ABB Aurora Power One Inverter

LED-2 Video Projector

Digital Video Converter

NBOX media player

Samsung ML-1640

Battery Manager Ultra

Profords DC-18 Electronic Alarm Clock

Imtech NTSC decoder

Sharp PC1245
pocket computer

NEC PC8201a laptop computer
EO tablet (1993)
Yarvik TAB250
Apple iPhone 5G clone
Also see:
Cycling Holidays

Profords DC-18 Electronic Alarm Clock

Back in the 1970s, digital was cool. My father bought this Profords DC-18 Electronic Alarm Clock in 1978 and it was a futuristic fixture in my parents' home from then onwards, always telling the time accurately.

The last time when I visited my mother, the clock was nowhere to be seen. My mother told me that it had gone wrong such that it was no longer possible to adjust the time. Luckily, Mum had kept it so I brought it home to fix it.

An easy diagnosis

Before: The damage to the push-button switches is obvious

On removing the cover the problem with setting the time was instantly obvious. The switches underneath the nice dual injection moulded switch tops had begun to crumble to dust.

Time had turned the crumpling rubber switches
into crumbling rubber switches.

I tried fitting a single new crumpling rubber switch into the holes in the piece of plastic which held them onto the PCB, but the distance between the metal pads on the PCB was too great so this didn't work.

Clearly new switches would be required, so I went looking for some from our local electronics hobby shop. The PCB mount switches which I found unfortunately didn't fit exactly in the holes on the PCB so I had to solder short lengths of solid wire through the PCB and make connections above. Also, they were a little more than 4 mm thick and therefore I had to use a file to remove material from both the switches and the key tops before correct switch action was restored.

Any clock using these switches is likely to require a similar repair now.

After: New switches look a bit strange, but they do work and they won't move.

Afterwards, the clock works perfectly again:

In the video you can hear the remarkably loud alarm sound


Looks cool from the side too. Note the figure 8 wire.

The case of the clock is discoloured. It was originally white, but now has a slightly brown appearance. This is due to bromide in flame retardent chemicals in the plastic. There are apparently good ways of converting the plastic back to its original colour, but I think I'll first try that on something which is less important to me than this clock.

Another thing which I considered replacing was the mains cable. It's a figure 8 cable with a single layer of insulation which would not pass modern safety standards. However, it's still in good condition, not hardened at all and still plastic. Therefore I've left it in place.

Texas Instruments TMS3834 IC

The workings of this clock are contained entirely within a single integrated circuit - a TI TMS3834, with a date-code of the 11th week of 1978. This chip is a custom version of Texas Instruments' TMS1000 four bit processor - one of several contenders considered as the very first microcontroller.

Many millions of TMS1000 variants found their way into many varied products. This particular version of the TMS1000 was first used in Texas Instruments' own digital clocks beginning in 1974. TI's own clocks used LED displays, and the chip needed assistance even to drive those.


On the base of the clock there are three switches. One is to choose between a 12 hour and 24 hour display, the second is for choosing between 110 V and 220 V operation, the third for 50 or 60 Hz power supply.

You'll find the equivalent of the 12/24 hour switch on any modern clock.

The reason why voltage needs to be selected is also obvious: switched mode power supplies were not so common or so inexpensive in the 1970s as they are now. Therefore this clock uses a linear power supply (with high quality shielded transformer) and it's necessary to choose the correct transformer primary winding to result in the correct voltage for operation.

The function of the 50 / 60 Hz switch may not be so obvious to some readers in the 21st century: it saved the price of a quartz crystal. What's more, it actually made the clock more accurate than was possible with a quartz crystal.

Well before the advent of digital electronic clocks, electrically driven analogue clocks were already synchronized with the mains electricity frequency simply because they used an AC synchronous motor to drive the hands on the clock face. In the 1970s, quartz crystal oscillators were relatively expensive and digital clocks therefore also often derived their timing from the mains voltage frequency. In this case, there's an input to the microprocessor which taps the unfiltered mains secondary voltage in order that the processor can count 50ths or 60ths of a second.

How accurate is a mains synchronized clock ?

When demand on mains power is high, both the voltage and frequency of mains electricity supplies drops slightly and clocks which work in this way run slightly slowly. When demand is low both voltage and frequency rise slightly and such clocks run slightly quickly. As a result, you might imagine that synchronizing with the mains would result in inferior time keeping to using an internal crystal oscillator but that would be an incorrect assumption.

The frequency of mains electricity is regulated by law and therefore overall time keeping is extremely good with any mains synchronized electric clock. In Europe, mains synchronized clocks have perfect accuracy. We are guaranteed an average of exactly 4,320,000 cycles per day (50 Hz * 60 seconds * 60 minutes * 24 hours ). As a result, clocks which synchronize themselves with the mains have the same long term accuracy as International Atomic Time.

I sell bicycle components and lead cycle tours for a living.