Kill LED lamp flicker...

[Fred Bloggs]

On Sunday, September 10, 2023 at 11:28:53 AM UTC-4, Martin Brown wrote:

On 10/09/2023 16:18, Fred Bloggs wrote:
On Sunday, September 10, 2023 at 10:57:40 AM UTC-4, Martin Brown wrote:
On 09/09/2023 15:11, Fred Bloggs wrote:
On Saturday, September 9, 2023 at 9:01:28 AM UTC-4, Lasse Langwadt Christensen wrote:
lørdag den 9. september 2023 kl. 14.46.48 UTC+2 skrev Fred Bloggs:
On Friday, September 8, 2023 at 11:25:43 AM UTC-4, John Larkin wrote:

Maybe you have cheap flourescent-replacement tubes.
Uh-huh. LEDs don\'t flicker. They\'re all powered by regulated switching ***current*** sources operating in the 20kHz-30kHz range. LEDs are not directly powered by voltage, and the current sources are immune to voltage fluctuation.

some are linear, enough LEDs in series to add up to most of the rectified line voltage

I\'m pretty sure they\'ve never used that method for the commodity lighting bulb market. Maybe for signs and indicator bulb types of applications where it has to be dirt cheap.
You are wrong. I have a dead one sat in a drawer somewhere nearby.
60 LEDs in series across rectified UK 240v mains.

No one is interested in an example of degenerate ad hoc engineering that was completely abandoned, and for good reason.
They were made like that presumably to be as cheap and nasty as
possible. When they worked they were fine and instant on with true rated
brightness unlike the previous generation of CFLs which came on dimly
and almost never reached the brightness that their packaging claimed.
One single LED in the chain has failed. It was the first LED bulb
failure that I ever saw so I dismantled it to see why.

They are the cheapest and nastiest on the market, but at the time it was
bought they sold for premium prices with exaggerated MTBF based on the
expected failure time of a single LED. True MTBF is claimed/60.

If you have a single LED I-V, then stacking them in series in effect creates a composite I-V that is the same with V-axis multiplied by the number in the stack. That could be quite a softening effect. For them to use 60 makes me think they were counting on that effect. It should work pretty well or not depending upon sensitivity of light output to differential I. If they were going for max lumens, they were probably working the LEDs too hard. Apparently significant derating of the operating power dissipation is key to longevity.

That was their undoing. I expect you can still buy them on fleaBay.

--
Martin Brown

 

[Lasse Langwadt Christensen]

søndag den 10. september 2023 kl. 17.49.22 UTC+2 skrev Fred Bloggs:

On Sunday, September 10, 2023 at 11:28:53 AM UTC-4, Martin Brown wrote:
On 10/09/2023 16:18, Fred Bloggs wrote:
On Sunday, September 10, 2023 at 10:57:40 AM UTC-4, Martin Brown wrote:
On 09/09/2023 15:11, Fred Bloggs wrote:
On Saturday, September 9, 2023 at 9:01:28 AM UTC-4, Lasse Langwadt Christensen wrote:
lørdag den 9. september 2023 kl. 14.46.48 UTC+2 skrev Fred Bloggs:
On Friday, September 8, 2023 at 11:25:43 AM UTC-4, John Larkin wrote:

Maybe you have cheap flourescent-replacement tubes.
Uh-huh. LEDs don\'t flicker. They\'re all powered by regulated switching ***current*** sources operating in the 20kHz-30kHz range. LEDs are not directly powered by voltage, and the current sources are immune to voltage fluctuation.

some are linear, enough LEDs in series to add up to most of the rectified line voltage

I\'m pretty sure they\'ve never used that method for the commodity lighting bulb market. Maybe for signs and indicator bulb types of applications where it has to be dirt cheap.
You are wrong. I have a dead one sat in a drawer somewhere nearby.
60 LEDs in series across rectified UK 240v mains.

No one is interested in an example of degenerate ad hoc engineering that was completely abandoned, and for good reason.
They were made like that presumably to be as cheap and nasty as
possible. When they worked they were fine and instant on with true rated
brightness unlike the previous generation of CFLs which came on dimly
and almost never reached the brightness that their packaging claimed.
One single LED in the chain has failed. It was the first LED bulb
failure that I ever saw so I dismantled it to see why.

They are the cheapest and nastiest on the market, but at the time it was
bought they sold for premium prices with exaggerated MTBF based on the
expected failure time of a single LED. True MTBF is claimed/60.
If you have a single LED I-V, then stacking them in series in effect creates a composite I-V that is the same with V-axis multiplied by the number in the stack. That could be quite a softening effect. For them to use 60 makes me think they were counting on that effect. It should work pretty well or not depending upon sensitivity of light output to differential I. If they were going for max lumens, they were probably working the LEDs too hard. Apparently significant derating of the operating power dissipation is key to longevity.

derating also significantly increases efficiency, at the cost of needing more leds for the same light output

 

[Don Y]

On 9/10/2023 7:57 AM, Martin Brown wrote:

You are wrong. I have a dead one sat in a drawer somewhere nearby.
60 LEDs in series across rectified UK 240v mains.

I suspect that\'s designed with cost as the sole factor.

Most of my LED lights have just a few (e.g., less than 10)
emitters. Hard to imagine a whopping dropping resistor
to soak up ~100V of potential!

One single LED in the chain has failed. It was the first LED bulb failure that
I ever saw so I dismantled it to see why.

I should take one of mine apart (they can be disassembled
but likely would be a nightmare to REassemble -- and have
working!)

They are the cheapest and nastiest on the market, but at the time it was bought
they sold for premium prices with exaggerated MTBF based on the expected
failure time of a single LED. True MTBF is claimed/60.

That\'s been the case with all of the newer lighting technologies.
I have \"commercial\" (130V) incandescent flood lights in the
living room which I can\'t recall replacing after their initial
installation. They are much preferred to any of the other
technologies because they can be dimmed to a level that
can\'t be detected UNLESS your eyes have acclimated to total
darkness (i.e., a house guest awakening in the middle of the
night could navigate without having to figure out where
the nearest light switch was located -- and risk blinding himself
by the sudden onset of full light)

The CFLs, OTOH, never dimmed and would fail in relatively short
order -- regardless of their life expectancy claims! (I have
a CFL uplight in the office that seems like it will fail RSN...
yay! one less!!)

 

[Lasse Langwadt Christensen]

søndag den 10. september 2023 kl. 20.08.52 UTC+2 skrev Don Y:

On 9/10/2023 7:57 AM, Martin Brown wrote:
You are wrong. I have a dead one sat in a drawer somewhere nearby.
60 LEDs in series across rectified UK 240v mains.
I suspect that\'s designed with cost as the sole factor.

Most of my LED lights have just a few (e.g., less than 10)
emitters. Hard to imagine a whopping dropping resistor
to soak up ~100V of potential!

what looks like just an LED can have internally multiple LEDs in series
some have a forward voltage of upto 18V, i.e. 6 LEDs in series

 

[Don Y]

On 9/10/2023 8:28 AM, Martin Brown wrote:

They were made like that presumably to be as cheap and nasty as possible. When
they worked they were fine and instant on with true rated brightness unlike the
previous generation of CFLs which came on dimly and almost never reached the
brightness that their packaging claimed.

Our LED lights are (RF) noisier than the CFLs were.
Unlikely that a simple 60Hz rectifier would be throwing
out enough hash to piss off the (HiFi) radio!

We also have some (permanent) \"night lights\" that
have a short string of diodes behind a ballast
and a \"controller chip\" (allows the light to be
set to 4 intensity levels). I can\'t recall what
else was in there (tore it down some time ago
to see about making it\'s lowest setting, lower)

 

[Fred Bloggs]

On Sunday, September 10, 2023 at 12:18:43 PM UTC-4, Lasse Langwadt Christensen wrote:

søndag den 10. september 2023 kl. 17.49.22 UTC+2 skrev Fred Bloggs:
On Sunday, September 10, 2023 at 11:28:53 AM UTC-4, Martin Brown wrote:
On 10/09/2023 16:18, Fred Bloggs wrote:
On Sunday, September 10, 2023 at 10:57:40 AM UTC-4, Martin Brown wrote:
On 09/09/2023 15:11, Fred Bloggs wrote:
On Saturday, September 9, 2023 at 9:01:28 AM UTC-4, Lasse Langwadt Christensen wrote:
lørdag den 9. september 2023 kl. 14.46.48 UTC+2 skrev Fred Bloggs:
On Friday, September 8, 2023 at 11:25:43 AM UTC-4, John Larkin wrote:

