I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith

"homebuilt aircraft"... any radio interference concerns from the
switching currents?

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith


Make one for a bike that runs on 2.4 Volts (two rechargeables in series)

Instead of a dimmer function, make it a flasher of various patterns.

Charlie Smith wrote:
I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith



Since it is for a flying machine, how about a little redundancy?




--

Richard Lamb
http://www.home.earthlink.net/~cavelamb/


"The clock of life is wound but once, and no man has the power
to tell just when the hands will stop, at late or early hour...
Now is the only time you own. Live, love, toil with a will.
Place no faith in time. For the clock may soon be still."

Charlie Smith wrote:
I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith

You'll generate a lot of heat in the L78S05 regulator - worst
case, about 4.2 watts. You could use a L78S09 instead, and reduce
the worst case heating in the regulator to about 2.4 watts. You'd
need to change the 68 ohm resistors to 270 ohms.

Ed

"Charlie Smith" wrote in message
...
I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary.

Well, it looks like it'll work fine. If you care, you can gain some
efficiency by just using a IC meant for the task (regulating LED current)
directly -- these are really just a switching power supply controller IC
that's been tweaked to regulate current rather than voltage; you can do such
tweaking yourself if it's a cost sensitive design (using, e.g., an LT1070 or
one of its many variants). If you look at, e.g., Linear Tech's web site,
you'll find plenty of example circuits and application notes on this sort of
thing -- LED controllers are a Big Deal these days since of course pretty much
every cell phone, laptop, and TV needs one.

---Joel

"cavelamb" wrote in message
...
Since it is for a flying machine, how about a little redundancy?

I'd be tempted to toss in a switch that connects +14V-Big resistor-All the
LEDs (with their individual current-limiting resistors) -- so if the 7805 or
555 fail, you can still have full brightness backlights.

But perhaps just having a flashlight around is just as good, if there's a
mount such that it can be pointed at the display. :-)

Will it work as drawn? Yes. Is it the most efficient design for the
job? No. Is it pretty simple and easy to build? Yes. However,
there are several single-point failure points, but if this aircraft
isn't designed for hard-IFR or if you choose to carry a backup
flashlight you should be OK.

Jim




On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith

Safety- it definitely needs fusing and, as Jim suggests, EMI from the
switching could conceivably screw up your radio communications or
other instruments. Maybe not enough DC current to worry about, but you
also should be aware that heavy DC currents can affect compasses, so
wiring layout may need a bit of care.

You could also consider a premade solution like this one:

http://www.digikey.com/scripts/DkSea...me=788-1023-ND

The 350mA constant-current dimmable version is less than $20 at
Digikey. There still be some EMI, so it has to be checked that it
doesn't cause interference (eg. with weak radio signals). You could
probably reduce the 68 ohm resistors to 33R without any problems.

You also need some kind of reliable backup for when this thing fails
and it's dark. Consider an on/off/bypass switch that would replace the
electronics with a resistor.

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith

It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John

On Tue, 12 Jan 2010 18:46:13 -0800, the renowned John Larkin
wrote:

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith

It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John

Good point, also I don't think running 20mA through 365-1467-ND 3mm
white LEDs is a very good idea. That's the abs. max. current at 25C.
If you want it not to fail at high panel temperatures/brightness (ever
come in out of the sun into a dark hangar?) and don't want to be one
of those folks who complains about crappy LEDs (because they abuse
them and they die fast) you should derate to something like 10mA. Two
in series at 10mA will give you about the same amount of light, but
draw 1/2 the current (at twice the voltage), but you need twice as
many LEDs.

So, say the voltage regulator goes from 5V to 12V, the resistors in
series with each series LED pair will be 560 ohms. A TO-220 LM317
should not need hardly any heatsink (and there will be no RFI
generated). Say 180R from LM317 out to sense input, and a 1K pot in
series with 510R to ground from the sense input, to give about a 5V to
12V adjustment range. Maybe a mechanical switch to give off/bypass
(full bright)/dimmed.

Disadvantage of this over the PWM is that the apparent brightness will
not vary nearly as linearly (most of the action will be down near
where it just comes on), so check it out on the bench to make sure you
can live with it if you decide to go this way.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

Spehro Pefhany wrote:
On Tue, 12 Jan 2010 18:46:13 -0800, the renowned John Larkin
wrote:

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering if
those qualified in this group would mind reviewing the schematic and
making constructive commentary. I'd like this thing to be reliable and
not cause a host of other electronic gremlins. Just keep in mind, I'm a
chemist, not an electronics professional. But, I do learn quickly.


