favurite boost regulator?

Ofcourse. There are loads of them. Even ones that will work with 1 AA cel.

TPS61200 ?

-Lasse

I'm interested in the transistor thingy. Do you have a reference or could post the schematic?

Many thanks in advance.

Marcel

On a sunny day (Sat, 19 Nov 2011 08:31:02 -0800 (PST)) it happened " snipped-for-privacy@fonz.dk" wrote in :

it.

That is a very nice chip, high efficiency. I get about 50%, this is much better. Thanks. :-)

it.

It is an extremely nice chip. How will you connect to it? Will you lay out a PCB?

John S

On a sunny day (Sat, 19 Nov 2011 19:04:15 +0100) it happened Marcel wrote in :

it.

Here is the old schematic for 1 AA cell, to 5V, to drive some LEDs:

----------------------------------- | | | |100u/6V e | | -------------- b PNP | === | | c Q1 | --- | | | | ( | | | R2 [ ] 10 | ( 33uH with 8 turns | [ ] 680 | |_____ | ( feedback wound over it | | R1 | | )| ( | | | | | | 1N5822 Schottky /// /// | === | |---------|>|-------------------- ~ +5V | --- | c | | | | | 1u |_ b NPN 5k6 e === | /// e Q2 |---b PNP --- | | zener --- c Q3 | 10uF/6V | /// 4V7 \ | /// | / \ | | --- | | | | | /// | |_________________________________________| NPN: BC547-B PNP: BC557-B

The losses are mostly in the diode, (voltage drop), and the transistor switch Q2 (not all the way to zero volt), and also some goes into Q1 and the 10 Ohm resistor.

When output voltage rises above about +5V then Q3 start conducting, and that makes the base of Q1 positive relative to its emitter (at the low input voltage), that reduces the base drive of the switch Q2, completing the feedback loop. I got about 48 % efficiency at 60 mA output and 2.0 V input. Better at higher input voltages.

You may notice that in the web page I have a IRLZ34N MOSFET in series, it is driven from a PIC that senses if the voltage drops below 1V, and then removes the gate drive to that MOSFET: battery protection. It is started by simply shorting the drain-source of the MOSFET, so 2 push buttons, one for start, and one for stop in that LED thing. This one will drive the scintillator probe + LCD display, where the LCD backlight is the biggest power consumer. I will do the same MOSFET and PIC voltage sense thing, but here at 2 volt, for 2 AA cells in series. The main difference for 1 AA cell and lower currents is that R1 can be higher (3k9). R2 was higher too, 47 Ohm IIRC.

On a sunny day (Sat, 19 Nov 2011 13:31:59 -0600) it happened John S wrote in :

it.

I will first see if I can get it locally. Maybe I can solder a copper strip to the thermal pad, and some fine wires to the connections, upside down.

esigned it.

if you only need one:

-Lasse

On a sunny day (Sat, 19 Nov 2011 13:10:08 -0800 (PST)) it happened " snipped-for-privacy@fonz.dk" wrote in :

But it is in Colorado:-) I see mouser has those for 2 Euro 68 cents in single quantity. Will have to wait until I have build up enough stuff for their free shipping limit. Never seems to happen...

designed it.

But, according to the data sheet, the connections are only 0.5mm apart. You can do that by hand?

designed it.

I solder stuff like that for nearly every project. Piece of cake.

designed it.

Please tell me how! I'd like to experiment with some devices on a breadboard, but devices like this keep me from doing so.

I'm all eyes.

John S

designed it.

I'd like to know too.

The outside pins are doable by hand (around 19 mils) and especially with a good magnifier and/or microscope, but the inside thermal pad isn't easy to deal with without some kind of reflow oven or other heating method.

If you're just playing within low current, low input voltage and low dissipation, you could leave the pad unsoldered I suppose.

boB

te:

o d=3D

its sparkfun several of european distributors as well, just google lipower and board or something like that

-Lasse

ho designed it.

lay

t.

I'd still connect it though, if you do it upside down just use some thicker wire

if you hand solder it on a pcb make the pcb with a big via in the pad so you can solder it from behind

-Lasse

And you'll need all of them! :)

Jamie

esigned it.

this one looks even simpler:

-Lasse

I've used pad style GHz op-amps before, placing them on an erroneous footprint (yeah...oops).

First, jumpers were placed on the PCB to fan out the existing pads. The chip was placed top against the PCB, deadbug style. Bypass caps were soldered directly onto the thermal pad, which is ground on this particular chip. Jumpers connected the chip's supply pins to the caps on the pad with a length of about 100 mils, thus achieving bypass comparable or better than PCB construction alone would allow. Unfortunately, the input and output terminals still had to travel the extra distance to reach the board (off the chip, down and over, to reach the existing footprint), but that was well within tolerance and no oscillation was observed. Ultimately, it performed exactly as expected.

I've heard of others hand-soldering them to homemade PCBs by drilling an extremely large hole in the middle of the footprint, enough to basically do the solder equivalent of plug-welding the hole, from thermal pad to top copper to bottom copper.

For deadbug purposes, you could tack it to a hunk of suitably shaped copper, maybe a bent strip of copper shim stock, so you have wings to tack to the main PCB you're working on. That also provides a handy mount which holds the chip in mostly-free space, while conserving the pinout (it's not upside-down!).

Say, that would look a lot like a mostly-SMT version of those old DIP ICs, the ones for medium power audio amps and vertical deflection and such, the ones that had the middle couple pins fused together as a wide strip for heat dissipation.

Tim

On a sunny day (Sat, 19 Nov 2011 17:36:00 -0800 (PST)) it happened " snipped-for-privacy@fonz.dk" wrote in :

Yes, I already did that, but then typing the chip in the search windows gives a negative. For UK.

UK is OK normally as it is fast to the Netherlands and no import tax and handling. Mail is cheap too.