The Ware for December 2013 is shown below.
It’s been a while since I’ve had a proper die shot for a name that ware. Thanks to T. Holman for sharing this ware with us!
[Edit: You can download a copy of the original, unobfuscated image here (warning: 23MB).]
An accelerometer?
An SAW-filter.
Is this some form of timer such as a 555?
A PIC12F629/675?
http://ww1.microchip.com/downloads/en/DeviceDoc/41190c.pdf
Looks like some kind of microwave circuit. Most of the large structures look suspiciously like waveguide filters, but I’m not enough of a wizard to definitively identify any of them.
No, waveguides would be too big, millimeter sized structures are suited for the terahertz range. micrometric hollow waveguides on silicium would be suited for micrometric waves, e.g. light, and that would be photonic circuits.
MMICs have long folded strip lines structures, these components are often amplifiers and have parallel wired RF transistors that requires wilkinson dividers, whose lengths must be matched to quarter waves.
Here is an example: http://dehron.com/images/products/%233,%205W%20MMIC%202-4.4GHz%20AM204437.JPG
This one is even more typical
http://www.ausairpower.net/XIMG/figure-3.png
rf/if saw filter from a motorola pager – maybe something fancy like a pagewriter?
Re: the “SAW” look: If you look at the comb-like structures with the small dots on the fingers, you’ll realize that there’s always a finger connected to one side of the areay, the next finger will be donnected to the other side. And there is a small thin metal in between. These structures that some commenter classified as a “SAW” is actually a transistor, with the “big dotted fingers” being metal-layers with a lot of vias. Those devices are supposed to carry a lot of current.
I’m counting 6 of the very large transistors, the very large one at 11o’clock and at least the pretty central one (at 6 o’clock, just above the pin connect), but also the 2 other larger ones at 5 o’clock seem to share the gate (the bigger metal just below 1997). Strangely enough, they all seem to have different size, which equals different current rating.
If you look at the pin-pads, there are three distinct types. The lone one at 9 o’clock without additional circuitry (possibly GND/package/substrate), the type at 10, 11 and 6 o’clock (as far as my guesses go, it’s a single protectino diode) and a identical cell for the pins at 7:30, 12:00, 2 o’clock and 4 o’clock which seems to be more elaborate.
I see only very little “small signal” transistors (at 9 o’clock) and the pretty obscure structures with the circles in them.
I’d summarize that this is…
– definitely high-power (several amps)
– very well protected pins (over/undervoltage)
– at least two high-powered push/pull transistor pairs
– very little “small signal” logic (less than a dozen transistors)
– not a symmetrical parts, like e.g. a dual H-bridge, that would contain several transistors of the same size/rating
So, a guess, and I’m really feeling not good at this, because of the few low-power components) would be (parts of a) 2 channel switch-mode voltage regulator, maybe a buck/boost (which has the equivalent of a H-bridge between the switcher’s coil’s pins, have a look at the http://www.linear.com/product/LTC3531 datasheet, for a low-current version). Buth with a much higher power rating.
Definitively a Texas Instruments chip.
Power seems crucial, maybe a MOS driver, or a level converter.
how do you know it’s definitely a TI chip?
It’s probably the 8 visible Ti marks in a row. I’m not as convinced of it’s provenance.
Huh. The Ti marks, I see those as mask alignment/etch quality metrics. They do kind of have the shape of Ti, but each one is printed on a different layer and the shape of the feature is typical of what you’d use to make sure you don’t have a problem on a particular film layer.
There’s some other features that I think could be very distinctive in terms of identifying the process and fab used, but I haven’t seen anyone comment on those yet.
A TI MAX232A from 1990 with the little M logo.
http://upload.wikimedia.org/wikipedia/commons/2/2d/TI_2BC121M_MAX232N.jpg
Isn’t that a Motorola logo to the left of the date?
No. I was windering the same thing so I looked and found:
From: http://en.wikipedia.org/wiki/Copyright_symbol
“The non-obligatory symbol used in a mask work protection notice is Ⓜ (capital letter M enclosed in a circle.)”
It’s the symbol for mask copyrights, e.g. the patterns used to define the layers on a chip.
Mask works have a copyright term of 10 years, unlike Mickey Mouse, so this design’s mask set is out of copyright and in the public domain:
“17 USC 904 (b) Subject to subsection (c) and the provisions of this chapter, the protection provided under this chapter to a mask work shall end ten years after the date on which such protection commences under subsection (a).”
A tiny edifice of copyright law that still retains a semblance of sanity; perhaps because it’s a better use of funds to develop newer, faster chips rather than pay lobbyists to protect existing designs for over a hundred years.
Mask patents are a relatively recent piece of law (late 80s early 90s?) – when I entered the field people were stealing each other’s masks left right and center – with 1 layer of metal (like this chip) everything’s exposed and it was found that copyright law didn’t apply – it reached the point where designers spent time including bogus circuits in their designs, polygons laid out in such a way that a naive copy would fail (perhaps turn on both halves of an output pair and fry the die)
Big transistor, three-part resistor ladder right under it, a middle rung of the ladder tied to a pin that runs to two comparators, a logic output… it’s a 555 clone, isn’t it?
Definitely a TLC551Y from Texas.
The datasheet has the schematic on page 2, which matches exactly:
http://www.ti.com/litv/slfs044b
Nice shot, BTW !
It’s definitely a 555 timer, but it’s CMOS (or possibly BiCMOS, but looking again I think that’s unlikely). The three part resistor ladder from the 555 is implemented as 6 FETs in 3 epi pockets (two per pocket). The bond pads are (clockwise from the top left) DIS, THRESH, CV, RESET, OUT, TRIG, GND, VCC.
Notice the protection circuits on the input pins and the large FET output stages for OUT and DIS (push pull and open drain, respectively).
The schematic in this datasheet matches up pretty well: http://www.sophphx.caltech.edu/Physics_5/Data_sheets/tlc555.pdf
So I will guess it is a TI TLC555.
Google has some interesting ideas :-)
https://www.google.com/search?sa=X&hl=en&site=imghp&tbs=simg:CAESZBpiCxCo1NgEGgIIAwwLELCMpwgaOwo5CAESE4YHVI8B4gOEB5IH4wOHB4UHmwEaIMXKTLBmdsjV-LjF-E9quVnNVomToQWGsKPKppB8irS3DAsQjq7-CBoKCggIARIEJqdcQgw&tbm=isch&ei=DJOmUouGN8jtlAW20YGICw&ved=0CCkQsw4&biw=1600&bih=815&dpr=1.2
the pads:
– I can identify 4 pads (upper center, upper right, right center, lower left) with similar circuitry, maybe inputs?
– Then there are two outputs (upper left and lower center) with a larger grid structure, maybe those are driver circuits? many transistors in parallel?
– The center left is connected to the border, could be ground and the remaining pin needs to be VCC
The overall circuit is fairly small, so my guess goes to some dual opamp
1997 is way too late for a 555 clone and I agree it looks like it’s meant to handle some power. There isn’t enough symmetry for a dual opamp. I’m going to take a wild guess and say a small switcher like the MC3063 type. Given all eight pins and the date, maybe the NCP3064?
Agree. It has to be a digital part that drives some power. There are not enough transistors to make that a linear regulator or an op amp.
Maxim MAX232 or similar?
It certainly is a TI die:
http://visual6502.org/images/3843/TI_UC3843P_20x_1b_die_shot_1600w.jpg
The 1997 font and (M) give it away. At first looks, it looks to be a dual something, but the logic in it seems to converge towards the left of the chip, indicating that it does one combined function.
I’ve looked through the analog chips produced by TI in 1997, and found the UC3708 dual non-inverting power driver. It has dual outputs, and has some input logic that controls them both together.
Texas Instruments TLC551Y timer.
This is a NE555 compatible part, revised on 1997 (as the date on the mask).
Page 2 of the datasheet shows the layout of the chip, which matches exactly:
http://www.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=slfs044b
Datasheet also revised in 1997.
TI layout style very distinctive, many diagonal and round shapes.
I agree – it’s a fun die to look at, all nicely laid out in your face:
– the two things with the 6 circles (bipolar transistors) each in the opposite corners are the 2 comparators
– the box at the top left is a single DISCHARGE N-FET,
– while all the structures in the bottom right quadrant are 3 chained cmos inverters (N/P-FET pairs), each bigger than the other driving the TTL/CMOS output.
– The 3 resistors are in the mid left (GND end on the left, VCC on the right)
– the SR-flop is largely under the resistors
– All the input pads have latchup protection
actually I take it back they’re not bipolars, just very carefully matched FETs – in carefully bounded wells
I also thought initially those were bipolars but then realized later they were FETs.
Any ideas as to why circular, vs. doing say, a square or octagonal annulus, or even just two parallel banks of conventional horizontal and vertical finger transistors?
Well the goal here has to be to make as near identical transistors as possible – I’d guess that circular objects might make the best results across all corners of the process – I can imagine process variations that might occur linearly across a wafer that might sort of cancel out no matter which direction they occur in
re: circular gate.
-Smooth contours (no sharp corners) yields higher breakdown voltage.
-low stray capacitances = high bandwidth
-improved radiation hardness
I’m guessing a power amplifier for a mobile phone. The layout doesn’t make too much sense, but it could be a dual band or so.
Hello, i’m pretty interested in supporting this laptop for GNU.
Also, who’s holding the laptop in the picture, i’m interested in knowing, she has quite an exquisite rack.
Dr Richard Stallman
President, Free Software Foundation
51 Franklin St
Boston MA 02110
USA
http://www.fsf.org http://www.gnu.org
Skype: No way! That’s nonfree (freedom-denying) software.
Use Ekiga or an ordinary phone call
Having read this I believed it was very enlightening. I appreciate you taking the time and energy to put this content together. I once again find myself personally spending a lot of time both reading and leaving comments. But so what, it was still worthwhile!
I think it’s the truth.