Maybe you have cheap flourescent-replacement tubes.
Uh-huh. LEDs don\'t flicker. They\'re all powered by regulated switching ***current*** sources operating in the 20kHz-30kHz range. LEDs are not directly powered by voltage, and the current sources are immune to voltage fluctuation.

some are linear, enough LEDs in series to add up to most of the rectified line voltage

I\'m pretty sure they\'ve never used that method for the commodity lighting bulb market. Maybe for signs and indicator bulb types of applications where it has to be dirt cheap.
You are wrong. I have a dead one sat in a drawer somewhere nearby.
60 LEDs in series across rectified UK 240v mains.

No one is interested in an example of degenerate ad hoc engineering that was completely abandoned, and for good reason.
They were made like that presumably to be as cheap and nasty as
possible. When they worked they were fine and instant on with true rated
brightness unlike the previous generation of CFLs which came on dimly
and almost never reached the brightness that their packaging claimed.
One single LED in the chain has failed. It was the first LED bulb
failure that I ever saw so I dismantled it to see why.

They are the cheapest and nastiest on the market, but at the time it was
bought they sold for premium prices with exaggerated MTBF based on the
expected failure time of a single LED. True MTBF is claimed/60.
If you have a single LED I-V, then stacking them in series in effect creates a composite I-V that is the same with V-axis multiplied by the number in the stack. That could be quite a softening effect. For them to use 60 makes me think they were counting on that effect. It should work pretty well or not depending upon sensitivity of light output to differential I. If they were going for max lumens, they were probably working the LEDs too hard. Apparently significant derating of the operating power dissipation is key to longevity.
derating also significantly increases efficiency, at the cost of needing more leds for the same light output

It must work out in favor of longer operating life in consideration of total carbon footprint.

One aspect of this flicker effect not mentioned is the phosphor and its decay time. Obviously phosphor with longer decay time will attenuate the perception of flicker. Blurb from the LRC at RPI. I wasn\'t expecting a tutorial, but he still could have done a better write-up than this:

https://www.lrc.rpi.edu/resources/newsroom/pdf/2015/ACLEDFlicker_8511.pdf

 

[none]

In article <87d7569f-932b-4430-a34c-858307af260bn@googlegroups.com>,
Lasse Langwadt Christensen <langwadt@fonz.dk> wrote:

derating also significantly increases efficiency, at the cost of
needing more leds for the same light output

No it is not, as long as efficiency means lumens/watt.
Two times as many leds cost two times as much obviously, but
the leds components are dirt cheap and vanish completely with
the price of total energy consumption.


Groetjes Albert
--
Don\'t praise the day before the evening. One swallow doesn\'t make spring.
You must not say \"hey\" before you have crossed the bridge. Don\'t sell the
hide of the bear until you shot it. Better one bird in the hand than ten in
the air. First gain is a cat spinning. - the Wise from Antrim -

 

[Martin Brown]

On 11/09/2023 10:13, albert wrote:

In article <87d7569f-932b-4430-a34c-858307af260bn@googlegroups.com>,
Lasse Langwadt Christensen <langwadt@fonz.dk> wrote:

derating also significantly increases efficiency, at the cost of
needing more leds for the same light output

No it is not, as long as efficiency means lumens/watt.
Two times as many leds cost two times as much obviously, but
the leds components are dirt cheap and vanish completely with
the price of total energy consumption.

Unfortunately consumers look at the headline initial price to buy and
ignore the running costs. That means that the cheapest and nastiest ones
sell in the greatest quantities even if they are less reliable.

