Amiga Technical Resource

A guide to repairing the Amiga real time clock circuit

This section will cover how to diagnose and repair common faults with the RTC circuit, as well as other common problems caused by a leaking battery.

PLEASE NOTE that this information is specifically for the A4000D, but the theory and repair notes can also be applied to other models.
The exact devices used and component pin functions can change between other models!

You can jump directly to the relevant section:

Disclaimer and caution

While the information on this page has been checked and is correct to the best of my knowledge, there is still the possibility of unintentional errors.
Please report any errors directly to this address so they can be corrected.

As with any repair work to delicate electronic equipment, you risk causing further damage to your system or yourself. I cannot be held responsible for any equipment damage or personal injury.

It is strongly advised that you possess some good common sense and preferably have previous experience in working with electronics before undertaking any repair on your system.
As always, you should take all anti-static precautions when working with semiconductor devices.

Theory of operation

In order to repair the circuit, we need to have a basic understanding of how it all works, so that we actually know what we're doing.
Below is the schematic section of the RTC circuit. Click on the picture for an enlarged version.

RTC schematic

U178 in the centre is the actual real time clock device, a RP5C01 made by Ricoh. It permanently holds the time and date and tracks the time both when the computer is on and off using crystal Y176 for timing.
Trimming capacitor VC190 is for setting the crystal's frequency to 32.768kHz. This will be factory set, so do not alter this unless you have an accurate frequency counter.
The time/date is only ever updated when you manually "save" the time from the Time Preferences program, or from an automatic time setting program like FACTS.

In the centre right hand side of the picture, you can see the battery BT176. This is the corrosive beast in question which keeps power supplied to the RTC when the system is switched off. It's charged via the +12V rail via resistor R178 and R179 which limits the maximum charge current to around 4mA. This means a dead flat battery will take longer than 15 hours to be fully charged.
A fully charged battery keeps the data for approx 5 months with the RTC drawing 15µA.
Specifications of the RP5C01 can be found on this datasheet.

Diode D177 and D178 are to limit the voltage to approx 5V across the RTC and battery and to prevent current being drawn from the +5V rail.

In the top centre is the 10k ohm resistor pack RP176. This is simply a network of 10k resistors used to bias or "pull up" the various I/O and data lines so they are always in a known state, rather than left "floating".

On the left side, we have U177 which is a 74HCT174 D-type flip flop, being used here as a latch on the address lines.
This acts like a 4 pole switch, which connects U178 to the address bus lines 2, 3, 4, 5 when the LATCH_ADR signal is received. Datasheet here.
At the same time this happens, either the _RTCW line is pulled low for a clock write (save) process, or the _RTCR line is pulled low for a read process. The time data is written/read via the 4 data lines D16-D19, to which the data access position is controlled using the 4 address lines.
Note that read and write operations occur in a few microseconds, so you will need a digital storage oscilloscope if you want to look at this in detail.

Below is the section of circuit as it appears on the A4000D motherboard.

RTC overview

RTC not seen by system

Corrosion from the battery can eat through PCB tracks and vias (the small holes which are used to make tracks change between different layers in the PCB).
If the RTC tracks are damaged, the system can no longer read or write to the clock. The corrosion can also damage ICs, causing them to operate incorrectly or not at all.

On start-up, the system attempts to read from the RTC. If it cannot "see" it, the system time will default to January 1st 1978, or perhaps some other random date.
Some programs can read the RTC directly, like the Workbench Time Preferences, Scout, etc. If the Time Preferences program cannot see the RTC, the "save" button will be ghosted, meaning you can't save the time (as there is no RTC to save to) and Scout simply reports that the RTC is not found.

Begin by visually inspecting the corroded area, looking for obvious breaks in the copper tracks.
Usually these are hard to spot as breaks tend to occur inside vias themselves, or at the point where PCB tracks join onto pads.

Next get the digital multimeter and using the appropriate ohms range, measure the following pins on the latch U177. Pins are arranged as per the following picture. Pin 1 is located next to the white lettering "U177" on the motherboard:

U177 SO16 package
Resistance measurements on U177 (latch)
Pin 1 1k ohm between pin 1 and 16 (+5V)
Pin 2 10k to ground, connects to U178 pin 7
Pin 3 2.5-3k to ground, connects to U175 pin 6
Pin 4 2.5-3k to ground, connects to U175 pin 7
Pin 5 10k to ground, connects to U178 pin 6
Pin 6 2.5-3k to ground, connects to U175 pin 8
Pin 7 10k to ground, connects to U178 pin 5
Pin 8 less than 1 ohm (approx) to ground
Pin 9 High resistance (greater than 1M ohm), connects to U150 pin 72 (via 47 ohms)
Pin 10 10k to ground, connects to U178 pin 4
Pin 11 2.5-3k to ground, connects to U175 pin 9
Pin 12 not connected
Pin 13 not connected
Pin 14 not connected
Pin 15 not connected
Pin 16 Less than 1 ohm (approx) between U975/U976 (located next to U177) pin 16 (+5V)

Exact measurements may vary slightly with the type of meter you have, but needless to say, pins 3, 4, 6 and 11 should have a similar reading as they are all address lines on the same bus.
Pins 2, 5, 7 and 10 should all measure about the same, as they all connect to U178 and have 10k pull up resistors on them.

