MOSCOW
Home
to Iskra 112
Museum of Soviet Calculators on the Web


Stats...



1972
1974

size

365mm
305mm
120mm

10kg


?


220V, 50Hz, 30W


Factories...

factory mark



Iskra 111/111I/111M/111T

This russian calculator uses Nixie tubes and 13 PCBs full of diodes and simple ICs. There are several variants known, including the 111, which uses square ICs, and the 111M which uses DIP ICs.

A copy of this machine has also been found, used in an East German company (VEB Volkseigener Betrieb) and made by Elektron Org Technika. One thing about this German machine is really crazy: It has 12 Nixie Tubes which all work for calculations but through the window in the cover only the 10 Tubes on the right side are visible. Two on the left side and the lamp for the minus sign can be seen only without cover!

Sergei Frolov points out that... "...the calculator Iskra 111M with fluorescent tubes. These tubes have a smaller size, than nixie, and consequently the designers have brought in reasonable change in design of a cover... the size of the window for digits occupies a place just for show 10 nixie tubes. Most likely that ...the calculator Iskra 111 with a cover from that fluorescent Iskra 111M."

It is unlikely that the machine was produced with the display digits covered. This appears to be another instance of a machine which has been modified with non-original parts after a repair. See the B3-25 for another interesting example of this.

© Sergei Frolov
Iskra 111M
Iskra 111M has a single column of keys at the left.

The circular loop you see in the picture lower on this page is the delay-line used for storing all of the calculator's internal registers. This loop stores 240 bits of information, which was enough to store the machine's 5 registers (each 12 decimal digits + mantissa). How? It's quite fascinating; each bit circulates along the loop in the form of a sound wave (or equivalent). The circuitry is quick enough to send 240 pulses before the first one arrives back again. The data circulating in the loop is constantly refreshed. Delay-line memory was also quite common in the very early Western computers and calculators. I like to think of it as 4th-dimensional storage; using time as a basis for storing data!

© Sergei Frolov
Iskra 111I
Iskra 111I has two columns of keys at the left.

© Sergei Frolov
key usage
Touch calculating, anyone?

Magneto-restriction Delay Line Memory!

Translated by Sergei Frolov, from the book "Maintenance and repairing of electronic keyboard computers" - you don't have to understand this; it's way over my head! Just nod occasionally while reading, and mumble "Mmmmh, interesting..." when you're finished

Magneto-restrictive delay-lines (MDL) are simple to construct, and are produced from cheap materials. The wires used to propogate the sound waves, formed in loops, can be packaged in such a manner that even at large lengths they will occupy a small area. Electronic circuits using MDLs for storage are simple and can perform the same function as semiconducting memory. The MDLs can be manufactured with propogation (loop) delays from units of microseconds and up to tens of milliseconds. The pulse repetition rate can be boosted up to 5 MHz. As the delay increases, the frequency of signals diminishes, so that it is possible to store in one delay line no more than 5 thousand bits of data.

© Sergei Frolov
Iskra 111 delay line memory
The remarkable delay-line memory.

MDLs can work using longitudinal and torsional oscillations. The use of longitudinal oscillations for the energization require converters of a simple construction, and allows devices with continuously adjustable delay. MDLs using torsional oscillations have lower speed of distribution, allowing the use of short sound-wires. Additionally, the torsional oscillations have no a dispersion stipulated finite width of a sound wire, and therefore lines on torsional oscillations have higher resolution and provide larger delay capability. The winding of a sound-wire in a spiral is coupled to appearance of a dispersion and in torsional oscillations, however at observance of the defined technology of mounting of a sound wire the phenomenon of a dispersion in most cases can be eliminated.

The basis of conversion of a delayed electrical signal to an sound (audio) signal is the phenomenon of magneto-restriction (MR). This is, that the ferromagnetic bodies undergo distortion when inside in a magnetic field. The sign (elongation or truncation) and value of maximum MR deformation depends upon the material used, and properties of its mechanical and thermal response to that field. The dependence of MR deformation on field strength is an even function; the sign of deformation does not vary at change of a direction of a field. Usually operating point with the help of a fixed magnetic biasing install on the most abrupt site of the MR characteristic.1

The receiving converter, in which the delayed audible tone will be converted to a signal electrical, operates using the phenomenon of return MR. This phenomenon is, that the magnetic inductivity of ferromagnetic materials varies under operation of mechanical power or deformation. In the case where the ferromagnetic material is in a magnetic field, under operation of mechanical power the magnetic field varies. In calculators, the lines on torsional oscillations are used, but at energization of these oscillations the method of conversion of longitudinal oscillations in torsional is used.

© Sergei Frolov
Figure 1.

For more complete understanding of MDL, consider the device for energization and reception of longitudinal oscillations, Figure 1. This figure shows a simplified MDL for longitudinal oscillations. It consists of a sound-wire [1], on which the entry spools [5] and output converters [6] are attached. The spools of the sound-wire should have MR property. The spools are supplied with fixed magnets [3] and [4], whose fields become isolated through zones of conversion of a sound-wire. If on a spool of an input translator to skip a current of a delayed signal, the magnetic field, established by it, due to the phenomenon of MR deforms a zone of entry conversion. This deformation as a wave of longitudinal oscillations will be distributed to the right (direct mode) and to the left (backward wave) from an input translator with speed of distribution of longitudinal oscillations. The direct mode, having come up to a zone of an output translator, due to return magnetostrictive effect will change a magnetic inductivity of a zone of conversion. And as the zone of conversion is biased by a fixed magnet [4], a magnetic flux, varying at it, will be inducting in a spool [6] output translators of EMF of an output signal.

If not to undertake of any measures, the part of a ultrasonic signal of a direct mode, past through an output translator and signal of a backward wave, having mirrored from the end of a sound wire, will be excited in an output translator by(with) spurious signals. The level of these signals will be commensurable with a level of a useful signal, and the delay factor will be determined in length of paths of distribution of these signals. To prevent this undesirable phenomenon, the test leads of a sound-wire are supplied with special devices [2], whose operation is based either on absorption, or on mutual compensation reflected from test leads of a sound-wire of signals. Let's consider energization of torsional oscillations in MDL. There are two ways of energization of torsional oscillations in a cylindrical sound-wire with the help of MR effect. In the first way longitudinal oscillations in an auxiliary MR bar in the beginning are excited, and then these oscillations are used for creation of the spinning moment in a main cylindrical sound wire. At the second way ˜ the torsional moment in a magnetostrictive zone of conversion of a cylindrical sound wire forms at the expense of effect of a variable magnetic field of the screw configuration. Let's consider only first way, which is applied in operating in the calculator MDL.

© Sergei Frolov
Figure 2.

In Figure 2 the converter operating on the first way is figured. The converter represents two thin bars [3] of a magnetostrictive material hardened concerning to to a cylindrical surface of a sound wire [2]. The magnetostrictive bars are supplied with drive windings [4] and magnets of offset [1]. At passage on drive windings of a current of an input signal in bars due to magnetostrictive effect there will be longitudinal oscillations, which in turn, will create a torque in a site of a cylindrical sound wire. A rule(situation) of drive windings on bars, direction of a signal current in drive windings and polarity of magnets of offset are selected so that the operations of bars on sound wires added in a phase. Thus, in a considered(examined) converter not only the electrical signals will be converted in ultrasonic, but also there is a reconversion of ultrasonic oscillations in electrical. The magnetostrictive bars create a torque on length of a sound wire equal bandwidth. The impulse in a sound wire is lengthened approximately on this value. That the backward waves were not mirrored from free test leadss of magnetostrictive bars, last are supplied with absorbers.

2.2. Device and operation MDL

Let's consider in more detail device and operation MDL, used in keyboard computers of a series "Iskra". MDL consists of the following main sites (fig. 19): a sound wire; an input translator; an output translator; control unit. Sound wire represents an elastic wire by a diameter of 0,5 mm from a material 40 KHXM, stacked by a two-row cylindrical spiral in slots of the reinforced elastic headers manufactured of ink KLSE with hardener K-1.

Notes

  1. The meaning of this sentence is unclear to me, so left in original (machine) translated form.

Western Machines with MDL

Rick Bensene's fine site "The Old Calculators Web Museum" presents some Western machines which used delay lines for storage.

Here's some of his write-up of the Friden 130. I do recommend you visit Rick's site, because it has some great information about all of these old machines.

    The Friden 130 uses transistor and diode-resistor gate technology. It performs all operations in bit-serial form, using a recirculating acoustic delay line as the medium for storing its working registers. This is a very interesting method of providing working storage registers for a calculator. Back in those days, transistors were rather expensive. It takes at least two transistors to make a flip-flop, which a 1-bit storage register. With 13 digits to store, and each digit taking 5 bits (the Friden machine used a rather odd 5-bit representation of each digit) that means that there would have to be 65 flip flops, or a minimum of 130 transistors, to store one register in the stack, and the 130 has 4 registers in the stack, plus one for the memory register. This would have taken over 600 transistors alone just to provide the storage for the registers, which would have been prohibitive both in terms of cost and space required. So, the engineers at Friden leveraged technology used in early computers to store the bits. Old electronic computers, before the advent of ferrite-core magnetic memory devices, used tubes filled with mercury with transducers at each end. The 'bits' ended up taking the form of sonic disturbances which propogated through the mercury at a fixed rate. The bits were sent through the mercury a bit at a time in serial fashion, and were constantly circulated through the tube like a big shift-register. When bits were needed, they were 'spihoned' off into the arithmetic unit bit at a time, where the appropriate operations were performed, and the results pushed back into the bitstream circulating through the mercury. The Friden machine used a similar method, but rather than using exotic materials like mercury, a special type of wire was used which held the bits as slight 'twists' (torque variations) in the wire that move along the wire from one end to the other. A transducer at one end of the wire places a torque pulse on the wire to create 'twists' which are registered at the other end of the wire by a similar transducer. By continuously circulating the 'bits' through the wire, the wire becomes the storage medium for the bits, and far less circuitry is required to maintain all of the bits that the machine needs to operate. In the Friden 130, the wire takes the form of a loop of wire carefully strung inside an enclosure that takes up most of the bottom of the chassis. - Copyright © Rick Bensene












Stats...



1979

size

?
7-8kg


?


?


Factories...

factory mark

This is possibly of Bulgarian manufacture.


Notes...

This calculator reported by Frank Boehm. You can see this, and many more calculators, on his web-site ELEKTRON.




Iskra 111M

The Iskra 111M is an updated version of the Iskra 111, using integrated circuit (IC) memory instead of the delay line, for storage of registers.

Iskra 111(c)Frank Boehm

(c)Frank Boehm

(c)Frank Boehm

(c)Frank Boehm













Stats...



?

size

?



?


?



Iskra 111T

© Sergei Frolov
Iskra 111T


MOSCOWMuseum of Soviet Calculators on the Webe-mail webmaster
Home
Home
to Iskra 112
Iskra 112


The content of Museum of Soviet Calculators (on the Web) is copyright © 1997-2001 Andrew Davie & contributors.
Unless indicated to the contrary, permission is granted for private non-commercial use of images and text.
Last modified on