--
Martin Brown

 

[Lasse Langwadt Christensen]

mandag den 11. september 2023 kl. 12.45.11 UTC+2 skrev Martin Brown:

On 11/09/2023 10:13, albert wrote:
In article <87d7569f-932b-4430...@googlegroups.com>,
Lasse Langwadt Christensen <lang...@fonz.dk> wrote:

derating also significantly increases efficiency, at the cost of
needing more leds for the same light output

No it is not, as long as efficiency means lumens/watt.
Two times as many leds cost two times as much obviously, but
the leds components are dirt cheap and vanish completely with
the price of total energy consumption.
Unfortunately consumers look at the headline initial price to buy and
ignore the running costs. That means that the cheapest and nastiest ones
sell in the greatest quantities even if they are less reliable.

Dubia had Philips design LED bulbs using 2-3x as many LEDs to increase efficiency and life and madated that the only allowed type of bulb
aka. Dubai lamp

 

[Don Y]

On 9/11/2023 3:45 AM, Martin Brown wrote:

On 11/09/2023 10:13, albert wrote:
In article <87d7569f-932b-4430-a34c-858307af260bn@googlegroups.com>,
Lasse Langwadt Christensen  <langwadt@fonz.dk> wrote:

derating also significantly increases efficiency, at the cost of
needing more leds for the same light output

No it is not, as long as efficiency means lumens/watt.
Two times as many leds cost two times as much obviously, but
the leds components are dirt cheap and vanish completely with
the price of total energy consumption.

Unfortunately consumers look at the headline initial price to buy and ignore
the running costs. That means that the cheapest and nastiest ones sell in the
greatest quantities even if they are less reliable.

Here, the electric utility subsidizes the purchase of \"energy efficient\"
lighting (often completely covering the cost). So, you are left with
THEIR notion of \"what\'s best\"...

 

[Phil Hobbs]

On 2023-09-09 05:42, upsidedown@downunder.com wrote:

On Fri, 8 Sep 2023 08:10:41 -0000 (UTC), Mike Monett VE3BTI
spamme@not.com> wrote:


Note the 1N4007 diodes are replaced by a regular bridge rectifier with
suitable ratings. It is easier to model using 1N4007s than trying to find a
bridge rectifier in LTspice.

The 1N4007 works more like a PIN diode, the other 1N400x diodes are
normal diodes.

IIRC it\'s 1N4005 and up that are PIN structures.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[piglet]

On 08/09/2023 8:10 pm, Mike Monett VE3BTI wrote:

piglet <erichpwagner@hotmail.com> wrote:


The model you used for 1N4007 has something very wrong - if I replace
with one of the 600V types included with LTSpice then the crazy currents
disappear and the circuit does more as you describe. Have you checked
Pdiss in R4?

piglet

VII doesn\'t have the 1N4007, so I got the model from LTspice IV. It shows a
PIV of 1,500 v. There are no 600 V versions in IV.

I see no crazy currents in IV or VII. Can you tell me more about what you are
seeing?

I tried to plot the power in R4 but had no luck in IV or VII. I tried to
calculate pwr(V(x), V(a)) but got 9.65e+307 with no units. So I have to
guess. I\'ll try a 2W wirewound.

Problem solved! During line wrap demangling I had corrupted the 1N4007
model Ibv=10u to become bv=10u so the extraordinary diode currents in my
sim were it breaking down at microvolts. Now it sims more like you saw.

piglet

 

[John Larkin]

On Mon, 11 Sep 2023 12:33:29 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2023-09-09 05:42, upsidedown@downunder.com wrote:
On Fri, 8 Sep 2023 08:10:41 -0000 (UTC), Mike Monett VE3BTI
spamme@not.com> wrote:


Note the 1N4007 diodes are replaced by a regular bridge rectifier with
suitable ratings. It is easier to model using 1N4007s than trying to find a
bridge rectifier in LTspice.

The 1N4007 works more like a PIN diode, the other 1N400x diodes are
normal diodes.

IIRC it\'s 1N4005 and up that are PIN structures.

Cheers

Phil Hobbs

The higher voltage rectifiers often make great drift-step-recovery
(Grehkov) diodes.

