Page 1 of 2 12 LastLast
Results 1 to 10 of 14

Thread: neon dividers in old british tube organ emc........help!!!

  1. #1
    ppp Pianississmo wrongatron's Avatar
    Join Date
    Sep 2009
    Posts
    1

    neon dividers in old british tube organ emc........help!!!

    hi, i have a fantastic old double manual organ from the 50's made in england. i have been slowly repairing it and have hit a brick wall. it uses neon divider networks and tube master oscillators. there are several neons that need replacing, but finding them is impossible. they are not the NE-2 type, (tried em and not suitable) but they look similar, only a little thinner and longer, still with the two terminals. if anyone knows where in the world these can be bought, please tell me, and have the satisfaction of knowing that you helped keep a relic alive!!

  2. #2
    mf Mezzo-Forte fredy2's Avatar
    Join Date
    Feb 2005
    Posts
    683

    Re: neon dividers in old british tube organ emc........help!!!



    I would remove a good one and measure the characteristics. </P>


    Breakdown voltage, and operating voltage at several current values as a start.</P>


    I am not sure how these are being used in the divider circuits. If you can post a partial schematic maybe I can come up with a workaround.</P>


    Then try this site:</P>


    http://www.cml-it.com/cgi-bin/htmlos...39163514698813</P>
    <P mce_keep="true"></P>

  3. #3

    Re: neon dividers in old british tube organ emc........help!!!



    wrongatron (I hope it's just a clever name)</p>

    The neon divider circuit is described in a 60's publication I happen upon on the web (google search "neon divider" and you'll see it). The divider is rather nicely explained in an aptly titled section, using an organ top octave divider circuit, for which a schematic and description is provided. The neon bulbs used are the "component" type rather than the indicator type, and were identified as A078 types in the article. Even though the specific organ wasn't identified, I would guess there is very little variation on the theme. The NE-2 won't work because its igniting voltage is too high and the difference between igniting and holding voltage is too large. The A078 has a relatively low ignition voltage (60VDC) and the difference between ignition and holding voltage is only about 6V. I don't know where you'd find these nowadays; there may be flavors of NE-2 that may come close enough to be usable with some circuit adjustments.
    </p>

    How this circuit works is a facinating study of early digital logic using non-digital components. The circuit is essentially a cascaded string of relaxation oscillators built around neon bulbs. </p>

    First, neon 101: A neon bulb, resistor and capacitor forms a simple relaxation oscillator. Basically, the neon is connected in parallel with the capacitor and the resistor charges the pair from a high DC voltage. The neon is essentially an open circuit until it fires. Once the capacitor charges up to the ignition voltage of the neon, the neon fires, exhibiting negative resistance during that brief period. It rapidly discharges the capacitor down below the neon's holding voltage and then extinguishes. When off, the neon becomes an open circuit once again. The capacitor then begins charging back up and the cycle repeats continuously. By sizing the capacitor and resistor properly, the circuit will produce a sawtooth voltage on the capacitor at an arbitrarily chosen frequency, with a peak to peak amplitude of just a bit over the ignition to holding voltage difference of the neon, which in the case of the example, would be about 7-8Volts. Since the waveform is a sawtooth superimposed on the average 60VDC of the neon, the waveform has to be AC coupled into a high impedance buffer circuit for use in the organ.</p>

    Now, here's the really clever part. You take a neon oscillator as just described, designed to produce a sawtooth at, say, high C. You then take another, identical oscillator and tune it to <u>slightly lower</u>than one octave down from the first. Its important that the free running frequency is slightly lower than 1/2 the frequency of the first osc, reason for which will become clear in a moment. The output of the first (higher frequency) oscillator is then partially coupled into the charging voltage of the second, such that at the trigger point of the first osc, a brief, small voltage pulse is added to the capacitor voltage across the second oscillator's neon. If the voltage on that neon's capacitor is only half way towards the ignition voltage, the second neon won't fire. On the following pulse from the first osc, however, the second osc capacitor is just about at ignition point, but not yet because it is tuned down frequency a bit. The pulse coming from the first stage now is big enough to kick the second stage into ignition, and it fires, perfectly synchronized with every second (i.e.: one octave down) pulse coming from the first stage. </p>

