DC DC converter?

[yclo]

I've been reading a lot of topics that deal with small ic converters but haven't found anywhere that sells them in melbourne. But I did find a converter ic MC34063 for sale for about AUS $5.

Would anyone know how to implement this into a single cell step-up for led usage?

Thanks!!

Ying Chee

 

[Steelwolf]

Where did you find this IC? Are there any product data sheets for it? Who makes it (National Semiconductor, Maxim, etc.)?

 

[**DONOTDELETE**]

Hi, Manufacturer of MC33063A is Fairchild,data sheet at www.fairchildsemi.com/ds/mc/mc33063a.pdf
- but does not have any application data.

 

[MrAl]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by yclo:


Would anyone know how to implement this into a single cell step-up for led usage?

Thanks!!

Ying Chee<HR></BLOCKQUOTE>

Well, according to the data sheet
that chip requires a minimum of
3 volts input to power it. This means
it wont be able to be easily used for
a single cell LED flashlight.
In fact, it wont even work in a two
cell flashlight because the voltage of two
cells in series quickly dies down to below
3 volts after a short operation time.
Also, the input current is quite high
and the max operating frequency is quite
low, indicating to me that this is a chip
on its way out. That's probably why
there is no app sheet, because they will
discontinue this item soon or just not make
anymore of them.

Hopefully, i'll be able to do a multi-
manufacturer multi-chip comparison soon.

I already checked out the Maxim inventory and
found a very nice chip, but it too
requires 2.6v minimum input. There is a
REALLY good Zetex chip made just for white
LED's that operates current mode.
I'm currently trying to find out how to
get these in small quantities.
If anyone is interested in ordering, say,
500 pieces and we can get 4 other people to
order 500 pieces(cost about $250 to $300 US)
we can get some now

--Al

 

[yclo]

damnit, stupid @&^$% chip....

 

[Steelwolf]

Yclo, if you're in Melbourne and want to try some single cell LED flashlights to build yourself, head over to **** Smith's Electronics.

For about AU$17, you can get a kit for single LED, 1 AA battery flashlight. It actually uses a bunch of discrete components (transistors, resistors, etc.) instead of a do-it-all IC. So once you have the schematic, you can probably build more at a fraction of the cost of the original kit. (kit no. K3018)

The other kit is a 6 LED single D cell flashlight. Unfortunately it uses a Maxim 1676EUB chip at its heart. I only just got it, and haven't tried it out yet, but a quick check through some local electronics components suppliers have proved futile for the chip. It will set you back about AU$52. (kit no. K3019)

The advantage of K3019, according to a quick read of the accompanying documents, is that the output is regulated. K3018 is apparently not regulated, as are many of the homebrew schematics that have so far been posted that do not use ICs. But at least you can build those from cheap and typically easily available discrete components. The IC chips typically needed are harder to find.

I think someone once posted a schematic featuring the Maxim 1674 chip which apparently is quite similar to the MAX1676. Still, you have the problem of finding the correct chip.

 

[yclo]

I have already bought and built the **** smith single AA led torch kit. But after I built it one question came to mind. This circuit is not regulated, all it does is steps up the voltage. When built it's the size of an AA battery, so it'll fit into a 2AA torch.

So the question is why not just use 3 N cells (+ resistor) instead? It'll be the same size as 1 AA cell + circuit, and does exactly the same thing.

Ying Chee

 

[Steelwolf]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>So the question is why not just use 3 N cells (+ resistor) instead? It'll be the same size as 1 AA cell + circuit, and does exactly the same thing. <HR></BLOCKQUOTE>

Cost? Run time? Aren't N cells harder to find and quite a little more expensive?

Actually, if you really need to squeeze a booster in to a tight space, the smallest is the SatCure. Quite bright too.

But as for regulated??? We need help.

How does a regulating circuit work? Obviously there must be some form of feed back and comparison. How to do that? The booster must then be driven harder to deliver the same amount of power. How does that happen? How does one cause any booster circuit to "drive" harder? Increase/decrease switching frequency?

 

[MrAl]

I've used 3 N cells plus resistor in my AA Mini Mag and it works fine.
The problem is the cost of battery replacement.
The run time is pretty good, because the N cells have a decent
rating(although not anything to scream about).
I dont use this flashlight that much anymore though, because
i build other 2 led type lights that are much brighter.
I run the led's at 40ma each and get quite a bit of brightness
(used for short time periods). If your going to use your
light for long time periods, i would recommend only 20 to
30ma.

There are two types of regulating circuits that will provide
full brightness over the life of the battery:
1. The voltage feedback method:
regulates the output voltage of the circuit, and you then
use a dropping resistor to limit the current--similar to
using 3 cells in series.
2. The current feedback method:
regulates the output current of the circuit, and you dont
need a dropping resistor to limit current, so you get
a little more efficiency.

Presently, you cant find the combination of low input voltage
(0.8v or lower) and current feedback regulation in any Maxim
chips. This means if you want to use a Maxim chip and you
want to use one cell only then you have to use voltage
feedback and a dropping resistor. This isnt all that bad really.

What do you all think is better for flashlights:
Transistors or I.C. chips?

Currently the Brinkmann uses 2 transistors.

--Al

 

[Steelwolf]

MrAl, Definitely, when it comes to having a good circuit at a small size with simple layout and low component count, IC based circuits can't be beat. But many of them require specialised chips that we may not be able to find, or which are expensive. So it would be nice, and I have been collecting the schematics, to have circuit designs that feature ICs and do the required job well.

But I would really like to find some good transistor based circuits that would do the job as well. Certainly, many of the designs would probably be very complex with large layouts and high component count, but sometimes that can't be helped. It is not so much that I want transistor based circuits, but that the components be non-dedicated/non-specialised and easily found.

For example, I built a flip-flop based booster circuit in 2 ways. The first with transistors, the second (which doesn't quite work right yet) using the 555 timer chip. This is acceptable because the 555 timer IC is as easily available as any transistor.

yclo, I haven't built that many so far, and not all of them have worked (probably because of some bad breadboarding). But looking at all the schematics I have collected and especially at the ones I have built, most of them are quite large. So far, the most compact one that uses easily available components is still the SatCure. The Brinkmann circuit is probably a close second, but I haven't got it to work yet. The **** Smith circuit in K3018 is next in line for size, but that is an asynchronous flip-flop circuit which could possibly be replaced with a 555 timer IC, thus reducing it's overall size and component count. And I have built another asynchronous flip-flop circuit that has an even higher component count, but that one works.

 

[yclo]

So how many different single cell driven led circuits have you built? How big are they?

 

[MrAl]

Oh thats very interesting too Steelwolf,
so if you dont mind using the 555 timer
chip, how about the LM339 comparator chip?
Thats a pretty popular chip and widely
available and you can do so much with it.
The reason im asking is because using one
of these LM339 chips in a design might make
it a very good current regulating LED
driver. The min voltage is 2v, so it would
have to be for a 2 cell design. The cost of
the chip is under $1.00 US.
Also, if your interested in a 4 transistor
regulating LED driver ill start breadboarding
that and see how it works. The simulation
worked very well, regulating over a wide
range of input voltages, but of course you
have to put up with using 4 cheap transistors.

--Al

 

[Steelwolf]

MrAl, Thanks for your willingness to design all these circuits. You mentioned that you simulated these designs. May I know what software you use for this purpose?

I would be very interested in the 4 transistor regulated circuit. The LM339 comparator is not too difficult to find, I think. I know it is quite widespread in its use. A circuit schematic that requires such a chip would also be very useful. As I said, a circuit that doesn't require hard-to-find/specialised chips is always welcome.

 

[jeff1500]

I'd like to see the 4 transistor regulator circuit too.

 

[MrAl]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Steelwolf:
MrAl, Thanks for your willingness to design all these circuits. You mentioned that you simulated these designs. May I know what software you use for this purpose?

