Replacement LED?

[DaveC]

The LED in my flashlight is blinking (2/sec). It’s not the support
circuitry (constant current): the voltage across the LED is constant 4v. I
presume it’s a failure mode of the LED. It happens immediately upon
power-on.

http://imgur.com/a/rIRDG

The form-factor is close to a 5x5mm (h x diam). Standard through-hole leads.

But it’s a pretty high-intensity one. Don’t know what makes it so, but I
classify anything with a yellow square visible in the center as
“high-intensity”. Maybe not technically accurate, but there you are.

What I’ve found so far is either a standard 5mm LED but not very bright, or
SMD types requiring heat sinks.

It’s a great little light, and I’ve not found anything as small,
long-lasting, with single AA that I like as much. And this is a learning
experience, so there’s that.

Any pointers to a suitable replacement LED would be appreciated.

Thanks.

[pf...@aol.com]

On Monday, March 21, 2016 at 1:30:08 PM UTC-4, DaveC wrote:

> It's a great little light, and I've not found anything as small,
> long-lasting, with single AA that I like as much. And this is a learning
> experience, so there's that.
>
> Any pointers to a suitable replacement LED would be appreciated.
>
> Thanks.

So, its mot a Maglite? That about covers the "suitable replacement" question. That light was never meant to be serviced, and so your best bet in terms of time/cost/effort, together with the inevitable PITA factor is to simply replace the light altogether.

Peter Wieck
Melrose Park, PA

p.s.: Only Maglites permitted in this household.

[Gareth Magennis]

wrote in message
news:e658aab4-4e65-4e77...@googlegroups.com...

I used to use Maglites, til LED torches came on the scene.

No way I would buy a Maglite now.



Gareth.

[Bob Engelhardt]

On 3/21/2016 1:30 PM, DaveC wrote:
> ... the voltage across the LED is constant 4v.
> ... with single AA ...

So it has a boost SMPS?

[Cursitor Doom]

On Mon, 21 Mar 2016 13:10:09 -0700, pf...@aol.com wrote:

> p.s.: Only Maglites permitted in this household.

Make good billy clubs, the larger ones. :->

[Gareth Magennis]

"Gareth Magennis" wrote in message news:xUYHy.69876$jz.5...@fx41.am4...

Oops, just discovered Maglite do LED's.

My old MiniMaglite wouldn't do 2 gigs on one set of batteries, and I would
regularly have to replace the bulb too, and the switch would always need
maintenance.
Very expensive running costs, very time consuming, but good for hammering in
nails etc.


Gareth.

[pf...@aol.com]

Maglite makes LED units as well. The also stand behind their products....

[John Larkin]

On Mon, 21 Mar 2016 10:30:03 -0700, DaveC <n...@home.cow> wrote:

>The LED in my flashlight is blinking (2/sec). It’s not the support
>circuitry (constant current): the voltage across the LED is constant 4v. I
>presume it’s a failure mode of the LED. It happens immediately upon
>power-on.
>
>http://imgur.com/a/rIRDG
>
>The form-factor is close to a 5x5mm (h x diam). Standard through-hole leads.
>
>But it’s a pretty high-intensity one. Don’t know what makes it so, but I
>classify anything with a yellow square visible in the center as
>“high-intensity”. Maybe not technically accurate, but there you are.

That's probably a downconversion phosphor. White LEDs are really blue
LEDs with a blob of epoxy on top, loaded with yag powder that converts
some of the blue to yellow.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

[Jeff Liebermann]

On Mon, 21 Mar 2016 10:30:03 -0700, DaveC <n...@home.cow> wrote:

>The LED in my flashlight is blinking (2/sec). It’s not the support
>circuitry (constant current): the voltage across the LED is constant 4v. I
>presume it’s a failure mode of the LED. It happens immediately upon
>power-on.

I beg to differ with your diagnosis. Unless the LED has magically
turned into a thermostat controlled flasher, it's not going to do
that. More likely, the LED driver circuit is doing the flashing.
Based on zero detail about the actual flashlight, my wild guess(tm)
would be the big electrolytic that usually goes across the battery.
Broken or badly soldered connections on the driver PCB are also
likely.

>http://imgur.com/a/rIRDG
>The form-factor is close to a 5x5mm (h x diam). Standard through-hole leads.
>
>But it’s a pretty high-intensity one. Don’t know what makes it so, but I
>classify anything with a yellow square visible in the center as
>“high-intensity”. Maybe not technically accurate, but there you are.

If it's not too much trouble, could you disclose the maker and model
number of your flashlight? Extra credit for providing a link to the
manufactures web site or China source link. If there are no numbers
or sources, perhaps a photo of the assembled flashlight?

>What I’ve found so far is either a standard 5mm LED but not very bright, or
>SMD types requiring heat sinks.

Yep. You're looking at the wrong stuff. Maybe something by Cree:
<http://www.cree.com/LED-Components-and-Modules/Products>
Nope. Ok, I give up. What the photo looks like is a common dome
shaped LED (as in the Cree URL above) with a lens over it. I did some
Googling looking for the lens and couldn't find it.

>It’s a great little light, and I’ve not found anything as small,
>long-lasting, with single AA that I like as much. And this is a learning
>experience, so there’s that.

I have one that meets your requirements. L3 Illumination L10:
<http://www.candlepowerforums.com/vb/showthread.php?376058-L3-Illumination-L10-%28XP-G2-R5-or-Nichia-219-1xAA%29-Review-RUNTIMES-BEAMS-VIDEO>
Single cell, very bright, tolerably priced, small, adjustable
brightness, etc. However, there's a catch. It has 4 brightness
levels set by twisting the two sections of the body. No problem
except that it makes it a two handed operation. I would have
preferred a push button at the back end. I've also measured the
brightness (using my highly creative and non-standard procedure) at
about 90-110 lumens (varies with temperature).

>Any pointers to a suitable replacement LED would be appreciated.

Again, I don't think it's the LED. The driver board is a more likely
culprit. Put it under a magnifier and see if you can find the broken
trace. Put a light behind the PCB to make it easier to see the break.
If desperate, trace out the schematic and try to identify the LED
controller chip. If it uses an MLCC capacitor, use a hot air gun to
reflow, not a soldering iron tip.

You might also get some help in CandlePowerForums:
<http://www.candlepowerforums.com>

Good luck.

--
Jeff Liebermann je...@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

[Phil Allison]

Jeff Liebermann wrote:

>
> I beg to differ with your diagnosis. Unless the LED has magically
> turned into a thermostat controlled flasher, it's not going to do
> that.
>

** The OP is probably correct and I have seen the same symptoms with 3mm red LEDs. Over time, excess DC current damages the LED chips and they intermittently drop light output and blink on and off.

Caused havoc in a Mesa Boogie amplifier when most of the dropping resistors for the 20 odd LED/CdS cell opto-couplers were the wrong value, by a factor of 10, resulting in almost 100mA of drive current.


.... Phil

[DaveC]

> I beg to differ with your diagnosis. Unless the LED has magically
> turned into a thermostat controlled flasher, it's not going to do
> that. More likely, the LED driver circuit is doing the flashing.
> Based on zero detail about the actual flashlight, my wild guess(tm)
> would be the big electrolytic that usually goes across the battery.
> Broken or badly soldered connections on the driver PCB are also
> likely.

Nebo NU15J:

https://www.nebotools.com/prod_details.php?id=31&cid=16&subid=94&subsubcid=

I like this model for its good trade-off between brightness and battery life.
Might be “better”, but this fits my needs. (And it ft fits in my pocket
along with my micro Swiss army knife...)

No controller chip. Looks like a boost converter (inductor, BJT, schottky
diode, ceramic cap).

I scoped the voltage across the LED: it's constant 4v (with a very
small--20mv?--rise and fall as it switches on and off). Haven’t yet
measured current. Surely if there was a bad solder joint or other failure it
would show up here.

Thanks for the referrals of other lights, but I’m not buying a replacement.
I’m doing this for fun and to learn.

Cheers.

[Syd Rumpo]

On 22/03/2016 03:51, Jeff Liebermann wrote:

<snip>

> I beg to differ with your diagnosis. Unless the LED has magically
> turned into a thermostat controlled flasher, it's not going to do
> that.

Thermostat controlled flasher is probably the exact description. It
happens when LEDs are driven straight from batteries too, and is
probably the bond wire heating and cooling making intermittent contact
at a few Hz.