Thanks.
Charlie Smith
It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John

Good point, also I don't think running 20mA through 365-1467-ND 3mm
white LEDs is a very good idea. That's the abs. max. current at 25C.
If you want it not to fail at high panel temperatures/brightness (ever
come in out of the sun into a dark hangar?) and don't want to be one
of those folks who complains about crappy LEDs (because they abuse
them and they die fast) you should derate to something like 10mA. Two
in series at 10mA will give you about the same amount of light, but
draw 1/2 the current (at twice the voltage), but you need twice as
many LEDs.

So, say the voltage regulator goes from 5V to 12V, the resistors in
series with each series LED pair will be 560 ohms. A TO-220 LM317
should not need hardly any heatsink (and there will be no RFI
generated). Say 180R from LM317 out to sense input, and a 1K pot in
series with 510R to ground from the sense input, to give about a 5V to
12V adjustment range. Maybe a mechanical switch to give off/bypass
(full bright)/dimmed.

Disadvantage of this over the PWM is that the apparent brightness will
not vary nearly as linearly (most of the action will be down near
where it just comes on), so check it out on the bench to make sure you
can live with it if you decide to go this way.


Best regards,
Spehro Pefhany


An audio taper pot should help with the brightness question?


--

Richard Lamb
http://www.home.earthlink.net/~cavelamb/

cavelamb wrote:

An audio taper pot should help with the brightness question?


They really aren't intended for DC current. They will become noisy
in a hurry, with multiple dead spots in the carbon track. Delco found
out the hard way when they built some early transistor car radios with
DC on the volume control, then had to do a recall to repair them, and
modify them to prevent future failures. A wirewound Rheostat would be a
better choice, but wouldn't have a linear change in intensity.


--
Greed is the root of all eBay.

Sure they are, you just can't run a LOT of DC current through them,
especially on the ends of the tracks.

Jim



On Wed, 13 Jan 2010 09:07:25 -0500, "Michael A. Terrell"
wrote:


cavelamb wrote:

An audio taper pot should help with the brightness question?


They really aren't intended for DC current.

RST Engineering wrote:

Sure they are, you just can't run a LOT of DC current through them,
especially on the ends of the tracks.

Jim

On Wed, 13 Jan 2010 09:07:25 -0500, "Michael A. Terrell"
wrote:


cavelamb wrote:

An audio taper pot should help with the brightness question?


They really aren't intended for DC current.


Then explain why Delco had to replace several million audio taper
pots in their AM car radios? They had a few milliamps flowing, yet over
70% failed within two years. I got to where I could replace them
without looking at the chassis more than a few glances. Prewire the
replacement, bolt it in and move a few wires.


--
Greed is the root of all eBay.

In message , Michael A.
Terrell writes

RST Engineering wrote:

On Wed, 13 Jan 2010 09:07:25 -0500, "Michael A. Terrell"
wrote:


cavelamb wrote:

An audio taper pot should help with the brightness question?


They really aren't intended for DC current.


Sure they are, you just can't run a LOT of DC current through them,
especially on the ends of the tracks.

Then explain why Delco had to replace several million audio taper
pots in their AM car radios? They had a few milliamps flowing, yet over
70% failed within two years. I got to where I could replace them
without looking at the chassis more than a few glances. Prewire the
replacement, bolt it in and move a few wires.

Mechanically, how do pots with a linear tracks differ from those with
non-linear tracks (in a way which cause them to fail when they pass DC).
--
Ian

"RST Engineering" wrote in message
...
Sure they are, you just can't run a LOT of DC current through them,
especially on the ends of the tracks.

Hell, I've got ordinary 100k audio taper Alphas (the ~1" chassis mount
solder lugs style) that go scratchy just from the offset my computer's sound
card produces. Microamperes at best.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

Ian Jackson wrote:

In message , Michael A.
Terrell writes

RST Engineering wrote:

On Wed, 13 Jan 2010 09:07:25 -0500, "Michael A. Terrell"
wrote:


cavelamb wrote:

An audio taper pot should help with the brightness question?


They really aren't intended for DC current.


Sure they are, you just can't run a LOT of DC current through them,
especially on the ends of the tracks.

Then explain why Delco had to replace several million audio taper
pots in their AM car radios? They had a few milliamps flowing, yet over
70% failed within two years. I got to where I could replace them
without looking at the chassis more than a few glances. Prewire the
replacement, bolt it in and move a few wires.

Mechanically, how do pots with a linear tracks differ from those with
non-linear tracks (in a way which cause them to fail when they pass DC).


Carbon pots aren't build for rheostat service. That's what they make
wirewound for.

--
Greed is the root of all eBay.

In message , Michael A.
Terrell writes

Ian Jackson wrote:

In message , Michael A.
Terrell writes

RST Engineering wrote:

On Wed, 13 Jan 2010 09:07:25 -0500, "Michael A. Terrell"
wrote:


cavelamb wrote:

An audio taper pot should help with the brightness question?


They really aren't intended for DC current.


Sure they are, you just can't run a LOT of DC current through them,
especially on the ends of the tracks.

Then explain why Delco had to replace several million audio taper
pots in their AM car radios? They had a few milliamps flowing, yet over
70% failed within two years. I got to where I could replace them
without looking at the chassis more than a few glances. Prewire the
replacement, bolt it in and move a few wires.

Mechanically, how do pots with a linear tracks differ from those with
non-linear tracks (in a way which cause them to fail when they pass DC).


Carbon pots aren't build for rheostat service. That's what they make
wirewound for.

Maybe, but there are many carbon pots which do carry DC (albeit usually
not a lot).
--
Ian

On Wed, 13 Jan 2010 21:52:39 +0000, Ian Jackson
wrote:

In message , Michael A.
Terrell writes

Ian Jackson wrote:

In message , Michael A.
Terrell writes

RST Engineering wrote:

On Wed, 13 Jan 2010 09:07:25 -0500, "Michael A. Terrell"
wrote:


cavelamb wrote:

An audio taper pot should help with the brightness question?


They really aren't intended for DC current.


Sure they are, you just can't run a LOT of DC current through them,
especially on the ends of the tracks.

Then explain why Delco had to replace several million audio taper
pots in their AM car radios? They had a few milliamps flowing, yet over
70% failed within two years. I got to where I could replace them
without looking at the chassis more than a few glances. Prewire the
replacement, bolt it in and move a few wires.

Mechanically, how do pots with a linear tracks differ from those with
non-linear tracks (in a way which cause them to fail when they pass DC).


Carbon pots aren't build for rheostat service. That's what they make
wirewound for.

Maybe, but there are many carbon pots which do carry DC (albeit usually
not a lot).

Cermet doesn't seem to mind.

cavelamb wrote:
Spehro Pefhany wrote:
On Tue, 12 Jan 2010 18:46:13 -0800, the renowned John Larkin
wrote:

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering
if those qualified in this group would mind reviewing the schematic
and making constructive commentary. I'd like this thing to be
reliable and not cause a host of other electronic gremlins. Just
keep in mind, I'm a chemist, not an electronics professional. But,
I do learn quickly.


Thanks.
Charlie Smith
It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John

Good point, also I don't think running 20mA through 365-1467-ND 3mm
white LEDs is a very good idea. That's the abs. max. current at 25C.
If you want it not to fail at high panel temperatures/brightness (ever
come in out of the sun into a dark hangar?) and don't want to be one
of those folks who complains about crappy LEDs (because they abuse
them and they die fast) you should derate to something like 10mA. Two
in series at 10mA will give you about the same amount of light, but
draw 1/2 the current (at twice the voltage), but you need twice as
many LEDs.
So, say the voltage regulator goes from 5V to 12V, the resistors in
series with each series LED pair will be 560 ohms. A TO-220 LM317
should not need hardly any heatsink (and there will be no RFI
generated). Say 180R from LM317 out to sense input, and a 1K pot in
series with 510R to ground from the sense input, to give about a 5V to
12V adjustment range. Maybe a mechanical switch to give off/bypass
(full bright)/dimmed.
Disadvantage of this over the PWM is that the apparent brightness will
not vary nearly as linearly (most of the action will be down near
where it just comes on), so check it out on the bench to make sure you
can live with it if you decide to go this way.

Best regards, Spehro Pefhany


An audio taper pot should help with the brightness question?



Thanks to all who contributed comments. I appreciate all the suggestions
for improvement.

I am somewhat puzzled by the characterization of it being inefficient. I
guess it depends on perspective. I think this approach has to be a
whole lot more efficient than incandescent bulbs and a lot more
reliable. This was my main goal here.

I appreciate the insight on noise. This is something I had not even
considered but, in hindsight, is really obvious. The linear regulator
seems a great idea. What do you think about the LT3080? It looks like I
would need a 100K trim pot to dial in my voltage. The required 50K
should be right in the middle of the range for the pot.

I was also surprised about the comments regarding If. The lit advertised
20 mA as typical current with 50 mA max and that's why I designed around
20. But, there it was in Fig 4 showing a steady drop off hotter than
25°C and the interior could easily see 35-40°C peak temps. That's a good
catch. Thanks. What I may do here is de-rate down to 16 or 18 mA since I
won't be spending much time in a 40°C cockpit. Promise. At 5V and 16
mA, it will take 100 ohm resistors. Even RadioShack should have those on
the shelf.

The board will be fused with a 2A fuse at the buss. Backup is a
flashlight, same as with the rest of us flying behind incandescent
bulbs. If I can mount them easy enough, I can consider putting them in
in pairs for higher illumination.

Thanks again.
Charlie

On Wed, 13 Jan 2010 17:15:57 -0500, Charlie Smith
wrote:

cavelamb wrote:
Spehro Pefhany wrote:
On Tue, 12 Jan 2010 18:46:13 -0800, the renowned John Larkin
wrote:

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering
if those qualified in this group would mind reviewing the schematic
and making constructive commentary. I'd like this thing to be
reliable and not cause a host of other electronic gremlins. Just
keep in mind, I'm a chemist, not an electronics professional. But,
I do learn quickly.


Thanks.
Charlie Smith
It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John

Good point, also I don't think running 20mA through 365-1467-ND 3mm
white LEDs is a very good idea. That's the abs. max. current at 25C.
If you want it not to fail at high panel temperatures/brightness (ever
come in out of the sun into a dark hangar?) and don't want to be one
of those folks who complains about crappy LEDs (because they abuse
them and they die fast) you should derate to something like 10mA. Two
in series at 10mA will give you about the same amount of light, but
draw 1/2 the current (at twice the voltage), but you need twice as
many LEDs.
So, say the voltage regulator goes from 5V to 12V, the resistors in
series with each series LED pair will be 560 ohms. A TO-220 LM317
should not need hardly any heatsink (and there will be no RFI
generated). Say 180R from LM317 out to sense input, and a 1K pot in
series with 510R to ground from the sense input, to give about a 5V to
12V adjustment range. Maybe a mechanical switch to give off/bypass
(full bright)/dimmed.
Disadvantage of this over the PWM is that the apparent brightness will
not vary nearly as linearly (most of the action will be down near
where it just comes on), so check it out on the bench to make sure you
can live with it if you decide to go this way.

Best regards, Spehro Pefhany


An audio taper pot should help with the brightness question?



Thanks to all who contributed comments. I appreciate all the suggestions
for improvement.

I am somewhat puzzled by the characterization of it being inefficient. I
guess it depends on perspective.


No, it's quantifiable. If you're running a 2 volt LED from a 14 volt
supply, with only resistive-type (lossy) current limiting, the
electrical efficiency is about 15%. The electrical-to-optical
efficiency will be a lot less, worse than incandescents probably. Most
of the input power is being burned up in the resistors.

If you put pairs of LEDs in series, efficiency doubles. Three or four
per string is even better... if efficiency matters at all here.

John

John Larkin wrote:
On Wed, 13 Jan 2010 17:15:57 -0500, Charlie Smith
wrote:

cavelamb wrote:
Spehro Pefhany wrote:
On Tue, 12 Jan 2010 18:46:13 -0800, the renowned John Larkin
wrote:

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering
if those qualified in this group would mind reviewing the schematic
and making constructive commentary. I'd like this thing to be
reliable and not cause a host of other electronic gremlins. Just
keep in mind, I'm a chemist, not an electronics professional. But,
I do learn quickly.


Thanks.
Charlie Smith
It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John
Good point, also I don't think running 20mA through 365-1467-ND 3mm
white LEDs is a very good idea. That's the abs. max. current at 25C.
If you want it not to fail at high panel temperatures/brightness (ever
come in out of the sun into a dark hangar?) and don't want to be one
of those folks who complains about crappy LEDs (because they abuse
them and they die fast) you should derate to something like 10mA. Two
in series at 10mA will give you about the same amount of light, but
draw 1/2 the current (at twice the voltage), but you need twice as
many LEDs.
So, say the voltage regulator goes from 5V to 12V, the resistors in
series with each series LED pair will be 560 ohms. A TO-220 LM317
should not need hardly any heatsink (and there will be no RFI
generated). Say 180R from LM317 out to sense input, and a 1K pot in
series with 510R to ground from the sense input, to give about a 5V to
12V adjustment range. Maybe a mechanical switch to give off/bypass
(full bright)/dimmed.
Disadvantage of this over the PWM is that the apparent brightness will
not vary nearly as linearly (most of the action will be down near
where it just comes on), so check it out on the bench to make sure you
can live with it if you decide to go this way.

Best regards, Spehro Pefhany

An audio taper pot should help with the brightness question?


Thanks to all who contributed comments. I appreciate all the suggestions
for improvement.

I am somewhat puzzled by the characterization of it being inefficient. I
guess it depends on perspective.


No, it's quantifiable. If you're running a 2 volt LED from a 14 volt
supply, with only resistive-type (lossy) current limiting, the
electrical efficiency is about 15%. The electrical-to-optical
efficiency will be a lot less, worse than incandescents probably. Most
of the input power is being burned up in the resistors.

If you put pairs of LEDs in series, efficiency doubles. Three or four
per string is even better... if efficiency matters at all here.

John




Thanks. Its a 3.4 V LED from a 5 V supply. I have the resistors
dissipating 656 mW (32%) and the LEDs dissipating 1360 mW (68%) of the
input power of 2 Watts ( 5V * 400 mA). Sound right?

Charlie

On Wed, 13 Jan 2010 18:40:43 -0500, Charlie Smith
wrote:

John Larkin wrote:
On Wed, 13 Jan 2010 17:15:57 -0500, Charlie Smith
wrote:

cavelamb wrote:
Spehro Pefhany wrote:
On Tue, 12 Jan 2010 18:46:13 -0800, the renowned John Larkin
wrote:

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering
if those qualified in this group would mind reviewing the schematic
and making constructive commentary. I'd like this thing to be
reliable and not cause a host of other electronic gremlins. Just
keep in mind, I'm a chemist, not an electronics professional. But,
I do learn quickly.


Thanks.
Charlie Smith
It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John
Good point, also I don't think running 20mA through 365-1467-ND 3mm
white LEDs is a very good idea. That's the abs. max. current at 25C.
If you want it not to fail at high panel temperatures/brightness (ever
come in out of the sun into a dark hangar?) and don't want to be one
of those folks who complains about crappy LEDs (because they abuse
them and they die fast) you should derate to something like 10mA. Two
in series at 10mA will give you about the same amount of light, but
draw 1/2 the current (at twice the voltage), but you need twice as
many LEDs.
So, say the voltage regulator goes from 5V to 12V, the resistors in
series with each series LED pair will be 560 ohms. A TO-220 LM317
should not need hardly any heatsink (and there will be no RFI
generated). Say 180R from LM317 out to sense input, and a 1K pot in
series with 510R to ground from the sense input, to give about a 5V to
12V adjustment range. Maybe a mechanical switch to give off/bypass
(full bright)/dimmed.
Disadvantage of this over the PWM is that the apparent brightness will
not vary nearly as linearly (most of the action will be down near
where it just comes on), so check it out on the bench to make sure you
can live with it if you decide to go this way.

Best regards, Spehro Pefhany

An audio taper pot should help with the brightness question?


Thanks to all who contributed comments. I appreciate all the suggestions
for improvement.

I am somewhat puzzled by the characterization of it being inefficient. I
guess it depends on perspective.


No, it's quantifiable. If you're running a 2 volt LED from a 14 volt
supply, with only resistive-type (lossy) current limiting, the
electrical efficiency is about 15%. The electrical-to-optical
efficiency will be a lot less, worse than incandescents probably. Most
of the input power is being burned up in the resistors.

If you put pairs of LEDs in series, efficiency doubles. Three or four
per string is even better... if efficiency matters at all here.

John




Thanks. Its a 3.4 V LED from a 5 V supply. I have the resistors
dissipating 656 mW (32%) and the LEDs dissipating 1360 mW (68%) of the
input power of 2 Watts ( 5V * 400 mA). Sound right?

Charlie

Yes, but the 5-volt regulator is only 5/14 = 35% efficient, and *then*
you get to power the resistors and LEDs at 68%, so the product is
about 24%. That's better than the 15% I estimated because your LED
voltage is higher.

If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.

A switching regulator can approach 100%, but is probably not worth the
hassle here.

John

John Larkin wrote:
On Wed, 13 Jan 2010 18:40:43 -0500, Charlie Smith
wrote:

John Larkin wrote:
On Wed, 13 Jan 2010 17:15:57 -0500, Charlie Smith
wrote:

cavelamb wrote:
Spehro Pefhany wrote:
On Tue, 12 Jan 2010 18:46:13 -0800, the renowned John Larkin
wrote:

On Mon, 11 Jan 2010 21:06:32 -0500, Charlie Smith
wrote:

I have cobbled together a circuit from available internet sources to
illuminate a homebuilt aircraft instrument panel. I was wondering
if those qualified in this group would mind reviewing the schematic
and making constructive commentary. I'd like this thing to be
reliable and not cause a host of other electronic gremlins. Just
keep in mind, I'm a chemist, not an electronics professional. But,
I do learn quickly.


Thanks.
Charlie Smith
It's not very efficient. You could put the LEDs in series clusters and
run the substrings from a somewhat higher voltage.

A mosfet would be nice, instead of the antique 3055.

Come to think about it, PWM isn't any more efficient here than linear
regulation. A variable-voltage regulator, LM317 sort of thing with a
pot, could replace the 555 and all that stuff. Much simpler.

John
Good point, also I don't think running 20mA through 365-1467-ND 3mm
white LEDs is a very good idea. That's the abs. max. current at 25C.
If you want it not to fail at high panel temperatures/brightness (ever
come in out of the sun into a dark hangar?) and don't want to be one
of those folks who complains about crappy LEDs (because they abuse
them and they die fast) you should derate to something like 10mA. Two
in series at 10mA will give you about the same amount of light, but
draw 1/2 the current (at twice the voltage), but you need twice as
many LEDs.
So, say the voltage regulator goes from 5V to 12V, the resistors in
series with each series LED pair will be 560 ohms. A TO-220 LM317
should not need hardly any heatsink (and there will be no RFI
generated). Say 180R from LM317 out to sense input, and a 1K pot in
series with 510R to ground from the sense input, to give about a 5V to
12V adjustment range. Maybe a mechanical switch to give off/bypass
(full bright)/dimmed.
Disadvantage of this over the PWM is that the apparent brightness will
not vary nearly as linearly (most of the action will be down near
where it just comes on), so check it out on the bench to make sure you
can live with it if you decide to go this way.

Best regards, Spehro Pefhany
An audio taper pot should help with the brightness question?


Thanks to all who contributed comments. I appreciate all the suggestions
for improvement.

I am somewhat puzzled by the characterization of it being inefficient. I
guess it depends on perspective.

No, it's quantifiable. If you're running a 2 volt LED from a 14 volt
supply, with only resistive-type (lossy) current limiting, the
electrical efficiency is about 15%. The electrical-to-optical
efficiency will be a lot less, worse than incandescents probably. Most
of the input power is being burned up in the resistors.

If you put pairs of LEDs in series, efficiency doubles. Three or four
per string is even better... if efficiency matters at all here.

John



Thanks. Its a 3.4 V LED from a 5 V supply. I have the resistors
dissipating 656 mW (32%) and the LEDs dissipating 1360 mW (68%) of the
input power of 2 Watts ( 5V * 400 mA). Sound right?

Charlie

Yes, but the 5-volt regulator is only 5/14 = 35% efficient, and *then*
you get to power the resistors and LEDs at 68%, so the product is
about 24%. That's better than the 15% I estimated because your LED
voltage is higher.

If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.

A switching regulator can approach 100%, but is probably not worth the
hassle here.

John




Thanks John. Its making more sense; 68% of 35%. I also don't know how
much power the transistor will consume but probably a lot less than
either the regulator or LEDs. I did check and see how much power 20
typical (GE-57) instrument lights would consume and it is about 67 Watts
(4.8 A). That has to be more than what this puppy will burn. At least
I hope so. I have the circuit bread boarded here with one LED. I can
check current draw and then add a couple more LEDs and get some
projection as to what it may pull when running all-out.

This is making the linear regulator seem more practical. It will permit
me, if needed, to wire the LEDs in series and then adjust Vout up to
something like 8 or 8.5 volts to accommodate the new forward voltage
requirement. Shielding against noise works but not as well as not
having noise in the first place.

Thanks again.

Charlie

On Wed, 13 Jan 2010 19:15:56 -0800, John Larkin
wrote:



If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.

---
Yes, but if one fails shorted the second one will follow soon after if
its If(max) is exceeded, while if one fails open two will go out so
you're trading efficiency for reliability.

JF

On Thu, 14 Jan 2010 06:53:35 -0600, John Fields
wrote:

On Wed, 13 Jan 2010 19:15:56 -0800, John Larkin
wrote:



If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.

---
Yes, but if one fails shorted the second one will follow soon after if
its If(max) is exceeded, while if one fails open two will go out so
you're trading efficiency for reliability.

JF

Sure, except that LEDs rarely fail.

John

"John Larkin" wrote in message
...
A switching regulator can approach 100%, but is probably not worth the
hassle here.

Pffbt ;-)
http://webpages.charter.net/dawill/Images/CC_Buck.gif
With a regulated 5V rail, you can replace the two diodes with a resistor.
You can also drop the capacitor, at some expense to switching performance.

The feedback winding is so small, you can probably use an ordinary solenoid
choke and run a few loops of track around it on the PCB.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

"Tim Williams" wrote in message
...
"John Larkin" wrote in
message ...
A switching regulator can approach 100%, but is probably not worth the
hassle here.

Pffbt ;-)
http://webpages.charter.net/dawill/Images/CC_Buck.gif
With a regulated 5V rail, you can replace the two diodes with a resistor.
You can also drop the capacitor, at some expense to switching performance.

The feedback winding is so small, you can probably use an ordinary
solenoid choke and run a few loops of track around it on the PCB.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms


Does that LT8030 linear regulator seem a good sub here for the L78S
regulator? Seems to solve the noise issue and provide future flexibility to
put in two LED's in series. With the L78, no possiblity to do that as it
puts out 5V. Anything about that option I should know about before I order
one with a trim pot?

Thanks group.
Charlie

On Thu, 14 Jan 2010 07:16:21 -0800, John Larkin
wrote:

On Thu, 14 Jan 2010 06:53:35 -0600, John Fields
wrote:

On Wed, 13 Jan 2010 19:15:56 -0800, John Larkin
wrote:



If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.

---
Yes, but if one fails shorted the second one will follow soon after if
its If(max) is exceeded, while if one fails open two will go out so
you're trading efficiency for reliability.

JF

Sure, except that LEDs rarely fail.

---
"Rarely" ain't "never", bucko.

It's kinda the same thing as "a long time" ain't "forever".

JF

On Thu, 14 Jan 2010 09:25:09 -0700, Jim Thompson
/Snicker wrote:

Didn't the OP want dimming?

---
Yup...

Be back in a bit. :-)

JF

John Fields wrote:
On Wed, 13 Jan 2010 19:15:56 -0800, John Larkin
wrote:




If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.


---
Yes, but if one fails shorted the second one will follow soon after if
its If(max) is exceeded, while if one fails open two will go out so
you're trading efficiency for reliability.

JF

I have a problem - my lack of knowledge - with running the
LEDs in series. I don't know aircraft rules/regs, or whether
there is ever a time you would want individual on/off control
of instrument illumination. So I want each LED to be able
to be switched on/off to accomodate that possibility, which
rules out series connection.

My next problem is the same that you have in mind - if one
LED fails you lose two lights. Not understanding what I'll
call "cockpit dynamics", I would be concerned about that series
design. If I'm a passenger on that plane, I want the pilot
to have 100% of his instrumentation available, 100% of
the time.

Actually, I'd prefer two complete systems - regulators,
PWM, whatever with two LEDs per instrument, one fed by
system A and one fed by system B. That way, it would take
two failures for an instrument's lighting to fail completely.
I'm guessing that if the 14V bus goes down, you have bigger
problems. Maybe I'm just being a nervous Nelly because
I have no knowledge of flying a plane.

Ed

On Thu, 14 Jan 2010 11:11:03 -0600, John Fields
wrote:

On Thu, 14 Jan 2010 07:16:21 -0800, John Larkin
wrote:

On Thu, 14 Jan 2010 06:53:35 -0600, John Fields
wrote:

On Wed, 13 Jan 2010 19:15:56 -0800, John Larkin
wrote:



If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.

---
Yes, but if one fails shorted the second one will follow soon after if
its If(max) is exceeded, while if one fails open two will go out so
you're trading efficiency for reliability.

JF

Sure, except that LEDs rarely fail.

---
"Rarely" ain't "never", bucko.

It's kinda the same thing as "a long time" ain't "forever".

JF

Whatever.

John

On Thu, 14 Jan 2010 12:39:35 -0500, ehsjr
wrote:

John Fields wrote:
On Wed, 13 Jan 2010 19:15:56 -0800, John Larkin
wrote:




If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.


---
Yes, but if one fails shorted the second one will follow soon after if
its If(max) is exceeded, while if one fails open two will go out so
you're trading efficiency for reliability.

JF

I have a problem - my lack of knowledge - with running the
LEDs in series. I don't know aircraft rules/regs, or whether
there is ever a time you would want individual on/off control
of instrument illumination. So I want each LED to be able
to be switched on/off to accomodate that possibility, which
rules out series connection.

His posted schematic didn't have switches.


My next problem is the same that you have in mind - if one
LED fails you lose two lights. Not understanding what I'll
call "cockpit dynamics", I would be concerned about that series
design. If I'm a passenger on that plane, I want the pilot
to have 100% of his instrumentation available, 100% of
the time.

Panel backlights usually have overlapping so that any region is lit by
at least two lamps. If one goes out, a zone may be a bit dimmer but
not invisible. If led's are used in series, just locate them so that a
string failure doesn't blank any region, just dims some.

One common structure is a clear plastic plate with a white reflective
layer on both sides, then a black layer on top. Lettering is engraved
through the black on top. Multiple lamps are recessed into the back
side, and they scatter light all over the place, so no area is lit by
just one lamp.


Actually, I'd prefer two complete systems - regulators,
PWM, whatever with two LEDs per instrument, one fed by
system A and one fed by system B. That way, it would take
two failures for an instrument's lighting to fail completely.
I'm guessing that if the 14V bus goes down, you have bigger
problems. Maybe I'm just being a nervous Nelly because
I have no knowledge of flying a plane.

A flashlight is the backup.

Do small planes still use dual magneto ignitions?

John

Jim Thompson wrote:

JF
Didn't the OP want dimming?

...Jim Thompson

A "quicky" is attached ;-)

...Jim Thompson


I might try that one on my sailboat...


--

Richard Lamb
http://www.home.earthlink.net/~cavelamb/


"The clock of life is wound but once, and no man has the power
to tell just when the hands will stop, at late or early hour...
Now is the only time you own. Live, love, toil with a will.
Place no faith in time. For the clock may soon be still."

On Thu, 14 Jan 2010 14:08:04 -0600, cavelamb
wrote:

Jim Thompson wrote:

JF
Didn't the OP want dimming?

...Jim Thompson

A "quicky" is attached ;-)

...Jim Thompson


I might try that one on my sailboat...

Caution: I didn't have time to elaborate on protection schemes. If
an alternator source is involved, add a series diode (to keep negative
transients from creaming the LM317), plus a series R and a zener or
TVS to ground to catch the positive ones.

(Dropping resistors were calculated based on white LED's (3.4V nominal
forward). Adjust accordingly for other colors.)

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

Jim Thompson wrote:
On Thu, 14 Jan 2010 14:08:04 -0600, cavelamb
wrote:

Jim Thompson wrote:
JF
Didn't the OP want dimming?

...Jim Thompson
A "quicky" is attached ;-)

...Jim Thompson

I might try that one on my sailboat...

Caution: I didn't have time to elaborate on protection schemes. If
an alternator source is involved, add a series diode (to keep negative
transients from creaming the LM317), plus a series R and a zener or
TVS to ground to catch the positive ones.

(Dropping resistors were calculated based on white LED's (3.4V nominal
forward). Adjust accordingly for other colors.)

...Jim Thompson

My electrical panel is backlit.
A translucent plastic panel with opaque covering with legends and lettering
engraved.

Stock was a single dim bulb buried under the switches, breakers and wiring.

I found a flexible white LED strip at AutoZone that worked out real well.
Three LEDs in a group, and four groups to a strip.
They can be cut apart and wires soldered to each group of three.
Self-adhesive backing too.

I cut up one and stuck it behind the panel - works great!
The legends and lettering are sharp, crisp, clear, easy to read.

But maybe a bit too bright on a dark night.
Rather than adding a red set (which to my eye look fuzzy), just dimming the
white would be fine.

And yes, we have a flashlight for backup



--

Richard Lamb
http://www.home.earthlink.net/~cavelamb/

On Thu, 14 Jan 2010 17:07:06 -0600, cavelamb
wrote:

Jim Thompson wrote:
On Thu, 14 Jan 2010 14:08:04 -0600, cavelamb
wrote:

Jim Thompson wrote:
JF
Didn't the OP want dimming?

...Jim Thompson
A "quicky" is attached ;-)

...Jim Thompson

I might try that one on my sailboat...

Caution: I didn't have time to elaborate on protection schemes. If
an alternator source is involved, add a series diode (to keep negative
transients from creaming the LM317), plus a series R and a zener or
TVS to ground to catch the positive ones.

(Dropping resistors were calculated based on white LED's (3.4V nominal
forward). Adjust accordingly for other colors.)

...Jim Thompson

My electrical panel is backlit.
A translucent plastic panel with opaque covering with legends and lettering
engraved.

Stock was a single dim bulb buried under the switches, breakers and wiring.

I found a flexible white LED strip at AutoZone that worked out real well.
Three LEDs in a group, and four groups to a strip.
They can be cut apart and wires soldered to each group of three.
Self-adhesive backing too.

I cut up one and stuck it behind the panel - works great!
The legends and lettering are sharp, crisp, clear, easy to read.

But maybe a bit too bright on a dark night.
Rather than adding a red set (which to my eye look fuzzy), just dimming the
white would be fine.

And yes, we have a flashlight for backup

I just installed some under-cabinet LED strips in my wife's office
because she was complaining of shadows. Absolutely delightful! So
bright I worry over lifetime :-(

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

On Thu, 14 Jan 2010 11:11:03 -0600, the renowned John Fields
wrote:

On Thu, 14 Jan 2010 07:16:21 -0800, John Larkin
wrote:

On Thu, 14 Jan 2010 06:53:35 -0600, John Fields
wrote:

On Wed, 13 Jan 2010 19:15:56 -0800, John Larkin
wrote:



If you use 14 volts to run a 3.4 volt LED, with a dissipative system,
the efficiency is always around 24%. Different circuits just move the
heat around.

With pairs of LEDs in series, efficiency becomes 6.8/14 = 48%.

---
Yes, but if one fails shorted the second one will follow soon after if
its If(max) is exceeded, while if one fails open two will go out so
you're trading efficiency for reliability.

JF

Sure, except that LEDs rarely fail.

---
"Rarely" ain't "never", bucko.

It's kinda the same thing as "a long time" ain't "forever".

JF

So "love you long time" is different from "love you forever"?


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com