Hey All,

I just got a "The Cricket!" sound/speech/clock peripheral for my //c. It didn't come with a power supply, but I was hoping to reverse engineer one. The specs are 5.5x2.5mm center positive 9 volts BUT the clock section relies on a 120hz pulse in the DC to advance the clock. Pulses above 9 volts are converted to a square wave that's fed to pin 39 of the on-board microcontroller. Presumably it uses an interrupt on that chip to advance the clock timer.

I had the unit open, and was able to upload some specs and pictures to Asimov. I also took the liberty to replace the electrolytic capacitors while the unit was open, as several had drifted.

Would anyone happen to have an otiginal power supply that's been opened? Or be able to point me towards a circuit which would work? I'm guessing 12vac transformer into a full-wave rectifier with a 12v zener to provide the clock signal which is then fed back into the filtered 9v output. The PS output seems to swing between 9v and 11v at 120 hz.

Frank

Hi Frank,

congratulations on your new acquisition! I think the Cricket! is the most
exciting peripheral ever made for the Apple //c. Mine didn't come with a
power supply either, but it would have been of no use in the EU anyways.
Blake Patterson was so kind to take some pictures of the original supply on
my request to get the original specs:
https://www.flickr.com/photos/blakespot/14764818904/in/photolist-owFsW4-2odCjbm-ouDnk8-odqJeJ-q1t2uZ-kDaXD8-owFtpP-pYx9Ds-pYx9Bo-j7EYdj-ouVCut-ouHygb-odr4wF-ouHyRQ-odqJGs-ouTZK5-kDaW9V-kDbp7v-kDd7yC-2j4QZXu

It is not extremely sharp but it confirms what you have guessed: it is a 9v
500mA brick with a center positive plug. You can replace it with any generic
wall wart with the same rating as long as it is not a switching power
supply. I use a multi-voltage transformer for mine and it works... just of
course a bit slower since we have 50Hz here.

The Cricket contains all necessary voltage regulators so there is no need
for them in the power supply. The clock base is derived through a clever
circuit based on a high-pass and a comparator that will extract the 120Hz
ripple common to a full-bridge rectified DC supply. No need to take any
extra precautions there. It should simply work.

I have collected some documentation and disk images related to it in the
past. Feel free to browse my archive.org collection:
https://archive.org/details/StreetElectronicsTheCricket

Have fun!

On Sunday, June 18, 2023 at 7:30:03 AM UTC-7, cybernesto wrote:
> Hi Frank,
>
> congratulations on your new acquisition! I think the Cricket! is the most
> exciting peripheral ever made for the Apple //c. Mine didn't come with a
> power supply either, but it would have been of no use in the EU anyways.
> Blake Patterson was so kind to take some pictures of the original supply on
> my request to get the original specs:
> https://www.flickr.com/photos/blakespot/14764818904/in/photolist-owFsW4-2odCjbm-ouDnk8-odqJeJ-q1t2uZ-kDaXD8-owFtpP-pYx9Ds-pYx9Bo-j7EYdj-ouVCut-ouHygb-odr4wF-ouHyRQ-odqJGs-ouTZK5-kDaW9V-kDbp7v-kDd7yC-2j4QZXu
>
> It is not extremely sharp but it confirms what you have guessed: it is a 9v
> 500mA brick with a center positive plug. You can replace it with any generic
> wall wart with the same rating as long as it is not a switching power
> supply. I use a multi-voltage transformer for mine and it works... just of
> course a bit slower since we have 50Hz here.
>
> The Cricket contains all necessary voltage regulators so there is no need
> for them in the power supply. The clock base is derived through a clever
> circuit based on a high-pass and a comparator that will extract the 120Hz
> ripple common to a full-bridge rectified DC supply. No need to take any
> extra precautions there. It should simply work.
>
> I have collected some documentation and disk images related to it in the
> past. Feel free to browse my archive.org collection:
> https://archive.org/details/StreetElectronicsTheCricket
>
> Have fun!

Hey Cybernesto!

Yes it will run from a generic 9vdc supply, but doesn't keep time correctly that way. I think I have enough spare parts to try the circuit you suggest, using a 9v (AC) transformer and bridge rectifier. It may need a bit of additional smoothing in the power supply. I think there's only a 100uf cap or two in the Cricket! itself. My oscilloscope showed the signal 9v at the lowest and 12v at the highest, oscillating at 120hz. I think I can just put a 1000 uf cap across the bridge, and connect one of the transformer outputs directly on the other side of the cap, so one piece of the output is filtered and one piece is not. I see a slight amount of slew in the sine wave so I assume the original power supply has a filter cap in there.

Just to note, the Cricket! has two/stereo AY-3-8913s, a speech chip, and a clock, with a pair of LM386 audio amps. Pretty nice design. The microcontroller in there is handling a lot of action (serial, 3 parallel chips, and timekeeping).

There is a 555 chip as well, and if I hook it up to 14vdc or so it will keep reasonable time. The speed of the clock seems to fluctuate a bit with the input voltage. I wondered if it was a backup time base or something. Based on the regulator it should be ok up to 18 volts but I didn't want to go that high. There's also a small potentiometer which I suspect may be related to a secondary time base. It's near the regulator and comparator you talked about. I put new electrolytic capacitors in mine so don't really want to open it back up again.

Frank

> Yes it will run from a generic 9vdc supply, but doesn't keep time
> correctly
> that way.
That's strange. You are sure that it is not a switching power supply?

I think I have enough spare parts to try the circuit you suggest,
> using a 9v (AC) transformer and bridge rectifier. It may need a bit of
> additional smoothing in the power supply. I think there's only a 100uf cap
> or two in the Cricket! itself.
Indeed. 1000uF are actually better.

> There is a 555 chip as well, and if I hook it up to 14vdc or so it will
> keep reasonable time. The speed of the clock seems to fluctuate a bit with
> the input voltage. I wondered if it was a backup time base or something.
The 555 oscillator is used together with some inductors to generate the -5V
required by the speech chip. The frequency is not critical for this
application.

> Based on the regulator it should be ok up to 18 volts but I didn't want to
> go that high. There's also a small potentiometer which I suspect may be
> related to a secondary time base.
The potentiometer is used by the speech chip to set its RC oscillator
frequency to about 160kHz.

> you talked about. I put new electrolytic capacitors in mine so don't
> really
> want to open it back up again.
Good idea. I have two and both stopped working because of some leaky
capacitors.

On Monday, June 19, 2023 at 9:55:06 AM UTC-7, cybernesto wrote:
> > Yes it will run from a generic 9vdc supply, but doesn't keep time
> > correctly
> > that way.
> That's strange. You are sure that it is not a switching power supply?
> I think I have enough spare parts to try the circuit you suggest,
> > using a 9v (AC) transformer and bridge rectifier. It may need a bit of
> > additional smoothing in the power supply. I think there's only a 100uf cap
> > or two in the Cricket! itself.
> Indeed. 1000uF are actually better.
> > There is a 555 chip as well, and if I hook it up to 14vdc or so it will
> > keep reasonable time. The speed of the clock seems to fluctuate a bit with
> > the input voltage. I wondered if it was a backup time base or something..
> The 555 oscillator is used together with some inductors to generate the -5V
> required by the speech chip. The frequency is not critical for this
> application.
> > Based on the regulator it should be ok up to 18 volts but I didn't want to
> > go that high. There's also a small potentiometer which I suspect may be
> > related to a secondary time base.
> The potentiometer is used by the speech chip to set its RC oscillator
> frequency to about 160kHz.
> > you talked about. I put new electrolytic capacitors in mine so don't
> > really
> > want to open it back up again.
> Good idea. I have two and both stopped working because of some leaky
> capacitors.

I was able to wire up a working linear, unregulated power supply using a 12vac transformer into a full-wave rectifier then finally a 560 uf 35v capacitor in series. The lower rating leaves enough ripple in the output to create the 120hz timebase clock signal. The connector is a 5.5mm x 2.5mm center positive barrel jack.

I'm wondering if Euro units are hard-coded for 50hz instead of 60hz line frequency? Or if the microcontroller is smart enough to tell based on the pulse widths which speed it's running at? Does yours normally keep time correctly?

BTW the caps to replace are five 100uf 25v and four 10uf 25v. Both Crickets I've worked on needed a touch of new solder to be able to desolder, and both had pretty dry caps. One unit was failed, one working but I could see leakage. All marked "Liberty".

A list of components inside:

U1 - custom Street Electronics microcontroller 03-801-00
U2 - SN74LS02N - 2-input NOR gate
U3 - TL084CN 340B (TI) - J-FET Op-Amp
U4 - UA723CN 406E (TI) - Linear Voltage Regulator
U5 - TLC555CP - Timer/Oscillator
U6 - TMS5220NL (TI) - Speech Synthesizer
U7 - AY-3-8913 - 3-voice Programmable Sound Generator
U8 - AY-3-8913 - 3-voice Programmable Sound Generator
U9 - LM386N - Audio Power Amplifier
U10 - LM386N - Audio Power Amplifier

Q1 - KS2N4401 - NPN transistor TO-92
Q2 - KS2N4403 - PNP transistor TO-92
Q3 - KS2N4403 - PNP transistor TO-92

REG - 3 pin TO-92 voltage regulator? (can't read markings)

RN1 - SCCO 10-9-5-R 47K(?) resistor network (Stackpole?)

XTAL - 4.9152 MHz (microcontroller clock / UART-friendly frequency)

On Monday, June 19, 2023 at 9:55:06 AM UTC-7, cybernesto wrote:
> > Yes it will run from a generic 9vdc supply, but doesn't keep time
> > correctly
> > that way.
> That's strange. You are sure that it is not a switching power supply?
> I think I have enough spare parts to try the circuit you suggest,
> > using a 9v (AC) transformer and bridge rectifier. It may need a bit of
> > additional smoothing in the power supply. I think there's only a 100uf cap
> > or two in the Cricket! itself.
> Indeed. 1000uF are actually better.
> > There is a 555 chip as well, and if I hook it up to 14vdc or so it will
> > keep reasonable time. The speed of the clock seems to fluctuate a bit with
> > the input voltage. I wondered if it was a backup time base or something..
> The 555 oscillator is used together with some inductors to generate the -5V
> required by the speech chip. The frequency is not critical for this
> application.
> > Based on the regulator it should be ok up to 18 volts but I didn't want to
> > go that high. There's also a small potentiometer which I suspect may be
> > related to a secondary time base.
> The potentiometer is used by the speech chip to set its RC oscillator
> frequency to about 160kHz.
> > you talked about. I put new electrolytic capacitors in mine so don't
> > really
> > want to open it back up again.
> Good idea. I have two and both stopped working because of some leaky
> capacitors.

Also, rather than remove the rubber feet, you can just peel from the outside corners and slip a phillips screwdriver underneath to get at the 4 mounting screws.