CREATOY TRAINS

I found these two on Ebay in early November of 2024. They are somewhat generic looking train keyboards by the brand Creatoy from 1995. Many toys made by Creatoy from this era, and similar toys like them have a special place for me. As a child of the late eighties and early nineties, I would see a lot of these lesser known toys at friends' houses, and school growing up. At the time, it seemed like there was no shortage of these off-brand obnoxious toys everywhere. Now they seem to be harder to find. When I have a little extra cash, I love to scour Ebay for these types of forgotten relics. 

This particular model was especially obnoxious. The sounds would come out blaring with no possibility of volume control. I don't imagine these were played with very long before kids(and parents) got sick of them. There are 8 keys that play standard notes, and seven multi-function sound effects buttons. A rotary knob selected between four modes: vehicles, animals, blip rhythms, and short melodies. 

The initial goal for modifying these keyboards was to send the toy's audio output through a PLL frequency multiplier with adjustable tracking. The black blob Chip-On-Board(COB) output audio to a pin with a low value resistor to ground. My educated guess is that it is a 4-bit DAC. The sample quality is terrible, but good enough to send through a comparator. I played around with a window-comparator to try to hone in on certain bandwidths, but it really didn't blow me away, and the PLL with multiplier would be doing the more fun stuff, so I settled with a Schmitt trigger instead. I really liked the way the PLL chewed up the audio signal. It sounded especially nice to mix the "Chaos" voice back with the original. 

I decided to get rid of the 7 sound-effects buttons, and instead rewire the keyboard with a three-way switch that allow the top four keys to be set to four of the seven "VEHICLE", "ANIMAL" sounds, or the standard notes. 

The PLL "Chaos" voice consists of a CD4046 PLL chip that tracks the frequency of the incoming signal. The audio signal from the COB is sent to a comparator(schmitt-trigger), and converted to a square-wave. The square-wave is then sent to a CD4013 dual flip-flop chip to create two sub-divisions of the frequency. One of the three can be selected by the three-way "OCT" switch, and sent to the PLL to be tracked. The tracking of the PLL can be adjusted with the "TRACK" knob to adjust the efficiency of the tracking filter. Lowering the efficiency will cause the output signal to slide, distort, and warble in chaotic ways. Furthermore, the tracking-filter capacitor can be set to different values with the three-way "RANGE" switch, giving even more possibilities for chaotic modulation of the signal. The output from the PLL is sent to a CD4017 counter chip before returning to the phase-comparator pin of the PLL. 8 of the outputs from the CD4017 and the RESET pin are sent to a CD4051 multiplexer chip. Depending on which output pin is connected to the CD4017's RESET pin, the frequency on the phase-comparator pin of the PLL will be divided by that interval(3, 4, 5, 6, 7, 8, 9, 10), and will increase the frequency of the PLL's VCO to try to match, effectively multiplying the input frequency. The CD4017 interval is determined by the 3-bit input on the CD4051 select pins. In the past I have used microcontrollers to create a simple 3-bit ADC to smoothly shift through the 8 intervals with a single potentiometer. Since the range accuracy was not so important for this project, I used a crude resistor-summing logic-gate ADC circuit that I found some years ago. I was never able to get it to work before, but with a proper rail-to-rail CV buffer in front of it, I was pleasantly surprised with how well it worked. I will definitely be using it again. ADC

After the PLL stage, the signal can be mixed back together with the original sounds from the COB using the "CHAOS" switch. The voices(s) are then sent to a PT2399 echo circuit that can be turned on or off using the "WET" switch. The "DELAY" knob controls the delay time, and the "REGEN" knob controls the number of repeats. The "100%" push-button switch adds extra signal to the feedback path. The depth of the "100%" switch can be offset by the "REGEN" knob, depending on where it is set.

The signal is then sent to a 12db lowpass filter with fixed resonance. 

For modulation, there are three separate LFO's. One is a dedicated key-trigger LFO that gates the keyboard keys repetitively. It doesn't actually gate the keys as much as it resets the key-press at a regular interval, so at extremely low frequencies the keys can be played freely without needing to wait for the next gate to activate them. The other two LFO's are for modulating the the keyboard's "PITCH", "SHIFT"(Chaos voice multiplier), and "FILTER". Each of the parameters have a three-way switch to select between Triangle, Off, or Sample/Hold. The Triangle LFO can be set to Ramp, Triangle, or Saw-wave. The frequency is controlled with the "LFO" knob. The Sample/Hold LFO samples the Triangle wave at a frequency determined by the "GATE" knob, and the gate time can be set to 50/50, or 99/1 with the S/H switch. In 1/99 mode, the Triangle-wave will be sampled and held quickly with little, if any slide. In 50/50 mode, the sample time will be 1/2 of the S&H LFO's frequency, and the entirety of the Triangle wave will play through until the the LFO's next half-cycle, when the last voltage from the Triangle-wave is held. This gives a periodic LFO or slide effect, depending on the frequency of the Triangle LFO.

Each of the three "PITCH", "SHIFT", and "FILTER" parameters can also be modulated externally using the three 0-5 volt CV input jacks. 

The process of finishing the keyboards took a backseat to numerous other projects over the 2024 holiday season, including a total overhaul of my workshop, which I am very happy about. There was really no rush to finish these, but once I had finished the first(Blue) one, the second one was a snap. I would like to say that these were quick and easy to make, but the truth is that they took a ton of hours of measuring and drawing alone. Modifying existing toys like this is always a challenge when trying to fit so much inside. It was definitely worth it, though.