16-11-2012, 10:39 PM
Although not strictly 'vintage' this clock is dated 2003 so is ten years old.
This quirky and complex clock was a freebie, given to me as 'not working' which was hardly surprising in view of the large number of mechanical parts in the design.
I've posted this in case it helps others who may be asked to repair one of these in future years. They are likely to become even more problematic with time.
The clock has a seven-segment display comprised of pins which cast a shadow, giving the time display a 3-D effect. The problem with this clock was that some of the segments would pop in and out several times before settiling down at the correct time.
The 12V DC power supply was routinely exchanged for one with the proper UK plug rather than the continental design and adaptor as supplied. The swap made no difference and the fault was still present.
First clue to the problem was that the minute 'units' display was worst affected with the minute 'tens' also exhibiting the problem to a much lesser extent. These digits work hardest and if the clock has been in continuous use since 2003, then a lot of minutes have been clocked up in that time. Wear and tear was the most likely cause of the problem, exacerbated by the fault itself failing to lock the segments first time after a digit change.
I removed the six screws holding the case together, then tried to remove the rear casing. This was held by the 12V supply socket and the internal reset switch, so these were released (one small screw for the supply socket and two for the reset switch). The back half of the case was removed and put to one side.
The internal construction is of high quality with high density components packed on a pair of back-to-back printed circuit boards labelled 'solenoid pcb' and 'contact pcb'.
The pcbs are connected together by a pair of short ribbon cables without connectors, presumably to aid reliability.
Clock pcb.
The actual clock appears to be simply a microcontroller embedded on a much smaller pcb by epoxy dipping. This is mounted to the side of, and separate to the main assembly.
Solenoid pcb.
The solenoid pcb carries a solenoid for each of the 7-segment display character, each being driven by a discrete transistor circuit. There are a lot of discrete components mounted on the board, in the form of 24 identical drive circuits.
Contact pcb
The solenoids drive an arrangement of mechanical latches on the contact pcb which lock on the first pulse of the solenoid and release again when a second pulse is applied. Each latch enables or disables a block of pins corresponding to one segment in the 7-segment display. The timing is such that when displaying for example, the figure '8' needing all segments to be latched, the solenoids operate at timed intervals presumably to prevent a high load on the power supply.
Pulsing a solenoid once will display a segment by latching it and pulsing again will make that segment disappear. The timing to remove the time 11:59 and replace it by 12:00 is therefore quite appreciable and much longer than that for 12:00 to 12:01 as some segments need to be made invisible by unlatching, and others made visble by latching. Each digit is handled individually by multiplexing presumably to prevent a high load on the power supply.
The fault itself.
There is a feedback arrangement consisting of gold-plated switches for each segment which appears to provide the microcontroller with the status of all segments at any time. The switches operate every time a segment changes status and the 'minutes' switches will get quite a hammering over the years.
The switches are mutiplexed onto 7 x 4 wires using a diode matrix presumably to keep the lead count to the microcontoller low.
Again, I'm presuming that if the wrong status is reutrned for a segment, then the controller will flip it again in an attempt to correct the display, causing a problem if the switches are worn or dirty. If one of the switches does not operate correctly, then the associated segment will flip backwards and forwards until it is correct.
The problem was found to be due to dirty switch contacts.
Repair progression
To access the switches, the six screws were removed and the pcbs assemblies separated to the extent of the ribbon cables. The operating pin and cam for each segment that is not displayed will now fall out so BE CAREFUL.
The remaining pins and cams for the segments being displayed were removed by depressing the cams manually to change the segment dispaly status. Lots of small pieces fall out, but don't worry, they go back with care.
Once all the pins and cams had been removed (one cam and one pin for each segment), then the switches were cleaned using a small cotton bud and methylated spirits. I did not use switch cleaner as it leaves an oily residue and this mechanism was obviously assembled dry with no lubricant.
Once clean, a trick was needed to get it back together again.
Firstly, each pin and cam was replaced, then the cam depressed to display the segment. This locked each cam and pin in position so they didn't fall out again.
The display showed '18:88' when all segments were locked as this was a 12 hour display clock.
The solenoid pcb was aligned and fitted to the contact pcb and all screws refitted.
The clock was tested for correct operation before refitting the power and reset connectors to the rear half of the case and bolting it together again.
Also noticed at this time was a small compartment containing three small mercury cells for remembering the correct time with power removed.
The cells were replaced as they were approaching 10 years old.
The clock now works perfectly well, but by design is both noisy and irritating in operation. The repair was not difficult, but time consuming due to the large number of fiddly, small parts in the design. It is not necessary to remove the front perspex window or the segments themselves.
I've posted some pictures indicating the overcrowded component layout of this design and hope this helps somebody who wishes to tackle a similar problem with this type of clock.
