The first mass-produced DRAM of the Soviet Union
cpushack.com
cpushack.com
wikipedia says Toshiba invented DRAM cells in 1965, and Dennard invented MOS DRAM in 1966
also:
> Elektronika BK-0010 is one of the first 16-bit personal computers in the world.
The TI/99 was actually the first 16-bit personal computer, predating this by 5 years.
I really can't trust anything else this article has to say.
After the DRAM made with discrete capacitors that was used in the Atanasoff-Berry electronic computer, the next invention was the Williams–Kilburn tube, invented in 1946 at the University of Manchester, which replaced the array of discrete capacitors with a cathode-ray tube. Such memories with CRT's have replaced the slower memories with delay lines, before being replaced themselves by the memories with magnetic cores.
For example, the first commercial IBM electronic computers, IBM 701 (1952) and IBM 702 (1953), had such DRAM made with CRT's, while their successors from 1954 had static non-volatile memories made with magnetic cores.
After more than a decade of neglect, the DRAM has been revived in the form of MOS integrated circuits, after the inventions by Toshiba and Dennard. Neither of them has invented DRAM, but they have just invented new methods of implementing the capacitors required by a DRAM, by using transistors.
According to Wikipedia, Dennard has invented at IBM the one transistor memory cell, which has been used in later DRAM products, not in the first generation launched by Intel in 1970 and followed by similar products from other US companies and from other countries, like the Soviet Union IC discussed in the parent article. The first Intel DRAM was based on a less efficient 3-transistor memory cell invented at Honeywell.
The earlier Toshiba invention was for the first kind of DRAM using bipolar transistors, which was used only in their calculators, while the later Dennard invention was for the improved MOS memory cell, from which all later DRAM cells were derived.
The actually unique feature of the Elektronika is that it is PDP-11 compatible though.
https://en.wikipedia.org/wiki/Electronics_industry_in_East_G...
Fundamentally the Soviets doctrine of only cloning meant they were fucked to begin with. MVII started development in 1982, came out in 1985, K1820 could only start by 1986 and was finished by 1990.
Being bankrupt was one factor in allowing the wall to fall peacefully.
https://www.cpushack.com/2023/12/21/the-first-mass-produced-...
It does not make contact with any pins. The ceramic package has a depressed cavity with a grounded metallized base, on which the semiconductor die is soldered. So the ceramic body is made of at least two layers that are bonded one over the other. On the inferior layer, there is the large pad on which the die is soldered and many small pads that are connected to the pins by metal traces, which remain enclosed inside the ceramic body after its layers are bonded. The pads on the semiconductor die are connected to the pads on the ceramic body by wire bonding, typically with gold wires. The die, the pads that lead to the pins and the bonding wires remain at a lower height than the upper surface of the ceramic body.
Then the metal lid is soldered on the ceramic body, closing hermetically the cavity, without touching the wires or the pads. All this is done because no plastic (i.e. polymeric material) is completely hermetic to gases, so if there are contaminants in the air they will pass slowly through the package, reaching the semiconductor die. Later, there have been developed methods of depositing protective layers directly on the semiconductor die, which have diminished the need for hermetic packages, at least in most milder environments.
The actual DRAM integrated circuit is only the one from the top of the picture.
The other 2 packages are like the one from the top, but they appear to have an additional plastic cover put over the package, to prevent touching the pins.
That plastic cover might have been intended to avoid the damaging of the IC by technicians who did not use proper ESD protection measures when handling the IC. Perhaps the cover is not from conductive plastic, but from metal (though that would be an extravagant expense), but in any case it will be removed when the integrated circuit will be inserted in a PCB or in a socket. Its purpose is only avoiding electrostatic discharge damage during the previous handling.
On the IC from the top, the only one where the metal lid and the pins are visible, it can be seen that the top of the ceramic body had a metal trace that formed a closed square, on which the metal lid had been soldered. The lid covers a cavity with the silicon die, the bonding pads and the bonding wires.
As seen in the photo, the square trace used for soldering the lid is connected by a metal trace to one pin in the corner of the package, which is the ground pin, ensuring that the metal lid is grounded.
In the middle of the ceramic body there are similar metal traces that connect the bonding pads to pads on the lateral side of the ceramic package.
The pins of the IC are brazed to the pads on the lateral side of the package, before soldering the silicon die.
The image is not clear, because the lateral side is in a shadow, but the lateral side has one metallized pad for each pin that will be brazed over it, and the metallized pads are separated by non-metallized ceramic surface, so there are no short circuits.
I assume this cover is supposed to be removed after installation. So I think you are right in assuming it's purpose is there to avoid ESD damage during handling.
Wasting so much copper sheet for a cover that will be dumped during the PCB assembly is not wise, but in the Soviet Union they were rather careless about wasting materials, so this would not be unusual.
For other early MOS ICs or transistors, which were similarly sensitive to ESD damage, the vendors did not use so big covers, but only a metallic clip made of thin elastic wire, which was applied on the pins and which pressed on them until removed, due to its elasticity, providing thus a short circuit to avoid discharges.
However any such metallic shorts raised the cost, so they were replaced with delivering the ICs either contained in tubes made of conductive plastic or inserted in black sponges made of conductive foam.
I agree, it does look like it's shorting out all the pins. Maybe it's intentional, to avoid any weird currents or whatever forming while the chips are not installed? Possibly an early form of ESD protection?
Are any pictures available? I would like to know how @ could look like a dog
At first, I was thinking it might be power, so that A and B should be the same but calculated differently, but the final output shows that B is a bit less than double that of A. Or maybe it's just losing more precision because of log conversions, but I'm surprised that would error towards a bigger number. FWIW, I don't think either answer is correct, so I think the program is just demonstrating different rounding errors due to loss of precision.
Dog is just one example, there is also a monkey, a mouse, a worm, etc.
[1]: https://en.wikipedia.org/wiki/At_sign#Names_in_other_languag...
This one is pretty cool!
~___@
/\ /\
"a with a long tail curling upwards, like that of a Laika dog[1]"
[1] https://www.wisdompanel.com/en-us/dog-breeds/east-siberian-l...