I have a cheap LED torch which uses nine parallel connected white LEDs
running directly from 3 AAs. After a battery replacement, one is out,
four flash at different rates, and four are on.


Cheers
--
Syd

[Phil Allison]

Syd Rumpo wrote:

>
> Thermostat controlled flasher is probably the exact description. It
> happens when LEDs are driven straight from batteries too, and is
> probably the bond wire heating and cooling making intermittent contact
> at a few Hz.
>
> I have a cheap LED torch which uses nine parallel connected white LEDs
> running directly from 3 AAs. After a battery replacement, one is out,
> four flash at different rates, and four are on.
>
>

** Ahhh - so you have seen it really happen too !!!

Bit like sighting the Loch Ness Monster.

Bonding certainly seems to be involved and LEDs are very sensitive to high temps - way more than silicon semiconductors.

Gallium arsenide is even whimpier than Germanium, lifespan deteriorating quickly above room temp.


.... Phil

[N_Cook]

Desolder the original LED, carefully , as balance of probability it will
be fine.
Jumper in any old high power LED and see if that flashes too.

[DaveC]

Circuit:

http://imgur.com/YeELiHI

which is identical to the application note in the datasheet (scroll down to
the ME2108A diagrams):

http://img.ozdisan.com/ETicaret_Dosya/431770_1622138.pdf

(easily translated via Google, but a bit more confusing...)

Measured current at the battery terminals is 275 mA (due to meter insertion
loss it’s difficult to get an accurate current reading at the LED’s
terminals--the LED dims). Voltage across the LED is a pretty-constant 4v.
Estimating the efficiency of the converter at (roughly) 75 percent the LED is
using 825 mW.

I guess I’m looking for a 1W replacement?

How’s my math?

Thanks.

[Jeff Liebermann]

On Mon, 21 Mar 2016 22:56:51 -0700, DaveC <n...@home.cow> wrote:

>Nebo NU15J:
>https://www.nebotools.com/prod_details.php?id=31&cid=16&subid=94&subsubcid=

Looks like the flashing LED is a common problem:
<http://www.candlepowerforums.com/vb/showthread.php?417142-Replacement-LED>

>I like this model for its good trade-off between brightness and battery life.

Some day, someone will design a flashlight with an automatic PWM light
dimmer. Shine the light at something bright, and the flashlight runs
at full brightness. Shine it at something in the dark, and it goes to
fairly dim.

>No controller chip. Looks like a boost converter (inductor, BJT, schottky
>diode, ceramic cap).

Yep, that's about it. No controller.

>I scoped the voltage across the LED: it's constant 4v (with a very
>small--20mv?--rise and fall as it switches on and off). Haven’t yet
>measured current. Surely if there was a bad solder joint or other failure it
>would show up here.

Ok, I'm wrong. The flashing is not caused by the non-existent
controller chip.

>Thanks for the referrals of other lights, but I’m not buying a replacement.
>I’m doing this for fun and to learn.

Like I mumbled previously, see if there's a dome type LED under a
plastic lens. You'll probably need to unsolder the LED leads to do
this. If they really are two parts, it should be possible to find a
replacement LED.

If you can't identify the LED by the power output, try measuring the
current drain with a new battery. Multiply the current with your
measured 4V, and you should get the power in watts. That should tell
you if you should be looking for a 1/2, 1, or 1.5v LED.

[DaveC]

Jeff Liebermann sez:

> Looks like the flashing LED is a common problem:
> <http://www.candlepowerforums.com/vb/showthread.php?417142-Replacement-LED>

That’s my post on canblepowerforums.com. (c:

> Like I mumbled previously, see if there's a dome type LED under a plastic
lens. You'll probably need to unsolder the LED leads to do this. If they
really are two parts, it should be possible to find a replacement LED.

Will do.

> If you can't identify the LED by the power output, try measuring the
> current drain with a new battery. Multiply the current with your
> measured 4V, and you should get the power in watts. That should tell
> you if you should be looking for a 1/2, 1, or 1.5v LED.

You mean 1/2, 1, or 1.5 *watt* LED, yes?

Thanks.

[Jeff Liebermann]

On Tue, 22 Mar 2016 16:56:48 -0700, DaveC <n...@home.cow> wrote:

>Jeff Liebermann sez:
>
>> Looks like the flashing LED is a common problem:
>> <http://www.candlepowerforums.com/vb/showthread.php?417142-Replacement-LED>
>
>That’s my post on canblepowerforums.com. (c:

Oops(tm).

>> Like I mumbled previously, see if there's a dome type LED under a plastic
>> lens. You'll probably need to unsolder the LED leads to do this. If they
>> really are two parts, it should be possible to find a replacement LED.
>
>Will do.

>> If you can't identify the LED by the power output, try measuring the
>> current drain with a new battery. Multiply the current with your
>> measured 4V, and you should get the power in watts. That should tell
>> you if you should be looking for a 1/2, 1, or 1.5v LED.
>
>You mean 1/2, 1, or 1.5 *watt* LED, yes?

Oops 2.0(tm). I was in a rush to get out of door. Sorry. Also, I
just noticed that there is no such thing as an individual 1.5w LED.

[rickman]

Didn't you say it used a *single* 1.5 volt AA cell? I think that would
make it 412 milliwatts not counting the conversion efficiency, so more
likely a 1/2 watt LED.

--

Rick

[DaveC]

> Also, I just noticed that there is no such thing as an individual 1.5w LED.

I think this is an under-used(?) 1W LED.

If I can’t find a direct replacement I’m considering flipping
the PCB it’s soldered to and using the copper side to solder a SMD
unit. That way can dissipate some W.

But that presents its own set of problems. Optics will need to be matched to
the new LED.

Hmm...

[DaveC]

On 22 Mar 2016, rickman wrote
(in article <ncssb4$c6l$1...@dont-email.me>):

> Didn't you say it used a *single* 1.5 volt AA cell? I think that would
> make it 412 milliwatts not counting the conversion efficiency, so more
> likely a 1/2 watt LED.

1.5v boosted to 4v (measured) output from the converter. 275 mA (measured) at
the battery terminals. I make that to be 1100 mW. If presume 75 percent
efficiency, 825 mW.

No?

[rickman]

Perhaps I am missing something. The battery terminals will be at the
battery voltage, no? So why would you use 4 volts which is at the
output? It would be 4 volts times the LED current or the battery
voltage times the input current.

--

Rick

[DaveC]

> Perhaps I am missing something. The battery terminals will be at the
> battery voltage, no? So why would you use 4 volts which is at the
> output? It would be 4 volts times the LED current or the battery
> voltage times the input current.

Rick

Yikes! Of course you’re right.

Thanks.

[Ralph Mowery]

"rickman" <gnu...@gmail.com> wrote in message
news:nct5ah$v2h$1...@dont-email.me...

From what small ammount of playing with the voltage converters, they are
almost like a transformer in action. If you go to a higher voltage, the low
voltage current will be a lot more than is used by the load, and if going to
a lower voltage , the current from the battery will be less than the actual
current used by the load.
That keeps the total power drawn the same minus the efficency of the
converter.

[Jeff Liebermann]

On Tue, 22 Mar 2016 18:56:51 -0700, DaveC <n...@home.cow> wrote:

>> Also, I just noticed that there is no such thing as an individual 1.5w LED.
>
>I think this is an under-used(?) 1W LED.

Sounds about right. If you can light it up, just measure the battery
current drain, and calculate or estimate the power dissipation.

>If I can’t find a direct replacement I’m considering flipping
>the PCB it’s soldered to and using the copper side to solder a SMD

>unit. That way can dissipate some W.
>But that presents its own set of problems. Optics will need to be matched to
>the new LED.

If you can remove what I believe to be a lens, I think you'll find
that the actual LED is rather conventional and can be found in the
Cree catalog.
<http://www.cree.com/LED-Components-and-Modules/Products>
If you sort the above list by power output, there are only 3ea 1w
LED's listed. Just find the right die size and good luck soldering
the tiny chip. You can also dig throught the current flashlight
offerings and see which 1w chips are popular.

>Hmm...

Learn by Destroying.... then buy a new flashlight.

[DaveC]

On 22 Mar 2016, DaveC wrote
(in article<0001HW.1CA25FD30...@news.eternal-september.org>):

Just measured working light of same model: using power supply @1.5v (measured
at flashlight battery terminals), the current from the ps is 400 mA.
Estimating 75 percent efficiency that makes about 450 mW.

So a 1/2W replacement LED might be what I’m looking for.

Thanks.

[DaveC]

> If you can remove what I believe to be a lens, I think you'll find
> that the actual LED is rather conventional and can be found in the
> Cree catalog.
> <http://www.cree.com/LED-Components-and-Modules/Products>
> If you sort the above list by power output, there are only 3ea 1w
> LED's listed. Just find the right die size and good luck soldering
> the tiny chip. You can also dig throught the current flashlight
> offerings and see which 1w chips are popular.

[Jeff Lieberman]

I found these:

https://www.led-tech.de/en/0.5W-Power-LEDs_DB-86.pdf

Looks good, no?

[Jeff Liebermann]

On Wed, 23 Mar 2016 21:50:18 -0700, DaveC <n...@home.cow> wrote:

>> If you can remove what I believe to be a lens, I think you'll find
>> that the actual LED is rather conventional and can be found in the
>> Cree catalog.
>> <http://www.cree.com/LED-Components-and-Modules/Products>
>> If you sort the above list by power output, there are only 3ea 1w
>> LED's listed. Just find the right die size and good luck soldering
>> the tiny chip. You can also dig throught the current flashlight
>> offerings and see which 1w chips are popular.
>[Jeff Lieberman]

Please deduct one point for spelling my name wrong.

>I found these:
>https://www.led-tech.de/en/0.5W-Power-LEDs_DB-86.pdf
>Looks good, no?

No. I thought you said your flashlight was bright and possibly 1
watt. Those are common 5mm LED's with 0.5 watts maximum.
<https://www.google.com/search?q=5mm+white+led+0.5+watts>
These photos look very much like what I would expect to be hiding
under what I believe to be a lens of some sorts:
<https://www.google.com/search?q=5mm+white+led+0.5+watts&tbm=isch>
This looks a bit closer:
<http://www.ebay.com/itm/161246348498>

That data sheet is also slightly insane, specifying the luminous flux
as:
Lumen typ.: 19 mcd
Lumens are measured in umm.... lumens, not millicandelas. It is
possible to convert between lumens and mcd's using the viewing angle:
<http://www.rapidtables.com/calc/light/mcd-to-lumen-calculator.htm>
Anyway, 19 lumens is not very bright but probably good enough for a
pocket flashlight.

[DaveC]

> > [Jeff Lieberman]
>
> Please deduct one point for spelling my name wrong.

And I tried so hard to get the “Lei/Lie” bit right! (c; 2 steps forward,
1 step back...

> These photos look very much like what I would expect to be hiding
> under what I believe to be a lens of some sorts:
> <https://www.google.com/search?q=5mm+white+led+0.5+watts&tbm=isch>

One of those images (located here):

http://tinyurl.com/j4pwvuw

looks exactly like the one in my light. It’s squat, with very (relatively)
large yellow die(?) in the center. I don’t think there’s a separate lens.
Maybe it’s molded to focus the light (integral lens)?
Wish I could find that LED other than @ Alibaba...

Ah-HAH! From that Alibaba image page: “strawhat LED”. Search turns up
similar-looking LEDs. Searching on the terms “strawhat” and “dimple”
tells me that this form-factor is for radial distribution of the light, not
throwing a beam. Which means the flashlight designer wanted the reflector to
shape the beam, not the LED.

So I’m looking for one of these types.

Thanks.

(Why are my search skills so crap? Google keeps returning Manga images and
cartoon faces...)

[Computer Nerd Kev]

In sci.electronics.components Jeff Liebermann <je...@cruzio.com> wrote:
> On Mon, 21 Mar 2016 22:56:51 -0700, DaveC <n...@home.cow> wrote:
>
>>I like this model for its good trade-off between brightness and battery life.
>
> Some day, someone will design a flashlight with an automatic PWM light
> dimmer. Shine the light at something bright, and the flashlight runs
> at full brightness. Shine it at something in the dark, and it goes to
> fairly dim.

Hmm, that sounds like a good idea to put in my planned LED torch project,
which was going to have a manual dimmer anyway. I guess an LDR would be
suitable? Maybe mount it in a small tube to make it directional?

I guess I'll have to find an LDR and wave it around with my multimeter
attached. Ahh, the dignity of research. :)

--
__ __
#_ < |\| |< _#

[Andre Majorel]

On 2016-03-22, Jeff Liebermann <je...@cruzio.com> wrote:

> If it uses an MLCC capacitor, use a hot air gun to reflow, not
> a soldering iron tip.

News to me. In what way are multi-layer ceramic caps and
soldering irons incompatible ?

--
André Majorel http://www.teaser.fr/~amajorel/
J'ai des vrais problèmes, vous avez des faux problèmes.

[John Larkin]

On Sat, 26 Mar 2016 19:56:12 +0000 (UTC), Andre Majorel
<che...@halliburton.com> wrote:

>On 2016-03-22, Jeff Liebermann <je...@cruzio.com> wrote:
>
>> If it uses an MLCC capacitor, use a hot air gun to reflow, not
>> a soldering iron tip.
>
>News to me. In what way are multi-layer ceramic caps and
>soldering irons incompatible ?

Not a bit.

[Jon Elson]

Andre Majorel wrote:

> On 2016-03-22, Jeff Liebermann <je...@cruzio.com> wrote:
>
>> If it uses an MLCC capacitor, use a hot air gun to reflow, not
>> a soldering iron tip.
>
> News to me. In what way are multi-layer ceramic caps and
> soldering irons incompatible ?
>

Really, before I got my pick and place machine, I hand-soldered about 25,000
.1uF 0805 capacitors. NEVER ONCE had a bad one. I still hand-solder a fair
number of low production boards and prototypes, and have never seen a
problem with MLCCs.

Jon

[Jeff Liebermann]

On Sat, 26 Mar 2016 19:56:12 +0000 (UTC), Andre Majorel
<che...@halliburton.com> wrote:

Sorry for the delay but I missed the followups to my comment.

>On 2016-03-22, Jeff Liebermann <je...@cruzio.com> wrote:
>
>> If it uses an MLCC capacitor, use a hot air gun to reflow, not
>> a soldering iron tip.
>
>News to me. In what way are multi-layer ceramic caps and
>soldering irons incompatible ?

Thermal shock easily cracks MLCC caps. I learned that the hard way
while fixing several Apple Mac Mini computahs, which feature a
collection of MLCC on the bottom of the main board.
<https://web.archive.org/web/20130606062903/http://blog.helpmymac.ru/?p=3585>
The original failure mode was shorted MLCC caps caused by either
thermal shock or board flex. The bad ones were easy to find with an
ESR meter. However, when I tried to install replacements (and
guessing the part value because Apple doesn't supply service
information to non-authorized repair shops), I managed to crack and
short several known good MLCC caps with a soldering iron. Having
learned the lesson, I used some solder paste and a hot air SMT reflow
gun to do the soldering. I also pre-heated the PCB and let the caps
cool down slowly. I don't know if that was necessary, but it worked
every time. I'm told that two solding irons used as a tweezer also
works, but I haven't tried that yet.

You'll find some more details under:
<https://www.google.com/#q=mlcc+capacitor+crack>
<https://www.google.com/#q=mlcc+capacitor+soldering+iron>
Lots of articles and guidelines on handling and soldering these caps,
some of which warn about using a soldering iron. For example:

<http://www.vishay.com/docs/45034/soldrec.pdf>
6. Soldering with a Solder Iron
Attachment by soldering iron is not recommended. A
heat shock may cause a crack in the MLCC chip capacitors,
however, if solder iron is used, the following precautions
should be taken: ... (etc)

Damage Prevention When Soldering Ceramic Chip Capacitors
<http://www.eptac.com/webinars/presentations/eptac_09_17_14.pdf>
Hand Soldering
- A pencil type soldering of 30 watts
maximum and with a diameter of 3 mm
maximum should be used.
- The soldering iron tip temperature should
be less than 300C [572F] and maximum
contact time should be 5 seconds.
- The soldering iron tip should never come in
contact with the component body.
Ever try to solder a small MLCC cap without touching the body with the
soldering iron tip? Good luck.

<http://www.murata.com/en-us/support/faqs/products/capacitor/mlcc/mnt/0001>
In order to prevent damage (cracks) to the component that
can be caused by localized rapid heating and heat shock, preheat
the chip, for example, to prevent it from being subjected to
heat shock.

[Jon Elson]

Jeff Liebermann wrote:


> The original failure mode was shorted MLCC caps caused by either
> thermal shock or board flex. The bad ones were easy to find with an
> ESR meter. However, when I tried to install replacements (and
> guessing the part value because Apple doesn't supply service
> information to non-authorized repair shops), I managed to crack and
> short several known good MLCC caps with a soldering iron. Having
> learned the lesson, I used some solder paste and a hot air SMT reflow
> gun to do the soldering. I also pre-heated the PCB and let the caps
> cool down slowly. I don't know if that was necessary, but it worked
> every time. I'm told that two solding irons used as a tweezer also
> works, but I haven't tried that yet.
>

I've heard these stories a number of times. And, yes, maybe some people use
insanely hot irons or in some other way cause this problem. And, some
really large caps are prone to this damage. But, as I say, I have hand-
soldered over 25,000 0805 MLCC caps of modest value without seeing this
problem. I use a very good Weller temperature-controlled iron, and run it
at a modest temperature. Much better to use an iron with really good
thermal conductivity at a lower temperature than one with poor conductivity
at a very high temperature.

Jon

[Jeff Liebermann]

On Mon, 28 Mar 2016 22:15:42 -0500, Jon Elson <el...@pico-systems.com>
wrote:

What does this suppose to mean, do, or prove?
Newsgroups: Choose,an,appropriate,group,to,redirect,replies...
If you want to black hole replies, just redirect to /dev/null

Newsgroup changed to: sci.electronics.design,sci.electronics.repair

>Jeff Liebermann wrote:
>> The original failure mode was shorted MLCC caps caused by either
>> thermal shock or board flex. The bad ones were easy to find with an
>> ESR meter. However, when I tried to install replacements (and
>> guessing the part value because Apple doesn't supply service
>> information to non-authorized repair shops), I managed to crack and
>> short several known good MLCC caps with a soldering iron. Having
>> learned the lesson, I used some solder paste and a hot air SMT reflow
>> gun to do the soldering. I also pre-heated the PCB and let the caps
>> cool down slowly. I don't know if that was necessary, but it worked
>> every time. I'm told that two solding irons used as a tweezer also
>> works, but I haven't tried that yet.

>I've heard these stories a number of times. And, yes, maybe some people use
>insanely hot irons or in some other way cause this problem. And, some
>really large caps are prone to this damage.

Yep. As James Arthur mentioned, it's not a problem with the
physically small caps, which come to a uniform thermal equilibrium
rather rapidly. It's only the big caps, with lots of plates, thin low
voltage ceramic dielectric, and larger thermal mass, that were a
problem for me.

>But, as I say, I have hand-
>soldered over 25,000 0805 MLCC caps of modest value without seeing this
>problem.

One exception will break any such rule. Dig out a larger and higher
capacitance MLCC cap in a 1210 or larger package. Try soldering it
with a single soldering iron. Unless you're very good, it won't take
much to trash the cap.

<http://www.avx.com/products/ceramic-capacitors/high-voltage/high-voltage-mlc-chips/>
"Chip sizes 1210 and larger should be reflow soldered only."

>I use a very good Weller temperature-controlled iron, and run it
>at a modest temperature.

I use several ancient Weller WTCP TC201 and TC202 (because they were
free) soldering stations. Most of my tips are 750F with a few at
850F.

>Much better to use an iron with really good
>thermal conductivity at a lower temperature than one with poor conductivity
>at a very high temperature.

I don't understand. All the tips are made of plated iron. There's
little difference in thermal conductivity between tips, unless you
want to throw in copper tips heated with a gas burner. What part of
the soldering iron varies in thermal conductivity?

[Michael A. Terrell]

Jeff Liebermann wrote:
>
> On Sat, 26 Mar 2016 19:56:12 +0000 (UTC), Andre Majorel
> <che...@halliburton.com> wrote:
>
> Sorry for the delay but I missed the followups to my comment.
>
> >On 2016-03-22, Jeff Liebermann <je...@cruzio.com> wrote:
> >
> >> If it uses an MLCC capacitor, use a hot air gun to reflow, not
> >> a soldering iron tip.
> >
> >News to me. In what way are multi-layer ceramic caps and
> >soldering irons incompatible ?
>
> Thermal shock easily cracks MLCC caps. I learned that the hard way
> while fixing several Apple Mac Mini computahs, which feature a
> collection of MLCC on the bottom of the main board.
> <https://web.archive.org/web/20130606062903/http://blog.helpmymac.ru/?p=3585>
> The original failure mode was shorted MLCC caps caused by either
> thermal shock or board flex. The bad ones were easy to find with an
> ESR meter. However, when I tried to install replacements (and
> guessing the part value because Apple doesn't supply service
> information to non-authorized repair shops), I managed to crack and
> short several known good MLCC caps with a soldering iron. Having
> learned the lesson, I used some solder paste and a hot air SMT reflow
> gun to do the soldering. I also pre-heated the PCB and let the caps
> cool down slowly. I don't know if that was necessary, but it worked
> every time. I'm told that two solding irons used as a tweezer also
> works, but I haven't tried that yet.

How were the new caps stored? If it is where they can adsorb
moisture, you can damage them with an iron that is too hot.

[Andre Majorel]

On 2016-03-29, Jeff Liebermann <je...@cruzio.com> wrote:

> Sorry for the delay but I missed the followups to my comment.

Not at all. That was very interesting. Thanks Jeff and everyone.

[Jeff Liebermann]

Well, I must admit that I didn't take any special precautions.
Methinks that the relative humidity in my office runs between 40% and
60% but is not monitored or recorded. Occasionally, it gets low
enough to where static electricity becomes a problem, or high enough
to where I'm rather uncomfortable, but those are rare. The caps a mix
of cut tape and loose bags stored in Ziploc bags (mostly pink
anti-static) and in paper coin envelopes. Nothing in hard plastic or
metal drawers that might chip or crack them. Although I know that
these caps make tolerable hydrometers (and microphones), I don't think
they can absorb enough moisture from the air to where a steam
explosion would be a problem.

I did some digging to see if humidity might be a problem in storage
conditions. There were plenty of notes on how a cracked capacitor
might allow water to enter the dielectric. Soft (solder) termination
is the recommended fix. Some suggests pre-heating the capacitors
before soldering to drive off any moisture. One demands that the caps
be used within 12 months. I didn't see humidity as being a problem
until AFTER the capacitors had cracked. It would take some time for
the moisture to alter the capacitor characteristics. With my hand
soldering technique, I managed to instantly produce shorted
capacitors, which methinks was more likely due to uneven thermal
expansion, than to moisture incursion.

[krw]

On Thu, 31 Mar 2016 04:20:59 -0400, "Michael A. Terrell"
<mike.t...@earthlink.net> wrote:

>

I haven't had any problems with cracking but I used to have problems
with end caps falling off. I haven't seen the issue for some time,
though perhaps it was a problem with the manufacturer. Our purchasing
group prefers Murata, so that's what I use (GRM series).

[Michael A. Terrell]

krw wrote:

>
> Michael A. Terrell wrote:
> >
> > How were the new caps stored? If it is where they can adsorb
> >moisture, you can damage them with an iron that is too hot.
>
> I haven't had any problems with cracking but I used to have problems
> with end caps falling off. I haven't seen the issue for some time,
> though perhaps it was a problem with the manufacturer. Our purchasing
> group prefers Murata, so that's what I use (GRM series).

We pre baked some boards and other components before assembly and
reflow. That eliminated cracked multilayer SMD capacitors, tombstoning
of two lead components and losing end caps. That was in N Central
Florida which has plenty of humidity problems. The so called HVAC
'engineers' were clueless about how to control the humidity, so we had
to resort to baking. LSI SMD ICs were backed and heat sealed into
antistatic bags with moisture adsorbing packets.

[krw]

Sure, baking humidity sensitive parts is common before pick-n-place.
It's a bit less common when hand (de)soldering, though. ;-)

[Michael A. Terrell]

We had a high failure rate on hand soldered FIR chips, before we
started baking them. The bottom of the packaging was the thinnest, and
it would bow out as it released steam during hand soldering. No one in
EE or ME believed me, until I finally got them to try it for themselves.