Joule Thief - still not working....

[Jon Kirwan]

On Sun, 26 Jul 2009 17:29:56 -0700 (PDT), fungus
<openglMYSOCKS@artlum.com> wrote:

On Jul 26, 9:43 pm, fungus <openglMYSO...@artlum.com> wrote:

I don't know where I'd get one around here

As an aside, how do you guys buy parts? Do you have
big electronics stores with catalogs or is it through
distributors?

I get parts by: (1) Various distributors, such as digikey, mouser,
newark, allied, arrow, etc.; and, (2) anything electronic I'm throwing
away. Once upon a time, it was from many local suppliers such as
Allied Radio (before it went through business changes), John Meshna
(remember?) surplus, Lafayette, Radio Shack (before and slightly after
the Tandy [leather products?] merger), Olson electronics, polypaks...

Now there are just one or two very small, local stores (except Radio
Shack, which only has a few drawers of only a few things) in a metro
area of almost two million people. Used to be I could walk to four
huge warehouse stores in 15-20 minutes along one street in downtown
when the population was half as much. Gone, gone, gone. So one buys
by mail. One could do that, back then, too. But the outlets still
existed, as well. Nice, because you could handle the stuff or look at
the surplus items to see if they were interesting enough to try out.

Jon

 

[Jon Kirwan]

On Sun, 26 Jul 2009 12:43:10 -0700 (PDT), fungus
<openglMYSOCKS@artlum.com> wrote:

snip of question about where to get a TPS61160

Maybe there's lots where you live but I don't know where
I'd get one around here, and there's none on eBay.
snip

Mouser, Arrow, Newark, and Digikey all come up, for example.

Try:
http://www.findchips.com/
http://octopart.com/

when looking. Others may have even better suggestions for searching,
too. Watch for minimum order sizes and shipping/handling charges when
you order, though.

Jon

 

[fungus]

On Jul 26, 9:43 pm, fungus <openglMYSO...@artlum.com> wrote:

I don't know where I'd get one around here

As an aside, how do you guys buy parts? Do you have
big electronics stores with catalogs or is it through
distributors?

 

[David Eather]

Jon Kirwan wrote:

On Sun, 26 Jul 2009 22:24:34 +1000, David Eather <eather@tpg.com.au
wrote:

fungus wrote:
On Jul 26, 8:57 am, ehsjr <eh...@NOSPAMverizon.net> wrote:
I don't _know_ if it qualifies as "a whole lot" better, but
available one chip solutions can meet the op's stated requirement
of keeping the current at 15-20 mA, and the joule thief cannot.

Can you maybe recommend one...?
LM3909 - but they are hard to get hold of now

David, is this your recommendation for "one chip solutions can meet
the op's stated requirement of keeping the current at 15-20 mA" for
six 3.3V LEDs?

It stacks the battery voltage with only one capacitor, to double the
voltage. That's about it. And it doesn't control current over
voltage source variations.

Here is an LM3909 equivalent that works. R1 and C1 set timing. R8 is
a current limiter, such as it is:

: +1.5 +1.5 +1.5 +1.5 +1.5
: | | | | |
: | | | | |
: | \ | | |
: | / R7 | Q2 e>| |<e Q1
: | \ 410 | 2N5401 |---+---| 2N3906
: \ / | c/| | |\c
: / R8 | | | | |
: \ 12 | | | | |
: / | |/c Q4 | | |
: | +-------+---------| 2N3904 | | |
: | | | |>e | | |
: | | | | '-----+ |
: | | \ | | |
: | | / R3 | | |
: | \ \ 22k | | |
: --- / R6 / | | |
: \ / D2 \ 410 | | | |
: --- LED / | | |/c Q3 |
: | | +------------------------| |
: | | | | |>e |
: | | | | 2N3904 | |
: | C1 | \ | | |
: | || 150uF | / R4 | | |
: +------||-------+ \ 10k | | |
: | || | / | | |
: | | | | | |
: | | | | | |
: | | +-----------' | |
: | | | \ |
: | | | / R2 |
: | | \ \ 100 |
: | | / R5 / |
: | | \ 22k | |
: | | / | |
: | | | | |
: | | | | |
: | | gnd | |
: +--------------------------------------------------' |
: | | |
: | | |
: | ,--------+ |
: \ | | |
: / R1 | | 2N3904 |
: \ 10k | Q5 c\| |
: / | |--------------------------------------'
: | _|_ D1 e<|
: | /_\ BAT54 |
: | | |
: | | |
: | | |
: gnd gnd gnd

Lots of parts and it doesn't meet with Ed's comment.

Jon

I like the LM3909, and it would efficiently do the job required (maybe
not what ED was thinking) with few additional parts. 1 IC, 1 cap and a
couple of resistors.

 

[David Eather]

Jon Kirwan wrote:

On Sun, 26 Jul 2009 17:06:24 GMT, ehsjr <ehsjr@NOSPAMverizon.net
wrote:

fungus wrote:
On Jul 26, 8:57 am, ehsjr <eh...@NOSPAMverizon.net> wrote:

I don't _know_ if it qualifies as "a whole lot" better, but
available one chip solutions can meet the op's stated requirement
of keeping the current at 15-20 mA, and the joule thief cannot.
Can you maybe recommend one...?
Manufacturer chips posted below are just the first few found by a
Google search with "led boost drivers" in the search box.

National recommends their LM3410X for this.
http://www.national.com/ds/LM/LM3410.pdf

$2.50-$3 each. Lots around.

TI shows the TPS61160 meeting the requirements.
http://focus.ti.com/lit/ds/symlink/tps61161a.pdf

Hmm. Cheaper. $2 each. Lots around.

Onsemi has the CAT3606-D
http://www.onsemi.com/pub_link/Collateral/CAT3606-D.PDF

Couldn't find the -D around anywhere. But did find CAT3606HV4-T2 at
Digikey for $1 (and at only two other places.) This device cannot
handle more than 4.2V input and must have at least 3V. It's designed
for Li-ion sources and can run in either 1X or 1.5X mode. I'm not
hyped on this as a 'solution.' It's a charge pump with regulation on
the current, I think.

Linear's LT3598 will do it:
http://cds.linear.com/docs/Datasheet/3598fa.pdf

Mucho expensive. I found them for over $7 each! (Some at under $5,
too.) Only a few places carry them.

....

TI seems to be the one out of the above I'd focus more on. Looks nice
and seems to do the right job for a reasonable price and is at various
stores, as well.

Jon

I'm not recommending any one of those over any other,
and there are other chips from those manufactures and
others that may suit your needs.

Ed

Um, all those chips are in surface mount packages - the OP has bugger
all chance of being able to solder them.

 

[greg]

fungus wrote:

Looks like this: http://www.artlum.com/jt/b_to_e.gif

I'm confused. Does "b_to_e" mean base-to-emitter?
I want to see *collector*-to-emitter.

--
Greg

 

[greg]

fungus wrote:

Here's a nicer one one across base/emitter:

http://www.artlum.com/jt/b_to_e_2.gif

And I can get this at the transistor base (relative to ground):

http://www.artlum.com/jt/base_ring.gif

I'm even more confused now. Isn't the emitter connected
to ground? How can base-to-emitter be different from
base-to-ground?

--
Greg

 

[David Eather]

fungus wrote:

On Jul 26, 7:51 pm, Jon Kirwan <j...@infinitefactors.org> wrote:
National recommends their LM3410X for this.
http://www.national.com/ds/LM/LM3410.pdf
$2.50-$3 each. Lots around.

TI shows the TPS61160 meeting the requirements.
http://focus.ti.com/lit/ds/symlink/tps61161a.pdf
Hmm. Cheaper. $2 each. Lots around.


Maybe there's lots where you live but I don't know where
I'd get one around here, and there's none on eBay.

