Came here to say the same thing. There are about 29 of the beefy transistors. This lines up well with a 24-pin dot matrix printer header (has 24 solenoids to operate) with a few extras transistors for other tasks.
Modules are still used today - my former employer produced system on modules for sale to customers. The reason for this is the SoC requires a 8 layer PCB and special processing due to the BGA density. Customers wanting to cut out the middleman to save costs would spend a lot more on having to invest in a more expensive PCB process to handle the fine details necessary to use the SoC. Using the module means they can switch to a more common 4 layer PCB process. This is cheaper and much easier to handle - at the expense of having to pay us for a module and some high density connectors.
Centronics, very typical back in the '90s with LPT printers. As a matter of fact, I never really saw a printer with a DB25 port - all of them had Centronics. A nice look into the STK for sure! Almost like magic.
The DB25 printers are more likely with RS232 serial interface. Not very common in home or office but in the logistics sector, such as shipping label printer or in the industrial sector as a status printer connected to a PLC...
I often wondered why the standard used a centronics connector on the printer, and a d-sub 25 on the PC side? There's probably some history there I don't know about. I know scsi also used centronics connectors, but 50 pin, vs 36 for printers... wikipedia says Wang Labs, who was building a low cost centronics compatible printer used that connector because they had a lot of them left over from another product, and it became the de-facto standard! Huh. Probably was also a good thing so people didn't plug in 25-pin serial cables, and fry the printers with RS-232's +/- 12V signalling, when the printer is expecting 5V signalling.
Pretty cool STK thing, probably made on purpose only for that series of printers. BTW, those Epson Smps's are quite good quality. I have one supplying +38V from an old printer, now powering a single supply audio power amplifier.
1:19 While a font card add-on is a good guess, this printer instead used add-on daughter cards to add additional interfaces for serial (with and with 32KB buffering) and parallel (with 32KB buffering as an upgrade).
Back in the day, Epson used to make Centronics, Serial, and other protocol printers. They kept the hardware platform consisting the STKxxxx common and they just used an EPROM for each protocol. Then the era came were they build Centronics with serial protocol capability printers and now we have just USB / WiFi
KiCad can do that. Set your Text Object to "knockout" and it will be inversed. If you put it over a metal fill then just recalculate the fill (shortcut b) and you are done.
so that's what that card slot is for. I believe I have a card that fits that- it's an RS-232 interface for an epson printer I've had for the better part of 3 decades in my junk bin. that giant hybrid in there is pretty interesting
12:25 I bet all that text in the EPROM is for the in-store demo mode. The non-text in between is probably control codes for moving the heads and changing fonts.
3:25 seen those connectors in several VCR's and DVD players. Connecting the SMPS to the main board. Probably easy assembly during manufacturing when you got those connectors. Could it even be a standard pinout for these connectors? Like these pins are ground, these are +5V, these are +12V etc... The SMPS is probably of the shelf units.
They're just larger transistors in an H bridge for the platen motor with current sensing provided by the all the passives and the bare die opamps nearby. The four smaller transistors in a horizontal line above those drive the print head indexing motor, and the 24 transistors around the top and right sides are for the 24 pins in the print head. The single big transistor in the lower right quadrant may be for the paper-out buzzer but that's just a guess. Along the bottom are a line of 4 small transistors which may be for the status LEDs on the panel. The field of resistors along the lower right probably works as the business end of a DAC to set motor current.
You could use a multimeter to ID the components, even the cheap component identifier if some leads were fitted. I just ordered some BAT85 diodes to attempt a 1MHz to 500 MHz SWR meter build, comparing the transformer design with the coupler / transmission line options. (#749 - 749b).
Sanken made these for Sanyo, along with Sanyo themselves. A way to cheaply make a power block, using bare dies, printed resistors, and thus make an amplifier that was easy to assemble, but not limited to 40V rails like most power IC's are. This one has the stepper drive for the carriage, the yellow coated darlington transistors, complete with the high power emitter resistors to allow current control, and then the catch diodes needed in regular SMD packages. At the bottom you see the analogue amplifiers for the current sense, used by the main controller to ramp up motor current for a smooth control of the head position. Top is the drivers, and catch diodes, for the print head, as there you need that 35V supply to drive the heads, and switch them on and off fast. so the integrated control IC that handles this, serial data in and a latch, then controlling the 24 head solenoids that each drive a pin used to print. Depending on model you could habe 8, 9, 16 or 24 pins on the head, and often Epson also included colour capability as well, allowing 3 colours and black to be printed, though it really was not great for making more than low resolution colour images, but great with vector art and graphs, not photo realistic. Top missing board is the serial interface, plugging in, and either having it's own interface microcontroller, allowing you to set baud rates and control from the front panel, or it had a set of DIP switches to configure it, depending on which board you got, as they all were the same size on most Epson printers, and used the same pinouts mostly, but not always interchangeable among models because of firmware. Then the microcontroller, Z80 core in the square package, along with an entire set of peripherals as well, and, because it also has in it a bit of RAM and Eprom (not erasable unless you had the development ceramic package, or access to an X ray machine otherwise) thus the special Epson version with their proprietary driver code in it. External ROM and RAM used to store model specific information, character sets, configuration of the printer mechanism features and such, and the RAM is the print buffer, using the Z80 DMA to read characters in from the serial interface, and then convert to a bitmap line of data before printing, using the hybrid as power output. Pretty smart printers, and with bidirectional control as well, and some even came with the option of swapping the print head (removeable to replace it with no tools, and a connector on the flex cable) and using a scan head to scan sheet documents, though limited in not having a sheet feeder, and also limited resolution, and slow.
@@VEC7ORlt Looked it up, SanKen also used STK. A motor driver I salvaged from an old Epson matrix printer was marked STR (and not STK as I thought). So this might very well be Sanyo.
I wonder if those yellow things were MOSFETs. (?) Also, Did anyone else suffer with screen tearing and frame rate issues when there was movement under the microscope ?
I have no idea what happened there. my video editing software has always done well with the microscope files, but not this time. after a software 'update'
This hybrid board is made up of Chip On Board technology (COB) and SMD parts. If you have the equipment and clean room, you can buy the dies and protect them with covers. The small wires coming off the dies give it away. Thanks for the video.
Would not negative printings within the copper pour cause additional impediments to the ground path/flow? I thought the wire bonds (Isn't that the name for the micro wire connections?) directly to the board were amazing. But, was that a good idea, for a unit that has motor vibration? Dunno...
If you look closely it appears to have a clear conformal coating which is pretty typical. Since they were already being conformal coated and are physically protected there isn't need to coat the chips multiple times.
Back in the 1980s and 90s Centronics ports were the norm for PC printers (other than SCSI for the purists :-) Then they started coming out with USB with the slogan Plug and Play, which we called Plug and Pray.
Plug and play worked even with parallel printers. I remember using a DeskJet 540 with centronics parallel connector in Win95. I think all printers back then had parallel, and some had serial as an option. Since then I've only used business class laser printers with ethernet - much less hassle to setup in Linux, and these days I hardly print anything anyway.
The four yellow squares are the transistors of the H bridge for the platen motor. The large IC does appear to be constructed from an off the shelf gate array but the circuitry appears to be nothing more than two or three 8 bit transfer registers, a 24 bit holding register, and a little I/O decoding. It appears to have 24 outputs, an 8 bit input data bus, 2 address bits, and 3 control lines.
@ 2:50 I have a Sony AV amplifier from the 2000's (STR DE-485E) that has a very similar connector from the power supply to the main board, never failed but i don't like it to much.
You see it a lot in consumer electronics - receivers, vcrs, sometimes dvd/blu ray players, etc etc. I've never seen it in commercial gear - they use wires in commercial stuff
I have seen that connector before, in a Samsung DVD player between the power supply and main board. Maybe one particular psu company likes it? Unless this is a samsung printer...
@@user-ku9cq8kk6i What would you expect to see looking into a mosfet or BJT when the device is off? You're not going to see much with a VNA - it will probably just have some capactive coupling to the power supply rails without any bias to turn the transistors on
@@gorak9000 That’s a good point, I was actually just curious if it was possible to reverse engineer something like that by treating it as a black box on a component level, And then power it up afterwards once it was mapped out as a test…After all wouldn’t that be the point, To analyze a network?
@@user-ku9cq8kk6i If you wanted to reverse engineer this, you'd be better off monitoring all the pins with a logic analyzer, and watching how the CPU controls it - you'd have to be careful to not blow up your analyzer though, logic analyzers usually don't like 35V signals that these outputs would be. I'd also be really careful about inductive flyback spikes, seeing as everything this module is driving is inductive (motor windings, or solenoid windings that make the pins strike the ribbon/paper). I'm guessing they already have flyback suppression diodes on each driver circuit already, but you never know. You'd probably be fine monitoring the interface between this and the CPU though. A vna is about measuring electrical characteristics (specifically complex impedance) of a device (or network) at various frequencies. A logic analyzer is more suited for analyzing functionality of digital circuits like this is.
The STK chip is the driver for the 2 stepper motors ( line and line feed) and the dot matrix solenoid head
Came here to say the same thing. There are about 29 of the beefy transistors. This lines up well with a 24-pin dot matrix printer header (has 24 solenoids to operate) with a few extras transistors for other tasks.
I think this is the largest IMST (Insulated Metal Substrate Technology) hybrid circuit from Sanyo.
I love the symmetry of the traces with their 90 degree angles.
I often see those connectors on the insides of DVD players (power supply control board). BTW, the STK internals are gorgeous, I'd frame it!
Modules are still used today - my former employer produced system on modules for sale to customers. The reason for this is the SoC requires a 8 layer PCB and special processing due to the BGA density. Customers wanting to cut out the middleman to save costs would spend a lot more on having to invest in a more expensive PCB process to handle the fine details necessary to use the SoC. Using the module means they can switch to a more common 4 layer PCB process. This is cheaper and much easier to handle - at the expense of having to pay us for a module and some high density connectors.
Centronics, very typical back in the '90s with LPT printers. As a matter of fact, I never really saw a printer with a DB25 port - all of them had Centronics.
A nice look into the STK for sure! Almost like magic.
The DB25 printers are more likely with RS232 serial interface. Not very common in home or office but in the logistics sector, such as shipping label printer or in the industrial sector as a status printer connected to a PLC...
I often wondered why the standard used a centronics connector on the printer, and a d-sub 25 on the PC side? There's probably some history there I don't know about. I know scsi also used centronics connectors, but 50 pin, vs 36 for printers... wikipedia says Wang Labs, who was building a low cost centronics compatible printer used that connector because they had a lot of them left over from another product, and it became the de-facto standard! Huh. Probably was also a good thing so people didn't plug in 25-pin serial cables, and fry the printers with RS-232's +/- 12V signalling, when the printer is expecting 5V signalling.
A work of art that hybrid board... Nice find!
Cool hybrid IC for sure. Clean it up and it would be suitible for framing and displayed in a nerd room.
Pretty cool STK thing, probably made on purpose only for that series of printers.
BTW, those Epson Smps's are quite good quality. I have one supplying +38V from an old printer, now powering a single supply audio power amplifier.
STK ( Sanken?) hyrids were common in a few applications like audio power amps and PSUs - never seen one that big though
I've seen those leadframes in consumer DVRs - I suspect they start off folded up for PCB assembly then get stretched out when the PCBs are fitted
You put the “tear” in “tear-down”. 😂
1:19 While a font card add-on is a good guess, this printer instead used add-on daughter cards to add additional interfaces for serial (with and with 32KB buffering) and parallel (with 32KB buffering as an upgrade).
Back in the day, Epson used to make Centronics, Serial, and other protocol printers. They kept the hardware platform consisting the STKxxxx common and they just used an EPROM for each protocol. Then the era came were they build Centronics with serial protocol capability printers and now we have just USB / WiFi
That board with all the output transistors might be the driver for the 24 pin dot matrix print head of the Epson LQ 570 printer.
KiCad can do that. Set your Text Object to "knockout" and it will be inversed. If you put it over a metal fill then just recalculate the fill (shortcut b) and you are done.
I was guessing surface mount LEDs also :). But I didn't know any better :).
That “lead frame” connector quite often turned up in some video recorder models.
I think that large yellow IC is a PLD or a gate array device... those yellow components to the left are probably part of an H-Bridge driver...
Unpackaged IC's & transistors, amazing!
Mount the die right to the board and stitch the connections.
it seems like a step from just discrete transistors on PCB to the first LSI microchip
but in 1990 ? very strange
so that's what that card slot is for. I believe I have a card that fits that- it's an RS-232 interface for an epson printer I've had for the better part of 3 decades in my junk bin. that giant hybrid in there is pretty interesting
That used PS board by itself sold for $60 on ebay last month.
I've come across that type of connectors on old crt televisions, many times.
12:25 I bet all that text in the EPROM is for the in-store demo mode. The non-text in between is probably control codes for moving the heads and changing fonts.
Reminds me of strip-line RF just after SMD parts started to apear
That's from a Epson LQ570 24pin dot matrix printer.
3:25 seen those connectors in several VCR's and DVD players. Connecting the SMPS to the main board.
Probably easy assembly during manufacturing when you got those connectors.
Could it even be a standard pinout for these connectors?
Like these pins are ground, these are +5V, these are +12V etc...
The SMPS is probably of the shelf units.
Replaced a lot of stk hybrid audio amps back in the day.
If I had a dollar for every STKxxxx amp module I changed in home receivers and car stereo amps back in the 80’s and early 90’s, I could retire now.
Pioneer SX780 if memory serves STK0050.
Any chance you might revisit PAL programming in the future? I enjoyed it very much!
I love a good bin find. So do you have to wait another year until you can throw all that stuff out - lol?
Could those yellow things be mosfet drivers?
They're just larger transistors in an H bridge for the platen motor with current sensing provided by the all the passives and the bare die opamps nearby. The four smaller transistors in a horizontal line above those drive the print head indexing motor, and the 24 transistors around the top and right sides are for the 24 pins in the print head. The single big transistor in the lower right quadrant may be for the paper-out buzzer but that's just a guess. Along the bottom are a line of 4 small transistors which may be for the status LEDs on the panel. The field of resistors along the lower right probably works as the business end of a DAC to set motor current.
You could use a multimeter to ID the components, even the cheap component identifier if some leads were fitted.
I just ordered some BAT85 diodes to attempt a 1MHz to 500 MHz SWR meter build, comparing the transformer design with the coupler / transmission line options. (#749 - 749b).
Really great episode!
If its weird and it says STK = hybrid IC by Sanyo.
Also (in)famous for audio amp modules.
Yes, that was my first though!
Doesan't SanKen use STK for motor drivers?
@@Peter_A1466 AFAIR was always Sanyo?
Sanken made these for Sanyo, along with Sanyo themselves. A way to cheaply make a power block, using bare dies, printed resistors, and thus make an amplifier that was easy to assemble, but not limited to 40V rails like most power IC's are. This one has the stepper drive for the carriage, the yellow coated darlington transistors, complete with the high power emitter resistors to allow current control, and then the catch diodes needed in regular SMD packages. At the bottom you see the analogue amplifiers for the current sense, used by the main controller to ramp up motor current for a smooth control of the head position. Top is the drivers, and catch diodes, for the print head, as there you need that 35V supply to drive the heads, and switch them on and off fast. so the integrated control IC that handles this, serial data in and a latch, then controlling the 24 head solenoids that each drive a pin used to print. Depending on model you could habe 8, 9, 16 or 24 pins on the head, and often Epson also included colour capability as well, allowing 3 colours and black to be printed, though it really was not great for making more than low resolution colour images, but great with vector art and graphs, not photo realistic.
Top missing board is the serial interface, plugging in, and either having it's own interface microcontroller, allowing you to set baud rates and control from the front panel, or it had a set of DIP switches to configure it, depending on which board you got, as they all were the same size on most Epson printers, and used the same pinouts mostly, but not always interchangeable among models because of firmware.
Then the microcontroller, Z80 core in the square package, along with an entire set of peripherals as well, and, because it also has in it a bit of RAM and Eprom (not erasable unless you had the development ceramic package, or access to an X ray machine otherwise) thus the special Epson version with their proprietary driver code in it. External ROM and RAM used to store model specific information, character sets, configuration of the printer mechanism features and such, and the RAM is the print buffer, using the Z80 DMA to read characters in from the serial interface, and then convert to a bitmap line of data before printing, using the hybrid as power output. Pretty smart printers, and with bidirectional control as well, and some even came with the option of swapping the print head (removeable to replace it with no tools, and a connector on the flex cable) and using a scan head to scan sheet documents, though limited in not having a sheet feeder, and also limited resolution, and slow.
@@VEC7ORlt Looked it up, SanKen also used STK.
A motor driver I salvaged from an old Epson matrix printer was marked STR (and not STK as I thought). So this might very well be Sanyo.
I wonder if those yellow things were MOSFETs. (?)
Also, Did anyone else suffer with screen tearing and frame rate issues when there was movement under the microscope ?
I have no idea what happened there. my video editing software has always done well with the microscope files, but not this time. after a software 'update'
@@IMSAIGuy No worries, I just wondered if it was my crappy PC.
It does appear those are MOSFETs in an H bridge for driving the platen motor.
Yes. Is the output from the microscope interlaced? The update to your video editing SW has probably changed a setting.
@@Peter_S_ WOW, what a lucky guess. Yay me. LOL.
Cut the external pins off, and frame it, really cool
This hybrid board is made up of Chip On Board technology (COB) and SMD parts. If you have the equipment and clean room, you can buy the dies and protect them with covers. The small wires coming off the dies give it away. Thanks for the video.
I have seen those kind of connectors in vhs vcr's. Sony SLV series. They connect the power board to main board as is in here.
Would not negative printings within the copper pour cause additional impediments to the ground path/flow? I thought the wire bonds (Isn't that the name for the micro wire connections?) directly to the board were amazing. But, was that a good idea, for a unit that has motor vibration? Dunno...
so kind of bizarre LSI board
chip-on-board with no black epoxy?
If you look closely it appears to have a clear conformal coating which is pretty typical. Since they were already being conformal coated and are physically protected there isn't need to coat the chips multiple times.
Back in the 1980s and 90s Centronics ports were the norm for PC printers (other than SCSI for the purists :-) Then they started coming out with USB with the slogan Plug and Play, which we called Plug and Pray.
Plug and play worked even with parallel printers. I remember using a DeskJet 540 with centronics parallel connector in Win95. I think all printers back then had parallel, and some had serial as an option. Since then I've only used business class laser printers with ethernet - much less hassle to setup in Linux, and these days I hardly print anything anyway.
@@gorak9000 Me too. I have a printer but it has been out of ink for about 2 years. It was last used to print birthday cards.
Yup, I remember always connecting the parallel port of a PC to a printer via a 25-pin D to Centronics cable 👴
Wonder what is between the traces that look like tiny solder blobs, nearly sorting at places
Is the large IC a gate array? Are the four large yellow things a bridge rectifier?
The four yellow squares are the transistors of the H bridge for the platen motor. The large IC does appear to be constructed from an off the shelf gate array but the circuitry appears to be nothing more than two or three 8 bit transfer registers, a 24 bit holding register, and a little I/O decoding. It appears to have 24 outputs, an 8 bit input data bus, 2 address bits, and 3 control lines.
omg how much did that cost to manufacture!
@ 2:50 I have a Sony AV amplifier from the 2000's (STR DE-485E) that has a very similar connector from the power supply to the main board, never failed but i don't like it to much.
You see it a lot in consumer electronics - receivers, vcrs, sometimes dvd/blu ray players, etc etc. I've never seen it in commercial gear - they use wires in commercial stuff
I have seen that connector before, in a Samsung DVD player between the power supply and main board. Maybe one particular psu company likes it? Unless this is a samsung printer...
Could you put your VNA on it and check it for S perimeters as a “black box”… Just curious.
uh, no
That actually surprises me, I was kinda hoping that kind of thing was possible… I am building a HP 89441a and had high hopes for it…Thx though.
@@user-ku9cq8kk6i What would you expect to see looking into a mosfet or BJT when the device is off? You're not going to see much with a VNA - it will probably just have some capactive coupling to the power supply rails without any bias to turn the transistors on
@@gorak9000 That’s a good point, I was actually just curious if it was possible to reverse engineer something like that by treating it as a black box on a component level, And then power it up afterwards once it was mapped out as a test…After all wouldn’t that be the point, To analyze a network?
@@user-ku9cq8kk6i If you wanted to reverse engineer this, you'd be better off monitoring all the pins with a logic analyzer, and watching how the CPU controls it - you'd have to be careful to not blow up your analyzer though, logic analyzers usually don't like 35V signals that these outputs would be. I'd also be really careful about inductive flyback spikes, seeing as everything this module is driving is inductive (motor windings, or solenoid windings that make the pins strike the ribbon/paper). I'm guessing they already have flyback suppression diodes on each driver circuit already, but you never know. You'd probably be fine monitoring the interface between this and the CPU though. A vna is about measuring electrical characteristics (specifically complex impedance) of a device (or network) at various frequencies. A logic analyzer is more suited for analyzing functionality of digital circuits like this is.
👌👌👌👍👍👍....!!!