Two transistor white LED driver circuit (con't)

[MrAl]

Two transistor white LED driver circuit (con\'t)

Hi again Jeff and others,

This is the circuit as it stands so far,
and also im trying out a little ascii
schematic drawing

Good luck with it,

--Al

 

[Mike]

Re: Two transistor white LED driver circuit (con\'t)

MrAl, what is the approximate efficiency of this circuit?

Thanks.

 

[Steelwolf]

Re: Two transistor white LED driver circuit (con\'t)

Oh boy! Another circuit to try out. Thanks MrAl.

How is the brightness of this circuit? Can it provide constant brightness over the life of the battery?

 

[Mercator]

Re: Two transistor white LED driver circuit (con\'t)

MrAl,

Thanks for the enhanced Brinkman circuit. I've been following this topic in the other thread.

I've built the original brinkman circuit and was wondering how to go about increasing the output. Your efforts along with jeff1500 are greatly appreciated. Now... if someone could somehow further enhance that simple circuit and power a white LS at a decent current.

BTW - I liked the way that KenBar built his little Satcure circuit and managed to build the basic brinkman with nothing but the componets closely connected in one little very compact cluster. One thing that greatly helped decrease the overall size was the fact that I found a handful of extreamy small 870ph caps. They are about 1/8 inch in dia.

Anyways... Thanks again for the research and the excellent tips on improving the circuit. I'm about ready to pull out my breadboard and give it a try for myself.

Mercator

 

[MrAl]

Re: Two transistor white LED driver circuit (con\'t)

Hello again to all of you interested
in LED drive circuits.

Mike:
I'm getting a theoretical 75% or so. I noticed that
a better transistor for the NPN produces a little
better efficiency, as does making the 10 ohm
series resistors 5 ohms if you have fairly
closely matched white LED's. I dont think
we will see 95% out of this circuit though.
I even went as far as to parallel a HEXFET
with Q2, getting its gate drive after the circuit
starts up with the normal NPN, but im not sure
if i can stabilize the gate drive enough to
insure reliable operation, although im sure
there is some way to do it
Perhaps the best transistor for Q2 is a Zetex
low sat type.

Steelwolf:
This particular circuit wasnt designed to provide
constant current/brightness over the battery life,
however im working on another circuit that will.
Just not sure if anyone wants to bother with
using 4 transistors instead of 2, even though
they are small and cheap.
There are of course ic chips out there already
that will do this, but thats if you want to get
some and try them out. There are also quite a
few points to note when choosing an ic for this
job. The efficiency claimed by the manufacturers
doesnt mean you can get that when driving a
white LED either. They normally spec that for
a dc to dc converter, which doesnt normally
have to use a dropping resistor on the output
like the voltage regulator types driving an LED.
There are a few current regulator chips out there
also. These would be the choice. If you can
set the output current with one or two resistors,
that would do the trick. You will have to watch
things like minimum input voltage and such though.
I might do a comparison on ic chips available soon,
but i have to wait for some feed back from some of
the ic manufacturers. I wanted to get what i think
is one of the best chips on the market for this kind
of application but the distributor normally sells in
3000 piece quantities! Im not ready to order that
many, yet

Mercator:
You should be able to increase output current throught the LED
by simply decreasing the value of R2.
Yeah, i like that little compact circuit too I have been
trying to keep parts count down in my stuff too so that
it will fit into a small package.


Well, good luck to all of you with your LED circuits,
and i hope to see you all posting results of experiments
and or circuit/flashlight creations on here.

--Al

 

[jeff1500]

Re: Two transistor white LED driver circuit (con\'t)

Multiple LED, Voltage Step-up Circuit Diagram

On my board right now, that's lit nice and bright and steady I have:

R1 = 10k ohms
R2 = 3k ohms
R3 = 10 ohms
C1 = 680 pico Farads rated for 1000 volts (I don't like this high voltage rating)
C2 = 47 micro Farads
Q1 = NTE12
Q2 = NTE11

Input = Two Eveready "C" Nimh cells, v=2.52 volts, i=116 mA. They discharge at fairly stable voltage.

Output = Two Nichia white leds in parallel, LED1: 3.46 volts, 24.8 mA, LED2: 3.41 volts, 27.1 mA

If I touch the end of the power wire to the battery it starts right up. If I touch the side of the wire to the battery, I have to wiggle it a little sometimes to get it started.

I've been testing, looking for sensitive component values.

I tried replacing C2 with a 4.7 micro F instead of a 47 micro F capacitor. I tried 5k and 20k as alternates for R1. I varied R2 from 1 to 10k, and I tried a second 680 pF capacitor in parallel with C1.

Here's what I think I saw:

No difference between 4.7 and 47 micro F for C2.

No difference when varying R1.

R2 is very clearly what sets the LED current.
I tried using a variable 5k trim pot, but it would drift to a higher resistance in a couple of minutes and dim the light quite a bit. Real resistors were very steady.

Using two parallel 680 pico F C1 capacitors made it start easier at lower led current levels around 18 mA each, but it reduced led current by about 3 mA.

I wish I understood C1 a little better, but it is starting up okay. Maybe someday I'll try the original Brinkman transistors and solder the connections instead of using skinny wire in a breadboard.

 

[MrAl]

Re: Two transistor white LED driver circuit (con\'t)

Hi again Jeff, that's very interesting too.

I was looking at the transistor again, and
i noticed again that it was a VERY high
rated device. The current rating has to be
a big overkill for this little circuit, so
now im thinking that the nonlinear
Beta characteristic of that transistor is
such that it has much lower gain for lower
collector currents, meaning it would have
a hard time starting up because it cant
provide enough current through the inductor
initially. A lower collector current, high
gain transistor would work better i think.
If you have any other transistors on hand,
try another one just to see what happens.
I dont have the NTE11 on hand myself and
cant find a cross ref so i cant
find a spice model to test it out in a
simulation.

Oh, BTW, C1 provides a sort of AC feedback
to turn Q1 off when the voltage on the
collector of Q2 rises. It has to be able
to provide energy for a long enough time
period (discharging through R1) to keep
Q1 off long enough for enough energy in
L1 to discharge through the Schottky and
LED(s).

--Al

 

[jeff1500]

Re: Two transistor white LED driver circuit (con\'t)

I've got an npn ECG 128P but I don't have the complementary pnp (ECG 129P).

The ECG 128P npn numbers are:
Vcbo 100v, Vceo 80v, Vebo 7v, IC 1A, PD 1W, FT 100mHz, HFE 100 min

I'll stick it in with the NTE 12 pnp this week to see what happens. The ebc wires are in a different order so I've got to take a minute and think about which one goes where.

 

[yclo]

Re: Two transistor white LED driver circuit (con\'t)

Does anyone know what the difference is between a HF and LF ferrite bead?
I built the satcure circuit and tried both (i bought both beads cause I didn't know which one it was suppose to use). One works and the other doesn't. And I have no idea which one is the working one since there are no apparent differences on the ferrite.

Ying Chee

 

[MrAl]

Re: Two transistor white LED driver circuit (con\'t)

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by jeff1500:
I've got an npn ECG 128P but I don't have the complementary pnp (ECG 129P).
<HR></BLOCKQUOTE>

Oh ok sounds good. I'd like to know what
happens with this transistor.

Ying: I would guess HF and LF stand for
'high frequency' and 'low frequency'.
I would also guess that the low frequency
(LF) ferrite worked in this type of circuit.
Perhaps the best idea is to buy one more
LF type and see if this new one works as well.
If it does, you know its the LF type for sure.
If it doesnt, you know its the HF type.

Good luck with it,
--Al

 

[jeff1500]

Re: Two transistor white LED driver circuit (con\'t)

I'm using an FP ferrite bead. Hole=.184", length=1/8", OD=3/8". I don't know what kind of ferrite it is. I got it from Hosfelt Electronics for 15 cents.

I have more test results for the two transistor circuit.

I swapped the NTE11 for an ECG128P. It started well but the output led current dropped about in half. Then I switched to rechargable alkalines instead of Nimh. I got a little more output current but it started harder.

I switched back to NTE11 / NTE12 and made a new coil with about 40 turns instead of 20. I got about 15 more mA through the leds so I increased R2 to 5k to get 27 mA in each led.

Next I varied C2. With 47 micro F it almost doesn't run. I can jump start it but then it fades and goes out.

With C2=4.7 micro F it runs fine after a jump start by pulsing R2 to 3k.

Without C2 it starts perfectly.

Results:

NTE11/12
R2=5k
C2=4.7 micro F
LED current = 27 mA each (3.5 v)
Battery input = 3 volts, 122 mA with C2=4.7, 116 mA with no C2.
Light quality = high

There's some kind of pattern emerging here. The 47 micro F capacitor works fine with the one transistor circuit, then for the two transistor circuit, as the number of coil turns increases, there's a preference for a lower value C2.

A larger coil might mean lower frequency so maybe lower frequencies like smaller C2 values. There's a little more light, but not much more, with the 4.7 micro F C2 compared to no C2.

I guess I gotta go find a smaller C2 capacitor. What is another size C1 to try? Is there a size relationship between C1 and C2?

 

[MrAl]

Re: Two transistor white LED driver circuit (con\'t)

There isnt really any size relationship
between C1 and C2, except when ripple
is a concern.

For the purpose of understanding C1 a little better, Q2 has two
distinct states:
1. Q2 turned on
2. Q2 turned off

When Q2 is on, C1 charges very quickly though the base emitter of Q1
and through Q2 collector emitter.
When Q2 is off, C1 discharges through the 10k resistor.

At very low input voltages like 1.2 volts, R1=5k does appear to
work better then 10k.
A value of 500pf appears to work with inductor values from
200uH to 3200uH, but i would try 1000pf with larger inductances
also. Stick to 500pf or so for lower inductances.
Apparently, the lower the value of C1 the lower the ripple output.

You think seeing waveforms would help?

--Al

 

[jeff1500]

Re: Two transistor white LED driver circuit (con\'t)

I made some new jumper wires for the breadboard. Same components.

Instant start without C2.

I think it's starting better now with C2. I don't seem to have to jump R2 today. Just dragging the wire on the battery terminal a little is enough to start it up.

Wave forms would be interesting. What effect do different values of C1 & C2 have on the wave?

 

[MrAl]

Re: Two transistor white LED driver circuit (con\'t)

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by jeff1500:
I made some new jumper wires for the breadboard. Same components.

Instant start without C2.

I think it's starting better now with C2. I don't seem to have to jump R2 today. Just dragging the wire on the battery terminal a little is enough to start it up.

Wave forms would be interesting. What effect do different values of C1 & C2 have on the wave?<HR></BLOCKQUOTE>

Well increasing C1 slows the frequency,
while increasing C2 decreases the ripple
on the output dc voltage. Since we
arent really building a power supply
that needs good clean dc, we can use small
values of C2. We just dont want 80ma
current pulses through the LED.

--Al

 

[PeLu]

Re: Two transistor white LED driver circuit (con\'t)

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by yclo:
Does anyone know what the difference is between a HF and LF ferrite bead?
<HR></BLOCKQUOTE>

LF ferrites have too high losses at higher frequencies (eddy currents, but other advantages).

 

[yclo]

Re: Two transistor white LED driver circuit (con\'t)

Thanks a lot Peter.

So will the single transistor and double transistor circuit work for a single AAA? What are the exact components I should use in these two circuits to have the led running at approx 40ma (or more?)?

Ying Chee

 

[jeff1500]

Re: Two transistor white LED driver circuit (con\'t)

I got this to work for one transistor and one battery: http://elm-chan.org/reports/led1/report_e.html

It was a little dim but still okay. Probably would work for 2 batteries and multiple leds by adjusting the base resistor and using R3 like below on the two transistor circuit. The double coil needs to run in opposite directions. I got it wrong at first. The opposing fields are what makes it oscillate.

And this is the latest configuration I've got set up and working/testing for two transistors.
http://edusite10.tripod.com/led2/brink/multi_led.htm

I just need to try (like Al suggested) a smaller C2 maybe of about 1 micro F for two leds.

 

[MrAl]

Re: Two transistor white LED driver circuit (con\'t)

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by jeff1500:
I got this to work for one transistor and one battery: http://elm-chan.org/reports/led1/report_e.html

The double coil needs to run in opposite directions. I got it wrong at first. The opposing fields are what makes it oscillate.
<HR></BLOCKQUOTE>

Hi again Jeff,

I meant to mention something about the
single transistor coil that may be of use.
Since the coil used in that design is really
a center tapped transformer/inductor, it
can be wound with that in mind and might
make it a little easier to get the leads
correct on the first try.

The primary can be wound on the core first
as the first layer, then create a center tap
by extending the end of the primary wire and
soldering a second length of wire to that
end. Then simply wind that length of wire
as the second layer (secondary) in the same
direction as the primary was wound.
When this
is complete, there will be only two free ends
and one center tap with two leads soldered
together. Simply run the center tap to the
plus (+) supply and run either of the two
remaining free ends to the other parts of
the circuit, and it doesnt matter which one
of the two ends runs to the collector or
base resistor.
This might make that circuit a little easier
to make.

I'll probably be using that circuit from
time to time for some voltage step up
circuits. I was thinking of possibly using
that circuit as a pre-regulator for a nice
LED driver which uses very common parts
but works very very well for regulating
LED brightness over the life of the
battery. This would work out even better
then using 4 transistors, as that circuit
gets a little fussy about parts values.
With a little pre-kick in voltage, there
are a wide variety of circuits that will
work well and reliably. Since it doesnt
have to drive the LED directly, the
efficiency in that part of the circuit wont
matter as much either.
The drawback is that any design
with this circuit as pre-regulator will
need an extra coil wound and an extra
transistor, but of course almost any
small power transistor would work.

I think ill look into this more.

What do you think about using this circuit
as a pre-regulator to solve the problem of
low voltage inputs with only one cell?

--Al

 

[jeff1500]

Re: Two transistor white LED driver circuit (con\'t)

My test board seems to work a little better each day. I've been polishing the connections and am using C2=4.7 micro F. It's starting well now with R2=5k. There's always a little magic involved.

Sounds like time to move to voltage regulation. I'd like to have a circuit that started with a range of voltage between about 1 and 3 volts and provided a stable 3.6 volt or so output.

Maybe something like this:
http://www.mitedu.freeserve.co.uk/Circuits/Power/lm317.htm

With, like you say, this kind of step-up pre-regulator. Can you propose a circuit sketch to think about?

I'm thinking that LM317T chips are big enough to solder connections to. I've got a couple of surface mount chips someplace but I can't imagine how I could ever solder anything to something that small.

 

[MrAl]

Re: Two transistor white LED driver circuit (con\'t)

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by jeff1500:
My test board seems to work a little better each day. I've been polishing the connections and am using C2=4.7 micro F. It's starting well now with R2=5k. There's always a little magic involved.

Sounds like time to move to voltage regulation. I'd like to have a circuit that started with a range of voltage between about 1 and 3 volts and provided a stable 3.6 volt or so output.

Maybe something like this:
http://www.mitedu.freeserve.co.uk/Circuits/Power/lm317.htm

With, like you say, this kind of step-up pre-regulator. Can you propose a circuit sketch to think about?

I'm thinking that LM317T chips are big enough to solder connections to. I've got a couple of surface mount chips someplace but I can't imagine how I could ever solder anything to something that small.<HR></BLOCKQUOTE>

Hi again Jeff,

I have been looking at some regulating
circuits recently, and i have found
that the only way to regulate over that
wide a range of voltage input is with
feedback. The problem with feedback is
that to do it right, you need a voltage
reference and at least one comparator.
This complicates the circuit. The original
four transistor circuit i designed works
fine with inputs of 2 to 3 (maybe higher)
volts input, but as soon as you go down
to 1 volt, problems start to show up.
This means im not sure if i can get it to
work down to 1 volt like i originally thought.
Now im leaning toward using a pre-regulator.
The kind of pre-regulator cant be a linear
type though like the LM317. It has to be
a boost switching regulator. While the 317
can be used as a switching reg also, the
parts count gets kind of high for this.
Since it only has to supply a little
current, i dont think efficiency will be
a problem, so i was thinking of using the
one transistor boost circuit we have been
talking about for the pre-regulator, and
either combine that with some easy to get
switching reg ic's or built up a switching
reg using a common ic like the LM339 and
a transistor or two. This would require
two coils.
The alternative is to get ahold of some of
the Zetex chips we have been discussing in
the other threads and see if they really
work down to 0.7 volts like they say.

Since i have found other chips that claim
the same fame and yet dont really put out,
i tend to be wary of these claims, but i
wont pass judgement untill i have tryed
one for myself
Since they are using a bipolar transistor,
i tend to believe the spec's though.

I'm not sure if i want to put out $25 for
parts just to be able to test one of
the Zetex chips either though. I'll have
to think about this or try to get a sample.

Untill then, it looks like either non
current regulation or pre-regulator.

The pre-regulator could possibly be wired in
parallel to the 'normal' boost regulator
also, so that during very low voltage inputs
the pre-regulator supplies some current
untill the normal regulator starts up.
Once the normal regulator starts up, it
takes over the entire process and shuts down
the pre-regulator. I think a number of
higher input voltage chips would work with
this method. Using a second schottky
in series with the output of the pre-reg
would accomplish this. The two would then
be paralleled at the output capacitor, so
that either converter can supply current to
the output. Should be simple to implement.
Only one output cap is needed also.
Chips that require bootstrapping should work
like this also. I dont know why i didnt
think of this before, but it will of course
require an extra light duty converter
built from one transistor, one coil, one
resistor, and one schottky. Actually,
the second schottky can be a normal diode,
since the pre-regulator circuit doesnt
supply much current once the normal
regulator starts up.
This would be a better all around solution
then the 4 transistor circuit.
I think ill draw this up and post it.

--Al