What is this electric clock mechanism

Hello all

Can anyone tell me anything about this clock mechanism? Is it useable? I think it will look great in my workshop.

Simon.

That looks like a repeater clock of the type that was fitted in factories and offices. This receives a signal from a master clock in some central location that operates the solenoid to flick the ratchet to the next minute.

Unfortunately, you need a master clock to operate it!

Personally, I would remove the mechanism and replace it with a quartz movement but retaining the case, dial and hands.

Brian

I've got a slave clock in my shed with the original mechanism removed and replaced with a quartz movement. I think if I had my time again, I'd probably drive the original via a Raspberry Pi and an IGBT.

Thanks for your replies. I have got a quartz movement, but I like the idea from Mark Rand, and if its a pulse every minute a 555 and a relay may be enough. Do any of you know what supply it needs? There are 3 wires, I am guessing 2 supply and one signal, is it 240V or something else (110V as IBM is US? 48 V for telephone exchange use?) And any guess at the age? The case is plywood and it looks maybe early 60s to me. I don't think Antiques Roadshow will have much interest, but its a nice piece of industrial history.

Thanks again, Simon

Please don't destroy this artefact by putting a normal movement in it. They are fairly simple to drive and if you don't want to DIY, have a look at this

**LINK**

Off the shelf module available, but low cost is relative. For DIY you could hide a quartz clock with a disk with holes in place of a hand and a slotted optocoupler detecting the holes and driving the coil. just adjust the holes for the pulse rate needed. (for mains use a synchronous motor and gearbox with a microswitch or if you are that way inclined a PIC Arduino or RPi

Robert G8RPI

Hi Simon, I used to run one of these using a GPO clock No 36. To drive the clock mechanism will require quite a bit of current for the oomph needed at 48 Volt. I used a system of a slave relay to protect the master clock movement contacts to drive a nicad on float charge to give the neccessary oomph. Very simple and worked for years. So just saying I would advise as Simon mentions above use the relay. Be aware they are quite noisy aith a noticable clunk every 30 seconds. I presume if you operate the mech yours moves on half a minute?

I sold my clock 36 Mechanism and converted the clock to a quiet german quartz mech. Finding one with the right hands was a fiddle. Alan

P.S just so I don't drive Robert into apoplexy with such heresy, I kept the old drive so it can be restored should anyone wish.

Edited By Norfolk Boy on 07/02/2019 07:46:46

Edited By Norfolk Boy on 07/02/2019 07:49:53

Edited By Norfolk Boy on 07/02/2019 07:52:12

$89 plus shipping and possibly import duty is a bit steep, I do like the idea of a quartz mechanism driving a slotted disc with an optocoupler, i may even do this for the heath-robinson value. It seems to have some sort of hour synchronisation mechanism. It has 3 wires, and it looks like the black-red drive it normally with a 1 minute pulse, but as the hour ticks over, a snail cam moves a switch such that it can only be driven by the white-black pair. I am guessing that this pair are active (with 1-minute pulses) from teh hour, for say 10 minutes. So any clock that is out of time will tick normally out of time until the hour is passed, then will stay still until this white-black pair start up on the hour, and this has to carry on long enough for the snail-cam switch to move back to the red-black contacts.

Thinking as I type - i don't need this, I just set the time and leave it.

Does anyone know the volts used? It looks like low voltage, maybe 12? I will start out with a box of AA cells and a switch and see how I get on.

Thanks

Simon

A few years back I converted a very similar clock for an acquaintance. I seem to remember that the solenoid needed 24 volt pulses. I build a driver based on a watch crystal, divided down to give 1 pulse/sec. If I was doing it again I would simply divide 50 Hz mains. Be aware that these mechanisms are quite noisey. OK in the workshop, but the clunk every minute might be irritating in a quiet domestic situation.

Correction. I meant 1 pulse/minute of course.

Simon,

You may find useful information here: **LINK**

https://www.ibm.com/ibm/history/exhibits/cc/cc_room.html

MichaelG.

.

P.S. ... I note there is a number stamped on your movement, but I can't read it.

