Casio PT20

[pianokeyjoe]

I just bought my second ever PT20 from ebay. This one is a as is for parts unit. It was not turning on as per the description. I fixed the issue that prevented from turning on:(input jack solder pad broke from the circuit board due to some one plugging and wiggling the plug around to force it to work). There is a small crack on the keyboard section of the control/keyboard PCB that only killed one key(middle D). But the real issue is that even though I have power to the unit and the unit is ON and making sound, the sound is all wrong. It sounds like slow attack and short decay for ALL sounds including the bass notes of the chord section, and the beats sound like arcade game blips with down pitching arpeggio(8bit laser sound?). Does anyone have a schematic? I am sick to my stomach right now so every 10 minutes I have to attend to my stomach and can not concentrate on the Casio so I had to close it up and give up for now. But while I recover, I would like to know if any one else has come across this issue maybe on a PT1 or such keyboard? It sounds like if the unit was running on dying batteries or like if you were playing it while turning it off. Seems like there is not enough voltage going to the unit type issue.  Shame I sold my first one since it was in good shape and now I regret it. From what I am looking at and from what I can hear, this model seems to be the flagship version of the PT1 like form factor PT keyboards since it packs ALOT into the same small sharp edged boxy case of the PT1 line! I almost think Casio could have made a CZ-PT in this SAME small case lol(I had such a dream once about such a keyboard). I have talked about this model before in this forum but I only went on my memory of what I once had years ago in 2010.. It is 2022 and I see this thing has ALOT more features than I remember! I would love to get one that works or make this one work. Any help would be greatly appreciated!

[IanB]

Sorry to hear you aren't well @pianokeyjoe, hope you will feel better soon!

 

I cant find a schematic for the PT20 so far. But I was reading through the Casio vowel/consonant patent the other day and though that isn't model specific, it seems that volume envelopes are done at the digital level with the music LSI. Can we have a photo or two of the PCB to see what's on there?

[pianokeyjoe]

Thankyou @IanB! I will take it apart again in a few days and take photos yes. Especially since I want to fix it. The unit is sounding like it is not getting enough power but the LED is bright so I do not know. From my past experiences with all things Casio and fleamarkets is that who ever had it used a wrong power supply with center positive polarity and the thing would not turn on so they wiggled the heck out of the jack trying to force it to work and then gave up and in the process broke the solder pad off the center negative jack.As soon as I put a wire from the jack to the next solder point on the PCB, the unit sprang to life since I do have that very RARE Casio branded 7.5vdc 600mA PSU. But the keyboard sounds bad. My guess is the main capacitors which are electrolytic, have been worn down. Remember in the Forum thread for the Casio CT310, I mention I had a CT310 that had a broken power cap, and I replaced it with a different one of the same voltage? That when I turned on the CT310, the volume on the CT310 would fade IN every time I turned it on? Same thing is happening here only it is not sounding right unlike the CT310, and it is still too low a volume. If I could post mp3 files here I could upload a sound example. I do not have video equipment for a youtube video.

[CYBERYOGI =CO=Windler]

The Casio PT-20 is a castrated PT-30 with less features (29 instead of 31 keys, no LCD). Either something went wrong with envelope capacitors or your CPU badly suffers of bitrot.

 

multipulse squarewave & timbre filter

The main voice is based on multipulse squarewave sound synthesis and strongly resembles Casio VL-1. The digital oscillator outputs for each timbre an up to 16 steps long repeating bit loop of square blocks of equal height. The waveform is multiplied in a VCA with the coarse digital envelope control voltage from a separate CPU pin. The envelope height differs among sounds to match the volume (and shorten certain sounds?). The main voice is routed through 2 fixed timbre filters, those are controlled by CPU outputs pins 68 O4 (highpass) and 67 O5 (lowpass) for 3 timbre settings (both lo = unfiltered). Casio PT-20 likely lacks the highpass and thus has no 'harpsichord' button.

