Step-up from ~3V to 24V

[glatocha]

Hi,

I am searching for a DC-DC converter that can produce for me 24V from 2 or 3 AA batteries. So best would be starting from 2V as someone might use rechargeables.
Current that I need should be ~30mA MAX. Mot continuous so heat should not be an issue.

I was searching some pages, but very difficult to find something like this.

If somebody knows something please help.

[gman4925]

Here you go:
http://www.pololu.com/catalog/product/799

[Psi]

Digikey have one that could work and is only US$5.83 in qty =1

However it's only 3.0V -> 3.6V input. So you'd need a 3.3V linear reg on the input from 3 cells which isn't ideal but is still quite a cheap option.

http://www.digikey.com/product-detail/en/RO-3.324S/945-1939-5-ND/2315043

* 3.0 - 3.6V input
* 24V output
* 42mA max current
* 85% efficient

[glatocha]

Thank you both.

I was thinking more like get an IC and make it on my PCB. Should be maybe cheaper.
But if not possible, will use one of your proposals.

[Psi]

There's 1000s of ICs you could use to make your own circuit for this.

Some semiconductor manufactures have a tool on their website for powersupply design.
You enter your input/output/current specs and it gives you a list of suitable switchmode controller ICs.
Then all you need to do is adapt the datasheet reference circuit to match what you need.

[glatocha]

I know.

The problem is, that they list Input Range and Output Range. When you look at datasheet it turns out that for my output the input needs to be higher.
But I will dig it deeper.

Any leading IC manufacturer for DC-DC?

[DutchGert]

I know.

The problem is, that they list Input Range and Output Range. When you look at datasheet it turns out that for my output the input needs to be higher.
But I will dig it deeper.

Any leading IC manufacturer for DC-DC?

Linear and TI

[Paul Price]

If you have a bipolar or FET version of the popular 555 timer on hand, a 30V n-chan logic level gate MOSFET on hand, you could make a DIY SMPS controller that works quite well.

A 555 timer configured as a free-running multivibrator circuit, the 555 output driving a low-voltage gate type of  n-chan MOSFET could be used to create a very good SMPS flyback circuit. Use a 12V zener/NPN pass transistor  to drop the filtered flyback output of 24V to bootstrap the 3V that starts up the 555 timer. That is to say, the power pin for the 555 is in in series with the Schottky diode trough a NPN(2N3904 and 12-V zener regulator) to power the 555. You use a schottky diode from the flyback coil to create a 24V filtered output, a small 22uH choke for the flyback coil. 555 to drive the MOSFET gate with a 3uS to 5uS on time, 20uSec off time, ( to start with), tweak on/off 555 timing adj. resistors for on/off time for best efficiency. A 24V zener through a small-valued resistor(100ohm) into the base of a npn small signal transistor(2N3904), collector to 555's Reset, using a 2.2meg pullup resistor to the +12V bootstrap, emitter to ground will provide good 24V regulation. Another npn transistor is needed for current-mode operation, (i.e., a 2N3904), connect its base through a 100-ohm to  a 1-ohm current sense resistor to the source of the MOSFET to ground, its collector or'ed with the other npn to the 555 Reset, emitter to ground, will complete the circuit, except for a timing capacitor (50pF approx and two resistors for on/off time), altogether making it a very good current-mode switching controller but requires more discrete parts than are used by a SMPS specific IC, but a 555 timer might be on hand and cheap if you want the job get done fast. The ones I've designed will start at approx 2.8V ( 555C FET version) battery V and continue to work reliably down to 1.9V once started with approx 75%-80% overall efficiency at small current supply output. Adj resistor values for the correct pulse widths at 3V through the Schottky diode and filter cap before installing the MOSFET. A simple voltage divider to the base of the voltage regulating npn transistor can be used if you don't have a 24V zener on hand to set the output voltage with somewhat sloppier regulation. A .1uf bypass on pin 5 is also required as well as a 1uf or greater bypass from pin 4 (VDD) to VSS.

Ok, a schematic would have been better to post.

[Zero999]

I know.

The problem is, that they list Input Range and Output Range. When you look at datasheet it turns out that for my output the input needs to be higher.
But I will dig it deeper.

That's probably because the only configurations you've looked at use a single inductor fly back topology which don't work very well when the output voltage is over 5 times the input. A fly back transformer will make it easy to get 24V from 3V.

[mariush]

I recently worked with LT1307 : http://cds.linear.com/docs/en/datasheet/1307fa.pdf

The datasheet has an example circuit using flyback, and it can do 24v easily (up to 85v actually)..don't know how much current though.

[Psi]

I recently worked with LT1307 : http://cds.linear.com/docs/en/datasheet/1307fa.pdf

The datasheet has an example circuit using flyback, and it can do 24v easily (up to 85v actually)..don't know how much current though.

There's also an example for a LCD Bias Generator at the end of the datasheet which produces 24V from 1, 2 or 3cell (5/15/35ma output).
Can remove the -Vout and pwm input sections and it should be pretty close to what he needs.

[glatocha]

Thank you all guys.
Have to schedule the LTSpice weekend