If you measure anything much different from these results, most likely there is an open circuit track to be repaired.
Note down where the faults were and continue below.

Use the multimeter on the lowest resistance (ohms) range for the following continuity checks on U178 between U177 (latch), U175 (ROM #2). If the comment is "connects to", then the measurement should be approx 1 ohm or less between the two connection points.

U178 DIL18 package U175 DIL40 package

Continuity on U178 (RTC)
Pin 1 1k ohm to ground
Pin 2 Connects to U150 (Gary) pin 56
Pin 3 Connects to ground
Pin 4 Connects to U177 (latch) pin 10
Pin 5 Connects to U177 (latch) pin 7
Pin 6 Connects to U177 (latch) pin 5
Pin 7 Connects to U177 (latch) pin 2
Pin 8 Connects to U150 (Gary) pin 52
Pin 9 Connects to ground
Pin 10 Connects to U150 (Gary) pin 50
Pin 11 Connects to U175 (ROM#2) pin 13
Pin 12 Connects to U175 (ROM#2) pin 15
Pin 13 Connects to U175 (ROM#2) pin 17
Pin 14 Connects to U175 (ROM#2) pin 19
Pin 15 10k between U975/U976 pin 16 (+5V)
Pin 16 Connects to 1 leg of crystal Y176
Pin 17 100k to other leg of crystal Y176
Pin 18 1.2k to battery + (Refer to side note!)
Some of these pins connect to Gary (U150), refer to the image below for the pin numbering details.

If the battery is fitted to the motherboard, take care not to touch the multimeter probes to pin 18 of U178 as this will have 3V on it, which may damage the meter in resistance mode.
The check on pin 18 cannot be done with the battery fitted as the voltage accross the 1.2k resistor will effect the resistance check.

Note down any differences you see and continue below.

The following pins of Gary (U150) connect to 4 places in the RTC circuit, some of these have been mentioned above.
All 4 connections are listed in the table below. Use the multimeter on the lowest resistance (ohms) range to check there is less than 1 ohm (approx) between them.

U150 PLCC84 package
Continuity checks to Gary (U150)
Pin 50 Connects to U178 (RTC) pin 10
Pin 52 Connects to U178 (RTC) pin 8
Pin 56 Connects to U178 (RTC) pin 2
Pin 72 47 ohms to U177 (latch) pin 9

From the schematic and the above connection tables, you should have worked out if there are any open circuit tracks.
The next thing is to find where the break is. The tracks could be open circuit due to a small scratch in the underside of the motherboard, but if there is battery corrosion, then most likely the fault will physically be in the RTC area.

Find the track which is open circuit and physically inspect it all the way between it's origin and destination. Though the A4000D motherboard is a 4 layer board (2 layers of tracks inside the board), pretty much all of the RTC circuitry is on the top or bottom layer, making it quite easy to repair.
Some tracks run under components and resistors, especially under the latch U177. As this component is close to the battery, it is highly likely that track damage has occurred underneath, so it may need to be removed. Refer to the U177 replacement section for details.

To repair a track where it connects to a pad, use a very small screwdriver to scrape away the green solder resist along 3-4mm the track from where it joins to the pad.
Clean the area using a fibreglass pen if available and with isopropyl alcohol.
Using a soldering iron, "tin" the bare copper track. Find some small insulated hook-up wire (about 0.2mm²) and extract a single strand of wire from inside. Solder this between the pad and the tinned copper track, thus repairing the break. If done very carefully, this kind of repair can be done underneath U177, so long as you keep the join as low in height as possible.

Afterwards, use the solder resist repair pen or clear nail varnish to coat the repair work.

An easier, but less attractive way to do the same repair is to use the green enamel wire-wrap wire.
Prepare the damaged tracks as above. Before attaching the wire to the board, apply a lot of heat to the end of the enamel wire along with some solder. The insulation will partially melt away and the solder will tin the end of the wire.
Scrape some green solder resist from the damaged track and solder the enamel wire to it. Cut it to length and do the same to the other end so that the wire joins the broken track, electrically bridging the damaged area.
The enamel wire is ideal for running a long length between two points where a large section of track has corroded in between.

To repair a track which is open circuit at a via (vias are the copper plated holes which electrically joins the tracks between different PCB layers), then scrape away the green solder resist along 3-4mm of the track from where it joins to the via, on both sides of the board.
Clean the bare copper track.
Using a single strand of the 0.2mm² hook-up wire, poke it though the via hole and solder it to the bare copper track on each side of the board.

Afterwards, use the solder resist repair pen or clear nail varnish to coat the repair work.

This diagram shows a cross section of the PCB and how it's made, as well as illustrating how to solder a wire to top and bottom side tracks in order to repair an open circuit via.

After the tracks have all been repaired and tested, temporarily power up the system and see if the RTC is now seen by the system.
If so, the job is complete!
If not, there is probably a faulty component. Most likely the latch U177, or the RTC itself. It's common for U177 to fail if it has been subjected to corrosion.

If the RTC can be seen by the system (Save button in time prefs not ghosted), yet the time is not stored when the power is off, or the clock does not "tick" while the power is off, it's possibly a problem with the Ricoh RP5C01 RTC.
I have new RP5C01 available in stock, Email me to organise a purchase nationally or outside of New Zealand.

Replacing U177

Refer to the guide on replacing SOIC devices to see how to correctly remove and replace an SOIC (such as U177) using surface mount soldering techniques.

Replacement 74HCT174 can be obtained from:
element14 (formerly Farnell), part number 110-3139
Or by emailing me with a request, as I keep a small stock.

Replacing the Ricoh RP5C01 Real Time Clock (U178)

The real time clock device, U178, is an 18-pin leaded Dual In-Line package (DIP) which is usually soldered into the A4000D motherboard. To remove a soldered in device, it's advisable to use a hot air surface mount rework station to minimise the possibility of damage to the circuit board.
Traditional de-soldering methods, such as de-soldering wick and vacuum solder suckers may be used, though caution should be used to prevent damage to the circuit boards tracks or plated through hole pads.

After a faulty device is removed, it is recommended to fit an 18-pin DIP socket to accommodate the replacement device. A suitable DIP-18 socket is available from element14 (formerly Farnell), part number 110-3847

Replacement Ricoh RP5C01 can be obtained from:
Anthony Hoffman, international customers are welcome. Preferred method of payment is via PayPal. Contact me for pricing and availability.
Components are supplied with fitting instructions for the Amiga A4000D.

And also available from
AmigaKit_logo.gif (2340 bytes)

Oscillator Calibration

If the RP5C01 Real Time Clock IC is replaced, or any of the components in the oscillator section are replaced, including crystal Y176 or trimming capacitor VC190, the oscillator circuit needs to be recalibrated for accurate time keeping, else you may find that the clock has lost or gained seconds/minutes while the Amiga has been powered off.

Either of the two procedures below can be used for calibration. The first one requires a frequency counter or oscilloscope with an accurate frequency display. The second procedure uses an accurate frequency source as a reference and any dual-trace oscilloscope to 'zero beat' the Amiga RTC oscillator frequency.

Procedure A:
  1. You need either a frequency counter or oscilloscope with a 1Hz resolution frequency counter, it must have a 10M Ohm input impedance (most x10 probes are suitable)
  2. Power off the Amiga so that the RTC is powered from the backup battery
  3. Connect the 10M Ohm probe from the frequency counter or oscilloscope to RP5C01 pin 17
  4. Using a plastic or ceramic trimming tool, adjust VC190 for exactly 32.7680kHz on the counter/scope
Calibrate_RTC_osc_procedureA.jpg (53116 bytes)
Oscilloscope with frequency counter display of 1Hz resolution

Procedure B:
  1. You need a computer with a sound card, and a dual trace oscilloscope (with 10M Ohm probe)
  2. Download the 16.384kHz sound file as AIFF or WAVE, set all sound equaliser settings to neutral, and play on repeat at maximum volume
  3. Connect the output of the computer sound card to channel A of the oscilloscope, trigger off this channel for a stable "not moving" display of the reference sine wave signal
  4. Power off the Amiga so that the RTC is powered from the backup battery
  5. Connect the 10M Ohm probe from the oscilloscope channel B to RP5C01 pin 17
  6. Adjust the oscilloscope channel B setting so that the 32kHz signal from the Amiga is visible on the oscilloscope display
  7. Using a plastic or ceramic trimming tool, adjust VC190 so that the two signals are not moving backwards/forwards with respect to each other
Calibrate_RTC_osc_procedureB.jpg (59809 bytes)
32.768kHz RTC oscillator input on channel B (top) and 16.384kHz frequency reference on channel A (bottom)
VC190 is adjusted until trace B is not moving horizontally

Repair Service

I also offer a comprehensive repair service down to SMD component level on Amiga main boards for customers from both in New Zealand and internationally.
Repair times are typically 1-2 weeks (not including time in transit). Items are returned with a full product test sheet.
Contact me via Email for an estimate and current repair timeframes.