Please don't debase it by fitting a quartz movement. I've got several slave clocks driven by home-brewed electronics. 555 is not go not accurate enough. The simplest is to use the crystal out of a quartz clock to give you a 1 second pulse and feed this to an arduino to further subdivide it to the (probably) 30 second pulse required, but if you're not into Arduino then you could use decade counter chips. You can also use decade counters from the mains, divide by 10, 10, 10 and 3. Arduino is simpler!

If you're interested send me a pm and I'll send further details

Various interesting stuff on similar clock mechanisms over at Light Straw

Light Straw, being the official name for the cream colour that was used on much GPO/BT equipment

Bill

Bear in mind that there are going to be a million pulses a year so quite a strain on a microswitch. I looked at this some 30 years ago for a church hall and concluded it would be hard to make a mechanism that had the long term reliability and servicing opportunity of a master clock. It is all very well relieving the switch contacts with a transistor but who would repair a specialist one off circuit. A master clock had a support infrastructure of trained clock repairers, now rather less common.

Also wrt slotted discs you need a circuit that limits the pulse length to about half a second. It might be that a trickle charged capacitor would provide the punch.

There was a construction series for a master clock and slave in one of the mags many years ago but I think it was EIM. College Engineering may have done a backplate casting.

I found a description of the IBM system here on the National Association of Watch and Clock Collector forum in a post by tom11443. He says:

IBM/Simplex Synchronous Correction cycle.

IBM developed this system in the late 1940s to replace the older ‘minute impulse’ systems. They sold their time equipment division to Simplex in 1958. Therefore, any clocks with the IBM logo were manufactured prior to 1958. The same basic protocol & mechanisms remain unchanged to the present. Several other manufacturers, including Latham, Cincinnati, Dukane, Standard Electric Time, use the exact same protocols.

Synchronous Systems

All consist of a master clock & a number of secondary clocks in remote locations. These secondary clocks appear to run like an ordinary plug-in electric clock & run on their own synchronous electric motor and will maintain correct time unless there is an interruption in electric service, depending on the master clock only for the periodic synchronizing impulse. These clocks differ from the ‘impulse’ type, which are electrically advanced each minute by the master clock. Impulse type clocks are usually called ‘slave’ clocks since they cannot function at all without the master clock.

There are two basic variations in synchronous systems, ’wired’ & ‘electronic’. In a wired system, the secondary clocks are ‘hard-wired’ to the master clock with their own network of wiring. This dedicated wiring consists of 3 wires & requires that all secondary clocks be connected to this network in order to receive corrections from the master clock. These systems are usually 110 volts but some are 24 volts. The master clock energizes the third wire to apply voltage to the correction solenoids (called ‘clutch magnets’ by the manufacturer) in the secondary clocks.

In an ‘electronic’ system, the correction impulse is applied to the normal building wiring, in the form of a high-frequency audio signal (typically 3510 hz) applied to the normal 60 hz frequency of the electrical system. Special receivers in the secondary clocks receive this impulse & apply voltage to the correction solenoid. In this type system, secondary clocks need no special wiring & may be plugged into any 110 volt outlet in the building.

The cycle can also be initiated manually for testing & adjustment as follows:
With clock running, rotate 90 degrees to the right, so that the ‘9’ is at the top. Hold in this position approximately 8 seconds (you may hear a clicking sound during this time), then return the clock to it’s normal upright position. Within 60 seconds, the hands will slowly advance to the next hour as described below in the third paragraph ** under ‘Hourly Correction’.

The actual correction that follows is the same in both types of systems. The only difference (described above) is the method by which the correction solenoid is energized.

Hourly Correction

The correction impulse is issued by the system master clock at 57 minutes 54 seconds after each hour & lasts for 8 seconds (until 58 minutes 02 seconds). The correction solenoids in the secondary clocks are energized by this pulse. This causes the attached armature & cam to engage a gear & rotate upwards to lift a latch allowing the correction cycle to begin. This takes 6 seconds & causes the secondary clocks to begin their correction cycle at precisely 58 minutes 0 seconds.

The correction cycle lasts precisely 60 seconds, ending at 59 minutes 0 seconds at which time, all clocks in the system will display the same time as the master clock.

** The cycle causes an internal clock-wise sweep, beginning at the 59 minute position & takes 60 seconds, rotating around the dial, ending at the 59 minute position. During this sweep, if the minute hand is not in proper position, when the sweep rotation reaches it, the minute hand is slowly rotated at 1 rpm until it reaches the 59 minute mark. At this point it is now in sync with the master clock.

During the cycle, the red second hand continues its normal 1 rpm motion but is held in place when it reaches the 60 second mark. It is released when the cycle is complete (minute hand reaches 59 minutes) & now minutes & seconds are both in sync with the master clock.

12 Hour Correction

This impulse is issued by the system master clock at 5:57:54 AM & PM & is a longer version of the hourly correction impulse. Instead of lasting 8 seconds, this impulse is of 14 seconds duration, lasting until 58 minutes, 08 seconds. The cams in the secondary clocks are rotated further & lock the correction cycle latch open, causing the cycle to continue until the clocks reach 5:59:00. Depending on how many hours the secondary clocks were slow, they will now be several minutes slower than the master. This error is then corrected by the next normal hourly correction at 6:57:54.

IF I read Tom's description correctly, the clock gets 110VAC at 60Hz on the red and black wire and runs on the 60Hz signal. I believe that the mechanical version gets the signal from the master clock, not from the mains. (Unlike the electronic version.) In both cases remote clocks are periodically synchonised with the master by pulsing the white wire (once per hour, resync every 12 hours.)

Looks like a good job for an Arduino and some simple electronics. A GPS reference module would make time keeping much more accurate than the original.

Dave

One possible source of information is The Clockworks, they specialise in historic electrical clocks. I have been told that there are quartz-controlled circuit boxes available that can directly impulse slave clocks, whether they can operate the IBM ones I don't know.

People over at the Synchronome group have quite wide interests and may also be able to help.

Someone mentioned above the series on making a master clock - actually it was a Synchronome - this was in EIM some years back, by Elliott Isaacs, between Jan '87n and May '88. Copies are around. The backplate casting is no longer available, but the articles included instructions to fabricate one, or it could be made with some flt plate with a few blocks bolted on. However whether the pulases from Synchronome would operate your clock I don't know.

Here you go, no experience of them myself as I have an ex-GPO pendulum master clock to generate my 30 second pulses.

N.B. if someone intends to drive multiple slaves from one master, then a relay is required to save overloading the master's pulsing contacts.

I can remember tales of a colleague at work who wired up his home-made, well work-time-made, Christmas lights; seem he thought it would be a good idea to have them flashing off the 1 second earth pulse, without a slave relay.

It suffices to say it wasn't one of his better plans.

This is not a "IBM/Simplex Synchronous Correction cycle" clock as suggested by SOD, it's a minute impulse with sync. Normally 1 ppm on red back pair but if out of sync a (normally manually initiated) series of rapid pulses on the white / black fast forwards until the cam switch opens. A capacitor discharge as suggested by Bazyle is a good idea. charged to between 12 and 24 V by a series resistor and discharged by a transistor through the coil.

Note that the IBM system is different in detail to the British ones.

I could be persuaded to do a circuit for you if you wanted. For a little extra I could put an rubidium standard in it so you would have your own atomic clock with 1 minute resolution

Robert G8RPI

Perhaps the hidden gem in the link that Peak4 provided is this picture of a Gent's movement

.

Which bears an uncanny resemblance to Simon's "IBM"

.

MichaelG.

Edited By Michael Gilligan on 07/02/2019 12:42:35

Now that's interesting. Presumably someone fitted an IBM clock face to a Gent movement as a repair. I bet there's a story behind that!

It's true the movement doesn't look like a synchronous motor - I assumed, oh dear - that the drum was hidden under the big central cam.  Wrong again.

Good news really - easier to fake a Master electronically for an impulse type clock I think.

Dave

Edited By SillyOldDuffer on 07/02/2019 13:16:53