Dennis.
This quirky and complex clock was a freebie, given to me as 'not working' which was hardly surprising in view of the large number of mechanical parts in the design.
I've posted this in case it helps others who may be asked to repair one of these in future years. They are likely to become even more problematic with time.
The clock has a seven-segment display comprised of pins which cast a shadow, giving the time display a 3-D effect. The problem with this clock was that some of the segments would pop in and out several times before settiling down at the correct time.
The 12V DC power supply was routinely exchanged for one with the proper UK plug rather than the continental design and adaptor as supplied. The swap made no difference and the fault was still present.
First clue to the problem was that the minute 'units' display was worst affected with the minute 'tens' also exhibiting the problem to a much lesser extent. These digits work hardest and if the clock has been in continuous use since 2003, then a lot of minutes have been clocked up in that time. Wear and tear was the most likely cause of the problem, exacerbated by the fault itself failing to lock the segments first time after a digit change.
I removed the six screws holding the case together, then tried to remove the rear casing. This was held by the 12V supply socket and the internal reset switch, so these were released (one small screw for the supply socket and two for the reset switch). The back half of the case was removed and put to one side.
The internal construction is of high quality with high density components packed on a pair of back-to-back printed circuit boards labelled 'solenoid pcb' and 'contact pcb'.
The pcbs are connected together by a pair of short ribbon cables without connectors, presumably to aid reliability.
Clock pcb.
The actual clock appears to be simply a microcontroller embedded on a much smaller pcb by epoxy dipping. This is mounted to the side of, and separate to the main assembly.
Solenoid pcb.
The solenoid pcb carries a solenoid for each of the 7-segment display character, each being driven by a discrete transistor circuit. There are a lot of discrete components mounted on the board, in the form of 24 identical drive circuits.
Contact pcb
The solenoids drive an arrangement of mechanical latches on the contact pcb which lock on the first pulse of the solenoid and release again when a second pulse is applied. Each latch enables or disables a block of pins corresponding to one segment in the 7-segment display. The timing is such that when displaying for example, the figure '8' needing all segments to be latched, the solenoids operate at timed intervals presumably to prevent a high load on the power supply.
Pulsing a solenoid once will display a segment by latching it and pulsing again will make that segment disappear. The timing to remove the time 11:59 and replace it by 12:00 is therefore quite appreciable and much longer than that for 12:00 to 12:01 as some segments need to be made invisible by unlatching, and others made visble by latching. Each digit is handled individually by multiplexing presumably to prevent a high load on the power supply.
The fault itself.
There is a feedback arrangement consisting of gold-plated switches for each segment which appears to provide the microcontroller with the status of all segments at any time. The switches operate every time a segment changes status and the 'minutes' switches will get quite a hammering over the years.
The switches are mutiplexed onto 7 x 4 wires using a diode matrix presumably to keep the lead count to the microcontoller low.
Again, I'm presuming that if the wrong status is reutrned for a segment, then the controller will flip it again in an attempt to correct the display, causing a problem if the switches are worn or dirty. If one of the switches does not operate correctly, then the associated segment will flip backwards and forwards until it is correct.
The problem was found to be due to dirty switch contacts.
Repair progression
To access the switches, the six screws were removed and the pcbs assemblies separated to the extent of the ribbon cables. The operating pin and cam for each segment that is not displayed will now fall out so BE CAREFUL.
The remaining pins and cams for the segments being displayed were removed by depressing the cams manually to change the segment dispaly status. Lots of small pieces fall out, but don't worry, they go back with care.
Once all the pins and cams had been removed (one cam and one pin for each segment), then the switches were cleaned using a small cotton bud and methylated spirits. I did not use switch cleaner as it leaves an oily residue and this mechanism was obviously assembled dry with no lubricant.
Once clean, a trick was needed to get it back together again.
Firstly, each pin and cam was replaced, then the cam depressed to display the segment. This locked each cam and pin in position so they didn't fall out again.
The display showed '18:88' when all segments were locked as this was a 12 hour display clock.
The solenoid pcb was aligned and fitted to the contact pcb and all screws refitted.
The clock was tested for correct operation before refitting the power and reset connectors to the rear half of the case and bolting it together again.
Also noticed at this time was a small compartment containing three small mercury cells for remembering the correct time with power removed.
The cells were replaced as they were approaching 10 years old.
The clock now works perfectly well, but by design is both noisy and irritating in operation. The repair was not difficult, but time consuming due to the large number of fiddly, small parts in the design. It is not necessary to remove the front perspex window or the segments themselves.
I've posted some pictures indicating the overcrowded component layout of this design and hope this helps somebody who wishes to tackle a similar problem with this type of clock.
Dennis.