Author Topic: Optoelectronics - die pictures  (Read 17652 times)

0 Members and 1 Guest are viewing this topic.

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #25 on: August 01, 2021, 11:39:51 am »


L133C, a 1x1024 CCD sensor built by "Werk für Fernsehelektronik Berlin".
It´s quite similar to the Fairchild CCD113.




The package consists of two parts that are glued together.




The datasheet explains the construction of the L133C.
It consists of one optical line and two CCD lines that hold and transport the generated charges.
Around the CCD lines there are two lines that reduce noise. The upper one generates a end of line signal.




The die is 14,5mm x 1,5mm. The whole circuit is protected with a metal layer so the light doesn´t interfere with it.




This bondwire connects the frame with the upper metal layer. The frame is connected to the negative supply. It seems like they wanted to have an option to connect the metal layer to a different potential.
It seems there was a first failed bond event.








There are three areas with mask revision but one is different than the other two.  :-//




Some structures to adjust the masks and monitor the process quality.




I/O protection




Input stage






The CCD lines operate with only one clock (Vgt). Instead of a second clock it uses a mid-level voltage (Vt).
You can spot the transfergate (red) and the four CCD lines (green and blue). The CCD line for the optical data is wider so it is able to store more electrons.
Over each CCD line there are overlapping gate electrodes which makes the structure confusing.
Ve is the input potential of the CCD lines.
There is a covered photogate area that gives the user a dark reference (purple). The dark reference is isolated from the rest of the photogate by four pixel (white). It seems at the beginning of the line there are four more isolation cells (pink).
The circuit gives you a withe reference at the end of the data stream. I´m not sure how that happens.




This structure modifies the photogate potential nb with respect to the clocks Vgt and Vt. Probably it supports the charge transfer from the photogate to the CCD lines.




The output stage.
Interesting: SHoutA, VshA, SHoutB and VshB are not connected on the die.  :o






At the end of the photogate there are four more isolation pixel and four more black reference pixel. It seems like there are 1,5 Pixel covered with metal that normally should not be covered.  :-/O
At the end of the CCD lines there are three output stages. The first part in such a stage is a reset transistor that conducts the charge to the positive supply before the next charge is delivered. The second part is a capacitor holding the charge. The third part is the output transistor of the CCD line. This output controls a bigger push-pull-stage driving the output of the L133C.
The uppermost CCD line is equipped with a similar output stage. The charges from the lower CCD line are conducted to the supply directly.








At the upper edge there is a circuit using the transport clock and generating the clocks necessary to control the output stage.
It seems like the bigger circuits are modifying the duty cycle and the smaller parts are adding some delay.


Some more pictures here:
https://www.richis-lab.de/Opto08.htm

 :-/O
 
The following users thanked this post: doktor pyta, RoGeorge, Miyuki, mawyatt, prisar

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3192
  • Country: us
Re: Optoelectronics - die pictures
« Reply #26 on: August 01, 2021, 01:52:42 pm »
Nice images and circuit/image analysis!!

Way back in the very early 80s we developed our own CCDs for a different use. Back then CCDs were mostly for imaging and some were used as delays, like the audio "reverb" use. Vaguely remember Tektronix using a very fast CCD in a scope to "capture" a high speed "event" then play it back at a slower speed for the ADC to digitize, high speed and resolution ADCs weren't available then.

Our use was different, the CCDs were used as complex R + jX convolvers in the Chirp Z Transform (and other proprietary DTCA transforms). The CZT was another method of Fourier Transforms, and the basis of our Real Time Spectrum Analyzer. CCDs are very tricky to get working properly, however when working they are remarkable devices.

Keep these nice old chip images coming, they are really fun to view and discuss, and bring back memories :-+

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #27 on: August 01, 2021, 02:46:12 pm »
Nice images and circuit/image analysis!!
[...]
Keep these nice old chip images coming, they are really fun to view and discuss, and bring back memories :-+

Thanks I still have some really interesting parts in stock. :)


Way back in the very early 80s we developed our own CCDs for a different use. Back then CCDs were mostly for imaging and some were used as delays, like the audio "reverb" use. Vaguely remember Tektronix using a very fast CCD in a scope to "capture" a high speed "event" then play it back at a slower speed for the ADC to digitize, high speed and resolution ADCs weren't available then.

Our use was different, the CCDs were used as complex R + jX convolvers in the Chirp Z Transform (and other proprietary DTCA transforms). The CZT was another method of Fourier Transforms, and the basis of our Real Time Spectrum Analyzer. CCDs are very tricky to get working properly, however when working they are remarkable devices.

I have done quite some (private) research on CCDs. I'm a big fan of the Gould 4074 oscilloscope. 1987 it was the fastest available real time sampling oscilloscope with 400MS/s (8Bit). The HP 54111D was able to do 1GS/s but just with reduced resolution (6Bit). 250MS/s with 8Bit.
The Gould 4074 uses a Plessey MS1007A CCD for each channel. It has eight lines each containing 128 points and working with 50MHz.
Gould integrated quite some magic to get the CCDs and the oscilloscope working as good as possible at 400MS/s. And they used mostly "normal" parts not much unobtanium.

I have heard CCDs were also used for signal processing in radar systems.
 
The following users thanked this post: edavid

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3192
  • Country: us
Re: Optoelectronics - die pictures
« Reply #28 on: August 01, 2021, 03:22:59 pm »

I have done quite some (private) research on CCDs. I'm a big fan of the Gould 4074 oscilloscope. 1987 it was the fastest available real time sampling oscilloscope with 400MS/s (8Bit). The HP 54111D was able to do 1GS/s but just with reduced resolution (6Bit). 250MS/s with 8Bit.
The Gould 4074 uses a Plessey MS1007A CCD for each channel. It has eight lines each containing 128 points and working with 50MHz.
Gould integrated quite some magic to get the CCDs and the oscilloscope working as good as possible at 400MS/s. And they used mostly "normal" parts not much unobtanium.


I was not aware of the Gould Scopes, thanks for the info :-+

Quote
I have heard CCDs were also used for signal processing in radar systems.

Yes, and anti-radar "stealth" applications ::)

Many additional signal processing applications as well.

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: Miyuki

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #29 on: August 01, 2021, 05:24:49 pm »
Up to now I have 16 of the Gould 40xx oscilloscopes in different shapes.  ;D

https://www.richis-lab.de/Gould407X.htm

Sometimes I hopefully will post my complete extreme teardown report. It has round about 400 pages but the proof reading takes soooo long only the first 60 pages are online... ...and it will be available in german only... ...but with a lot of pictures.  ::)

Offline Renate

  • Super Contributor
  • ***
  • Posts: 1460
  • Country: us
Re: Optoelectronics - die pictures
« Reply #30 on: August 01, 2021, 05:37:24 pm »
That's a pretty device.
I guess that they didn't like knobs that much!

Unrelated: Here's my half-baked re-write of Hantek software.
The ranges (voltage, timebase & everything else) can be operated with the mouse scroll.
 

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #31 on: August 01, 2021, 05:59:25 pm »
The only real drawback is the old CRT (3kV). The figures aren´t really sharp.  :-[


Unrelated: Here's my half-baked re-write of Hantek software.
The ranges (voltage, timebase & everything else) can be operated with the mouse scroll.

A nice update!  :-+

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #32 on: September 16, 2021, 01:17:38 pm »


L211C, a 190x244 CCD sensor built by the Werk für Fernsehelektronik. It is similar to the Fairchild CCD211 but there are some differences.
M1 stand for January 1980.




There are quite some testpins!  :o








The package consists of two plastic parts that are put together with some chewy glue.




The die is placed in a shallow pit.
Interesting: The GND pin is connected to the gold plating on the bottom of the pit not to the die itself. (The die is connected to the plating on the left.) I assume that´s better because in the die the light generates free charges. The negative charges are used to get a signal the positive charges probably go through the substrate. If you would connect the die to GND on a single point there would be an uneven current flow and the potential of the substrate wouldn´t be uniform distorting the charge accumulation and in the end giving you bad pictures.




There are scratches telling you where to place the die.  :-+ ;D




edge length 6,3mm






Well that´s not a perfect die. These two lines probably give you corrupted information.




I don´t know what the C in L211C tells us but the Fairchild CCD211 was binned and the CCD211C was the worst one. It was possible to get a part with up to three defect adjacent columns.




Eight masks...




(Picture from Fairchild CCD The Solid State Imaging Technology 1981)
The L211C works similar to the CCD211 for which we have a block diagram. There are photosensitive pixel in which charges are generated. All the charges are transferred to vertical CCDs which transports them to a horizontal CCD which transports them to the output.
This construction is called Frame Interline Transfer CCD and it takes the whole picture in one step.
Under the horizontal CCD there is a anti blooming gate. If you put too much light on one pixel there are too much charges for the one slot and the charges travel to the surrounding pixel (vertical). The same problem occurs in the horizontal CCD. In the first place that gives you bright lines where the overwhelmed pixel is. Later on it generates bright bars in the horizontal CCD. In the CCD211 (and the L221C) there is a horizontal anti blooming gate. It is a path for these charges so they don´t interfere with the other columns. There is still vertical blooming but no horizontal blooming.
At the output there is a floating gate reading the charges in the horizontal CCD and controlling the output stage.








Surprise! At the bottom of the die there is an additional CCD-line probably for testing.
On the left side you see the connection and distribution of the clock lines for the vertical CCDs.
In the matrix you see the elements in which light generates charges. On the right side of every pixel row there are the vertical CCDs.




The edge length of the active area seems to be something around 15x15µm. I don´t know what the short red lines are.  :-// Well there are a lot of structures one above the other...






In the upper part of the die there is the horizontal CCD. On the left side of the horizontal CCD there are some buckets covered with metal to give the user a dark reference.
Between the sensor matrix and the horizontal CCD you can spot the anti blooming gate.






On the right side of the die the horizontal CCD makes a 180° turn. The first floating gate controls the output stage but the CCD travels further to the left where you can see some more floating gates...




The output stage is quite simple.






There is a big circuit that is connected to the additional floating gates. From outside it is just connected with testpins.
It seems to be a wider CCD that splits into two CCDs at the left end. On the left side there are two output stages which are connected to the two CCDs. One of the CCDs is shorter than the other. It seems like a output stage as we have seen it in the BBDs (https://www.richis-lab.de/bbd01.htm) where just every second bucket contained information. Using the BBDs you have to connect the two outputs.




(Patent US3806772A)
The wide CCD (and the folded horizontal CCD) seems to be a so called Charge Coupled Amplifier a circuit like a distributed amplifier. You can probably use the amplifier to take pictures in low light.
The datasheet of the Fairchild CCD211 shows a square called "amplifier circuit" but beside that hint you don´t find anything about an additional amplifier in the CCD211 and L211C documents. Perhaps the amplifier was not implemented for the "normal" user?


https://www.richis-lab.de/Opto09.htm

 :-/O
 
The following users thanked this post: doktor pyta, RoGeorge

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #33 on: November 06, 2021, 05:03:46 am »




DLSF1414, a strange 4* 16 segment display built by Siemens. You can´t find any information regarding the DLSF1414. There was just a DL1414 but Siemens did customer specific displays too. Perhaps the DLSF1414 is such a customer specific display.




The DL1414 contains a RAM which is adressed with two adress bits and a write input. Through a 7Bit interface you put the code of the character you want to see into one of the four slots of the RAM.
There is an oscillator that activates one digit after the other. While activating a digit the circuit activates one of the slots in the RAM. The RAM outputs the saved code to the ROM and the ROM outputs 17Bits that activates the segments that show the character we wanted to see.






There are other displays that provide you with more functions. Perhaps the DLSF1414 is a DL1414 with some of these functions added.




In the DLSF1414 there is a board potted in clear potting.




On the backside we see the control circuit that is potted with some dark potting. There seems to be a second circuit in the upper left corner. Perhaps that´s a small logic circuit and that is the difference between the DL1414 and the DLSF1414. (This circuit is lost.  :-[)




9041, probably the datecode.




The digits are dies with the dimensions 2,1mm x 1,8mm. There are 16 segments and a dot.
Since there is a dot at the upper edge and at the lower edge you can bond the die rotated 180°.
Thin metal lines distribute the current over the light emitting segments. The substrate is the common anode.




The control circuit is 3,1mm x 2,7mm.








Siemens designed the control circuit in 1988. SMC4621 probably is a internal naming.
TONY? BE?  ;D




We can find five mask revisions: 1C, 2C, 4D, 7C, 8C




Some areas are easy to spot. At the upper edge there are 17 driver for 16 segments and a dot (red). At the lower edge there are four driver for the four digits (green).
At the left edge we can find the data interface (blue). There are seven bondpads side by side but D0 is placed at the bottom of the die. There are buffer stages as shown in the block diagram.
The data interface is connected to a 4x7 RAM (yellow). The control lines for the four columns of the RAM are connected to the digit drivers (green). You activate one column and the digit drivers activate the right digit (over a small logic).
The output of the RAM is the input for the ROM with a quite complex signal conditioning (orange). The size of the ROM is best explained with 17x17x2x3. The output of the ROM is connected to the segment drivers.
In the middle of the die we can spot an oscillator with a big capacitor (pink). The clock is fed to a bondpad and from there to something like a clock divider and distribution. It seems like the control circuit can be used for other displays with more functionality. As seen in the block diagrams there are displays with a clock input/output. One bondpad is directly connected to the segment drivers. Perhaps that one was used for dimming the display. There are some more used and unused bondpads eleven of them are equipped with buffer stages as the data interface.




Here we see a segment driver. There is a big transistor right of the bondpad and a control circuit right of the transistor. There has to be some current limiting.




The power transistor for the small dot is smaller than for the other segments as it has to sink less current for the same brightness.




The digit drivers are bigger but less complex.




Above and below the input bondpads there are protection circuits.




Here we have the 4x7 RAM. Left the input, right the output.






The ROM has to contain 63 characters. So you need a 17x63 ROM. But that ROM looks quite strange...




From left and right there are 8 and 9 control lines (yellow/red). But that would give us just a 17x17 ROM, not enough for the character set.
On the upper edge you can spot three transistors. One transistor (ON) allows us to switch every segment on, probably for a cursor functionality. Two transistors (OUT1/OUT2) switch the segment driver to one of two vertical lines (green). That expands the memory to 17x17x2 or 17x34, still not enough.
There are three more control lines below the memory array (blue). These control lines switch three transistors (C1/C2/C3) that activate three vertical lines around the two green output lines. That expands the memory to 17x17x3 (17x51), still not enough in the first place. There has to be 12 characters which are generated by adding two other characters, i.e. by activating more than two vertical lines.


https://www.richis-lab.de/Opto10.htm

 :-/O

 
The following users thanked this post: edavid, Vgkid, doktor pyta, RoGeorge

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #34 on: December 23, 2021, 11:01:09 pm »




The VQB76 is a 7 segment display built by the Werk für Fernsehelektronik. The segments are formed with long LEDs. There was also a variant with a red glass for better contrast (VQB76-1).






The package is a metal tray with a glass element glued onto it. In the metal tray there is a ceramic plate carrying the circuit.




Each segment is built with two LEDs connected in series. You must pay attention that the dot consists of just one LED and so you need a different resistor for this one.
Another important point: There are three independent anode pins.




The LEDs are 1,2mm x 0,3mm. They are split in three segments. Here the segment in the middle of the die isn´t formed accurate.
The thin metal line distributes the current over the segments.
The element in the upper right corner is probably for checking the mask alignment.




The left LED is damaged. The splintered part is stuck in the solder at the bottom of the die.




The dot LED doesn´t look very good.






It´s Christmas time! Let´s turn on some light.  8)
At 20mA the forward voltage is 3,24V (2x 1,62V).
You can see the incomplete segment in the middle of the die.
The upper epitaxial layer (GaAsP) is translucent while the lower substrate (GaAs) blocks the light.






At 20mA (1,67V) you can see the higher current density around the metal contact. In this area the light is a little brighter.




Let´s drive the LED into reverse breakdown. At 52V (26V for one LED) the current is 0,1mA and you can see the first red dots.




1,2mA, more small lights.




5mA




10mA




Switching from 10mA to 20mA there are more light dots but some areas got darker.
I assume the light intensity dropped with the higher temperature of the die. At 10mA the power dissipation is 0,29W. At 20mA the power dissipation is 0,62W.






I have another VQB76 without a marking.
The layout on the ceramic plate is a little different and in the bond areas there is no gold plating.




The big LEDs are the same but the metal layer is a little thicker.




For the dot they used a different type of LED. We know this one from the VQC10: https://www.richis-lab.de/Opto05.htm


https://www.richis-lab.de/Opto11.htm

 :-/O
 
The following users thanked this post: daqq, doktor pyta, RoGeorge, capt bullshot

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #35 on: March 26, 2022, 08:59:22 am »








SP107 is a photodiode built by the Werk für Fernsehelektronik. Combined with the IR emitting diode VQ170 you can built an optical transmission line <1km.

Besides the printed name the green dot marks the SP107. A red dot would show that it is a VQ170.




There is a plastic protection cap for the SP107.




We can already see the photodiode.






The photodiode is placed on a quite big heatspreader. It is protected with some transparent potting. Dirt on this potting obscures the pictures a little.




The edge length of the photodiode is 1mm. It´s a PIN-diode. In the lower right corner there is an alternative bondpad.


https://www.richis-lab.de/Opto12.htm

 :-/O
« Last Edit: March 26, 2022, 10:28:58 am by Noopy »
 
The following users thanked this post: RoGeorge

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: Optoelectronics - die pictures
« Reply #36 on: March 26, 2022, 09:24:46 am »



The housing uses a conductive coating. Here you can see the second lens in the "screw".

Just in case someone is interested in. The technology used is called MID (Molded Interconnect Device). By the looks of it it uses the LPKF-LDS process, which means the non-conductive plastic is injection molded and contains additives that are activated using 3D laser structuring, you can tell by the vertical structure. Afterwards a wet chemical process is used to plate copper on the activated areas, with a final finish of nickel and gold. See also LPKF-LDS or 3D-MIDs

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 
The following users thanked this post: edavid, RoGeorge

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #37 on: April 07, 2022, 06:25:36 pm »
I have and update to the L211C CCD sensor.

There is a paper "Konzeption Fertigungseinführung L220C" describing the manufacturing of the bigger L220C in detail. It´s highly probably that the L211C is constructed very similar. Here you see some pictures showing the construction of the CCD sensor. I have added some color.




The first steps is to integrate a n-doped shift register channel and the highly p-doped channel stopper into the p-doped substrate.






Two polysilicon layers represent the electrodes for the CCD shift register. The two supply lines run above each other to the side of the matrix so they  don´t shadow the active area too much.




There is no pictures showing the metal layer so I had to add it.  ;) The metal lines on top of the shift register act as a photogate and protect the shift register against light.




The dark squares are probably the active areas. The short red looking lines could come from the fact that the surface is quite uneven.


https://www.richis-lab.de/Opto09.htm#L220

 :-/O
 
The following users thanked this post: RoGeorge

Offline Ivan7enych

  • Regular Contributor
  • *
  • Posts: 158
  • Country: ru
    • My astronomy projects
Re: Optoelectronics - die pictures
« Reply #38 on: June 28, 2022, 09:29:42 am »
Thing of beauty... Costs like my house...
Came to my hands for 2 days, Gsense 6060 CMOS sensor

Bright field illumination, Nikon 5x BD plan and Nikon 20x ELWD
 
The following users thanked this post: edavid, RoGeorge, magic

Offline Ivan7enych

  • Regular Contributor
  • *
  • Posts: 158
  • Country: ru
    • My astronomy projects
Re: Optoelectronics - die pictures
« Reply #39 on: June 28, 2022, 09:45:29 am »
On last picture many test pins looks as being used for testing the chip.
 
The following users thanked this post: RoGeorge

Online magic

  • Super Contributor
  • ***
  • Posts: 6733
  • Country: pl
Re: Optoelectronics - die pictures
« Reply #40 on: June 28, 2022, 03:00:30 pm »
Thing of beauty... Costs like my house...
Came to my hands for 2 days, Gsense 6060 CMOS sensor

Bright field illumination, Nikon 5x BD plan and Nikon 20x ELWD
Good stuff :-+

I suppose those blue balls are the microlenses array? How are they even fabricating this stuff :scared:
 

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #41 on: September 19, 2022, 07:36:37 pm »




The IL-C6 is a CCD image sensor consisting of a single line with 2048 pixels. The device is manufactured by the Canadian company DALSA, which now belongs to Teledyne.

The CCD image sensor is in a special DIL package with dimensions of 42,4mm x 7,5mm. The datasheet highlights the high readout speed, which can be up to 15MHz. The dynamic range is 1:6000.




Viewed from the side the glass lid on the ceramic case is clearly visible.




The VBB potential (Substrate Bias Voltage) is fed to the sensor via four pins. Three of the four potentials are connected to the housing with low impedance by two bondwires each. The signal information is located as charge packets in the sensor areas and in the CCD line. These charges are processed and moved by electric fields. Local potential fluctuations in the substrate could accordingly have a negative effect on signal recording and signal transmission. For this reason it is important to keep the substrate potential low impedance at a fixed potential.






The black sensor area is clearly visible on the die. The datasheet shows the basic operation of the CCD image sensor. It is 2048 pixels with a width of 13µm and a height of 500µm. At the beginning and at the end of each line two darkened pixels are integrated, which can be used as reference value. Two further pixels each represent a buffer to the active area.

Via the potential PR the sensor areas can be reseted to the potential VPR. After exposure TCK ensures that the charges generated in proportion to the light incidence are transferred to the CCD shift register. The four phase-shifted clock signals CR1, CR2, CR3, and CR4 generate an electric field with potential wells that move from left to right, shifting the charge packets toward the output. This is described in more detail in the L133C (https://www.richis-lab.de/Opto08.htm).

On the far right is a two-stage amplifier circuit that serves the output OS. The electrode VSET causes the charge packets to flow from the CCD array to the amplifier input. The amplifier circuit converts each charge packet of the CCD array into a voltage which is proportional to the light incidence into the respective pixel. The evaluated charge must then be neutralized into the potential VOD via the pin RST.




In the datasheet there is another blockdiagram that shows the function of the CCD image sensor in yet another way.




The glass on the ceramic case is clear enough that you can take pictures of the integrated circuit without opening the case.

Apart from the sensor area, the whole die is covered with a metal surface. This ensures that no free charges are generated in the other circuit parts that would lead to unwanted current flows. The metal layer makes it difficult to analyze the circuit, but does not make it impossible.




In the lower left corner of the die there is a copyright notice of the company DALSA from 1990 and the name of the device.




The individual elements are clearly visible in the sensor area. The distribution of the potential VPR is clearly visible too. Controlled by the potential PR free charges in the sensor area can be diverted and thus neutralized.

The sensor area is framed with strips of the lower metal layer. It´s not 100% clear whether this frame is connected to the VB or the VBB potential. VB is the "Bias Voltage" which is usually connected to Vdd. VBB is the "Substrate Bias Voltage" to be connected to -3V-0V. The structures indicate that the inner area around the sensor is connected to the positive VB potential.






Controlled by the potential TCK the charges of the individual pixels flow into the CCD row below. Below the CCD row there are four wide lines of the lower metal layer, which carry the four phase-shifted transport clocks. Especially for the lower three lines, the contacts and the contours of the lines carrying the clock signals upwards are clearly visible.

Small square elements are connected to the lines carrying the outer potentials to the sensor. Most likely these are diodes that protect the circuit from problematic voltages. Since the gate structures have to be protected against too high circuits, it can be assumed that they are Z-diodes.




The electrodes on the CCD line can only be guessed at. This is not only due to the small structures, but also because the electrodes overlap and show many contours in the metal layer.




At the right end of the CCD row is the output amplifier, which receives an exclusive supply there. The VSS potential at the top right seems to serve just as shielding of the output signal.






The charge packets of the CCD array control the two-stage output amplifier and are then diverted to the right.


https://www.richis-lab.de/Opto13.htm

 :-/O
 
The following users thanked this post: RoGeorge

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #42 on: October 21, 2022, 07:20:58 pm »




Kyocera sells the 910-00011-IT, a light source that is described as "LaserLight". The 910-00011-IT already includes a star-shaped metal core PCB. Alternatively, the light source itself can be purchased under the designation 910-00010-TR.

With 9V and 2,5A the LaserLight module delivers a luminous flux of 1000lm. The special feature is the small radiation area, which has a diameter of just 0,5mm. The luminance is 1300Mcd/m² and so is significantly higher than that of a LED or a high pressure gas discharge lamp. A small light source is very helpful when you want to create sharp light cones with small optics. One application are vehicle headlights that are intended to illuminate just certain areas of the road.

The datasheet shows the wavelengths contained in the light. The large, narrow-band blue component is striking. The CRI is accordingly just 70.




The small circuit diagram in the datasheet describes a series connection of two LEDs. A zener diode connected in parallel keeps negative voltages at a very low level and limits overvoltages, such as those that occur during ESD events.




In fact the 910-00011-IT does not contain normal LEDs, but so-called superluminescent diodes (SLD). The IEEE article "A Stripe-Geometry Double-Heterostructure Amplified-Spontaneous-Emission (Superluminescent) Diode" (IEEE Journal of Quantum Electronics, August 1973) shows a possible construction of an SLD.

The design strongly resembles a semiconductor laser (edge emitter). The light from a pn junction is amplified in a channel and coupled out at the edge of the device. In a semiconductor laser there are semi-transparent mirrors at the edges. Optical resonance sets in between the mirrors.

A superluminescent diode also uses the effect that the light beam is amplified in the channel. In contrast to a laser diode, however, the aim is to prevent reflections at the edges. This goes so far as to make the edges slightly slanted so that reflected components are not directed back into the channel. Even reflections at optical fibers, into which one wants to couple the light, can be problematic. Sometimes the rear area of the channel is designed as an optical termination. The IEEE article describes that this is easier to implement than a surface with very low reflections.

The emitted light beam offers a very high luminance like a laser. At the same time, the bandwidth of the light is wider, similar to a light emitting diode. A higher bandwidth is advantageous if white light is to be generated via a luminescent material. Strictly speaking, the term "laser" is wrong here, because the laser effect is not supposed to start with a superluminescent diode.






The pictures of the 910-00011-IT are not mine. They were taken by the user "dominic_m833" from the mikrocontroller.net forum. He allowed me to use them.

On the right and on the left side there is a superluminescent diode. The optical axes are slightly offset vertically. The areas above the SLDs have been blackened. In the center is the yellow luminescent material used to generate the white light. The Z-diode is mounted in the upper right corner.




Viewed from the side, it can be clearly seen that the SLDs are located on wedge-shaped structures that ensure that the light rays strike the luminescent material from above.




The SLDs has very low impedance contacts. The upper metal contacts the channel from above. The potential of the lower metal is conducted to the lower layers just before the channel begins.




In normal operation hardly any light from the lasers can be seen.




If you adjust the exposure appropriately, you can see where the light rays from the superluminescent diodes hit.




You can enhance the light of the superluminescent diodes too.  8)


https://www.richis-lab.de/Opto14.htm

 :-/O
 
The following users thanked this post: edavid, RoGeorge

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #43 on: October 27, 2022, 08:33:40 pm »


The guy who took the pictures of the Kyocera 910-00011-IT compared it with one of the brightest LEDs, the OSRAM KW CELMM1.TG.  :o


https://www.richis-lab.de/Opto14.htm#Vergleich

 :-/O

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #44 on: October 28, 2022, 03:01:00 pm »
It seems I was wrong. The 910-00011-IT uses laser diodes, not superluminescent diodes.  :palm:

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 6146
  • Country: ro
Re: Optoelectronics - die pictures
« Reply #45 on: October 28, 2022, 05:09:54 pm »
LASER is supposed to produce coherent light.
Fig.1 reads "incoherent output beam".

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #46 on: October 28, 2022, 05:16:04 pm »
The output light is incoherent but the initial source can be coherent. It is transformed in the phosphor.

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 6146
  • Country: ro
Re: Optoelectronics - die pictures
« Reply #47 on: October 28, 2022, 05:36:26 pm »


From this image I understand the beam is not coherent, and it looks like that's the beam before entering into the phosphorus.  So the SLD (Super Luminiscent Diode) LED is not a L.A.S.E.R. device.  Isn't that so?
« Last Edit: October 28, 2022, 05:47:42 pm by RoGeorge »
 

Offline NoopyTopic starter

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Optoelectronics - die pictures
« Reply #48 on: October 28, 2022, 06:02:15 pm »
You are right, a SLD gives incoherent light.

But in every datasheet of these Kyocera LaserLights you just find the word laser. There is sometimes the acronym SLD but that doesn't mean Superluminescent Diode but Soraa Laser Diode Inc., a company that was bought by Kyocera (and again "Laser").

I somehow wanted to believe they use Superluminescent Diodes but it looks like that are Laser diodes. There is no hint that they use SLD.  :-//
 
The following users thanked this post: RoGeorge

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16544
  • Country: us
  • DavidH
Re: Optoelectronics - die pictures
« Reply #49 on: October 28, 2022, 11:01:41 pm »
LASER is supposed to produce coherent light.
Fig.1 reads "incoherent output beam".

The coherence length of most lasers is very small, so the light is no longer coherent where it gets used.  Special steps have to be taken to get a useful coherence length, like special construction and temperature control.  For instance you cannot make an interferometer from any common laser source.

Laser diodes routinely have a coherence length of millimeters.
 
The following users thanked this post: boB, RoGeorge


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf