1v2 to 5v very efficient DC-DC converter needed

[neazoi]

Having tested 3-4 schematics of 1v2 to 5v using discrete components, I have been disappointed by their efficiency and I would like to know if an efficient DIP chip exists for that purpose.
Output current is low I would say under 100mA maybe much less, but It must be a DIP packade or a three terminal device, no SMD
Any help on this?

 

[andre_luis]

Generally SMPS controller is not properly the root of the problem, but mainly switching devices and transformer.
Can you evaluate what component of these schematics is the one which dissipate more heat ?


+++

 

[neazoi]

most of the circuits (joule thief variations) failed to even give 5v, they mostly give less than 4v with a single rechargeable cell.
When I connect a PSU source and increase the input voltage, the transistor of the inverted gets too hot, and it dissipates 200mA.

I need to power a micro and an atmel serial eprom as well as a PC keyboard controller, so The needed current is low.

 

[andre_luis]

Could you attach here this circuit ?


+++

 

[dick_freebird]

Seems like you just want a boost converter with about a
10V output device (for luck) and a very, very low Vgs to
make full current - but you'd be in a bind with "off" switch
leakage and "on" resistance having only 1.2V (typ) to
separate them.

And at that stepup ratio the 10V device needs to be
capable of 500mA or more, at the corners. Probably 1A
at typ VIN and 25C. That's a big switch to fit in a little
package.

Not to say it can't be done. Or isn't worth doing. But
my gut says, not easy, not small and not really wanting
a high thermal resistance package style (like you request)
either.

 

[neazoi]

so isn't out there any 1.2v booster to 5v at all (even on a chip)? I do not believe that.
the circuits I have tried are these short of **broken link removed**
https://www.eleccircuit.com/step-up-dc-converter-1-2v-to-5v-5v-for-micro-computer/

 

[BradtheRad]

If you want to get 100 mA output at 5V, your 1.2V supply must provide more than 1A current bursts.



The switching device needs sufficient bias.

The diode is necessary.

 

[betwixt]

The MAX1797 is the best device I've found for doing this. It has adjustable output but also has pre-set 3.3V and 5V voltages. It can supply 1A at 5V and with inputs even lower then 1.2V. I have had them working down to 0.7V!
The only drawback is it's size, they only make them in uMax8 packages which makes them tricky to solder by hand but certainly not impossible.

Brian.

 

[FvM]

Having tested 3-4 schematics of 1v2 to 5v using discrete components, I have been disappointed by their efficiency and I would like to know if an efficient DIP chip exists for that purpose.
Output current is low I would say under 100mA maybe much less, but It must be a DIP packade or a three terminal device, no SMD

Chips with acceptable efficiency (0.85 - 0.9) are available from major manufacturers (Maxim as said, also Linear and TI), but definitely no DIP device.

Why not get involved with surface mount technology?

 

[neazoi]

Thanks all for your replies,
Avoiding SMD is basically an issue of easiness of construction from other inexperienced users. I am building this https://translate.google.com/transl...fram.de/projekte/avr/chipbasic2/bedienung.php and inside a keyboard **broken link removed**
As you can see it is more an amateurish construction and smd should be avoided.
The max1797 you propose is perfect for the purpose. Maybe you could propose other not so efficient but DIP ones from linear technology or maxim etc? That would be helpful.
I have max1771 here but this starts at 2v.
Merry Christmas!

- - - Updated - - -

If you want to get 100 mA output at 5V, your 1.2V supply must provide more than 1A current bursts.



The switching device needs sufficient bias.

The diode is necessary.

Could you remake the calculation of input current requirement, for just 20mA output?

 

[betwixt]

Nice! And I agree about keyboards. I'm using an old compact 102 key 'standard' keyboard. Most of the ones on sale these days are cluttered with a multitude of keys for working in Windows which are a waste of space when you spend most of your time programming and in Linux!

Brian.

 

[BradtheRad]

Could you remake the calculation of input current requirement, for just 20mA output?

Input current peaks at 217 mA (simulated).

You adjust current by changing the bias resistor value.