EPSON SMD-400 Repair - Amiga 500 Floppy Drive

The old Amiga 500 I got had a non working floppy drive. The span up, abnormally fast, and then stopped. When there was no disk in the drive, the magnetic head twitched back and forth, which I think is normal behaviour.

First thing to try is changing electrolytic capacitors. To get to the caps, the drive's disk tray needs to be dismantled. It took me a while to figure it out, but its done like this:

Mounting bracket and casing of drive has been removed (easy). If you need to use force, you are doing it wrong.




Step 1.
Remove lock-screw and spring, marked in the first image.
Push the disk ejection button to release the upper disk holder mechanism. Push it forwards-downwards to release it.






Step 2.
Remove springs marked in second image.
Push disk eject mechanism forward to release it.






Step 3.

Replace the five caps, marked C08, C09, C10, C11 and C05. I used EMLF160ADA100MD73G as replacement caps.

After this the drive works as expected again.

Caps removed

New caps in place

Amiga 500 power supply repair

I bought a cheap Amiga 500 as non functioning. I knew that the power supply was non-functioning, so I started there.

Amiga 500 PSU

The inside looked OK, but there where several signs of somebody tampering with it.
The bottom part of the PCB had some funny looking solderings with a lot of gunky-looking flux residue.

Bottom side of PSU PCB. Some gunky looking flux residue can be seen to the bottom right, top right and top left.

The board is marked 871, 6L78, A 500.
When I removed the strange looking solder I discovered that one of the pads to the transformer didn't have a connection at all! Maybe it was loose and I sucked it into the solder sucker.

Anyway, I soldered some wire to the transformer and touched up the bad looking solder joints, and the PSU worked nicely after that.

Repaired transformer connection behind the fuse

Transformer pads with solder removed. Pad with no connection to it to the right of the large mounting hole in the middle.

I made a note of the electrolytic capacitors on the board, so that I can order new ones and recap it.

No caps	C(uF)	U(V)	diam (mm)
3	22	25	5
1	4700	6.3	15
2	330	25	10
1	4700	25	16

When I have recapped the PSU, I will properly clean off all the old flux.

40m Ham radio receiver using a NE602

I've been building a circuit I found in "Experimental methods in RF design". It's a 40m AM/SSB/CW receiver using a NE602. I've been doing some experimentation to get the receiver working, especially with the local oscillator. In the book, they recommend a toroid wound coil, but I couldn't get that to oscillate. Finally I made an air wound coil, which worked. But since the coil was wound "mid air", it picked up even the smallest vibration in the surroundings, leading to a very unstable oscillation. I solved it by shoving a roll of paper inside the coil. I also worked some on getting the antenna input filter better matched.
It receives between 6.9 and 7.4MHz, but it hasn't got a lot of selectivity, which I will try to solve somehow. Also, I want to interface the LO to a frequency counter, so I need to build a buffer amplifier to not load down the oscillator.
Here are some images of the receiver and the schematics so far:

Receiver circuit board. Antenna input, with attenuator pot seen to the left. Tuning caps on the right.

Receiver with shielding in place. Tuning knobs to the left.

Receiver diagram.

Interfacing old telephone to PIC

The old telephone has two ingoing lines. These are balanced inputs, one at about 48V on  hook, in respect to the other. I will call the 48V line positive and the other negative, even though they sometimes are the other way around.
To make a simple and quick conversion circuit, I want to use a PICS general I/O, with a max input of 5V. The phone's positive line has three levels when dialing with pulses. 48V (hook on) , ~10V (hook off) and 0V low level during dialing. I want the PIC to see the two low levels as 0, and the high level as 1. I also need to protect the PIC against over voltage  during ringing, when the line goes through 110V AC with 30Hz frequency. During ringing I also need to protect against negative voltages.
Since the PIC has ESD protection in the form of diodes on the IO's, I could rely on them to do the voltage protection. This seems like a bad idea though, since they are poorly speced in the datasheet and the data sheet advice against using them this way. The built in diodes has a voltage drop of about 0.6V. To beat these diodes I put in a BAT54 in the same fashion This is a Schottky diode with 0.35V forward voltage drop, so that it shunts overvoltage before the internal diodes.
Here's what the interface looks like:

Phone line to PIC I/O interface

and the board so far:

Interface and PIC along with a row of LEDs. Phone line connectors to the left, top and bottom.

I've tested having the PIC flash an LED when detecting high/low level, and there seems to be no spurious errors with this setup. Next step is to start counting pulses with the right time spacing, and register what number is dialed.

Old telephone conversion

I found an old rotary phone on a flea market and decided toplug it in. My land line goes through the cable modem, which doesn't support this old way of entering the phone number. What I need todo is convert the pulses of the phone to dual tone multi frequency tones. Step one is to reverse engineer the old phone interface. I connected a scope to the connected phone plug and got these signals:

On hook.

Off hook, with speach.

Receiving ringing.

Dialling a "4".

My initial thoughts is to connect these signals to a microcontroller to count pulses, and then let the microcontroller inject the right dialling tones on the line. First thing to do is to solder phone plug connectors to a perfboard to see if I can convert the 48V signals to logic levels (3.3V on 5V tolerant ports is my thought).

Annoying Beeping Prank Circuit

Since i quit my old job last week, and I had the last day more or less alone in the office, I made a few loving pranks on my coworkers. One of them was a simple circuit based on a 555 timer. The circuit is simply a pulse oscillator with a couple of minutes between pulses. The circuit is seen here:

556 Pulse oscillator

The 1M Ohm resistor sets outputs high time together with the capacitor. The 100 Ohm resistor and capacitor sets the low time. This setup gives the inverse output of what I wanted to drive the buzzer (should be any buzzer with internal oscillator that can be driven by 9V), but it needed to be that way to keep power consumption low. If the resistors where swapped, the DIS-pin would discharge the power rail through the 100 Ohm resistor instead of the capacitor. To invert the output i used a second 555 as an inverter. In this configuration the circuit drew 8mA, which would give a couple of days use with a 9V, 580mAh battery.
Google told  me there was a CMOS-based low power version of the chip, TLC556. That chip was pin compatible and used up 0.3mA. This gave a run time of about 80 days which I was very happy with.
The circuit simply gave a short, loud beep every few minutes. I hid it in my colleagues flower pot and it took him quite a while to find it.. ;)

Implemented circuit

MIDI2VC+, a MIDI driven Voltage Controller

The MIDI2VC+, v2, is finally finished. These are the implemented features:

MIDI2VC+, v2, primary side.

MIDI2VC+, v2, secondary side.

• 5 octaves 1V/octave VC output
• 3V gate output
• Low note priority
• Filters channel number, plays any MIDI-note regardless of channel
• MIDI input and USB-MIDI input. Prioritise USB if both are connected.
• Reset button
• Pitch bend, 2 semitones in each direction
• Input voltage 8-35V (high input voltage may need heat sink on regulators)
• 5us trigger output
• Powers USB-keyboard
• USB interface tested on Korg MicroKEY 25
• RCA jacks as outputs. Used RCA wires are easy to find very cheap, and jacks can easily be salvaged from scrapped audio equipment.
• Both code and hardware is open source licensed to CERN OHL v.1.2.

The PCB needs two straps. The MIDI input to the micro controller has no connection. Pin 4 on the optocoupler needs to be strapped to pin 24 on the micro controller. Also, the optocoupler needs to be powered from the 3V3 rail instead of the 5V. Before the optocoupler is mounted, one needs to cut out pin 6 and the pad from the pull-up resistor right besides it. When the optocoupler is mounted, strap a wire from pin 6 to the 3V3 rail.

Cutout below optocoupler.

Anyone is welcome to edit the layout files to fix the straps. I'm not going to do another version of the PCB at the moment. Eagle files and source code is available here.