But ... while I was trawling eBay I found a lot of people selling
LM3914s. It's a dedicated LED driver chip, can work with 3V
input and drive up to 10 LEDs with programmable current
(up to 30mA each).

http://www.national.com/mpf/LM/LM3914.html


Even better: It has an input sensor making LED bar graphs.
I assume I could just pull the sensor high to turn on all the
LEDs but it opens up lots of fun possibilities - eg. I might be
able to make the LEDs respond to sound, which would be
very cool for processions.

Almost seems too good to be true ...

LM3914 would seem to work a treat - it needs about .9 of a volt more
than the LED uses (page 5 "led current regulation drop out")
Congratulations on a flexiable view of the problem.

 

[greg]

Jon Kirwan wrote:

But then I'd argue that
it isn't all that important because _with_ only that current, the OFF
transition is still in the area of the reverse transit time of the BJT
and that with a 2N2222, that's only about 100ns or so.

Does the datasheet qualify that figure with any
conditions?

Seems to me it will vary with the amount of base
current available to suck charge out. If that's
not very high, the transition time could be
longer than the published figure.

--
Greg

 

[greg]

Jon Kirwan wrote:

LTSpice shows me exactly what I
expect from theory, so far

Does the Spice model take the stored charge in the
transistor and its effect on the turn-off time into
account?

--
Greg

 

[ehsjr]

Jon Kirwan wrote:

On Sun, 26 Jul 2009 17:06:24 GMT, ehsjr <ehsjr@NOSPAMverizon.net
wrote:


fungus wrote:

On Jul 26, 8:57 am, ehsjr <eh...@NOSPAMverizon.net> wrote:


I don't _know_ if it qualifies as "a whole lot" better, but
available one chip solutions can meet the op's stated requirement
of keeping the current at 15-20 mA, and the joule thief cannot.

Can you maybe recommend one...?

Manufacturer chips posted below are just the first few found by a
Google search with "led boost drivers" in the search box.

National recommends their LM3410X for this.
http://www.national.com/ds/LM/LM3410.pdf


$2.50-$3 each. Lots around.


TI shows the TPS61160 meeting the requirements.
http://focus.ti.com/lit/ds/symlink/tps61161a.pdf


Hmm. Cheaper. $2 each. Lots around.


Onsemi has the CAT3606-D
http://www.onsemi.com/pub_link/Collateral/CAT3606-D.PDF


Couldn't find the -D around anywhere. But did find CAT3606HV4-T2 at
Digikey for $1 (and at only two other places.)

That is the part number given in the datasheet for the CAT3606-D.
http://www.onsemi.com/pub_link/Collateral/CAT3606-D.PDF

This device cannot
handle more than 4.2V

Wrong. Typical Vin is 3 to 5.5, max Vin is 7. See
the datasheet.

input and must have at least 3V. It's designed
for Li-ion sources and can run in either 1X or 1.5X mode. I'm not
hyped on this as a 'solution.' It's a charge pump with regulation on
the current, I think.

Yes - you program the current with one resistor.

Linear's LT3598 will do it:
http://cds.linear.com/docs/Datasheet/3598fa.pdf


Mucho expensive. I found them for over $7 each! (Some at under $5,
too.) Only a few places carry them.

....

TI seems to be the one out of the above I'd focus more on. Looks nice
and seems to do the right job for a reasonable price and is at various
stores, as well.

Jon


I'm not recommending any one of those over any other,
and there are other chips from those manufactures and
others that may suit your needs.

Ed

 

[greg]

fungus wrote:

The problem seems to me to be that the transistor receives the
exact same current as the LEDs. As LED current rises, so does
transistor current.

With the output capacitor, it actually carries *more*
current than the LEDs.

But that's not a problem in itself, as long as Vce is
low enough for a large enough proportion of the on
period. (Power = voltage * current, so high current
with negligible voltage doesn't give much power.)

--
Greg

 

[greg]

ehsjr wrote:

It works by transistor saturation, not core saturation.
A saturated inductor will take more current, as long as
the transistor is capable of pumping more current into
the inductor. As long as the current continues to flow
in the same direction, there is no polarity reversal
and no pulse created.

Yes, if you were just to let the inductor saturate in
the standard Joule Thief circuit, it would still turn
off by Ic reaching the beta limit. The only difference
would be a big current spike at the end of the on
period, which wouldn't do any good.

So some kind of modification would be needed. In the
pulse transformer circuit I posted earlier, there is
a small resistor in series with the emitter. There's
also a capacitor connected to the base.

The way I think it works is that when saturation
occurs, the voltage across the resistor begins to
rise sharply. The capacitor prevents the base voltage
from being able to keep up, so Vbe drops and the
transistor starts to turn off, with feedback from
the base winding finishing the process.

I expect something similar could be done with the
Joule Thief to make it work in inductor-saturation
mode.

--
Greg

 

[Jon Kirwan]

On Mon, 27 Jul 2009 16:35:41 +1200, greg <greg@cosc.canterbury.ac.nz>
wrote:

Jon Kirwan wrote:
LTSpice shows me exactly what I
expect from theory, so far

Does the Spice model take the stored charge in the
transistor and its effect on the turn-off time into
account?

TR and TF model the base's charge storage and I can definitely 'see'
the reverse transit time in the displayed waveforms, no problem. So
it is clearly in the model. It also models various depletion layer
capacitances, which are non-linear.

If you want, I can send some of the LTSpice files I'm using.

Jon

 

[Jon Kirwan]

On Mon, 27 Jul 2009 16:33:05 +1200, greg <greg@cosc.canterbury.ac.nz>
wrote:

Jon Kirwan wrote:
But then I'd argue that
it isn't all that important because _with_ only that current, the OFF
transition is still in the area of the reverse transit time of the BJT
and that with a 2N2222, that's only about 100ns or so.

Does the datasheet qualify that figure with any
conditions?

TR is actually just a modeling parameter that determines one of the
diffusion capacitances associated with the mobile carriers in the BJT
(I can give you the equation, if you like, or you can look it up on
your own -- it is called C_DC and the charge is Q_DC.) But there is a
capacitance that is modeled that depends on this parameter. It's not
really a datasheet specification, except that some datasheets (like
Fairchild's) actually do include the Spice model parameter.

Seems to me it will vary with the amount of base
current available to suck charge out.

Yup. I agree.

If that's not very high, the transition time could be
longer than the published figure.

Yeah. However, I think there is enough current to make short work of
the charge. I do like John's point about using a Thevenin equivalent
at the base, instead, to provide a 'shorter' pathway for the reversing
current. It makes sense to me, now that I've given it a little more
thought. I think it's better, but not by such a huge difference.

If I didn't have too many things competing for time right now, I
should go and take some time to play around and measure it in some of
the devices I have around here. I've got plenty PN2222A devices. That
would tell me what reality says about it.

Jon

 

[Jon Kirwan]

On Mon, 27 Jul 2009 04:37:42 GMT, ehsjr <ehsjr@NOSPAMverizon.net>
wrote:

Jon Kirwan wrote:
On Sun, 26 Jul 2009 17:06:24 GMT, ehsjr <ehsjr@NOSPAMverizon.net
wrote:

snip
This device cannot
handle more than 4.2V

Wrong. Typical Vin is 3 to 5.5, max Vin is 7. See
the datasheet.
snip

Sorry. You are right. I went completely past the first page where it
was obvious and looked at page 4, stupidly. My apologies.

Jon

 

[Jon Kirwan]

On Mon, 27 Jul 2009 14:19:43 +1000, David Eather <eather@tpg.com.au>
wrote:

Jon Kirwan wrote:
On Sun, 26 Jul 2009 17:06:24 GMT, ehsjr <ehsjr@NOSPAMverizon.net
wrote:

fungus wrote:
On Jul 26, 8:57 am, ehsjr <eh...@NOSPAMverizon.net> wrote:

I don't _know_ if it qualifies as "a whole lot" better, but
available one chip solutions can meet the op's stated requirement
of keeping the current at 15-20 mA, and the joule thief cannot.
Can you maybe recommend one...?
Manufacturer chips posted below are just the first few found by a
Google search with "led boost drivers" in the search box.

