This is a small analog computer from 1968. The name "Tyrotek" is in small letters on the front panel, but there is no model number and no identification plate with a serial number or location of manufacture.
I received the unit through University of B.C. surplus sales around the year 2000. The smaller size and level of functionality provided suggest it was meant for the educational market. The electronic componentry is of high quality, suggesting it was intended for 'serious' teaching. Based simply on size and the front panel this computer would appear to be comparable to the Systron-Donner SD-3300 Educational Analog Computer from around the same period.
In early 2011, sometime after mentioning this unit on the web, I received a message from the son of the founder of Tyrotek. Apparently Tyrotek was a small California startup company. Jim Pepper, the founder, subsequently provided the following about the origins:
At the time I built this computer I had been laid off from a company that made analog components. I was formally employed by Beckman Instruments as a design engineer for large scale analog computers used for aircraft design. I decided to try making a small scale computer for schools. I had one other person with me who also worked for Beckman. We managed to sell 6 units. Unfortunately the funding for schools was cut from the government spending and we stopped building them. We had only sold 6 computers, two going to the University [of British Columbia] in Vancouver.
Another of Jim's sons recalled working on assembling the units:
My father, Jim Pepper (and a partner or two) assembled these computers initially in the garage of our family home in Orinda, California. In fact my brother and I actually assembled the PC boards that went into the unit.
As is typical with surplus equipment, there was no manual with the unit when I received it.
The unit has been reverse-engineered to produce a schematic.
The technical information presented here has been deduced from examination of the unit and the schematic.
Additionally, a few sources on the web provided some clues about the general practice around analog computers in the 1960's.
|Contents (this page):||Schematic (new window):||Sub-pages:||
The computer is based around a plugboard using 'standard' IBM plug-wires. The plugboard itself unplugs from the computer to permit easy change to another program. Two plugboards were received with this unit.
The plugboard is divided into four identical sections, each section being further sub-sectioned into a 2 by 3 matrix. Each of the 6 sub-sections is a functional unit to the programmer:
In total then, the model can accomodate up to 20 amplifiers. This particular unit has modules for two sections installed, thus 10 amplifiers: 4 summers, 4 integrator-summers and 2 multiplier-divider-summers.
The photo at right shows two sections or one-half of the plugboard (click on photo for a closer view). A simple test program to sum two values (the values being determined by switches and potentiometers) is wired up.
Along with the contacts for the functional units, there are contacts for the coefficient potentiometers, some switches, reference levels, and terminals on the back panel.
The following table presents descriptions of the plugboard contacts.
Operation of the multiplier-dividers has not been determined in depth.
|General & Summer:||X1||Times-1 summing input.|
|X10||Times-10 summing input. The input value will be multiplied by 10.|
|SJ||Sum junction (op-amp input).|
|Integrator:||[integral]||Jumper to the output to configure the amp as an integrator.|
|[sum]||Jumper to the output to configure the amp as a summer.|
|IC||Initial Condition input for an integrator.|
|N||Direct connection to the amp input.|
|Multiplier-Divider:||M||Jumper these contacts for multiplication.|
|D||Jumper these contacts for division.|
|FB||Feedback. Jumper to the output if the amp is to be used for multiplication or division.|
|[root]||Jumper these contacts for square root, along with jumpers for division. i.e.: solve |
|Miscellaneous:||+R, -R||Positive and negative reference levels (±50V).|
|P0x||Two or three contacts for each of the front panel potentiometers.|
|S0x||Three contacts for each of the front panel switches.|
|Txx||For connection to terminal points on the rear.|
Associated with the meter is a precision (0.15% linearity) wire-wound 10-turn potentiometer with a calibrated knob, referred to as the 'Reference Divider', and a Null switch. These can be used to more accurately determine an output value by seeking a null point and reading the knob calibration.
There is also a row of neon lamps, each op-amp in the unit connects to one of the lamps. The lamps function as 'overflow' indicators, to indicate that an op-amp output voltage level has exceeded it's useful maximum.
I will guess that an analog graph plotter was expected to be used as the primary output device. There are some terminals on the rear of the unit which would provide for connection to an external device.
Technology / Implementation
To implement the functional units, printed circuit board modules are mounted in the mainframe behind the plugboard receptacle. The modules have contacts which mate with the plugboard plugs. Each module implements two sub-sections (one column) on the plugboard. As such, there are two types of modules to correspond with the labeling on the plugboard:
One of these modules is shown on the left in the photo of boards.
The multiplier-divider is based on a time-division chopper circuit. The Y input controls the duty-cycle of an oscillator operating around 120KHz. The oscillator drives two MOSFETs driven in complement and configured into a variable-gain feedback path around an op-amp, along with the X input feeding into the op-amp.
The boards in the middle and right in the photo comprise one of these modules. The board in the middle contains the op-amp. The daughter board on the right contains the multiplier-divider circuitry.
In all cases the basic op-amps are essentially identical. All the circuitry is constructed from discrete components. JFETs are used for the op-amp differential-input pairs.
At the time this unit was designed or produced, the first IC op-amps - the Fairchild 702 (1963) and 709 (1965) were available - and the National Semiconductor LM101 (1967) was just being released.
Use of such ICs in this application would have required numerous discrete components to bring them up to the abilities of the discrete op-amps it uses. There would have been little or no benefit to the use of the relatively expensive ICs.
Tyrotek Analog Computer