2KV negative pulse:

https://www.dropbox.com/s/q82toc257fv43z8/DSRD_neg-2KV.JPG?raw=1

 

[Phil Hobbs]

On 2023-09-11 14:08, John Larkin wrote:

On Mon, 11 Sep 2023 12:33:29 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2023-09-09 05:42, upsidedown@downunder.com wrote:
On Fri, 8 Sep 2023 08:10:41 -0000 (UTC), Mike Monett VE3BTI
spamme@not.com> wrote:


Note the 1N4007 diodes are replaced by a regular bridge rectifier with
suitable ratings. It is easier to model using 1N4007s than trying to find a
bridge rectifier in LTspice.

The 1N4007 works more like a PIN diode, the other 1N400x diodes are
normal diodes.

IIRC it\'s 1N4005 and up that are PIN structures.

Cheers

Phil Hobbs

The higher voltage rectifiers often make great drift-step-recovery
(Grehkov) diodes.

2KV negative pulse:

https://www.dropbox.com/s/q82toc257fv43z8/DSRD_neg-2KV.JPG?raw=1

I\'m impressed that it survived enough pulses for a full trace.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[John Larkin]

On Mon, 11 Sep 2023 14:14:38 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2023-09-11 14:08, John Larkin wrote:
On Mon, 11 Sep 2023 12:33:29 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2023-09-09 05:42, upsidedown@downunder.com wrote:
On Fri, 8 Sep 2023 08:10:41 -0000 (UTC), Mike Monett VE3BTI
spamme@not.com> wrote:


Note the 1N4007 diodes are replaced by a regular bridge rectifier with
suitable ratings. It is easier to model using 1N4007s than trying to find a
bridge rectifier in LTspice.

The 1N4007 works more like a PIN diode, the other 1N400x diodes are
normal diodes.

IIRC it\'s 1N4005 and up that are PIN structures.

Cheers

Phil Hobbs

The higher voltage rectifiers often make great drift-step-recovery
(Grehkov) diodes.

2KV negative pulse:

https://www.dropbox.com/s/q82toc257fv43z8/DSRD_neg-2KV.JPG?raw=1


I\'m impressed that it survived enough pulses for a full trace.

Cheers

Phil Hobbs

The rep-rate was low, with about 50 amps forward and reverse current
into the DSRD.

Hers\'s a water-cooled Pockels Cell driver.

https://www.dropbox.com/scl/fi/5arhyamrp0bl3tgb2fasn/DSC02771.JPG?rlkey=3ttcc2yt6s9nrtdouuv3aneol&raw=1

The big problem was inductors. I kept frying them when I was making
1200 volt pulses at 4 MHz. Skin effect and such.

What worked was that air-core inductor, hand-wound on a selected
Sharpie, with a gap-pad to conduct heat through the board to the
water-cooled baseplate below.

 

[Mike Monett VE3BTI]

piglet <erichpwagner@hotmail.com> wrote:

Problem solved! During line wrap demangling I had corrupted the 1N4007
model Ibv=10u to become bv=10u so the extraordinary diode currents in my
sim were it breaking down at microvolts. Now it sims more like you saw.

piglet

I once tried to bypass the line wrap by zipping the files and uploading to
Google Drive. But I got so many complaints that people could not unzip them
that I gave up.

Now, after having good experience with bare unzipped files, I may try
again.

As a favor, try this and see how it works:

https://tinyurl.com/r6r2m9kc

It will say \"No Preview Available\", but give you the opportunity to
download the file. Do so, and separate the files into ASC and PLT.

Now, any wrap will happen in your text editor. Set the line width to the
maximum and you should have no trouble.

I\'m going to experiment with uploading ASC and PLT files directly. If I can
get Google Drive to stop using one of its editors and wrapping the lines,
then it should work.

--
MRM

 

[Mike Monett VE3BTI]

Mike Monett VE3BTI <spamme@not.com> wrote:

I\'m going to experiment with uploading ASC and PLT files directly. If I
can get Google Drive to stop using one of its editors and wrapping the
lines, then it should work.

It works!

I gave it gibberish extentions that it doesn\'t know what to do with. Just
move the last letter in the extension to the first place.

ASC = CAS
PLT = TPL

Here\'s the files:

https://tinyurl.com/mr32zynw
- cas

https://tinyurl.com/upr8yeap
- tpl

No unzipping, no editing, no line wrap, just rename and go.



--
MRM