    Take this structure and repeat it for the number of octaves the keyboard covers. Then repeat the whole chain for every note in the top octave, and viola, you have a full tone source. The trick will be calibrating the "one octave plus a bit down" tuning of each successive divider stage, and there is an order of tuning to get it right. What you need to do is first tune every primary oscillator, which is typically a tube oscillator. The dividers all cascade from each top octave tube circuit, and tuning would consist of walking down the octaves of each chain, tuning that critical "1/2 frequency minus a bit" in each stage, since that will also affect the amplitude of each stage output too. I can see that tuning one of these would require the skill and patience of a piano tuner. </p>

    Of course, all of this can be replaced with a single digital chip, but I guess it would be supremely satisfying to get this old lady running perfectly in all her vintage glory.</p>

    exp
    </p>

  4. #4
    mf Mezzo-Forte fredy2's Avatar
    Join Date
    Feb 2005
    Posts
    683

    Re: neon dividers in old british tube organ emc........help!!!



    Yes, kinda of what I expected... little like a starved oscilator that is triggered by previous stage.</P>


    Neon bulbs with that low an ignition voltage are near impossible to find... I know what I would do... I would build a CMOS divider circuit to replace it.</P>


    If they are factors of two dividers or MOST ANY integer divider, it should be easy to do this.</P>


    The lamps they used must be fairly close tolerance to be reliable.</P>

  5. #5
    fff Fortississimo davidecasteel's Avatar
    Join Date
    Oct 2003
    Location
    Dallas, Texas
    Posts
    2,693

    Re: neon dividers in old british tube organ emc........help!!!



    I Googled "electronic suppliers" and checked out a few for neon lamps. This link http://www.jameco.com/webapp/wcs/sto...20Neon%20Lampshas several lamps listed with a 65-110 volt rating, and at least one of those says it's a NE-51.</P>


    David</P>

  6. #6
    Moderator Brendon Wright's Avatar
    Join Date
    Sep 2008
    Location
    New Zealand
    Posts
    7,511

    Re: neon dividers in old british tube organ emc........help!!!



    Cor!</p>

    The innards of this baby must be huge!!!</p>

    </p>
    -1958 Hofner 550 archtop guitar -1959 C3 and PR40- -1964 Busillachio Harmonium- -1964 M101-
    -1967ish Leslie 122- -1975 T500 (modded..chopped, and reassembled!)-
    -DIY 760 FrankenLeslie/rat hideout-
    -1980 Electrokey Electric Piano- -Yamaha electric Harmonium (early 80's?)-
    -1990 Jansen GMF150 amp- -1992 Korg 01W/fd- -1992 G&L S-500 geetar.

  7. #7
    fff Fortississimo davidecasteel's Avatar
    Join Date
    Oct 2003
    Location
    Dallas, Texas
    Posts
    2,693

    Re: neon dividers in old british tube organ emc........help!!!



    Brendon, I don't think a well-designed neon divider system would take up any more space than discrete transistor circuitry to do the same does. The discrete transistor tone generators for the Schober Recital Model (2x61, 32p and 32 stops) occupy 12 identical boards mounted along the back of the instrument, each board being about 4"x12". The Recital Model uses a console that is more or less normal sized for a 2-manual AGO instrument and there is a lot of empty space in it. However, it is fully solid-state (even though the transistors are individual units, not ICs) and so the rest of the circuitry of the target organ, being vacuum tubes, would occupy more space. I'm only guessing, but it would not be unreasonable to suggest that each of the 12 notes has 7 octaves of tone (with maybe an extra "C"); the highest pitches would come from the driving oscillators and then there would be 6 divide-by-two stages for each note; that would result in a need for 73 neon lamps (one per divider) with their accompanying resistor/capacitor pairs. All those components are fairly small in size and would not, themselves, require a lot of space; the tube oscillators, however, are a different story. (Possibly the master oscillators are also neon-driven relaxation oscillators? Those would take up less space, but are notoriously unstable.) It ought not to be difficult to stuff all that electronics into a normal-sized organ console, although the vacuum tubes for the oscillators and amplifiers would probably produce a lot of heat and make some form of forced ventilation necessary. There is an advantage with the neon dividers--the output waveforms are already sawtooth in shape, so there is no need for an integrator on each pitch to achieve that end. (Sawtooth waves can be filtered to create more convincing open-pipe and reed voices than square waves can make.) In effect, each of the divider stages is its own integrator.</P>


    David</P>

  8. #8

    Re: neon dividers in old british tube organ emc........help!!!



    David</p>

    Actually, two neons per stage; the circuit in the accompanying text about neon dividers uses two series connected bulbs, with the junction driven by the previous stage pulse. This was done to provide a high impedance trigger input load to the capacitive divider from the driving stage. It could be done with a single neon, but you'd need a FET buffer to pick off and divide the trigger signal. As you know, FETs were kinda rare in the 50's
    </p>

    You're right about the neon dividers advantage; the output sawtooth
    contains a sweet mixture of harmonics, especially even ones, to produce
    the desired sounds. A true digital divider's output square waves
    contain almost exclusively odd-harmonics, and would have to be treated
    to make them asymmetrical to get a decent mix of even-odd harmonics to
    work with.</p>

    It's not the space for 146 neon bulbs and all the junk around them that I'd be worried about, however, its the 73 little trimpots that will all need adjusting systematically to tune the thing. Yikes!!</p>

    exp
    </p>

  9. #9
    mf Mezzo-Forte fredy2's Avatar
    Join Date
    Feb 2005
    Posts
    683

    Re: neon dividers in old british tube organ emc........help!!!



    It is easy to generate the sawtooth waves from the digital square waves. Pure sawtooth can be generated with simple R/C integrating circuit and a reset of the capacitor using a CMOS transmission gate.</P>


    If I were doing this, I would probably program up a PIC microcontroller to do all the dividing and come out with timed narrow reset pulses from the output pinsto the gates across the C's fed by R's. Whole thing would probably fit on 20 square inches of board space.</P>

  10. #10

    Re: neon dividers in old british tube organ emc........help!!!



    fredy</P>


    Thats a really good application for a PIC, as long as you can find one with enough IO pins and structure the code to run deterministically so that you don't get timing jitter. You still have to tune each RC, however, as the slopes from higher octave outputs have to be steeper than those from lower octaves. The idea issort of reverseto the way the old PC joystick position detector used to work, where the PC would discharge a capacitor and then measure the time it took the capacitor to charge up through the variable joystick pot resistance. </P>


    If I were doing this, I'd try to use an approach that either didn't need tuning or at least used the same fixed R-C values on all outputs. This would lead me to explore the use of a single low to mid size FPGA, implementing73 little PDM engines whose pulse density on each output follows a sawtooth shape of appropriate octave and note. That way, choosing the right "carrier" frequency, the outputs would only needsimple L-C or even R-C networks, all the same values, to render nice clean sawtooths. The waveshapes, phase and timing would be perfectsince it would use a crystal osc master source. Vibrato could be easy to implement by running it from a VCO whose frequencyis modulated by a low frequency sine wave of variable amplitude. If there's room in the FPGA, it may even be possible to implement the vibrato mechanism in digital form, producing a PDM or PWM sinewave that can be filtered and used to modulate the external VCO. The mechanical equivalent of this, if it were possible to do, would be to vary the rotor speed of a TWG with a modulating signal.</P>


    Ahem, getting back, I think that wrongatron wants to do a true-to-vintage restoration rather than just get a working organ. I believe that neons of the type he needs are still available, although rare. CML lists a "component" type that seems to have a 65-80V firing voltage and 55-65V holding voltage, which are supposedly available from Farnell in the UK. I suspect the circuit, if its anything like the one in the book excerpt I found, should be adjustable enough to accomodate a slightly different neon. If he only needs to fix a few stages and not the whole generator, it may work out for him.</P>


    exp</P>

Page 1 of 2 12 LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •