Forum Top Banner Ad


Ebay Classic organs



No announcement yet.

1980s digital organs - memory sizes?

This topic is closed.
  • Filter
  • Time
  • Show
Clear All
new posts

  • 1980s digital organs - memory sizes?


    I'm trying to understand the history; I'm new to organs, but very much not new to computers, I've been active since the late 80s.

    And so I'm trying to understand how the early digital organs worked. Like the Allen ADC 220 which my wife (the actual organist - I'm the techie) recently played, and it sounded very nice.

    It's just that I can't make this compute (pun intended) with what I know of memory sizes of the era. Somewhere in this Forum I found that the ADC 220 has 14 actual stops, so how would it have sufficient ROM to store all the samples?

    The standard for high quality sound in those days was 44k 16bit, "CD quality" - except unlike CD, you probably don't need stereo samples, as you work out stereo by positioning pipes. And you need at least a second of every note, right? So, 61 notes in a stop, pretty close to a minute, giving us 5 megabytes in this quality. Total storage required: 70 megabytes. And that if you do attack and release in pure software.

    That was the size of a 80s hard drive. How did an organ pack this much data in ROM? Or what tech did they use to reduce the ROM requirements?

    RAM and CPU power appear fine for the 80s, at least since the release of the 68020.

  • #2
    Early Allen digital organs used wavetable synthesis rather than recorded samples. One half of the stop's waveform was stored in ROM in either 7 or 8 bit resolution depending upon the model's generation. The complete waveform was produced by clocking out the samples forward for the first half of the waveform and then in reverse order for the remainder. A single waveform was used for the entire compass of the stop. This form of synthesis allowed Allen to store entire voices on a single punch card.

    Search for threads on Allen alterable voice cards for more detailed information.

    Allen 965
    Zuma Group Midi Keyboard Encoder
    Zuma Group DM Midi Stop Controller
    Hauptwerk 4.2


    • #3
      Wow, they produced really good sound for that sort of tech. I wonder - does a software implementation of this kind of synthesis exist, so one could get an "Allen sound" from a virtual organ?


      • #4
        That is a software implementation. :)


        • Admin
          Admin commented
          Editing a comment
          KC9UDX We can debate semantics forever. If you want to use your definitions, rather than those found in the dictionary and employed in the vernacular, you are of course correct.

          My objection to your statement is that these instruments were NOT software implementations as the term software is commonly understood. Why that distinction matters is that there are software implemented organs, VPOs like Hauptwerk, and classical instruments like the Viscount Physis, and others.

          These old Allens are vastly different. They're a collection of hardwired logic circuits. Allen called them Digital Computer Organs because the term digital and what it meant wasn't common and while people might not understand what digital meant, everybody knew, even then, that computers are magical. In the broadest sense of the meaning of computers, you could say that this collection of logic circuits is a computer, but it would not be considered a computer by contemporary standards. They have no firmware, they have no software, they are not executing a software program.
          Last edited by Admin; 06-13-2019, 08:40 PM.

        • Admin
          Admin commented
          Editing a comment
          ramendik jbird604 would probably know for sure, but I think it was the MDS line that introduced the software based console controller. Allen, and most electronic organ builders are secretive about their technology. I don't know what Allen is doing today, but I suspect that their sound engine is still more hardware than software based.

        • don60
          don60 commented
          Editing a comment
          I agree with Admin and would go farther to say that the Allen "Digital Computer" organ is not a digital computer even by 1950s standards. It is a synchronous sequential digital logic system.

      • #5
        I mean a software implementation available on modern PCs :) Allen is still selling its tone cards - would be nice to read them into a PC and get this sound from a MIDI console.

        EDIT. A google search turned up a thread on this very forum.


        • John Drabik
          John Drabik commented
          Editing a comment
          You could probably learn to write code to emulate the old logic circuits, using C, Python, or maybe even Java. For example, a common function was an "address counter" to walk up and down through a series of numbers that were used to access the sample memory. Early "digital" organs used a circuit board full of parts to do that - you could do that now, for dozens of software counters, on a simple controller today with just a few lines of code. A modern micro PC like a Raspberry Pi has far more than enough compute power to emulate pretty much anything on the older hardwired logic boards. For about $35 (on Amazon or eBay), such a machine also has enough RAM storage to hold a reasonable sample-set, although you could expand on that easily using solid state disk or flash cards. The biggest problem would probably be processing "jitter", i.e., slight timing variation caused by the multitasking operating system such cards usually (but not always) use. Some Raspberry OS alternatives include a real-time operating system (RTOS) with predictable latency and jitter. Or, you may want to consider one or more dedicated micro-microcontrollers like Arduino or ESP32 devices (available for $5 or less each. At that price, you could dedicate a micro per stop!) Those have no OS at all, have very predictable timing (they are often used for clocks or other accurate controls), and are a cinch to learn to use and extend with circuits.

      • #6
        The ADC series of Allen organs are based on the S-152 chip. This is the equivalent of what nowadays is described as an ASIC (application specific integrated circuit).
        These type of devices were originally called PLAs (programmable logic arrays) or PLCs (programmable logic controllers) and were only one time programmable. ASICs are re-programmable as they use flash technology.
        The most important attribute of these type of devices is that they can process multiple complex logic operations in a single clock cycle (hence why ASICs are used in Bitcoin mining)

        The S-152 is basically a cycling state machine that decodes the serial key-press and stop-tab data streams.

        As Admin stated the ADC organs used wavetables that by reverse addressing and sign flipping generate a very close approximation to a single frequency cycle from one quarter of the data. It is this addressing and flipping at the requisite frequency that the S-152 controls.
        The data values for the wavetables were obviously generated by complex algorithms that required significant non real-time processing.

        What I have recently realised is that on certain stops more than one frequency per individual key is processed - the fundamental frequency is generated separately from at least one of the harmonics.
        This was revealed when a faulty FG-1 card resulted in an 8' stop sounding as a 4'.

        I suspect this may be why it is often mentioned that an organ sounds much better than before when an FG card is replaced ie:- the harmonic generation had failed prior to the fundamental.


        • #7
          Thanks for the clear explanation of the role of that fundamental chip in the ADC organ. That does indeed give substance to the prior discussions of an organ sounding better with an FG-1 card replacement. I suspect that the only relationship between the previously mentioned S-151 (in MOS organs) and an S-152 chip may be that they are sequential Allen part numbers with no similarity in their programmed functions. Has anyone cloned this chip yet?

          I'd still love to see someone (you?, don60, Circa1949) who understands more about how the MA, KA, FG, EG and TG cards work together tie all of the previous discussions on here in one document. An exposition of the function and data flow in those boards would make troubleshooting much easier. There are still lots of these instruments out there and larger ones are now coming on the market for very reasonable prices.

          I created a distillation of information about the cards themselves that is now a sticky on the Allen Organ Owners Group but we are still lacking a good system functional description. A companion sticky would complete the ADC documentation to the best of our collective knowledge.

          And I think we are still in the dark about why there are two D>A chips for each channel on a TG card.
          Last edited by AllenAnalog; 01-06-2020, 02:06 PM.
          Larry is my name; Allen is an organ brand. Allen RMWTHEA.3 with RMI Electra-Piano; Allen 423-C+Gyro; Britson Opus OEM38; Steinway AR Duo-Art 7' grand piano, Mills Violano Virtuoso with MIDI; Hammond 9812H with roll player; Roland E-200; Mason&Hamlin AR Ampico grand piano, Allen ADC-5300-D with MIDI, Allen MADC-2110.


          • #8
            Double resolution?


            • #9
              Expression pedal positions are passed in the stop-tab serial data as 8 bit values. The extra D->A is used to generate a corresponding voltage level for input into the active three band equaliser op-amps to vary the gain.
              There are also D->As on the AP-1 card that can control the audio channel gains as opposed to division gains. I'm not sure were the digital control data for these is derived.