160W to 100W convertor

[Harshana]

Hi all,

I have a beam light of 160W running on 12V DC power. I want this to be
reduced to around 100W (reduce light). I made a switching cct with a
555, oscilating around 600~700Hz, and choping the supply with a C3055
transistor (just on and off driven by 555 followed by a D400).It
worked for about 5 minutes and went off. I found that 3055 has died
due to too much heat. Another solution is to have a 6V 90W bulb in
series with the beam. Can anyone suggest me a simple cct to do this.
Note that I can't have a huge heat sink.

Harshana

 

[Winfield Hill]

Harshana wrote...

I have a beam light of 160W running on 12V DC power. I want this to be
reduced to around 100W (reduce light). I made a switching cct with a
555, oscilating around 600~700Hz, and choping the supply with a C3055
transistor (just on and off driven by 555 followed by a D400). It
worked for about 5 minutes and went off. I found that 3055 has died
due to too much heat. Another solution is to have a 6V 90W bulb in
series with the beam. Can anyone suggest me a simple cct to do this.
Note that I can't have a huge heat sink.

Hi Harshana, it's nice to see you are experimenting with high-power
switching circuits. The 160W lamp draws about 13A from 12V when warm,
and much more when cold. There was nothing wrong with your idea to
reduce its light output with pulse-width modulation (PWM), but your
selection of the famously-wimpy '3055 transistor type was a problem.
That transistor used as a switch has a nasty rising voltage drop for
currents above 5 to 10A. This situation is exacerbated if you don't
use hefty 10% drive base currents, e.g., 1A base for 10A collector.
The 555, as we know, can only deliver 150 to 200mA, on a good day.

These days most folks would use a little power MOSFET for such a job,
and many good low-cost choices are available. The issue would become
selecting one you can easily get your hands on.

Where do you live? Where do you usually get your electronic parts?

 

[CFoley1064]

Subject: 160W to 100W convertor

From: [email protected] (Harshana)
Date: 9/20/2004 7:25 AM Central Daylight Time
Message-id: <[email protected]>

Hi all,

I have a beam light of 160W running on 12V DC power. I want this to be
reduced to around 100W (reduce light). I made a switching cct with a
555, oscilating around 600~700Hz, and choping the supply with a C3055
transistor (just on and off driven by 555 followed by a D400).It
worked for about 5 minutes and went off. I found that 3055 has died
due to too much heat. Another solution is to have a 6V 90W bulb in
series with the beam. Can anyone suggest me a simple cct to do this.
Note that I can't have a huge heat sink.

Harshana

Good morning, Harshana. All 555 questions are customarily posted on s.e.b.

You had a good idea in modulating the transistor to reduce power. The problem
is that transistors are current-driven devices, and at 13 amps the 2N3055 will
typically only have a current gain of 20 to 50. That means your poor 555 would
have to supply anywhere between 0.26A and 0.65A to do the job. In addition, if
you're using a transistor as a switch, you'll want to at least triple the
nominal base current to get hard saturation, which isn't even near what a 555
can do. Your poor 2N3055 probably had half the voltage across it when it was
on, leading to overheating and early expiration.

If you drive a darlington transistor with your 555, you'll have a current gain
of around 500, which puts you in the range of what a 555 can do. Note that,
since a darlington has 1V from collector to emitter when saturated, you'll have
to provide a heat sink that can dissipate 13 watts max. But your easiest
solution is to replace the transistor, and this will work.

A TO-3 darlington which might work for you is the 2N6284. It has a guaranteed
minimum current gain of 750 at 10A, and can stand off 100V. Try something like
this (view in fixed font or M$ Notepad):


VCC
+
|
12V |
+ .-.
| ( X )
| '-'
.----o----. |
| | |
| | 330 |
| |3 ___ |/
| Outo---|___|-o---|2N6284
| | | |>
| 555 | .-. |
| | 2.2K| | |
| | | | |
| | '-' |
| | | |
| | === ===
'----o----' GND GND
|
|
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Good Luck
Chris

 

[Tim Wescott]

Hi all,

I have a beam light of 160W running on 12V DC power. I want this to be
reduced to around 100W (reduce light). I made a switching cct with a
555, oscilating around 600~700Hz, and choping the supply with a C3055
transistor (just on and off driven by 555 followed by a D400).It
worked for about 5 minutes and went off. I found that 3055 has died
due to too much heat. Another solution is to have a 6V 90W bulb in
series with the beam. Can anyone suggest me a simple cct to do this.
Note that I can't have a huge heat sink.

Harshana

Do you have 160 watts in your light beam, or do you have a 160 watt
bulb? Is it an incandescent bulb or LED? If it's an incandescent bulb
when you reduce the average voltage the filament temperature will drop
which will change the color of the light emitted, reduce the bulb
efficiency, and lower the bulb's resistance, thereby increasing the
current that must be carried by your switch when it's on.

Your switching circuit idea sounds OK on the surface. I suspect that
the 3055 overheated either because it wasn't turning on fully, it wasn't
turning on (or off) fast enough, or just because you were exceeding it's
average dissipation.

If you want to stick with the switched bulb idea I would suggest using a
good low on-resistance MOSFET as the bulb driver. If you need "simple"
I'd make sure that I had a good strong transistor driving the MOSFET and
that I had a fairly low resistance gate pull-down resistor. Make sure
you're turning the thing on and off fast enough (this should _not_ be
rocket science at 700Hz, but shouldn't be ignored with such a simple
circuit). Make sure that you know the instantaneous current so you can
check that your I^2R losses in the driver aren't too severe.

Or you can just go with your 90W bulb .

 

[Tony Williams]

I have a beam light of 160W running on 12V DC power. I want this
to be reduced to around 100W (reduce light).

Incandescent bulbs are very non-linear devices.
Here are some useful sums for reference.

1.43 3.3
Pv ( Vv ) lmv ( Vv )
---- = (----) and ----- =(----)
Pnom (Vnom) lmnom (Vnom)

Pnom, Vnom, lmnom, are the nominal values for
Power, Voltage, and luminous flux (lumens).

Pv, Vv, and lmv, are what happens at other voltages.

You can see that to obtain a ratio of 100W/160W
requires a Vv of about 0.72x12V, or about 8.64V.

However, a 12V bulb running at 8.64V will have a
luminous flux that is about 34% of nominal (and will
have a pronounced shift towards red).

100W and 8.64V is a current of about 11.6A, which is
not that far below the normal 13.3A at 12V.

I made a switching cct with a 555, oscilating around
600~700Hz, and choping the supply with a C3055 transistor

[snip]

Keep the 555, but perhaps use a MOSFET instead, noting
that it will require to efficiently sink that 11.6A.
Note also that (say) a 0.1ohm MOSFET will have an On
voltage drop of about 1.16V and this must be accounted
for when setting the 555 duty cycle for an average
Vv across the bulb of 8.6V.

 

[Tam/WB2TT]

Good morning, Harshana. All 555 questions are customarily posted on
s.e.b.

You had a good idea in modulating the transistor to reduce power. The
problem
is that transistors are current-driven devices, and at 13 amps the 2N3055
will
typically only have a current gain of 20 to 50.

It gets worse than that. Minimum beta at 10A is only 5. Curve shows typical
at 10A of around 12. 10A is as high as the curves go. If you have your heart
set on the 555, use a low Rds FET. If not, how about an adjustable 50W power
resistor.

Tam

 

[Tony Williams]

Keep the 555, but perhaps use a MOSFET instead, noting
that it will require to efficiently sink that 11.6A.

Hmm... 11.6A is the average filament current at 8.64Vdc.
The MOSFET will actually be sinking 16A whilst ON.

 

[Harshana]

Hi all,

I have a beam light of 160W running on 12V DC power. I want this to be
reduced to around 100W (reduce light). I made a switching cct with a
555, oscilating around 600~700Hz, and choping the supply with a C3055
transistor (just on and off driven by 555 followed by a D400).It
worked for about 5 minutes and went off. I found that 3055 has died
due to too much heat. Another solution is to have a 6V 90W bulb in
series with the beam. Can anyone suggest me a simple cct to do this.
Note that I can't have a huge heat sink.

Harshana

Hi all,

Thanks for the replies. Sorry for not being clear earlier. I'm using a
555 which drvies D400, which drives the C3055. D400 and C3055 are in
darlington arrangement.

Actually, I'm from Sri Lanka. I buy my parts from local component
shops. They usually have very good collection. If the component U R
looking for is in SONY, LG or any other common TV model, then U R sure
u can buy it in SL

So please tell me good FET, BJT that I can use for this (Preferably,
from a popular TV set).

Thanks,
Harshana

 

[Winfield Hill]

Harshana wrote...

Harshana wrote ...

I'm using a 555 which drvies D400, which drives the C3055. D400 and
C3055 are in darlington arrangement.