I would be very interested in the 4 transistor regulated circuit. The LM339 comparator is not too difficult to find, I think. I know it is quite widespread in its use. A circuit schematic that requires such a chip would also be very useful. As I said, a circuit that doesn't require hard-to-find/specialised chips is always welcome.<HR></BLOCKQUOTE>

Oh ok great.
As far as simulations go, right now im
using my own cooked up software, but i am
also going to try out a ready made package
made by a company called "intusoft" at www.intusoft.com .
They make a whole package and they will send
you a demo cd for free to try out.
I havent gotten to try it yet, but i hope to
in the future.
The cheapest pro version they sell is
$900.00 US though, with the next model up
$2000.00 . It runs on Win95/98/NT platforms.
If you are interested in this, go to their
site and fill out the application. If
for some reason you cant find it, try to
download some spice models of some parts,
and you will be prompted for address and
telephone data. A few days later a sales
person will call and offer you a demo cd.
I dont know how good it is yet, but one
thing you should know is that the spice
models vary very much, so that a simulation
isnt worth its weight in gold. You still
have to interpret the results somewhat. The
nice thing is, it will tell you approximately
how the circuit behaves through waveform
outputs and such.
Right now if you think more waveforms of
the two transistor circuit would help i could
post more on my site too.

Good luck with it,
Al

 

[MrAl]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by jeff1500:
I'd like to see the 4 transistor regulator circuit too.<HR></BLOCKQUOTE>

Oh ok, i guess ill get to drawing it out.
I think i'll breadboard it too.
I guess i was trying to decide whether to
go with the LM339 control or the extra
two transistor control.

--Al

 

[Steelwolf]

If it isn't too much trouble, having both circuits would be great.

BTW, mostly unrelated. With the now obsolete TL496CP, the kit data sheet shows that pin 1 is for "feedback" and pin 4 is for "T input". In the specified setup given in the kit, all other pins are used except these two. Does anyone know if by connecting these pins to the rest of the circuit in a certain way one can ensure that the output of the chip is regulated?

 

[MrAl]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Steelwolf:
If it isn't too much trouble, having both circuits would be great.

BTW, mostly unrelated. With the now obsolete TL496CP, the kit data sheet shows that pin 1 is for "feedback" and pin 4 is for "T input". In the specified setup given in the kit, all other pins are used except these two. Does anyone know if by connecting these pins to the rest of the circuit in a certain way one can ensure that the output of the chip is regulated?<HR></BLOCKQUOTE>
http://www-s.ti.com/sc/psheets/slvs012b/slvs012b.pdf

That chip looks unsuitable for
regulated LED driving. Although it might work, it will be very inefficient because
it uses a linear 9v regulator.
It's basically a 9v power supply controller.
Take a look at the spec sheet. It's
got the pins outlined as well as applications. The switching supply part of
it looks like it should work, but it will
of course be unregulated. Its kind of
hard to tell EXACTLY how this chip works
because the functional block diagram on
the site is pretty poor. Perhaps there is
better documentation somewhere else?

Good luck with it,
--Al

 

[PeLu]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by MrAl:
That chip looks unsuitable for
regulated LED driving. Although it might work, it will be very inefficient because
it uses a linear 9v regulator.<HR></BLOCKQUOTE>

The linear regulator only works when outside (probably mains) power is maintained. Depending on your input voltage the efficieny may be well above 70%.

For 'battery' mode, the linear regulator's efficiency is of no significance.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR> The switching supply part of
it looks like it should work, but it will
of course be unregulated.<HR></BLOCKQUOTE>

Yes, we all hope that they sell working circuits .-)
Actually they write that the switching regulator maintains constant output voltage, it is definitively regulated.

But the circuit is pretty old, newer and better ones should be available.

 

[Steelwolf]

Thanks for the help guys. I know that better chips are available, but I have a couple of them lying around and wanted to find out if the output could be regulated.

At the moment, the current setup as defined in the kit sheet does not regulate the output voltage. The data sheet MrAl posted seems to say that regulation is only possible when there is an AC input.

It might be alright in some LED applications, and at the moment, I have one set up to run 4 LEDs in a twin series/parallel setup. It sure is brighter than most other circuits I've tried. The application I was planning for it was more for a wireless mic which chews through 9V batteries like it has a share in Energizer. I can buy AA batteries in such bulk that it is ridiculously cheap, and if the chip can regulate the output till the battery runs dry, it would then be more economical to run.