HBATEC international analogue squarewave keyboard with bizarre sound, accompaniment & programmable analogue rhythm

This strange analogue tablehooter looks (besides slightly larger drum pads) almost exactly like the digital Letron MC-3 and has many functions common, but it contains totally different hardware and makes a lot of bizarre sound glitches. The main difference to the MC-3 is that the HBATEC is only 2 note polyphonic and employs real analogue percussion (that can be nicely modified).

The PCB has the small numbers "89-12", those may be the initial manufacturing date (december 1989) of this instrument. Due to strange quirks better consider this a monophonic synth with the 2nd channel as a bonus, because it randomly truncates held notes. Bought with wrong expectation it may surely disappoint you. But when treated right, it is one of the craziest lofi sound machines you can imagine.

Likely this was an obscure predecessor of the MC-3, thus I describe here only the differences.

different main features:

eastereggs:

  • 8(!) additional OBS rhythms found.
  • alternative to these a rhythm bank switch button to select them through the normal 8 rhythm buttons.
  • arpeggio on/ off (very cute and cheesy doodling squarewave). The CPU has an own sound output pin for it which is even already connected through an own resistor to the accompaniment pot. ( =>Apparently this mainboard was designed to be equippable with different control panel PCBs.)
  • 2 similarly behaving buttons those activate a very strange mode, It turns accompaniment, demo and custom drummer function off, switch the main voice into a monophonic mode, distort the main voice slightly and change the sound envelopes of all preset OBS sounds. Sometimes (when keys are hold down during button press?) the keyboard ignores releasing keys, i. e. all notes toot until another key is pressed. One button turns this mode on, the other on and off.

    • modifications:


    Here you see the row of knobs I
    have added.    		 The brass coloured bar is the sensor contact. Blue knobs control the drum envelopes.

    notes:

    I really don't know what this strange instrument was intended to be. The case top has on its backside pre- moulded holes for the many LEDs of the Letron MC-3 (but round instead of rectangular ones), and 4 parallel wired OBS sound buttons exist to imitate the Letron's 12 sound button layout. But due to the CPU has more functions than the Letron and the analogue drums are costlier, it makes no sense to be a pirated Letron (besides that a Letron is no well- known brand product commercially worth to be imitated), though it can only be a predecessor. The HBATEC has a "tango" rhythm button where Letron has "rock" (despite that the rock rhythm exists as easteregg also in the HBATEC) and also the sound buttons are labelled differently.

    My theory is that either 2 different companies independently(??) imitated a 3rd this way looking name brand keyboard, or (likelier) the manufacturer had first designed the HBATEC hardware with its 28 pin "Zilog KZ 283" CPU for an older instrument, then started to develop the Letron MC-3 and already produced many plastic cases for the new model, but by a technical problem he couldn't finish the new model's ICs in time and though built the old hardware into the new cases, which would explain the pre- moulded LED holes and duplicate sound buttons etc. Strange is also that the manual booklet of my HBATEC (some pages are missing) describes instead an instrument of the name "HBATEC DX", which seems to be rather identical with my MC-3 (e. g. the "500 Miles Away From Home" demo melody is mentioned there). Even more mysterious is that even a 3rd instrument Fujitone 3-A with the same case design was made, which technology and behaviour is extremely similar like the HBATEC, but it has even a few less functions (e.g. no organ chord mode), only a mono amplifier and is based on a 40 pin CPU "KZ381" with the strange brand label "Angeltone". Either this was a predecessor of the HBATEC, or it was even a successor that attempted to replace the Zilog brand CPU (bought from an external company and thus possibly expensive) with a re-designed CPU made by the instrument manufacturer, but he omitted some functions because he didn't manage to get them to work in time. (On the box of my Fujitone 3-A is even a photo of an instrument with the same case but no brand label and button names corresponding to the MC-3.) Later I found the Elta KE-3 toy keyboard, which contains another Zilog CPU with 40 pins and similar main voice sounds (but no percussion), and the case photo on its box shows an Angeltone brand label. Thus it might be that Angeltone was the genuine designer of all these "KZ" keyboard CPUs.

    The main voice sounds of the HBATEC are made from squarewave with different pulse widths and generally resemble much the Letron MC-3, but it features a "banjo" with a (rather slow) mandolin ring effect and uses more different octave settings (e.g. "piccolo" has a very high pitch). The harsh "clarinet"/ "oboe" sound matches rather the latter name and is harsher than the MC-3 timbres. Also the "xylophone" is a quite harsh sound with the same timbre like the "guitar" - only with shorter envelope. The "harpsichord"/ "clavichord" otherwise uses a dull 1:1 pulse ratio and should be better named "music box" or "cowbell" or "celesta"; also the "banjo" uses this timbre. Like on MC-3, also here the OBS sound select buttons can be abused for realtime sound changes, although the HBATEC always restarts the envelope of held notes when any of these (and various other) buttons are pressed. A little nasty is that this instrument employs a "last note priority" scheme in that when (without accompaniment) 2 simultaneous notes are played and a 3rd key is accidentally touched (e. g. by pressing a key before completely releasing another one) the first of the 2 previous notes is shut off to play the note of the newly pressed key. This can be unpleasant when 1 key is hold down with the left hand in the intention to play a continuous bass tone while the right hand shall play a monophonic melody to it, because an accidentally touched 3rd key during this can stop the bass tone which is quite confusing, despite it may help a lot to train discipline to hit keys more precisely. Especially annoying is that also by very fast play the instrument truncates the note of a held key to re-use its channel to keep the release phase of a previous note played with the other hand. (Some small Casio and Bontempi keyboards have a similar flaw.)

    If you want an easily playable squarewave instrument, then one with MC-3 hardware is certainly better suited to fulfil this need; also its 4 note polyphony are an argument for it. But the HBATEC is (IMO) the more interesting beast - simply because its behaves so bizarre in many ways and one can do so many strange sounds on it. It also seems to sound somewhat warmer (almost a Minimoog- like timbre?), while the Letron is more like a C64. The HBATEC is also way more fascinating for circuit bending due to its analogue hardware and the lots of separate sound output lines of its CPU.

    Unlike the MC-3, the HBATEC has lots of funny sound glitches and in organ chord mode the tone pitch seems to be quite off when chords are played. There is also always a low tooting hum present, which gets higher for some seconds after a note is played. (Don't worry, the famous historical Hammond organ also featured a well known background hum.) The rhythm tempo of the HBATEC can be set so high that the rhythms turn into a drumroll. Particularly with rhythm volume muted, in this state the accompaniment patterns (especially with arpeggio) start to sounds much like the famous arpeggiator effects known from C64 homecomputer musics. The modified drum section of my HBATEC can create a variety of strange analogue percussion, from tom- tom and base drums or congas up to metallic stuff like (sort-of) gongs and triangles, and when drum decay is set too long, it results in howling self- oscillation. (The point where this happens is very depending on the actual drum pitch and distortion settings.) Also the hihat input sensor contact makes lots of funny tweeting or squelching sounds when touched. Its much fun to create tekkno- like musics on it by cranking the drum sounds knobs while the rhythms plays. (Remember: there is even one programmable rhythm pattern available.)
     

    circuit bending details

    The hardware contains only one digital IC: the CPU "KZ 283, R335, LIL9217ASIC, 9024  IW" (crystal clocked @ 7.7MHz) with only 28 pins, made by the well known brand "Zilog" (creator of the famous "Z80" 8 bit homecomputer CPU).

    This CPU makes nice analogue distortions and seems to have a VERY bizarre concept of implementing the 2 note polyphony. The first pressed key not only causes the generation of a squarewave tone, but also changes a sort of internal subclock speed (about up to 10%?) depending on which note is played (apparently as a correction term to fine tune the note due to an otherwise too coarse internal frequency lookup table or something similar). When a 2nd key is pressed, the CPU though has to generate a difference frequency to the previous note regarding the changed clock speed, which causes all kind of weird intermodulation distortions those are especially audible when the capacitor envelope of one note has faded silent while its tone generator still sounds through the 2nd note with a ring-modulation-like effect until the CPU finally shuts down the silent note completely. So keyboard matrix signals, rhythm tempo and a zillion of other things wobble around in bizarre ways by the changing clock speed. Despite the CPU itself is crystal clocked, also various short, howling glitches get audible when additional keys are pressed due to the CPU determines by the relative key position order which tone should be played by which oscillator; though a tone played by the 1st one may suddenly jump to the 2nd when a lower note key gets pressed; especially in accompaniment mode (where 3 voices exist for accompaniment) this is very audible.
    The squarewave sound generator uses analogue capacitor envelopes and all accompaniment voices have own pins. Beside CPU there are no digital ICs. The PCB is a mess of discrete components that rather resembles a 1970th transistor radio. They do fully analogue percussion (with transistor noise), main voice envelopes, and even some matrix demultiplexing is apparently done in discrete diode- transistor logics.
     
    sound pulse width octave
    piano 3:1  
    oboe 15:1 (?)  
    guitar 7:1  
    xylophone 7:1  
    piccolo 1:1 +2 
    banjo 1:1 +1
    organ 3:1  
    harpsichord 1:1 +1
         
    bass 1:1  
    chord 1:1 +1
    arpeggio 1:1 +1
    		Some preset sounds play 1 or 2 octaves higher, but the waveforms only differ in the pulse width ratio of the squarewave. (The ratios were estimated on an analogue oscilloscope at the sound output jack, which isn't very accurate.)

    keyboard matrix

    The keyboard matrix uses a special algorithm with keys wired by diodes in both direction to save IC pins. The "in" and "out" pins change their roll in turns to poll them, thus all matrix pins are listed here as "out" and "in". They are also multiplexed with the percussion trigger outputs and LEDs. Each key or button needs to be connected through a diode from an 'out' to an 'in' line. Due to higher forward voltage, LEDs don't trigger keys even when they share the same matrix place. All matrix pins can be accessed at the ribbon cable at the PCB's lower edge (pin 6 to 16) or in the same order at any of the 11 diodes rows (upper diode pins) on the keyboard PCB.

    CPU pin 6 is an additional matrix output that together with the normal matrix lines as inputs delivers the 5 percussions and the inverted panel LEDs. However the LEDs can be also found uninverted among the normal matrix lines; each output powers one LED (wired to any unused input and even some used ones). The chord trigger output is active while any chord section keys are held during "manual bass" or at the begin of accompaniment notes. So it may have been meant as an (unused) envelope trigger or timbre filter switch. Because these modes disable the 2nd main voice channel (monophonic), it also may have been initially intended as a signal to reuse its outputs to somehow control the bass or chord timbre, which however was not implemented. The percussion trigger outputs were designed to start 5 independent analogue percussion sounds (not "conga+hihat = snare" like implemented here). The matrix runs with about 333Hz, which crosstalks into the audio as a steady low toot.

    The "sustain on" & -"off" buttons are not in the matrix, but control an external flipflop of discrete transistors to change the envelope.
     
    13
    12
    11
    10
    9
    8
    26
    25
    24
    23
    22
     
    CPU pin
    in 1
    in 2
    in 3
    in 4
    in 5
    in 6
    in 7
    in 8
    in 9
    in 10
    in 11
    in / out
     
     
    T.
    chord
    R.
    stop
    R.
    slow rock
    R.
    tango
    R.
    tempo +
    C.
    Y-C
    o
    o
    o
    o
    out 1
    13
    L.
    tempo
     
    R.
    start
    R.
    salsa
    R.
    bossanova
    R.
    tempo -
    o
    o
    o
    o
    o
    out 2
    12
    P.
    hihat open
    O.
    organ
     
     L.
    chord
    L.
    chord
    L.
    chord
    o
    o
    o
    o
    o
    out 3
    11
    P.
    conga
    O.
    banjo
    L.
    program
     
    L.
    program
    L.
    program
    o
    o
    o
    o
    o
    out 4
    10
    P.
    hihat close
    O.
    picolo
    L.
    strange mode
    L.
    strange mode
     
    L.
    strange mode
    o
    o
    o
    o
    o
    out 5
    9
    P.
    snare
    O.
    xylophone
    L.
    R. bank
    L.
    R. bank
    L.
    R. bank
     
    o
    o
    o
    o
    o
    out 6
    8
    P.
    base
    O.
    harpsichord
    R.
    sync
    R.
    rock'n'roll
    R.
    pops
    C.
    off
     
    o
    o
    o
    o
    out 7
    26
    demo
    O.
    oboe
    R.
    waltz
    R.
    swing
     R.
    march
    C.
    single
    o
     
    o
    o
    o
    out 8
    25
    R.
    play/space
    O.
    piano
    R.
    big band
    R.
    samba
    R.
    rhumba
    C.
    fingered
    o
    o
    		  
    o
    o
    out 9
    24
    R.
    program
    C.
    arpeggio
    R.
    country
    R.
    16 beat
    strange mode 1
    C.
    manual bass
    o
    o
    o
    		  
    o
    out 10
    23
    O.
    guitar
    R.
    on/off
    R.
    heavy metal
    R.
    disco
    strange mode 2
    R.
    bank
    o
    o
    o
    o
    		  
    out 11
    22
    T.
    /chord
    L.
    /tempo
    L.
    /chord
    L.
    /program
    L.
    /strange mode
    L.
    /R. bank switch
    T.
    hihat open
    T.
    conga
    T.
    hihat close
    T.
    snare
    T.
    base
    out P
    6
     

    legend:
    "o"
    		 = keyboard key
    underlined
    		 = function needs locking switch (i.e. stays active only so long the switch is closed)
    R.
    		 = preset rhythm
    O.
    		 = preset sound ('orchestra')
    P.
    		 = drumpad
    L.
    		 = led out
    T.
    		 = trigger out (percussion or chord)
    orange
    background     		 = easteregg (unconnected feature)

    Any functions can be triggered by a non- locking switch in  series to a diode from an "out" to an "in" pin. Some functions disturb the analogue percussion (make it toot). To prevent this wire an 1k resistor in series.

    • 8 additional rhythms & bank switch button
      There are 8 unused preset rhythms buried in the matrix. Instead of individual OBS buttons you can add the rhythm bank switch button to select these through the normal rhythm buttons. (Rhythm names identified fromAngeltone DM-380; only the "slow rock" is very different, but it resembles that rhythm on later Medeli hardware (like MC-38, MC-5000).
    • arpeggio on/off
      Add this button to start arpeggio. The arpeggio output is CPU pin 21. The resistor is already there.
    • rhythm on/off
      uses a single button instead of 2.

      I initially thought there was also a rhythm mute button, but it turned out to be already there as"play/space" of the custom drummer. 1 press mutes rhythm percussion (accompaniment stays on), 2nd press stops also accompaniment. Muted rhythms can be continued by pressing the corresponding OBS rhythm button.
       

    • Y-C switch
      a locking switch here changes the single finger chord playing method from Yamaha to Casio (see here).
    • strange mode
      These are 2 similarly behaving buttons. When pressed, they turn accompaniment, demo and custom drummer function off (their buttons/ LEDs get completely inoperative), switch the main voice into a monophonic mode, set it to "piano", distort the main voice slightly and change the sound envelopes of all preset OBS sounds. The "banjo" in this mode doesn't ring anymore and gets inaudible unless many keys are pressed rapidly ( <=ultra- short envelope?). The "piccolo" turns into a music box. Sometimes (when keys are hold down during button press?) the keyboard ignores key release, i. e. all notes toot until another key is pressed. Both buttons toggle the mode on and off, but sometimes this seems to works only with the 2nd. Holding these buttons makes the drums howl by holding their trigger out high. (Both buttons only differ in the affected drum?)
      I don't know what this mode is good for; possibly it was a chip test mode or the CPU should communicate with an external IC or anything similar, or it is just a buggy IC function that was once intended to do something more reasonable.
    To avoid hum and crosstalk, use shielded cables for all long analogue cables.

    separate channel volume controls

    All main voice and chord channels are output through resistors from individual CPU pins through the "ABC volume" potentiometer to the amplifier input. Disconnect the chord resistors and feed the IC lines into the right end of a potentiometer (e.g. 25k log); wire the pot's left end to GND and the wiper through a resistor and possibly a mute switch back to the mixing point where the resistors were; bridge the given 10k resistor with 1k to make it louder. Wire the same way a pot from the disconnected arpeggio resistor output to the "ABC Volume" output (wiper) pin. So you can control the individual accompaniment voices and arpeggio volume and mute them individually.

    volume sensor contacts

    Connect the input of the master volume pot through a capacitor to a small metal rail at the rear case rim. Connect the individual chord channel outputs each through an 10nF capacitor in series with 4.7k resistor to a nearby touch sensor (e.g. the body of mute switches). For ESD protection of the CPU, install a diode from GND to each such CPU out pin and from that pin another diode to the CPU +5V supply voltage. Touching rail and sensors together will now play thin overtones of the individual chord/ accompaniment channels, which is particularly interesting when some channels are muted or turned low with the channel volume mod.

    percussion envelope

    On the PCB are 3 trimmers for base, conga and hihat decay time. Replace these with potentiometers. The oscillator will howl (self- oscillation) with decay set too long.

    percussion pitch

    The resistor R90 (100k) controls the base drum pitch, R96 (3.9k) the conga pitch. Disconnect their non-GND end. Replace them with each a 470k potentiometer; wire its right end through an 1k resistor to GND and the wiper to the still connected resistor end. To preserve the default pitch as preset at the leftmost position, cut the carbon trace inside the pot at the unused left end and wire it with the open end of the original resistor. 

    Adjusting the pitch knobs will strongly influence decay time and can make the drum tone howl. This is normal because the circuit was not designed as a drum synth. Use the envelope knobs to compensate.

    percussion distortion

    Below the PCB number "TC 079" is a 100k resistor. Connect at its capacitor end the wiper of a 1M potentiometer. Connect the other end if the resistor through an 1k resistor to the wiper of a 5M potentiometer. Wire the left ends of both pots to GND. So you can distort the rhythm with both pots differently.

    main voice envelope controls

    I have no good documentation of this, so it may be wrong. Above the CPU to the left is a discrete flipflop made from individual transistors to control sustain. Through one common resistor it connects in each channel through 2 further transistors 2 additional 10uF electrolytic caps to GND.

    The last of these transistors has the collector on +5V and emitter forms the envelope signal. Wire to that base a capacitor against GND to increase the envelope duration. Wire to that base the wiper of a 100k potentiometer with its right end through a 47uF lytic against GND. This will make the envelope longer. Wire to that base the wiper of another 100k pot with its left end connected through a resistor (1k?) against GND to shorten the attack phase. Because it is awkward to adjust the envelope of both polyphony channels separately, I recommend to install stereo potentiometers to control both channels parallel.
     

    pinout KZ283

    The "Zilog KZ 283, R335, LIL9217ASIC, 9024  IW" (28 pin DIL, crystal clocked @ 7.7MHz) is one of the most bizarre home keyboard CPUs ever made. It supports 2 note polyphonic main voice with external analogue envelopes, 3 note chords, arpeggio and 5 analogue percussion sounds. The whole thing is strictly squarewave based and most functions are like MC-3 (it even has more rhythms), but it has tons of quirks and glitches those make it sound very special but also disturb playability. Due to low pin count the keyboard matrix lines swap their rolls as inputs and outputs to double the supported matrix size. Even percussion trigger outputs are part of the matrix. Strangest is that it seems to compute the frequency of the 2nd polyphony channel somehow in ratio to the 1st by modulating an internal clock rate, which makes bass notes howl and rhythm tempo very unstable. Although note keys and controls respond reasonably fast, the wacky behaviour feels like a music program in BASIC on an 8 bit homecomputer. Likely it is based on a generic microcontroller that was never designed for polyphonic sound output, because none of the pins are analogue. (Even envelope controls output only a TTL level pulse of varying length.)

    I compared pinouts in the "Zilog - Z8® Family Design Handbook" of June 1988, which very strongly hints that it is one of the following:

    Z8600 MCU (2K ROM, 128 byte RAM, 8MHz, NMOS)
    Z86C00 MCU (2K ROM, 124 byte RAM, 12MHz, CMOS)
    Z86C10 MCU (4K ROM, dto.)
    Z86C20 MCU (8K ROM, dto.)

    This already leads to the interesting conclusion that the firmware is maximum 8KB. I would love to examine the machine code of this cacophonic baby. This pinout is based on my own hardware observation combined with above MCUs. Line names were chosen by me to describe their functions (partly inspired by Casio naming conventions).
     
    pin name line purpose
    1 +5V   supply voltage +5V
    2 XTAL2   crystal out
    3 XTAL1   crystal in / clock in (7.7 MHz)
    4 /RESET   (diode to +5V and cap to GND)
    5 /DS   data strobe (active during external memory transfer | stays hi, not used)
    6 P3.5 KP matrix percussion & led out
    7 GND   ground 0V
    8 P0.0 K6 key matrix
    9 P0.1 K5 key matrix
    10 P0.2 K4 key matrix
    11 P0.3 K3 key matrix
    12 P0.4 K2 key matrix
    13 P0.5 K1 key matrix
    14 P1.0 ME2 melody channel 2 envelope trigger out
    15 P1.1 M2 melody channel 2 tone out
    16 P1.2 ME1 melody channel 1 envelope trigger out
    17 P1.3 MEL melody envelope long
    18 P1.4 C3 chord tone 3 out
    19 P1.5 C2 chord tone 2 out
    20 P1.6 C1 chord tone 1 out
    21 P1.7 ARP arpeggio tone out
    22 P2.1 K11 key matrix
    23 P2.2 K10 key matrix
    24 P2.3 K9 key matrix
    25 P2.4 K8 key matrix
    26 P2.5 K7 key matrix
    27 P3.1   (wired to GND, a software test pin?)
    28 P3.6 M1 melody channel 1 tone out

    The KZ 283 CPU was likely the earliest predecessor of the later very successful MC-3 hardware. "KZ" may mean "Keyboard Zilog" and the "8" hint to a Zilog Z8 variant. Perhaps the engineer was even so proud of using only 28 pins, that this became the origin of the first 2 digits. The Angeltone DM-200 has a Zilog KZ282. "KZ" CPUs in other keyboards (e.g. KZ381 in Angletone DM-380 and Fujitone 3-A) lack the Zilog brand, so they were likely made later and possibly knock-offs produced by a different company.
     

    Another instrument with 28 pin Zilog CPU and similar clock rate was the ultra-rare Superb Sound EK-922.

    This keyboard was also released as Angeltone DM-280 (very rare - normal DM-280 contain MC-3 hardware, seen only once on eBay). I once saw there also a keyboard "Damff Intertone" which had the same case like the HBATEC, but was claimed to be 4 note polyphonic. I am not sure if both are genuinely technically identical, or if the Damff Intertone was just another variant of the Letron MC-3. At least the button field on the blurred photo looks extremely like the HBATEC. Either the direct predecessor or successor of the HBATEC was the Fujitone 3-A (not to confuse with Fujitone 3A) which has an Angeltone CPU. Unlike the widespread MC-3 instruments, these 2 keyboards seem to be generally very rare. Someone spelled it there "HBA TEC" (in 2 words); im am not sure if it was just misspelled or really a different variant, because I also saw a variant labelled H.B.A. International, which has the same case but apparently contains MC-3 hardware (according to button writing). Also the HBATEC DX (described in my manual) seems to be an MC-3. An initially more common and a bit easier to find successor of the HBATEC hardware class (with same great analogue drum section and 16 OBS rhythms) was the Angeltone DM-380, which could be cheaply found on German eBay. Also a different "HBATEC" keyboard was released with MC-3A case (likely MC-3 hardware), which also had no model name.
     

     removal of these screws voids warranty...    
    WarrantyVoid
    back to tablehooters collection
    		  
     
    back