pinout D1868G, HD61703

The LSI "NEC D1868G xxx" / "Hitachi HD61703xxx" (80 pin SMD, pins count anticlockwise, xxx = software number of internal ROM) is the CPU of cheap Casio mini keyboards with optional ROM-Pack port. It outputs multipulse squarewave waveforms on individual pins for monophonic main voice, monophonic obligato voice and a 4 note chord+bass. Each of these have an additional pin with coarse digital envelope (strong zipper noise, 10 stair steps?) those need to be combined in an external VCA (transistor circuit). Only the blip percussion (made from plain squarewave of several pitches + shift register feedback noise) already contains a 3-bit(?) logarithmic decay envelope in its output signal. The CPU can control key lighting LEDs (through key matrix outputs and 3 additional pins, using transistors) and has an LCD output. It is clocked by an LC oscillator with tuning trimmer; unlike modern keyboards, the clock can be stopped and continue from there, thus it is fully static without DRAM.

 

By its main voice sound engine this is the direct sucessor of the D1867G CPU in Casio VL-1, but percussion engine and other features differ. So it is way less calculator-like; the D1868G does not need 3V (obviously designed for one lithium or 2 alkaline button cells) anymore, and none of the versions contain a pocket calculator nor the ADSR synth (that was using its "M+" memory and number entry method).

While the PT-80 has a "NEC D1868G  007" CPU, its case variant MT-18 contains a "Hitachi HD61703A01" with the same functionality. Also the wiring of both looks the same, which strongly hints that HD61703 is identical or successor of the D1868G found in earlier and more complex mini keyboards like PT-50 and PT-30, those had sophisticated sequencers with LCD display and datasette storage. Despite the digital melodic drums always contain a grainy digital envelope, most versions add an external filter. In PT-80 the additional analogue snare has its own trigger pin and mixes some digital white noise from the CPU into its transistor circuit, and the main voice has 2 filter settings (lowpass, highpass). PT-30 and PT-50 apparently contain a similar snare (with softer timbre), and have a cymbal trigger and cymbal envelope shorten output; their main voice filter has even 4 settings. Later toy keyboards like PT-82 and EP-20 had very simplified hardware without external percussion envelopes and a poor sounding digital snare (noise modulated squarewave). Unlike in Casiotone 201, the unconnected LCD port of HD61703B01 (Casio PT-80, PT82) seems indeed unused; while pins {39..41} have characteristic multi-step waveforms, the pins {42..66} output an identical steady squarewave of some kHz. None of these ever change (as far my analogue scope can tell), thus likely there was no control software included in this cut down toy. But the existence proves the relationship with the older D1868G. In PT-80 schematics the CPU is labelled µPD1868AG-007 / µPD1868G-007, which hints that another alternative name may be D1868AG. Its ROM-Pack port is wired 1:1 to CPU pins 33..23.

 

In PT-30 the SRAM "Hitachi HD61914B" uses almost the same pinout like a ROM-Pack, but has 3 additional pins 14, 17, 19 multiplexed with keyboard matrix and datasette port. In PT-50 the wiring is more complex to handle 2x SRAM + ROM-Pack + datasette.
 

software number	hardware class	notes & features
001	PT-30	LCD support + datasette module, uses SRAM HD61914B
004	PT-50	LCD & ROM-Pack support + datasette module, uses 2x SRAM HD61914B
006	KX-101	combined with "NEC D1879G 002", uses 2x SRAM HD61914C, datasette storage
007	PT-80	ROM-Pack & key leds support
A01 (HD61703)	MT-18	same like 007
B01 (HD61703)	PT-82	ROM-Pack & key leds support, no chord buttons

 

By the lack of schematics, I first had made an EP-20 pinout (like PT-82) from my own observation. Later I filled the rest and identified pin names from the Casio PT-80 service manual. The PT-82 hardware has simplified digital blip percussion and no switchable filter.

caution: The PT-80 service manual indicates that this CPU uses "negative logic", i.e. technically +5V is its GND while 0V is its -5V supply voltage. So the voltages are not was the pin names suggest. I use the positive voltage naming convention (from 0V to +5V, not -5V to 0V).
 