With enough transistors in parallel you could likely use a '3055
to PWM dim your light. This is the configuration you should use:

.. ,---- +12
.. |
.. |/ '3055 (4 places)
.. 555 ----| ,======+======+===O LAMP O=== +12
.. (running |\V | | |
.. from 12V) | |/ | |
.. +---/\/\---| | |
.. | |\V |/ |
.. +---/\/\---- | ---| |
.. | | |\V |/
.. '---/\/\---- | ---- | ---|
.. 27 3W | | |\V
.. (3 places) | | |
.. '======+======+===== 12V return

At full load, as the die temperature goes up, the Vce(sat) voltage
goes up slightly (see AoE page 1062), helping to equalize current
sharing between the three transistors.

 

[Tony]

Harshana wrote...

With enough transistors in parallel you could likely use a '3055
to PWM dim your light. This is the configuration you should use:

. ,---- +12
. |
. |/ '3055 (4 places)
. 555 ----| ,======+======+===O LAMP O=== +12
. (running |\V | | |
. from 12V) | |/ | |
. +---/\/\---| | |
. | |\V |/ |
. +---/\/\---- | ---| |
. | | |\V |/
. '---/\/\---- | ---- | ---|
. 27 3W | | |\V
. (3 places) | | |
. '======+======+===== 12V return

At full load, as the die temperature goes up, the Vce(sat) voltage
goes up slightly (see AoE page 1062), helping to equalize current
sharing between the three transistors.

Just one caution - the relatively low PWM frequency mentioned anywhere
will minimize switching losses, but watch for filament resonance,
which can see off an expensive lamp quite quickly due to the
considerable oscillatory forces from 12A of current modulation. In the
extreme case you could lower the frequency to as low as 100-120Hz (the
same as it sees when running on AC).

Tony (remove the "_" to reply by email)

 

[Jeff]

Just one caution - the relatively low PWM frequency mentioned anywhere
will minimize switching losses, but watch for filament resonance,
which can see off an expensive lamp quite quickly due to the
considerable oscillatory forces from 12A of current modulation. In the
extreme case you could lower the frequency to as low as 100-120Hz (the
same as it sees when running on AC).

But the PWM is a square wave which is harsher then a sine wave.

 

[Tony Williams]

. ,---- +12
. |
. |/ '3055 (4 places)
. 555 ----| ,======+======+===O LAMP O=== +12
. (running |\V | | |
. from 12V) | |/ | |
. +---/\/\---| | |
. | |\V |/ |
. +---/\/\---- | ---| |
. | | |\V |/
. '---/\/\---- | ---- | ---|
. 27 3W | | |\V
. (3 places) | | |
. '======+======+===== 12V return

Incandescent lamps are awkward little beggars aren't they?

Note, the OP's alternative solution of 6V/90W in series is
pretty well spot on.

 

[Harshana]

Just one caution - the relatively low PWM frequency mentioned anywhere
will minimize switching losses, but watch for filament resonance,
which can see off an expensive lamp quite quickly due to the
considerable oscillatory forces from 12A of current modulation. In the
extreme case you could lower the frequency to as low as 100-120Hz (the
same as it sees when running on AC).

Tony (remove the "_" to reply by email)

watch for filament resonance,
which can see off an expensive lamp quite quickly due to the
considerable oscillatory forces from 12A of current modulation.

This is one concern I have, @ 600~700Hz, I can here the filament
singing (slight mmmmmm). Will it reduce the lifetime of the beam
considerably? How do I know whether its resonating or not ???


Harshana

 

[Tony]

But the PWM is a square wave which is harsher then a sine wave.

Certainly; but no worse than an AC dimmer output. And if you don't
have a big enough supply of lamps to empirically find out the total
wide frequency range of their filament resonances (which I couldn't
detect, and could only guess were a problem when another lamp blew),
the harmonics in a 120Hz square wave are at least at a much lower
level than the full amplitude modulation that you get if the
fundamental hits a resonance mode. Anyway, that was my rationale when
I ran out of lamps, copped out and knocked the frequency down (with
successful results).

Tony (remove the "_" to reply by email)

 

[Tony]

This is one concern I have, @ 600~700Hz, I can here the filament
singing (slight mmmmmm). Will it reduce the lifetime of the beam
considerably? How do I know whether its resonating or not ???

It was about 35 years ago I went through this exercise, but yes, a
couple of lamps lasted only 10 minutes or so (ASAIR a PAR36 lamp -
YMMV). I guess if your ears are good enough you could put a pot into
the 555 and wind it through the low frequency range to find the middle
of a broad frequency null where it made the least noise. I didn't
imagine the resonance Q could be very high with a hot filament, but
conversely it also wouldn't take much to rupture one.

Tony (remove the "_" to reply by email)