[OT] World's brightest LED

classic Classic list List threaded Threaded
16 messages Options
Reply | Threaded
Open this post in threaded view
|

[OT] World's brightest LED

Ryan O'Connor
I stumbled across the world's brightest single cell LED light on youtube
and immediately decided I had to share it with Russel, but then realised if
he is interested in LEDs then other people on here will be too:

https://www.youtube.com/watch?v=bBV-1VNWscA

Ryan
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

RussellMc
On Mon, 7 Sep 2020 at 08:58, Ryan O'Connor <[hidden email]> wrote:

> I stumbled across the world's brightest single cell LED light on youtube
> and immediately decided I had to share it with Russel, but then realised if
> he is interested in LEDs then other people on here will be too:
>
> https://www.youtube.com/watch?v=bBV-1VNWscA
>
> Fun & interesting.
He (surprisingly) does not show what the  beam is actually like.
I didn't note any claims on lumen output, but it should be well over
100,000 lumen.

Note to self: Find out brand of power supplies used and avoid buying them.
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Ryan O'Connor
Yeah it was a shame about the power supplies. If they didn't die I'm sure
he could have come up with many more demo ideas. He spent so much work
making the cooler though that I'm sure we will see a follow-up video
surface eventually.

On Mon, 7 Sep 2020 at 15:19, RussellMc <[hidden email]> wrote:

> On Mon, 7 Sep 2020 at 08:58, Ryan O'Connor <[hidden email]> wrote:
>
> > I stumbled across the world's brightest single cell LED light on youtube
> > and immediately decided I had to share it with Russel, but then realised
> if
> > he is interested in LEDs then other people on here will be too:
> >
> > https://www.youtube.com/watch?v=bBV-1VNWscA
> >
> > Fun & interesting.
> He (surprisingly) does not show what the  beam is actually like.
> I didn't note any claims on lumen output, but it should be well over
> 100,000 lumen.
>
> Note to self: Find out brand of power supplies used and avoid buying them.
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Clint Jay
In reply to this post by RussellMc
Well the server supplies were definitely HP branded, I think DPS800, the
boost converters were some no name Chinese brand ones

On Mon, 7 Sep 2020 at 04:19, RussellMc <[hidden email]> wrote:

> On Mon, 7 Sep 2020 at 08:58, Ryan O'Connor <[hidden email]> wrote:
>
> > I stumbled across the world's brightest single cell LED light on youtube
> > and immediately decided I had to share it with Russel, but then realised
> if
> > he is interested in LEDs then other people on here will be too:
> >
> > https://www.youtube.com/watch?v=bBV-1VNWscA
> >
> > Fun & interesting.
> He (surprisingly) does not show what the  beam is actually like.
> I didn't note any claims on lumen output, but it should be well over
> 100,000 lumen.
>
> Note to self: Find out brand of power supplies used and avoid buying them.
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>


--
Clint. M0UAW IO83

*No trees were harmed in the sending of this mail. However, a large number
of electrons were greatly inconvenienced.*
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Alan Pearce
In reply to this post by RussellMc
> He (surprisingly) does not show what the  beam is actually like.

Yeah, I was wondering how it would focus using the lens from a high
beam headlight from a vehicle which had separate low and high beam
lights, or maybe the lens from a driving headlamp.

These were about the only lens I could think of that would be large
enough to capture the beam.

I did like his use of aluminium tape on the clamps. I presume these
were plastic seeing he had 3D printed them, but you didn't actually
see them without the tape on.