National recommends their LM3410X for this.
http://www.national.com/ds/LM/LM3410.pdf

$2.50-$3 each. Lots around.

TI shows the TPS61160 meeting the requirements.
http://focus.ti.com/lit/ds/symlink/tps61161a.pdf

Hmm. Cheaper. $2 each. Lots around.

Onsemi has the CAT3606-D
http://www.onsemi.com/pub_link/Collateral/CAT3606-D.PDF

Couldn't find the -D around anywhere. But did find CAT3606HV4-T2 at
Digikey for $1 (and at only two other places.) This device cannot
handle more than 4.2V input and must have at least 3V. It's designed
for Li-ion sources and can run in either 1X or 1.5X mode. I'm not
hyped on this as a 'solution.' It's a charge pump with regulation on
the current, I think.

Linear's LT3598 will do it:
http://cds.linear.com/docs/Datasheet/3598fa.pdf

Mucho expensive. I found them for over $7 each! (Some at under $5,
too.) Only a few places carry them.

....

TI seems to be the one out of the above I'd focus more on. Looks nice
and seems to do the right job for a reasonable price and is at various
stores, as well.

Jon

I'm not recommending any one of those over any other,
and there are other chips from those manufactures and
others that may suit your needs.

Ed

Um, all those chips are in surface mount packages - the OP has bugger
all chance of being able to solder them.

Well, there's that, too. Good point.

Jon

 

[David Eather]

fungus wrote:

On Jul 27, 6:31 am, David Eather <eat...@tpg.com.au> wrote:
http://www.national.com/mpf/LM/LM3914.html
LM3914 would seem to work a treat - it needs about .9 of a volt more
than the LED uses (page 5 "led current regulation drop out")


I hadn't understood that part (in fact I'm still not sure I do)
but does it mean I need an extra volt from somewhere?

Will I have to use a joule thief to drive the LM3914? :-(

No. You will need a supply greater than the LED voltage + .9v for near

perfect current regulation. When the .9 volt is not available the
current regulation becomes a bit sloppy and the current drops a bit. If
you wanted to suck the batteries *really* dry then you might need a way
to get a bit more voltage.

As an aside have you thought about using 6 volts (should be
easier/longer lasting to use with white LED) or a small 9v battery
(would last a bit more than an hour)?

 

[Jon Kirwan]

On Sun, 26 Jul 2009 23:33:48 -0700 (PDT), fungus
<openglMYSOCKS@artlum.com> wrote:

On Jul 27, 6:28 am, greg <g...@cosc.canterbury.ac.nz> wrote:

I'm confused. Does "b_to_e" mean base-to-emitter?
I want to see *collector*-to-emitter.

This is collector->emitter:

http://www.artlum.com/jt/coll_to_em.gif

Can you provide some voltages for the y-axis and times for the x-axis?

Looks to me like it is spending a lot of time with the BJT in the OFF
state, as the collector appears to be high for quite a while. I'm
assuming that the bottom span there is the BJT ON time, where it does
make sense that there is a 'rise' occuring in Vce as Ic grows linearly
upwards. Looks like about a 3:10 ratio, eyeballing it with a piece of
paper here (or close to 25% ON and 75% OFF.)

If I didn't screw this up, it should approximately be true (this
avoids the freewheeling diode and a tiny saturation voltage):

t_on/t_off = (Vout-Vin)/Vin

We know from your scope display that:

t_on/t_off = 0.3 (roughly)

So,

Vout = 1.3*Vin

That doesn't look right to me if I assume you have a stack of LEDs
there.

Can you see if that's about right?

Jon

 

[fungus]

On Jul 27, 6:29 am, greg <g...@cosc.canterbury.ac.nz> wrote:

fungus wrote:
Here's a nicer one one across base/emitter:

http://www.artlum.com/jt/b_to_e_2.gif

I'm even more confused now. Isn't the emitter connected
to ground? How can base-to-emitter be different from
base-to-ground?

Oops, should be called b_to_c....