pin name purpose 1 /M1 power switch on 2 KI1 key matrix in 3 KI2 key matrix in 4 KI3 key matrix in 5 KI4 key matrix in 6 KI5 key matrix in 7 KI6 key matrix in 8 KI7 key matrix in 9 KI8 key matrix in 10 KO1 key matrix out | led drive out 11 KO2 key matrix out | led drive out 12 KO3 key matrix out | led drive out 13 KO4 key matrix out | led drive out 14 KO5 key matrix out | led drive out 15 KO7 key matrix out | led drive out 16 KO6 key matrix out | led drive out 17 KO8 key matrix out | led drive out 18 KO9 key matrix out | led drive out 19 KO10 key matrix out | led drive out 20 KO11 key matrix out | led drive out 21 P reset 22 INT PT-30, PT-50: sram pin 12 | PT-80: (wired to supply voltage GND) 23 D4 rom pack pin 11 | PT-30, PT-50: sram pin 11 | PT-50: MT pin 2 24 D3 rom pack pin 10 | PT-30, PT-50: sram pin 10 25 D2 rom pack pin 9 | PT-30, PT-50: sram pin 9 26 D1 rom pack pin 8 | PT-30, PT-50: sram pin 8 27 OP rom pack pin 7 | PT-30, PT-50: sram pin 7 | PT-50: MT pin 3 28 ø2 rom pack pin 6 | PT-30, PT-50: sram pin 6 29 ø1 rom pack pin 5 | PT-30, PT-50: sram pin 5 30 VDD2 rom pack pin 4 (auto-power-off APO out) | PT-30, PT-50: sram pin 4, MT pin 7 31 VDD1 ground 0V 32 CE1 rom pack pin 2 | PT-30, PT-50: sram pin 2 33 GND supply voltage +4.5V (4.6V in EP-20) 34 T test? (wired to ground VDD1) 35 CE2 PT-50: MT pin 5 | PT-80: snare trigger out GE2 | PT-82: NC (high resistance)  36 SP clock out (divided by something) | PT-50: MT pin 4 37 OS1 clock out 38 OS0 clock in 39   lcd common (?) 40   lcd common (?)

pin name purpose 41   lcd common (?) 42   lcd segment out 43   lcd segment out 44   lcd segment out 45   lcd segment out 46   lcd segment out 47   lcd segment out 48   lcd segment out 49   lcd segment out 50   lcd segment out 51   lcd segment out 52   lcd segment out 53   lcd segment out 54   lcd segment out 55   lcd segment out 56   lcd segment out 57   lcd segment out 58   lcd segment out 59   lcd segment out 60   lcd segment out 61   lcd segment out 62   lcd segment out 63   lcd segment out 64   lcd segment out 65   lcd segment out 66   lcd segment out 67 O5 PT-30, PT-50, PT-82: melody filter switch 2 out | PT-80: cymbal envelope shorten /out 68 O4 PT-30, PT50, PT-80, PT-82: melody filter switch 1 out 69 O3 PT-30, PT-50: snare trigger out | PT-80, PT-82: led common out 70 O2 PT-30, PT-50: cymbal envelope shorten /out | PT-80, PT-82: led common out 71 O1 PT-30, PT-50: cymbal trigger out | PT-80, PT-82: led common out 72 W white noise out (for cymbal/ hihat) 73 SPC auto-power-off /APO out 74 RH low conga, hi conga, base drum out | PT-82: all percussion audio out 75 E3 bass+chord envelope out | PT-30: chord envelope out 76 F3 bass+chord wave out | PT-30: chord wave out 77 E2 obligato envelope out (with crosstalk) | PT-30: bass envelope out 78 F2 obligato wave out | PT-30: bass wave out 79 E1 melody envelope out 80 F1 melody wave out

 

 

 

[IanB]

@CYBERYOGI =CO=WindlerYou deserve much kudos for this depth of investigation and elaboration!👏👏 👏👏 👏👏

[IanB]

By the way, this is one of the original Casio patents describing the square wave generator and processing etc.

 

https://patents.google.com/patent/US4590838A/en?q=synthesis&inventor=Toshio+Kashio

[CYBERYOGI =CO=Windler]

Thanks. This patent seems to have rather to do with the more complex consonant-vowel hardware of Casiotone 201 etc. (priority date 1978) and not the multipulse squarewave engine of PT-20 (same like VL-1). Nice is that it shows the data format of waveform and envelope parameters.

Edited July 15, 2022 by CYBERYOGI =CO=Windler

[IanB]

@CYBERYOGI =CO=WindlerYes, my mistake, you're right about the patent. I found it some weeks ago when researching vowel-consonant and failed to appreciate the difference with the more simple engines on the keyboards like the PT-20. Oops😕

[pianokeyjoe]

I have a second PT20 with no sound(YET), and am currently working on it as mentioned in a different thread about a dead PT31, but my original PT20 that DOES turn on and stays on for about 30 of 40 seconds but does shut off and would not come back on until I waited a few seconds unplugged. It still has the strange issue that when I play a key, the sound will play at say B flat, dims the power led, then whether I hold the key or not once the sound is about to die off(release or decay),the pitch goes up a little and then the LED brightens back up again. This is acting like there is not enough power going in and as you play notes, the power led dims and the sound sounds like dying batteries. Sadly, with no schematic to target the power section of this board, I have no clue what is going on. The CPU thankfully, works or there would be NO sound at all. All ribbon cable pins are intact and strong. No battery acid leak damage on any of the traces and believe you me, I have checked and rechecked. There IS a odd second negative - terminal at the battery compartment that connects to a set of diodes separate from the Wall power input jack. It is like the unit has 2 split negative voltage points on the battery compartment but only ONE positive terminal point for the battery/jack contact break and another for the metal bar under the keys(yellow wire) and foil paper under the main board(green wire). I have to determine why there is not enough power going in that causes the sound to cry like dying batteries. Also the envelopes of all the sounds are "pwaaammp" on all presets. You hear the different tones(fantasy, piano,organ,etc)but all have the same long pwaammp envelope. No attack/decay cycles for piano for example. A real stumper..

[CYBERYOGI =CO=Windler]

The APO (auto power-off) uses a transistor driven by an output pin of the CPU. To avoid waste of battery power, the circuit side that stays always connected to the battery has very high resistance. Thus if the PCB is dirty with anything conductive, or even became conductive itself by being soaked with battery acid, it may trigger the APO transistor either permanently on or off, or into an intermediate state (i.e. too low voltage). If soiled, it may help to thoroughly wash the PCB with dishwashing detergent (flush and rinse it well, towel it dry and wait a day or use hairdryer at medium heat). In soundtoys with slightly broken or poorly made electronics, the APO tends to fail most easily (i.e. turns on by itself or does not turn off properly). In PT-20/30 I could imagine that battery liquid seeping under the SMD CPU may cause all kinds of such trouble due to its very narrow pins, causing e.g. voltages from other pins to modulate the APO pin voltage on its way to the transistor.

 

I read that  battery leak residues soaked into PCB material can be neutralized by putting pure dishwashing detergent on the spot, let it react for a day and then rinse it off.

Edited December 6, 2022 by CYBERYOGI =CO=Windler

[pianokeyjoe]

9 hours ago, CYBERYOGI =CO=Windler said:

The APO (auto power-off) uses a transistor driven by an output pin of the CPU. To avoid waste of battery power, the circuit side that stays always connected to the battery has very high resistance. Thus if the PCB is dirty with anything conductive, or even became conductive itself by being soaked with battery acid, it may trigger the APO transistor either permanently on or off, or into an intermediate state (i.e. too low voltage). If soiled, it may help to thoroughly wash the PCB with dishwashing detergent (flush and rinse it well, towel it dry and wait a day or use hairdryer at medium heat). In soundtoys with slightly broken or poorly made electronics, the APO tends to fail most easily (i.e. turns on by itself or does not turn off properly). In PT-20/30 I could imagine that battery liquid seeping under the SMD CPU may cause all kinds of such trouble due to its very narrow pins, causing e.g. voltages from other pins to modulate the APO pin voltage on its way to the transistor.

 

I read that  battery leak residues soaked into PCB material can be neutralized by putting pure dishwashing detergent on the spot, let it react for a day and then rinse it off.

Thankyou for your advice and info! I have the main boards for both PT20s removed from the keyboard case so I can clean them. I did find out vinegar and baking soda neutralizes the acid/alkali too. But I do have Blue Dawn dish liquid so I can try that too.

[pianokeyjoe]

I have now become an official youtuber! I do not know exactly what the heck is going on and I did check power requirements for the Casio PT20 and it is only 7.5VDC at 133mA! The power adapter I have is AD1U 7.5VDC 600mA, so plenty of juice. And here is the link to my own youtube video. Maybe some one here or on YT can tell me what to do next. I do not feel like buying a 3rd unit..

 

[CYBERYOGI =CO=Windler]

If you cleaned with vinegar, be carefull to later rinse it properly, because it is corrosive by itself and else will produce copper rust.

 

The analogue envelope capacitor circuit definitely gets not enough voltage or is otherwise shorted. The supply voltage runs through a transistor that is controlled by the APO pin (73) of the CPU. If the transistor outputs too little current (by a high resistance short at its base or by itself being toast) the rest will go nuts. It also may be that someone inserted a wrong polarity power supply and so destroyed a voltage regulator. The CPU should get triggered to power on by a lo voltage at pin 1 when the power switch is moved (may be only a short lo pulse through a capacitor), which makes it wake up and change pin 73 to lo and so switches the APO transistor on that feeds the rest of the circuit (amplifier, envelope VCA etc.) with voltage. When a timer in the CPU runs out or power is switched off, the CPU sets pin 73 hi again and so makes the transistor switch off the voltage to the rest of the circuit. That's how AFAIK it basically works. It may be that it switches the ground 0V instead of 4.5V line because the whole thing has negative logics.

Edited December 7, 2022 by CYBERYOGI =CO=Windler

[pianokeyjoe]

@CYBERYOGI =CO=WindlerThankyou my good sir, for your input! I did clean off the vinegar. it was very little with a cotton swab to clean the affected areas but went back with water and then with isopropyl alcohol and cleaned more. I have since put away the PT20 for now as I need a better work area to see better and function better. I will tackle this again when I have setup a more appropriate work zone. In the mean time you may have hit the nail on the head with saying some one may have plugged in the wrong polarity. Indeed when I bought this first Casio PT20 it had the jack broken or loose like some one was trying hard to force it to work and wiggled it loose. I fixed that and thus was able to power it on and you already saw and heard it in the video. So maybe voltage regulator. Now as for the other unit that stays ON when the switch is OFF, and has no sound when the switch is then turned ON, that one has a low hum from the speaker when plugged in. I have determined the worse case for it. It DID have broken ribbon cable connections and I fixed those and the SAME issue still existed, so I gave up for now. Same cleaning procedure was performed on it and still nothing. Both units still exhibit their original faults despite my interventions. I may swap the voltage reg from the DEAD unit onto the power wavering unit that does sound, and see if that changes things at least for ONE PT20 lol! But.. later on that. Thanks again for your direction!

[CYBERYOGI =CO=Windler]

If a CPU died of wrong polarity or overvoltage, it often runs very hot. If it burns your fingers (or at least is warmer than normal - battery operated Casios barely produce heat) than it is likely toast. But a CPU can also play dead if the clock oscillator fails to run (dead or shorted capacitor if there is no quartz - an oscilloscope shows if there is a frequency of some MHz) or simply reset never ends by a shorted capacitor (which is the main fault in early Sharp talking clocks). Pulling reset pin hi or lo through a resistor should wake it up if this is the fault.

Edited December 8, 2022 by CYBERYOGI =CO=Windler