On Mon, 7 Sep 2020 at 04:19, RussellMc <[hidden email]> wrote:

>
> On Mon, 7 Sep 2020 at 08:58, Ryan O'Connor <[hidden email]> wrote:
>
> > I stumbled across the world's brightest single cell LED light on youtube
> > and immediately decided I had to share it with Russel, but then realised if
> > he is interested in LEDs then other people on here will be too:
> >
> > https://www.youtube.com/watch?v=bBV-1VNWscA
> >
> > Fun & interesting.
> He (surprisingly) does not show what the  beam is actually like.
> I didn't note any claims on lumen output, but it should be well over
> 100,000 lumen.
>
> Note to self: Find out brand of power supplies used and avoid buying them.
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Sean Breheny
This appears to be the manufacturer page for this LED:

https://store.yujiintl.com/products/yujileds-bc-series-high-cri-cob-led-900h-1500w-pack-1pcs

It has a forward voltage of between 30 and 41V which means that it must be
an array of LEDs. As such, I'm not sure what is meant by "world's brightest
single-unit LED". I find it hard to believe that nobody makes a more
powerful single module (containing multiple LED junctions) especially since
this company is one I've never heard of before and is not in the top ten
global manufacturers of either LED dies, LED modules, or LED lighting
assemblies.

Also, the manufacturer itself links to this guy's video on the page for
this LED :)

Sean


On Mon, Sep 7, 2020 at 6:53 AM Alan Pearce <[hidden email]>
wrote:

> > He (surprisingly) does not show what the  beam is actually like.
>
> Yeah, I was wondering how it would focus using the lens from a high
> beam headlight from a vehicle which had separate low and high beam
> lights, or maybe the lens from a driving headlamp.
>
> These were about the only lens I could think of that would be large
> enough to capture the beam.
>
> I did like his use of aluminium tape on the clamps. I presume these
> were plastic seeing he had 3D printed them, but you didn't actually
> see them without the tape on.
>
> On Mon, 7 Sep 2020 at 04:19, RussellMc <[hidden email]> wrote:
> >
> > On Mon, 7 Sep 2020 at 08:58, Ryan O'Connor <[hidden email]> wrote:
> >
> > > I stumbled across the world's brightest single cell LED light on
> youtube
> > > and immediately decided I had to share it with Russel, but then
> realised if
> > > he is interested in LEDs then other people on here will be too:
> > >
> > > https://www.youtube.com/watch?v=bBV-1VNWscA
> > >
> > > Fun & interesting.
> > He (surprisingly) does not show what the  beam is actually like.
> > I didn't note any claims on lumen output, but it should be well over
> > 100,000 lumen.
> >
> > Note to self: Find out brand of power supplies used and avoid buying
> them.
> > --
> > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> > View/change your membership options at
> > http://mailman.mit.edu/mailman/listinfo/piclist
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Mike Rigby-Jones
Luminous efficacy is ~80lm/W which is very ordinary by today's
standards.  Possibly of more interest to Russel would be the new Samsung
LM301B, a low-mid power surface mount LED with 220 lm/W efficacy  which
is not too far from the theoretical limit of a phosphor based white LED
(260-300lm/W).

https://www.samsung.com/led/lighting/mid-power-leds/3030-leds/lm301b/

Mike

On 07/09/2020 19:27, Sean Breheny wrote:

> This appears to be the manufacturer page for this LED:
>
> https://store.yujiintl.com/products/yujileds-bc-series-high-cri-cob-led-900h-1500w-pack-1pcs
>
> It has a forward voltage of between 30 and 41V which means that it must be
> an array of LEDs. As such, I'm not sure what is meant by "world's brightest
> single-unit LED". I find it hard to believe that nobody makes a more
> powerful single module (containing multiple LED junctions) especially since
> this company is one I've never heard of before and is not in the top ten
> global manufacturers of either LED dies, LED modules, or LED lighting
> assemblies.
>
> Also, the manufacturer itself links to this guy's video on the page for
> this LED :)
>
> Sean
>
>
> On Mon, Sep 7, 2020 at 6:53 AM Alan Pearce <[hidden email]>
> wrote:
>
>>> He (surprisingly) does not show what the  beam is actually like.
>> Yeah, I was wondering how it would focus using the lens from a high
>> beam headlight from a vehicle which had separate low and high beam
>> lights, or maybe the lens from a driving headlamp.
>>
>> These were about the only lens I could think of that would be large
>> enough to capture the beam.
>>
>> I did like his use of aluminium tape on the clamps. I presume these
>> were plastic seeing he had 3D printed them, but you didn't actually
>> see them without the tape on.
>>
>> On Mon, 7 Sep 2020 at 04:19, RussellMc <[hidden email]> wrote:
>>> On Mon, 7 Sep 2020 at 08:58, Ryan O'Connor <[hidden email]> wrote:
>>>
>>>> I stumbled across the world's brightest single cell LED light on
>> youtube
>>>> and immediately decided I had to share it with Russel, but then
>> realised if
>>>> he is interested in LEDs then other people on here will be too:
>>>>
>>>> https://www.youtube.com/watch?v=bBV-1VNWscA
>>>>
>>>> Fun & interesting.
>>> He (surprisingly) does not show what the  beam is actually like.
>>> I didn't note any claims on lumen output, but it should be well over
>>> 100,000 lumen.
>>>
>>> Note to self: Find out brand of power supplies used and avoid buying
>> them.
>>> --
>>> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
>>> View/change your membership options at
>>> http://mailman.mit.edu/mailman/listinfo/piclist
>> --
>> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
>> View/change your membership options at
>> http://mailman.mit.edu/mailman/listinfo/piclist
>>

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

RussellMc
On Tue, 8 Sep 2020 at 06:56, Mike <[hidden email]> wrote:

> Luminous efficacy is ~80lm/W which is very ordinary by today's
> standards.  Possibly of more interest to Russel would be the new Samsung
> LM301B, a low-mid power surface mount LED with 220 lm/W efficacy  which
> is not too far from the theoretical limit of a phosphor based white LED
> (260-300lm/W).
>
> https://www.samsung.com/led/lighting/mid-power-leds/3030-leds/lm301b/
>
>
> They look useful. Here we go again :-)

My "best" (Nichia) LEDS longish ago managed 165 l/W at about 100 mW and 130
l/W at 200 mW.
That was about as good as was available commercially then at sensible
prices.
When compared on equal terms (CCT, CRI, binning, ...) the Samsung LEDs
appear to be in the 150% - 200% better l/W range.
Half angle of illumination was 40 degrees against 60 for the Samsung which
was useful in my applications.
(1/2 brightness but far more than half light energy withing that angle).

Datasheet *here
<https://cdn.samsung.com/led/file/resource/2020/03/Data_Sheet_LM301B_CRI80_Rev.10.0.pdf>
*(26
pages)  ref1

Digikey offer SPMWHD32AMD5XAV0S0
Closest in datasheet is SPMWHD32AMD5XAV0S

This is a 3000 degrees K "whole bin"  part meaning you'll get low flux
items (and MAY get some high ones as well :-) ).
Digikey product page *here
<https://www.digikey.co.nz/product-detail/en/samsung-semiconductor,-inc./SPMWHD32AMD5XAV0S0/1510-2376-1-ND/8568392?gclid=CjwKCAjwtNf6BRAwEiwAkt6UQh2E-rBnsn67oP0BcI4Tqn64ao59y3Z3ku3Y6lDwosZ0b8Pt-YpAZRoCz0IQAvD_BwE>*
ref3
At $US0.18 / 1000 that's amazingly good.

The 1.5 kW super LED was CCT 5600K (colour temperature) BUT also CRI = 95
(Colour rendering index).
To get true high CRI you usually need a CCT somewhat under 3000 K to match
tungsten.
The Samsung has CRI=80 which is fine for use where you don't want colours
to look abnormal but where colour matching is not important.

