on a dynamic-load-dependent basis. But the good news is you can buy chips for a buck which do that control, and often have the switch built in.

Would it be hard to DIY?

Click to expand...

Not really.

Terrific explanation Andy - thank-you!

Any reason not to implement something like this to power an LED on a front panel? Say snake off a rectified output from a transformer and power a light?

Thanks for the brilliance!

plenty of youtube video's about how buck boost converters work.

There's a lot of scary talk about buck boosts in audio. Most of it from a time when they switched at near-audio rates. These days, modern buck boosts are running in the hundreds of kHz's, if not higher. most of the time, they are way beyond the audio range. In addition, many folks follow them with some aggressive filtering and additional linear regulation.

/R

Phrazemaster said:

Terrific explanation Andy - thank-you!

Any reason not to implement something like this to power an LED on a front panel? Say snake off a rectified output from a transformer and power a light?

Thanks for the brilliance!

Click to expand...

You only need a few volts and a handful of parts to light an LED.  Why would you need a SMPS?

> I'd still like to be able to DIY one

WHY??

It is three bucks *shipped*. Similar things for different up/down and outputs are also real cheap. That will not start to buy the coffee/beer you need to work-out the values, much less buy parts.

Phrazemaster said:

...
Any reason not to implement something like this to power an LED on a front panel? Say snake off a rectified output from a transformer and power a light?
...

Click to expand...

Many, usually you use a resistor for that, much simpler and only wastes a few mW at most.  If you have serous power in lights that worth a SMPS to feed them you'd probably better doing so directly from mains, and leave the transformer at the shelve (or only doing noise sensitive tasks) Also SMPS are well known source of noise, even if filtered correctly or only feeding misc stuff on your rig, you still have to be careful with EM emissions from the inductor working at HF.

JS

PRR said:

> I'd still like to be able to DIY one

WHY??

It is three bucks *shipped*. Similar things for different up/down and outputs are also real cheap. That will not start to buy the coffee/beer you need to work-out the values, much less buy parts.

Click to expand...

True but...isn't that cheating? Fun to learn stuff.

I appreciate all the thoughts gang.

PRR said:

> I'd still like to be able to DIY one

WHY??

It is three bucks *shipped*. Similar things for different up/down and outputs are also real cheap. That will not start to buy the coffee/beer you need to work-out the values, much less buy parts.

Click to expand...

He did say:

I want to take a AA and run an LED for a project.

Click to expand...

Okay, so this switcher does operate with 1.5 volts input, but to light a LED with a 1.5V battery you might as well use the simple "joule thief" circuit.  A joule thief is basically the bare minimum of the switching part of a switching regulator without the regulator part.  Googling shows schematics with five components, and that's including the battery and the LED.

Helpful. Thx!

You could also add a second AA and then you only need a resistor.

Matador said:

You could also add a second AA and then you only need a resistor.

Click to expand...

Of course, and it's been a consideration. This project calls for light weight, so if I could do it off one AA that'd be good. But I'm leaning towards 2 AAs, and also I may need to power up to 3 LEDs - haven't decided this part yet. So having flexibility is good.

But yeah, simpler's usually better. 

Phrazemaster said:

But yeah, simpler's usually better. 

Click to expand...

Probably not helpful, but there are those cheap keychain fobs with a white LED and a 3V lithium coin cell, no resistor, I guess the battery's wimpy output Z does the current limiting, Must be tough on the battery.

Gene

Nah it's fascinating how theory and reality commingle sometimes. "In the trenches" vs on the board.

Thx!

SMPS layouts can be very tricky.  I've have excellent luck with that MAX1771 chip, but you need to select an inductor, FET, and diode per application.  The chip switches at pretty high frequency, so layout around the feedback loop is critical to ensure stability (you can't clobber the ADJ/feedback pin's with bypass capacitance like you can a linear regulator).

Keep an eye out for chips that feature "self-controlled switch frequency", because with a low step-up and low current loads the switch frequency can drop down into the audio band to maintain efficiency.  Ideally you can set the switch frequency with a small cap if the chip supports it.  Also many chips are designed for Li-Ion cells, so the minimum input voltages may be too high for a single AA battery.

Lastly, essentially all of these chips are available in SMD only, in case that is a consideration.

As an example: the NCP1400 chip from OnSemi can go down to 0.8V input voltage, and has a fixed 5V out at up to 100mA.  You can get many different output voltages, and the switching frequency is fixed at 180kHz.  Seems ideal for audio use.

SparkFun has a $6 breakout board for the 3.3V version:  https://www.sparkfun.com/products/10967

There's also a huge array of modules with all elements included, from little potted bricks to open boards, e.g.

http://www.mouser.com/ProductDetail/Murata-Power-Solutions/CRE1S0505S3C/?qs=sGAEpiMZZMsc0tfZmXiUnehdnqY2%252beSYIniGrHHDhfkvSGqygV25ng%3d%3d

or

http://www.mouser.com/ProductDetail/Texas-Instruments/DCH010515DN7/?qs=sGAEpiMZZMsc0tfZmXiUnZtq28tdDzsSiBlckuM2tAg%3d


All sort of options in various form from complete roll-your-own with discretes, to a controller IC + full external parts (switching transistor + caps/inductors, etc), to controller IC + external inductor only, to full circuit module.

Many of the full modules are no different to implement than a linear regulator, though you should still be aware of things like their switching frequency (and pick something out of the audio band), intended load current, etc.

If you can try to define the particulars of the circuit as best as you can:-
- Vin (AA battery)
- Vout
- Nominal Load
- Worst case load
- Etc

Then that will lead to the short list of options and you can choose the right compromise between simplicity (full module) and optimization (probably a controller with an external inductor is about as deep as you'd want to go for such a thing.  External switching transistors would be of more use in higher power where you want to optimize that stuff for your particular case).
In the end it will probably be driven largely by form factor and price for your application.