Get best of the week

We look inside the Soviet integrated circuit with TTL

Translation of an article from Ken Shirriff's blog

In this article we will examine the chip of the 1980s used in watches for the Soyuz spacecraft. The photo shows through a microscope a silicon crystal inside the case with a clear geometric layout. Silicon in the photo looks pinkish-purple, and the layer with metal conductors - white. Along the edges of the chip, connecting conductors (black) connect the chip pads with its contacts. Tiny structures are resistors and transistors. IS 134LA8 on logical gates AND NOT The chip is used in the clock in the photo below. Recently, this watch flying into space on a Soyuz spacecraft (it is not known which flight - judging by the marking, the watch was made in 1984) came to our museum




. The upper left display shows the time, and the lower timer. The “alarm” function activates the external circuit at the set time. At first I thought that this watch would have a single chip inside, but it turned out to be unexpectedly complex, containing more than 100 ICs on ten boards.



The printed circuit boards of the clock open in the manner of a book, after which the IC and other components become visible - this allows you to make flexible fasteners for the wires connecting the boards. Among ICs, the most common are 14-pin flat chips in a metal housing with surface mounting. I wanted to learn more about these ICs, so I opened one of them, took a picture and reverse-engineered its circuit (do not worry, we did not destroy the chips from the clock - we just bought similar ones on eBay; it was unexpectedly easy to find).


The wiring harnesses are arranged so that the boards can open. A quartz crystal serving as a timer is visible at the top center. The power is located on the boards on the right, with several round inductors.

Soviet integrated circuits


The clock is assembled on an IC with TTL - this digital logic was popular from the 1970s to the 1990s, as it was reliable, inexpensive and easy to use (if you were engaged in amateur electronics at that time, you probably know the 7400 series ). The simplest TTL chip contained only a few logic gates - for example, 4 NAND gates or 6 inverters, and more complex chips could implement such functional modules as a 4-bit counter. As a result, TTLs gave way to CMOS chips used in modern computers that consume less energy and have a higher density.

The photo below shows the chip with the metal cover removed. In the middle, a tiny silicon crystal is visible, and connecting it with the contacts of the conductor. This is a fairly small IP - the case dimensions are 9.5 mm x 6.5 mm, noticeably smaller than the nail. To open such a chip, I usually put it in a vice and then hit the joint with a chisel. However, in this case, the chip opened itself - while I was looking for a hammer, the lid suddenly popped off due to the pressure exerted by the vice.


IC with a metal cover

removed. Chip marking - 134LA8 0684 (134 - low-power chip, L - logic, A - NAND gate, 8 - subtype of this category, 0684 - manufactured in the 6th month of 1984). It implements four NAND open-manifold valves.. The NAND gate is a standard logic gate, issuing 0 if both inputs are 1, otherwise - 1. The output of the open collector is slightly different from the standard.
In case 0, the voltage at the output contact will be low, and in case 1, it will be floating (“high impedance state”). An external pull-up resistor is required to tighten the output in case of result 1. Three such chips are used in the watch: one in the circuit with a quartz oscillator, and two in the role of inverters in other parts of the watch.


Logic 134LA8

According to the CIA, the USSR lagged behind the United States in the development of IP by about 9 years. And the lag would be much greater if the USSR had not copied many Western IPs. As a result, most of the Soviet TTL chips have Western equivalents. However, the 134LA8 chip I studied differs from the western ones in two features. Firstly, to reduce the number of external resistors on the chip, there are two pull-up resistors that can be connected as you like. Secondly, the chip has two common input pins, which frees up the two pins used by the resistors. So, although the USSR copied IP, he also creatively developed his own chips.

IP Components


Under the microscope, the components of the IC, transistors and resistors are visible. Sections of a silicon crystal, depending on impurities, have shades of pink, purple or green. By mixing other materials with silicon, it is possible to change its semiconductor properties to obtain n-type and p-type silicon. The white lines on top are metal paths connecting the components of the silicon layer.

The photo below shows a resistor on a silicon substrate. The resistor is formed by the addition of impurities to silicon, generating a path with high resistance - this is a reddish line in the photo. The longer the track, the greater the resistance, so resistors are often made in the form of zigzags to get the desired resistance. The resistor is connected to the metal layer on both sides, and the other track passes over it.


IC

chip resistor This chip, like other TTL chips, uses bipolar npn transistors. These transistors have an n-type emitter, a p-type base and an n-type collector. In IC transistors are made by adding impurities to silicon, forming layers with different properties. At the bottom of the stack, the collector, with the help of additives converted into n-type silicon, forms the majority of the transistor (large green area). Above it is a thin p-type silicon region forming the base; this is a reddish patch in the middle. Finally, a small n-type emitter rectangle is formed above the base. These layers form the npn structure. Note that the metal connection of the collector and base is on the side of the main part of the transistor.



TTL circuits usually used transistors with several emitters, one per output, which can be seen above. Such a transistor may seem strange, but it is quite simple to do in the IC. The transistor above has two emitters connected. If you look closely, you can see that there are four emitters, and unused ones are shorted to the base.

The output transistors on the chip emit an external signal from the chip, so they must support much higher currents compared to others. As a result, they themselves are larger than other transistors. As before, the transistor has a large n-type collector area (green) with the base higher (pink) and with an emitter at the very top. The output transistor has long contacts connecting the metal layer and silicon, instead of small square contacts, like the previous one. The emitter (with a conductor in the form of U) is also larger. This allows more current to pass through it. In the photo below, the transistor on the left does not have a metal layer, so its details are easier to consider. The transistor on the right shows metal conductors.



How does a TTL AND-NOT valve


The diagram below shows one of the NAND logic gates with an open collector. To understand how the circuit works (the most detailed description of the work can be found here), first suppose that it receives 0. It flows through the resistor R1 and the base of the transistor Q1 exits through the emitter of the transistor. Transistor Q2 will turn off, so R3 pulls Q3 base down and turns it off. Thus, the output will be floating (i.e., open collector 1 output). Now suppose that 1. is supplied to both inputs. Now, the current flowing through R1 cannot pass through the input, so it will exit through the collector Q1 (in the opposite direction) and into the base Q2, which will disconnect Q2. Q2 will pull the Q3 base up, including Q3 and pulling the low output voltage. Thus, the circuit implements the AND-NOT valve, giving out 0 if high voltage is applied to both inputs. Note that Q1 does not work like a normal transistor - instead, it “controls the current”, directing the current from R1 in one direction or another.



The diagram below shows the components of one of the NAND gates marked in accordance with the diagram above (the other three NAND gates on the chip are similar to this one). The wiring of the valve is simple compared to most ICs; metal tracks (white) can be compared with the conductors in the diagram. Pay attention to the winding path from the ground to Q3. Q1 has two emitters, and Q3 has a large output transistor. Two unused transistors are below Q2.



Conclusion


The Soviet chip in 1984 is simple enough to understand the operation of the circuit illustrating the design of the TTL valve AND-NOT. The downside of simple chips is that the clock from Soyuz took more than 100 chips to implement the simplest clock functionality. Even at that time, there were already chips that fully implemented the work of watches and alarms. Today, chips can contain billions of transistors, which is why they have a huge range of functions, but their work cannot be understood simply by looking at them.

Video in which CuriousMarc parses a space clock:

More articles:


All Articles