At 5700 K the Samsung gives flux of 38-40 lumen in the SL bin.
To see what you can really buy you need to analyse what the datasheet says
with some care as you have binnings and groupings for Vf (forward voltage),
CCT, flux and position relative to the black body axis on the
They do say
 Kitting Bin Concept
1. Under agreement between customer and SAMSUNG ELECTRONICS, SAMSUNG can
supply kitting bin (VF, Color, lm).
2. A forward voltage (VF) of kitting bin is combined by a pair of same VF
rank such as (AY+AY), (AZ+AZ) or (A1+A1)
3. A Chromaticity Coordinates of kitting bin is mixed by kitting
procedure.(below kitting simulation)
4. A luminous flux(lm) of kitting bin is combined by a pair of IV rank such
as (SH+SH), (SH+SJ), (SJ+SJ), (SJ+SK), (SK+SK), (SK+SL), (SL+SL), (SL+SM)
or (SM+SM)

You are going to pay more for a low Vf (high efficiency), specific CCT &
CRI version but the price still is liable to be good.
In many applications the CRI is not too critical and even changes of flux
(light out) of say 20% are quite likely not to be noticeable.

On page 15 is the Beam Angle curve which may make a significant difference
in some applications.
In my portable lighting applications it was useful if an LED had a low
enough cone angle to be usable wholly or partially without reflector or
lens. This (potentially) reduces both cost and additional light loss.
In many other applications (grow lights, room lights, ...) the angle may be
acceptable as is or the application not so cost sensitive that beam shaping
matters. (Street lights almost invariably use lenses, grow lights probably
don't,...)

The Samsung half intensity angle is +/- 60 degrees or 120 degrees full
angle.
This is typical of silicone encapsulated LEDs

More anon maybe.
Other calls ....

400 Watt growlight using them *here
<https://herbals.co.nz/products/quantum-board-led-grow-light-samsung-lm301b-with-meanwell-driver>
 ref2*




            Russell

Just in case links above break:

ref1:
https://cdn.samsung.com/led/file/resource/2020/03/Data_Sheet_LM301B_CRI80_Rev.10.0.pdf

ref2:
https://herbals.co.nz/products/quantum-board-led-grow-light-samsung-lm301b-with-meanwell-driver

ref3:
https://www.digikey.co.nz/product-detail/en/samsung-semiconductor,-inc./SPMWHD32AMD5XAV0S0/1510-2376-1-ND/8568392?gclid=CjwKCAjwtNf6BRAwEiwAkt6UQh2E-rBnsn67oP0BcI4Tqn64ao59y3Z3ku3Y6lDwosZ0b8Pt-YpAZRoCz0IQAvD_BwE
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

RussellMc
The previously mentioned *growlight
<ref2: https://herbals.co.nz/products/quantum-board-led-grow-light-samsung-lm301b-with-meanwell-driver>*
is
running the LEDs at about 160 mA/LED =  within spec

*This AliExpress unit
<https://www.aliexpress.com/item/4001205543524.html?src=google&albch=shopping&acnt=494-037-6276&isdl=y&slnk=&plac=&mtctp=&albbt=Gploogle_7_shopping&aff_atform=google&aff_short_key=UneMJZVf&&albagn=888888&albcp=9444695485&albag=99457316601&trgt=296904914040&crea=en4001205543524&netw=u&device=c&albpg=296904914040&albpd=en4001205543524&gclid=CjwKCAjwtNf6BRAwEiwAkt6UQtvAGjo1t0jA6SYDFITYRkywxrN0yq8oXYv0ch3sH8vWBPpQdpnTAxoCPzsQAvD_BwE&gclsrc=aw.ds'>*
operates 195 LEDs at about 600mW/LED or at or bove their rated absolute
maximum current. *Ref4*
Available as LED panel, heat sink and power supply.


   Russell


Ref4:
https://www.aliexpress.com/item/4001205543524.html?src=google&albch=shopping&acnt=494-037-6276&isdl=y&slnk=&plac=&mtctp=&albbt=Gploogle_7_shopping&aff_atform=google&aff_short_key=UneMJZVf&&albagn=888888&albcp=9444695485&albag=99457316601&trgt=296904914040&crea=en4001205543524&netw=u&device=c&albpg=296904914040&albpd=en4001205543524&gclid=CjwKCAjwtNf6BRAwEiwAkt6UQtvAGjo1t0jA6SYDFITYRkywxrN0yq8oXYv0ch3sH8vWBPpQdpnTAxoCPzsQAvD_BwE&gclsrc=aw.ds


Growlight  ref2:
https://herbals.co.nz/products/quantum-board-led-grow-light-samsung-lm301b-with-meanwell-driver



>
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Jim Ruxton-2
In reply to this post by RussellMc

> He (surprisingly) does not show what the  beam is actually like.
> I didn't note any claims on lumen output, but it should be well over
> 100,000 lumen.
>
> Note to self: Find out brand of power supplies used and avoid buying them.

I doubt that these supplies are meant to be run in series the way he
connected them.

jim

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Ryan O'Connor
Yeah true don't you need a special output stage on a constant current or
mixed mode SMPS in order to do that?

On Tue, 8 Sep 2020 at 17:27, Jim Ruxton <[hidden email]> wrote:

>
> > He (surprisingly) does not show what the  beam is actually like.
> > I didn't note any claims on lumen output, but it should be well over
> > 100,000 lumen.
> >
> > Note to self: Find out brand of power supplies used and avoid buying
> them.
>
> I doubt that these supplies are meant to be run in series the way he
> connected them.
>
> jim
>
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

madscientistatlarge
All you need to do is disconect the ground on one, which isn't great safety wise but will let you run them in series.  There's a page for RC guys who do this to recharge there cells quickly.  I think they fried due to the likely very high peak current demand of the boosters.  More filter caps or a heavier ac supply would be needed.  Obviously if used commercially you'd just buy/build a supply, But it's a good way to test them.  They are also oldish supplies, I have 2 old servers that use that form factor.  Obviously power supplies loose capacity as the caps age.


Sent with ProtonMail Secure Email.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Tuesday, September 8, 2020 12:21 AM, Ryan O'Connor <[hidden email]> wrote:

> Yeah true don't you need a special output stage on a constant current or
> mixed mode SMPS in order to do that?
>
> On Tue, 8 Sep 2020 at 17:27, Jim Ruxton [hidden email] wrote:
>
> > > He (surprisingly) does not show what the beam is actually like.
> > > I didn't note any claims on lumen output, but it should be well over
> > > 100,000 lumen.
> > > Note to self: Find out brand of power supplies used and avoid buying
> > > them.
> >
> > I doubt that these supplies are meant to be run in series the way he
> > connected them.
> > jim
> > --
> > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> > View/change your membership options at
> > http://mailman.mit.edu/mailman/listinfo/piclist
>
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist



--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Ryan O'Connor
Well the problem with doing it when they aren't designed for it is that one
can end up drawing much more current than the other since the voltage will
be a little different and output caps are generally pretty low tolerance.
You need an output stage that is designed to sense current across both of
them and adjust the voltage to allow the other supply to draw more to
balance it. That's why constant current mode works much better for joining
two outputs since a good design will basically do this automatically.

A lot of these older power supply designs are constant voltage, which vary
the current supplied to keep the voltage consistent. With current drawn
through two supplies and large voltage drops due to the load, this can
become a chaotic enough system that one or both supplies can get damaged,
and even the load.

On Tue, 8 Sep 2020 at 19:42, madscientistatlarge <
[hidden email]> wrote:

> All you need to do is disconect the ground on one, which isn't great
> safety wise but will let you run them in series.  There's a page for RC
> guys who do this to recharge there cells quickly.  I think they fried due
> to the likely very high peak current demand of the boosters.  More filter
> caps or a heavier ac supply would be needed.  Obviously if used
> commercially you'd just buy/build a supply, But it's a good way to test
> them.  They are also oldish supplies, I have 2 old servers that use that
> form factor.  Obviously power supplies loose capacity as the caps age.
>
>
> Sent with ProtonMail Secure Email.
>
> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
> On Tuesday, September 8, 2020 12:21 AM, Ryan O'Connor <[hidden email]>
> wrote:
>
> > Yeah true don't you need a special output stage on a constant current or
> > mixed mode SMPS in order to do that?
> >
> > On Tue, 8 Sep 2020 at 17:27, Jim Ruxton [hidden email] wrote:
> >
> > > > He (surprisingly) does not show what the beam is actually like.
> > > > I didn't note any claims on lumen output, but it should be well over
> > > > 100,000 lumen.
> > > > Note to self: Find out brand of power supplies used and avoid buying
> > > > them.
> > >
> > > I doubt that these supplies are meant to be run in series the way he
> > > connected them.
> > > jim
> > > --
> > > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> > > View/change your membership options at
> > > http://mailman.mit.edu/mailman/listinfo/piclist
> >
> > --
> > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> > View/change your membership options at
> > http://mailman.mit.edu/mailman/listinfo/piclist
>
>
>
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Allen Mulvey
In reply to this post by Jim Ruxton-2
Please pardon my ignorance. I know little, I think less and less every day, about power supplies. I have always just found one that works and gone with it. However, I had always presumed that they would behave, more or less, like batteries. Put them in series to increase the voltage, parallel to increase the current. What am I missing here? Series to deliver more current?

Thanks,
Allen

From: Jim Ruxton
Sent: Tuesday, September 8, 2020 1:27 AM
To: [hidden email]
Subject: [CAUTION: Failed DKIM Test]Re: [OT] World's brightest LED


> He (surprisingly) does not show what the  beam is actually like.
> I didn't note any claims on lumen output, but it should be well over
> 100,000 lumen.
>
> Note to self: Find out brand of power supplies used and avoid buying them.

I doubt that these supplies are meant to be run in series the way he
connected them.

jim

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

madscientistatlarge
On server supplies the negative side is always connected to the chasis/power ground.  To put them in series you have to insulate the cases from each other as they are tied to the -side, you also have to remove the power ground from one supply to do this so there are potential safety concerns (you would also have to remove any movs from the floating supply so an ac power surge doesn't get sent to the case).

Simple parallel wont work, one supply will always have a different voltage output from the other even if adjusted.  The higher voltage output will provide all the current and any slightly lower output supply will provide zero current.  If you don't mind degraded regulation they can also be connected with low value resistors, the voltage drop across the resistors tending to swamp the differences in output voltage.

The only easy way to get current sharing between multiple supplies without regulation suffering is to use supplies designed to do that.  They share a feedback signal and basically one supply is master and the rest are slaves  (Much the same way that in any high current power supply uses multiple pass elements paralleled to increase current capacity.  By sharing the drive to the pass element from one supply is sent to all the slaves (so the control circuitry in one supply drives all pass elements involved.  this way and one will track the other.  Hot swap redundant current supplies generally have connections between power supplies to accomplish that.

Note that server power supplies can often be had very cheaply once they go "obsolete".  Servers depreciate rapidly and server farms regularly trade out all the gear to save energy with newer designs.  This means there is a glut of used servers and parts available cheap as they approach obsolescence.  I.e. I can get older hp server supplies, 900W, 96% efficient, for around $20 (or less) on ebay.  Server supplies also usually have PMbus serial connections so you can monitor operation and current output of each supply, as well as telling you if one loses it's ac connection or other fault conditions.

Whole servers can also be had cheaply.  1.5 years ago I bought an hp 585 server with 4 processors, 12 cores each and newer processors that are 2 generations newer and 16 cores each (64 cores total!) to replace the original ones.  With rack slides and every thing (got a free half rack from craigslist) for about $750 including internal power supplies, and internal and external power cables on ebay, you just have to look for bargains and pounce when one is going cheap.

I bought this specifically for large compile jobs (Gentoo Linux).  It also has 11 pcie sockets, and can run 5 graphics cards if I wanted to.  Server hardware is awesome, you just need another room to put it in.  The one I have has 5 internal fans, collectively 200W of fan!  When it gets going it gets pretty noisy, though fairly quite at lower processor loads.  You also get server grade drives usually, mine came with 2 450G 10,000 rpm drives.  All the server hardware is much, much nicer than in a "normal" desktop.  With all 4 hotswap power supplies you can get 3,600 Watts total on 110V ac, 4,800 on 220V and similar.  Right now I'm only using 2 supplies, saving the others for backup as this isn't being commercially used and I can handle it if it shuts down due to one supply failing (the other 2 aren't connected to the ac supply so they don't see ongoing stress).  It's simply a beast!


Sent with ProtonMail Secure Email.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Tuesday, September 8, 2020 7:10 AM, Allen Mulvey <[hidden email]> wrote:

> Please pardon my ignorance. I know little, I think less and less every day, about power supplies. I have always just found one that works and gone with it. However, I had always presumed that they would behave, more or less, like batteries. Put them in series to increase the voltage, parallel to increase the current. What am I missing here? Series to deliver more current?
>
> Thanks,
> Allen
>
> From: Jim Ruxton
> Sent: Tuesday, September 8, 2020 1:27 AM
> To: [hidden email]
> Subject: [CAUTION: Failed DKIM Test]Re: [OT] World's brightest LED
>
> > He (surprisingly) does not show what the beam is actually like.
> > I didn't note any claims on lumen output, but it should be well over
> > 100,000 lumen.
> > Note to self: Find out brand of power supplies used and avoid buying them.
>
> I doubt that these supplies are meant to be run in series the way he
> connected them.
>
> jim
>
> --------------------------------------------------------------------------------------------
>
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>
> --------------------------------------------------------------------------------------------------------------------------------------------------------
>
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist



--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Reply | Threaded
Open this post in threaded view
|

Re: [OT] World's brightest LED

Clint Jay
In reply to this post by madscientistatlarge
Nope, those supplies are capable of their rating* but you do need to bear
in mind they're almost definitely second hand, at least 5-10 years old and
will have been worked
hard during their life and then abused by a hobbyist who's done god knows
what to them to get them to work in series and possibly make the fans
quieter or boost the output voltage to the bleeding edge of
reliability/stability

* In a well managed data center environment with guaranteed power, ambient
temperature, airflow etc...

They *are* designed for parallel load sharing...

On Tue, 8 Sep 2020 at 08:42, madscientistatlarge <
[hidden email]> wrote:

> All you need to do is disconect the ground on one, which isn't great
> safety wise but will let you run them in series.  There's a page for RC
> guys who do this to recharge there cells quickly.  I think they fried due
> to the likely very high peak current demand of the boosters.  More filter
> caps or a heavier ac supply would be needed.  Obviously if used
> commercially you'd just buy/build a supply, But it's a good way to test
> them.  They are also oldish supplies, I have 2 old servers that use that
> form factor.  Obviously power supplies loose capacity as the caps age.
>
>
> Sent with ProtonMail Secure Email.
>
> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
> On Tuesday, September 8, 2020 12:21 AM, Ryan O'Connor <[hidden email]>
> wrote:
>
> > Yeah true don't you need a special output stage on a constant current or
> > mixed mode SMPS in order to do that?
> >
> > On Tue, 8 Sep 2020 at 17:27, Jim Ruxton [hidden email] wrote:
> >
> > > > He (surprisingly) does not show what the beam is actually like.
> > > > I didn't note any claims on lumen output, but it should be well over
> > > > 100,000 lumen.
> > > > Note to self: Find out brand of power supplies used and avoid buying
> > > > them.
> > >
> > > I doubt that these supplies are meant to be run in series the way he
> > > connected them.
> > > jim
> > > --
> > > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> > > View/change your membership options at
> > > http://mailman.mit.edu/mailman/listinfo/piclist
> >
> > --
> > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> > View/change your membership options at
> > http://mailman.mit.edu/mailman/listinfo/piclist
>
>
>
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>


--
Clint. M0UAW IO83

*No trees were harmed in the sending of this mail. However, a large number
of electrons were greatly inconvenienced.*
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist