The Ware for June 2011 is below.
This one should be tougher than last month’s ware.
This entry was posted on Wednesday, June 22nd, 2011 at 7:16 am and is filed under Hacking. You can follow any responses to this entry through the RSS 2.0 feed.
Both comments and pings are currently closed.
Is it a TRON race-track?
Is it an early military-packaged IC which has been de-lidded, like the SN5451?
Not sure what’s around the outside, encapsulant or glue?
Looks like an early IC …
Symmetry makes me think of some dual logic gate… Reminds me of these early dual nand gates nor… Apollo guidance computer IC ?
Just the look of it reminds me of a differential amplifier. Lots of traces very close together, so I’m thinking FET based.
Very early IC, probably early ’70s or so. Bonding wires look like they were done by hand (lady w/microscope & bonder), before the advent of vision-based bonding robots.
I would guess a logic gate of some sort. First time I’ve ever seen intra-chip bonding wires on a single piece of silicon! I guess this was done in a process that only supported a single metal layer.
It does look relatively old, especially with the wire-bonding that goes point-to-point without hitting a pin and the huge process size.
The silicon has near-perfect left/right symmetry, but signals aren’t pulled out to pins in any way that makes me think differential signalling. If it’s a dual gate IC, I’d bet on intput/output pairs on pins 1/3 and 5/4, with pin 1 on the bottom-left. I also find it interesting that the top-left two pins connect to structures on the left-hand side whose duals on the RHS aren’t wire-bonded to anything.
Beyond those observations, I don’t have much to say, unfortunately.
Looks like a 1960s IC something like a triple input nor gate. Perhaps as used in Lunar mission?
Given the vintage look and the central use of pin 3, I’m going with a TI 51 series (Pin 3 is VCC). I can’t quite decide which one, but it seems like it could be an SN510A, R-S Flip-flop/Counter. I choose this one given that it has a slight asymmetry.
If I’m right about the series, hopefully someone can more closely guess the actual circuit.
SN514B? first commercial TI IC
Check out this link: http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19640015793_1964015793.pdf
Fig 4 (note the 2-9, 4-8 connections are external) – in the die photo I think pin 1 is top right – I think there are a few broken bonding wires (1&2)
TI SN7400 series in flat pack package. Maybe a flip flop (7473/7474).
The package seems to be a DQFN14 as is seen in this link:
http://ics.nxp.com/packaging/dqfn/
Seems to be a logic circuit with two equal blocks.
Looking at the 74HC74 datasheet, the DHVQFN14 pinout match the photo pin layout.
For example, the 1D input at pin 2 is simetrical of the 2D input at pin 12. And the central conections 1CP (pin 3) and 2CP (pin 11) match with a clock input in the die photo. There is some broken pins that makes it a bit difficult to follow.
I guess a 74HC74 in a DQFN14 package.
I’m not sure I agree with a DQFN package. Yes, the leads come in from the right places, but that’s probably common for all lead frames regardless of package.
I’d also bet against any of the 74-series logic. I don’t see the repetition I’d expect from a 7404 or 7414, the pairings I’d expect from 2-input gates, or the mirroring I’d expect from a dual-anything.
In fact, the connections just look strange to me. If we number the pins clockwise starting with the upper one coming in from the left side of the frame, there don’t seem to be any connections at all from pins 1 and 14.
Pins 2 and 3 connect to the features in the upper-left quadrant, and those features cross-connect to the symmetric features on the other side.
To me, that looks like device matching. Silicon is anisotropic: electrons behave differently when they move left-to-right versus up-and-down.. or right-to-left for that matter. There can also be gradients in the doping from one side of the chip to the other. The standard way to average out the errors is to divide devices into modules and arrange the modules around a common centrode. Having one pair upper-left/lower-right, and another pair lower-left/upper-right is a common pattern.
That leads to some assumptions: First, if you care about matching, you’re doing analog. Second, given that the two sets of features are arranged around a single centrode, I’m guessing the whole thing has to be matched. Third, since pins 2 and 3 connect to opposite sides of the matched features, I’d guess that we’re looking at some kind of differential signal.
Given the symmetry of the layout, pin 4 is my favorite candidate for Vcc. The pin is one of the usual suspects for power coming in, and the central feature is ‘up’ relative to the features on both sides.
Pin 5 connects to the large feature that spans the whole chip. Given its size, it’s either something that moves a whole lot of power, or something that needs to be big, like a capacitor.
Pins 6, 7, and 8 are just confusing.. I see what look like two leads from 6 and 7 that go off the right side of the image, and the only thing likely to be there would be the lead for pin 8.
Pins 9 and 13 connect to large and symmetric features below the probably-matched features. Signals that go several places tend to be more general purpose, like enable/disable, or bias. It doesn’t make sense to disable half of a matched transistor, but it might make sense to allow for offset compensation.
Pins 10 and 11 are my favorite candidates for input, for a handful of reasons: First, they connect to symmetric small features, and if big means general, small means specific. Second, there are dots that look like test points near both features, and it makes more sense to put test points near the input than the output. Third, the connections at pins 11 and 12 look awkward.. swapping the attachments would preserve symmetry, so the designers probably had a reason to keep those two pins together.
Pin 12, I dunno.. it connects to another feature at center, so it’s probably a whole-chip kind of thing, but I couldn’t tell you what. For want of a better choice, I’m going to call it GND.
So overall we have what’s probably an analog chip, two inputs, two outputs, probably differential output, possibly with compensation adjustment on either side, and one pin for an external programming component, possibly a capacitor.
I’m guessing a dual-output op amp that’s either uncompensated, or which offers the option to tune the compensation.
It could be a SN51x
The die photo is similar
http://www.vintchip.com/FLATPACK/DATA/TI51SERIES.JPG
but none of them has 14 pins, judging from the photos od every chip http://www.vintchip.com/FLATPACK/TEXASINSTRUMENTS.html
pin1 seems to be on the left (the lowest horizontal one)
pin 3 & 11 are power supplies, probably. Should rather be 4 & 11 !!
the power supplies don’t match SN5470/72
the left and right bonds are lacking, perhaps? damaged ? NCs ?
There are 4 “bond jumpers”. The crossed ones indicate a flip flop.
So my conclusion is simple : this is a TI SN51x chip, that has been custom bonded in a 14pin package (instead of 10) to provide a special logic function, probably essentially a single flip flop.
Very interesting catalog (NASA compilation) from 1968 : http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19680020054_1968020054.pdf
A lot of people are guessing the TI SN51x series, but I suspect it was slightly later than that since it has 14 pins rather than 10. Probably an early (1965?) device in the 5400 series, probably a flip flop. It’s likely that the same die was used for multiple devices, but having only one metal layer, it was cheaper to wire bond to “configure” the device than to have several possible metal masks.
There are broken bond wires which makes it hard to tell what pins went to what. It does have the VCC and GND pins in the “middle” as you can tell by the distribution metal in the middle.
It looks like it’s sitting in the original yellowish (non ESD!) foam that would have been used to ship it as a sample, rather than the plastic carriers that it would have been shipped with for production. You can see the reflection of light off the plastic lid of the sample case. It probably looks like this: http://www.vintchip.com/FLATPACK/texasinstrumentssn515.html.
some BBD-bucket brigade device or magnetic bubble test device.