[EE] Unknown LED current regulator

[John Coppens]

Hello all.

I've been really surprised at the variations of the circuitry inside
LED bulb replacements. A while ago, I bought a couple of bulbs ('Great
Value'-brand, a local Walmart product), which started showing failing
after just a couple of weeks. Opening them up showed that they were
fed by just a (linear) series regulator. So much for green energy.

Now I acquired a couple of OSRAM bulbs, and opened one up just to
investigate. It did have a switching current regulator, but it used
a 4-pin chip for which I could not find any data. In fact, I didn't
find _any_ 4-pin current regulator IC.

I've attached the circuit. I suspect the 1.8 Ohm resistor to be the
current-determining component. Strange is the (I suspect) transformer
which is actually across the LEDs (I'm not positive about it's
internal connections. I measure 10 Ohm across two pins, 14k across
two other pins, but without unsoldering). 12 LEDs (I suspect triplets)
are connected between + and -.

Does anyone have an idea which IC this could be? It's marked 9938F
and S1C01X. No logo as far as I can detect.

[Jean-Paul Louis]

Look for ZFS1009 from diode inc. They use only 4 connections and two resistors 
Just a thought,Jean-Paul N1JPL

Sent from Yahoo Mail on Android

On Thu, Nov 1, 2018 at 10:00 AM, John Coppens<***@jcoppens.com> wrote: Hello all.

I've been really surprised at the variations of the circuitry inside
LED bulb replacements. A while ago, I bought a couple of bulbs ('Great
Value'-brand, a local Walmart product), which started showing failing
after just a couple of weeks. Opening them up showed that they were
fed by just a (linear) series regulator. So much for green energy.

Now I acquired a couple of OSRAM bulbs, and opened one up just to
investigate. It did have a switching current regulator, but it used
a 4-pin chip for which I could not find any data. In fact, I didn't
find _any_ 4-pin current regulator IC.

I've attached the circuit. I suspect the 1.8 Ohm resistor to be the
current-determining component. Strange is the (I suspect) transformer
which is actually across the LEDs (I'm not positive about it's
internal connections. I measure 10 Ohm across two pins, 14k across
two other pins, but without unsoldering). 12 LEDs (I suspect triplets)
are connected between + and -.

Does anyone have an idea which IC this could be? It's marked 9938F
and S1C01X. No logo as far as I can detect.

[Harrison Cooper]

It would not surprise me if this was a custom ASIC, OsRAM is a huge company, and already doing ASIC's for other products, and if they are using a older technology...say 56nm for this, this chip would be pennies.

-----Original Message-----
From: piclist-***@mit.edu <piclist-***@mit.edu> On Behalf Of John Coppens
Sent: Thursday, November 01, 2018 7:58 AM
To: ***@mit.edu
Subject: [EE] Unknown LED current regulator

Hello all.

I've been really surprised at the variations of the circuitry inside LED bulb replacements. A while ago, I bought a couple of bulbs ('Great Value'-brand, a local Walmart product), which started showing failing after just a couple of weeks. Opening them up showed that they were fed by just a (linear) series regulator. So much for green energy.

Now I acquired a couple of OSRAM bulbs, and opened one up just to investigate. It did have a switching current regulator, but it used a 4-pin chip for which I could not find any data. In fact, I didn't find _any_ 4-pin current regulator IC.

I've attached the circuit. I suspect the 1.8 Ohm resistor to be the current-determining component. Strange is the (I suspect) transformer which is actually across the LEDs (I'm not positive about it's internal connections. I measure 10 Ohm across two pins, 14k across two other pins, but without unsoldering). 12 LEDs (I suspect triplets) are connected between + and -.

Does anyone have an idea which IC this could be? It's marked 9938F and S1C01X. No logo as far as I can detect.

[Sean Breheny]

56nm is an "older technology"? Sure, it is not even close to the latest for
high-end processors, but it was the state-of-the-art in 2006 and I would
think it would be way overkill for small mostly analog ASICs.

Post by Harrison Cooper
It would not surprise me if this was a custom ASIC, OsRAM is a huge
company, and already doing ASIC's for other products, and if they are using
a older technology...say 56nm for this, this chip would be pennies.
-----Original Message-----
Sent: Thursday, November 01, 2018 7:58 AM
Subject: [EE] Unknown LED current regulator
Hello all.
I've been really surprised at the variations of the circuitry inside LED
bulb replacements. A while ago, I bought a couple of bulbs ('Great
Value'-brand, a local Walmart product), which started showing failing after
just a couple of weeks. Opening them up showed that they were fed by just a
(linear) series regulator. So much for green energy.
Now I acquired a couple of OSRAM bulbs, and opened one up just to
investigate. It did have a switching current regulator, but it used a 4-pin
chip for which I could not find any data. In fact, I didn't find _any_
4-pin current regulator IC.
I've attached the circuit. I suspect the 1.8 Ohm resistor to be the
current-determining component. Strange is the (I suspect) transformer which
is actually across the LEDs (I'm not positive about it's internal
connections. I measure 10 Ohm across two pins, 14k across two other pins,
but without unsoldering). 12 LEDs (I suspect triplets) are connected
between + and -.
Does anyone have an idea which IC this could be? It's marked 9938F and
S1C01X. No logo as far as I can detect.
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Harrison Cooper]

The problem is many fabs (unless its an inhouse fab) are moving to the smaller geometries, so even tho its overkill (and I do not argue on that) you need someone to be able build it 😊

But John did find the answer. Its actually amazing how many devices are out there, with Asian origins, that we as "western" design engineers have very little insight into.



-----Original Message-----
From: piclist-***@mit.edu <piclist-***@mit.edu> On Behalf Of Sean Breheny
Sent: Thursday, November 01, 2018 2:07 PM
To: Microcontroller discussion list - Public. <***@mit.edu>
Subject: Re: [EE] Unknown LED current regulator

56nm is an "older technology"? Sure, it is not even close to the latest for high-end processors, but it was the state-of-the-art in 2006 and I would think it would be way overkill for small mostly analog ASICs.

Post by Harrison Cooper
It would not surprise me if this was a custom ASIC, OsRAM is a huge
company, and already doing ASIC's for other products, and if they are
using a older technology...say 56nm for this, this chip would be pennies.
-----Original Message-----
Sent: Thursday, November 01, 2018 7:58 AM
Subject: [EE] Unknown LED current regulator
Hello all.
I've been really surprised at the variations of the circuitry inside
LED bulb replacements. A while ago, I bought a couple of bulbs ('Great
Value'-brand, a local Walmart product), which started showing failing
after just a couple of weeks. Opening them up showed that they were
fed by just a
(linear) series regulator. So much for green energy.
Now I acquired a couple of OSRAM bulbs, and opened one up just to
investigate. It did have a switching current regulator, but it used a
4-pin chip for which I could not find any data. In fact, I didn't find
_any_ 4-pin current regulator IC.
I've attached the circuit. I suspect the 1.8 Ohm resistor to be the
current-determining component. Strange is the (I suspect) transformer
which is actually across the LEDs (I'm not positive about it's
internal connections. I measure 10 Ohm across two pins, 14k across two
other pins, but without unsoldering). 12 LEDs (I suspect triplets) are
connected between + and -.
Does anyone have an idea which IC this could be? It's marked 9938F and
S1C01X. No logo as far as I can detect.
--
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

[Denny Esterline]

Not to completely hijack the thread, but speaking of unknown Asian parts,
has anybody been following the adventure of the three-cent microcontroller?

Post by Harrison Cooper
The problem is many fabs (unless its an inhouse fab) are moving to the
smaller geometries, so even tho its overkill (and I do not argue on that)
you need someone to be able build it 😊
But John did find the answer. Its actually amazing how many devices are
out there, with Asian origins, that we as "western" design engineers have
very little insight into.
-----Original Message-----
Sent: Thursday, November 01, 2018 2:07 PM
Subject: Re: [EE] Unknown LED current regulator
56nm is an "older technology"? Sure, it is not even close to the latest
for high-end processors, but it was the state-of-the-art in 2006 and I
would think it would be way overkill for small mostly analog ASICs.

Post by Harrison Cooper
It would not surprise me if this was a custom ASIC, OsRAM is a huge
company, and already doing ASIC's for other products, and if they are
using a older technology...say 56nm for this, this chip would be pennies.
-----Original Message-----
Sent: Thursday, November 01, 2018 7:58 AM
Subject: [EE] Unknown LED current regulator
Hello all.
I've been really surprised at the variations of the circuitry inside
LED bulb replacements. A while ago, I bought a couple of bulbs ('Great
Value'-brand, a local Walmart product), which started showing failing
after just a couple of weeks. Opening them up showed that they were
fed by just a
(linear) series regulator. So much for green energy.
Now I acquired a couple of OSRAM bulbs, and opened one up just to
investigate. It did have a switching current regulator, but it used a
4-pin chip for which I could not find any data. In fact, I didn't find
_any_ 4-pin current regulator IC.
I've attached the circuit. I suspect the 1.8 Ohm resistor to be the
current-determining component. Strange is the (I suspect) transformer
which is actually across the LEDs (I'm not positive about it's
internal connections. I measure 10 Ohm across two pins, 14k across two
other pins, but without unsoldering). 12 LEDs (I suspect triplets) are
connected between + and -.
Does anyone have an idea which IC this could be? It's marked 9938F and
S1C01X. No logo as far as I can detect.
--
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

[RussellMc]

Post by Denny Esterline
Not to completely hijack the thread, but speaking of unknown Asian parts,
has anybody been following the adventure of the three-cent microcontroller?
http://youtu.be/Rixo78hv_lw
Padauk PMS171B

Datasheet
http://www.padauk.com.tw/upload/doc/PMS171B_datasheet_EN_v000_20181011.pdf

IDE user manual (Chinese only)
http://www.padauk.com.tw/upload/doc/IDE%20user%20guide_v1.01%20(CN).pdf

Package marking information (English only !)
http://www.padauk.com.tw/upload/doc/PADAUK_Package%20Information_V004_20171103.pdf

MCU selector guide http://www.padauk.com.tw/en/product/method.aspx?num=1

...

Russell

[s***@interlog.com]

Post by RussellMc

Post by Denny Esterline
Not to completely hijack the thread, but speaking of unknown Asian parts,
has anybody been following the adventure of the three-cent microcontroller?
http://youtu.be/Rixo78hv_lw
Padauk PMS171B

Datasheet
http://www.padauk.com.tw/upload/doc/PMS171B_datasheet_EN_v000_20181011.pdf
IDE user manual (Chinese only)
http://www.padauk.com.tw/upload/doc/IDE%20user%20guide_v1.01%20(CN).pdf
Package marking information (English only !)
http://www.padauk.com.tw/upload/doc/PADAUK_Package%20Information_V004_20171103.pdf
MCU selector guide http://www.padauk.com.tw/en/product/method.aspx?num=1
...
Russell

I vaguely considered buying 1000, an emulator and a programmer last time I
was in China (Oct), but found the equivalent (small) stack of Y100
notes more compelling.

For my current pursuits, there are some more interesting oddball ones
with good ADCs etc.


--Spehro Pefhany

[s***@interlog.com]

Quoting John Coppens <***@jcoppens.com>:


The chip is made in China by Bright Power (Shanghai).

Loading Image...


I doubt there's much point in them exporting the chips to the West.
China and SE Asia is pretty much the world market. 1 reel is 15,000 pieces.


--sp

Post by John Coppens
Hello all.
I've been really surprised at the variations of the circuitry inside
LED bulb replacements. A while ago, I bought a couple of bulbs ('Great
Value'-brand, a local Walmart product), which started showing failing
after just a couple of weeks. Opening them up showed that they were
fed by just a (linear) series regulator. So much for green energy.
Now I acquired a couple of OSRAM bulbs, and opened one up just to
investigate. It did have a switching current regulator, but it used
a 4-pin chip for which I could not find any data. In fact, I didn't
find _any_ 4-pin current regulator IC.
I've attached the circuit. I suspect the 1.8 Ohm resistor to be the
current-determining component. Strange is the (I suspect) transformer
which is actually across the LEDs (I'm not positive about it's
internal connections. I measure 10 Ohm across two pins, 14k across
two other pins, but without unsoldering). 12 LEDs (I suspect triplets)
are connected between + and -.
Does anyone have an idea which IC this could be? It's marked 9938F
and S1C01X. No logo as far as I can detect.

[John Coppens]

On Thu, 01 Nov 2018 17:11:16 -0400

Post by s***@interlog.com
The chip is made in China by Bright Power (Shanghai).
https://i.imgur.com/PC2pGzD.png

Hi sp.

Thanks for the hint. I downloaded the datasheet (very minimal!)
The circuit is slightly different because of the extra lowpass.

I suspect there is no second winding on the output coil, and
the 14k I measured is likely leakage current in the capacitor.

This is the simplest switcher I ever came across. Even the fact
it gets it power directly from the 300+ Volt line is amazing.

Again, much appreciated... And somewhat surprised and disappointed
Google didn't lead me to it.

BTW, an ad on the page at datasheetspdf.com, offered them in lots
of 25 at $0.1335/ea.

John

[s***@interlog.com]

Post by John Coppens
On Thu, 01 Nov 2018 17:11:16 -0400

Post by s***@interlog.com
The chip is made in China by Bright Power (Shanghai).
https://i.imgur.com/PC2pGzD.png

Hi sp.

I really should get around to adding a signature on this webmail nonsense.

Post by John Coppens
Thanks for the hint. I downloaded the datasheet (very minimal!)
The circuit is slightly different because of the extra lowpass.
I suspect there is no second winding on the output coil, and
the 14k I measured is likely leakage current in the capacitor.
This is the simplest switcher I ever came across. Even the fact
it gets it power directly from the 300+ Volt line is amazing.

There are some in TO-92 package! The bulbs have to be made very cheaply.

Post by John Coppens
Again, much appreciated... And somewhat surprised and disappointed
Google didn't lead me to it.

You're welcome.

Post by John Coppens
BTW, an ad on the page at datasheetspdf.com, offered them in lots
of 25 at $0.1335/ea.
John

Judging by other similar products, I suspect the real domestic China price
is a few cents in smallish quantity and more like a penny or two (negotiated
and secret) in actual production volume. Domestic courier package
shipping is maybe Y8-10 ($1 or $1.50).

--Spehro Pefhany

[RussellMc]

Post by John Coppens
...
Does anyone have an idea which IC this could be? It's marked 9938F
and S1C01X. No logo as far as I can detect.
I looked at the circuit and decided that they had been extravagant with

the pinout - you only need 3 leads for that functionality :-).
An eg SOT23 or smaller may suffice.

And then I found these:

3 leads (albeit 8 pin package)

http://www.fsjingfeng.com/product/show_product.php?id=443
http://www.fsjingfeng.com/product/show_product.php?id=454

4 leads

https://cn.made-in-china.com/tupian/didiwei0601-cvqJweuKCEVm.html

5 leads - fully isolated


https://cn.made-in-china.com/tupian/fszjy8888-hXvmKWulXxkD.html

external FET, current sense external


https://cn.made-in-china.com/tupian/didiwei0601-pveEdCgwAmcR.html

LEGION

https://cn.made-in-china.com/photo/hengliuqudongxinpian-1.html
\
_________________

"S9306/S9308 core flying non-isolated dual color temperature S4223 dual
chip combination version of the periphery is simple and low cost, excellent
quality"

https://cn.made-in-china.com/tupian/didiwei0601-YBMmwigzJnhA.html

Note links & prices at bottom of page!

And on and on and ...



R

[Richard Prosser]

You could review some of the "Big Clive" videos on youtube. He's taken
apart just about every LED lamp known & discusses the circuit, operation
and failings etc. In most cases he's drawn the schematic & seems to have a
good source of datasheets for the ICs.

RP

Post by John Coppens
Hello all.
I've been really surprised at the variations of the circuitry inside
LED bulb replacements. A while ago, I bought a couple of bulbs ('Great
Value'-brand, a local Walmart product), which started showing failing
after just a couple of weeks. Opening them up showed that they were
fed by just a (linear) series regulator. So much for green energy.
Now I acquired a couple of OSRAM bulbs, and opened one up just to
investigate. It did have a switching current regulator, but it used
a 4-pin chip for which I could not find any data. In fact, I didn't
find _any_ 4-pin current regulator IC.
I've attached the circuit. I suspect the 1.8 Ohm resistor to be the
current-determining component. Strange is the (I suspect) transformer
which is actually across the LEDs (I'm not positive about it's
internal connections. I measure 10 Ohm across two pins, 14k across
two other pins, but without unsoldering). 12 LEDs (I suspect triplets)
are connected between + and -.
Does anyone have an idea which IC this could be? It's marked 9938F
and S1C01X. No logo as far as I can detect.
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist