Siemens & Halske


Type HU 20/48* min

9 U Sk 2021/15/M

9 U Str 2011/2

9 U Ms 2011/2

F-Nr. 2 R 7554

* 48 might point at an operational voltage of 48 volt. Our first experiments have been commenced at an operational voltage of 60 V, but it is necessay to check whether the clock runs sound on 48 V as well.


A survey


Never have been engaged as a clock collector before, I have, therefore, to apologise first, for often not using the correct technical expressions in conjunction to this clock related subject.


Just after this new webpage became yesterday accessible, Alex Ponsen very kindly responded 'that he was afraid to correct me' in regard to technical clock terms. These will be corrected today instantly. There quite likely will be more expressions that need correction.



Clock mechanism is running, therefore the pendulum is reproduced a bit fuzzy


This page was initiated on 5 April 2015

Status: 8/9 April 2015


Some months ago Dick Zijlmans donated an electrically driven master clock  to our collection (in Germany known as Hauptuhr); because he possesses no longer space for displaying it. Following the objectives of our Foundation, it was evident that we should, at least trying, to get this clock operational again.

He possessed no documentation. I was told that such master clocks were used in former Dutch telephone exchange systems. Actually two were used at once being synchronised upon by timing signals originating from the former Dr. Neher Laboratory in Leidsendam ; the former Dutch PTT scientific centre. The Dutch PTT equals British GPO. These kind of centres do no longer exist in developed countries. However, they operated both in parallel, I guess, in a democratic circuitry controlling the slave-clocks somewhere in the exchange building as to provide the time-reference for the tariff-counters. I was told, that when matters ran regularly, that both pendulums operates as if they were mechanically fixed at one another.  

However, for safety reasons the clock pendulum has to be removed first before transporting it. For it Dick removed also the clock-module.

Dick Zijlmans never got this watch operational and knew that it once had been operated somewhere in Haarlem, which, in those days, relied upon Bell systems (BTM) their system operated on 48 volt (+ kept at ground level).

Since late 2011, I am also engaged as Secretary of the Dutch Foundation sTEN. Its Chairman Rob Timmermans has, among many things, a penchant to 'master clocks'. What a great luck. He was willing to come over and fixing the various wires which formerly, for practical reasons, had been disconnected.

Electrically he managed it quite easily, but we encountered a nuisance, as the clock, when it once was put out of order, was abruptly stopped in the middle of its operation sequence. We should hereby be aware that Siemens master clocks relied upon a regular clockwork, which relied fully on a classical 'loaded wire' mechanism. Electricity only was used for winding up the load (weight). When the mains failed, it continued un-interrupted operation, until the load came at its lowest position (no winding-off energy left). I do not yet understand, but was told, that when the power being available again the clock is winding-up its loading and, at the same time providing + and next - 60 volt pulses which were fed onto the slave clocks which by this means synchronised finally onto the master clock again. On Tuesday 7 April, Rob Timmermans explained during a telephone conversation, that the so-called memory signals (as to synchronise the slave clock again) fed onto the slave (Tochter) clock was meant for a power-failure-duration lasting, say, about an hour or so.

Rob Timmermans promised to keep us informed when a slave clock is on sale somewhere. Thus, patience is what we should have firstly.  

I thought to be clever and manually triggered the winding-up-mechanism in a certain position.

A learning process was starting, with many disappointments to be encountered first. It was soon noticed that the clock-winding-up or unloading system should be blocked by a special provision, when it reaches a limit.

Afterwards, it was discovered that the wire which was accompanied (attached) to our clock when we got it, was not having an appropriate length.   



Viewing one of the first experiments, the clock-weight hanging at its lowest position

Photo taken during clock operation; causing a fuzzy pendulum.

From this low angle perspective it is not well visible where the lowest-level-blocking mechanism is. 



The white arrow (pointer) is showing the movement of the nickel-coloured arrow that, when the wire-loading has reached its lowest position, is pushing the lever to turn some degrees anti-clockwise

By doing this the clock stops moving and electrical contacts being forced to initiate the electrical winding up modus. Which will not happen as long as electrical supply fails.

Very curious, the nickel arrow pointer is only visible at the two extreme clock loading situations (loaded and un-loaded).



Normally not visible from outside, please notice the drum which shows that the winding up- or de-winding procedure is somewhere in the middle of its operational sequence

When the drum is filled fully with nylon wire, this may indicate that the clock possesses its maximum winding-up power and/or it has reached its maximal capacity and some action is due to come; as to prevent that the winding-up wire can run off the drum or being 'torn apart'.



The clock winding-up has reached its ultimate level, and the nickel arrow is just pushing lever clockwise; by doing so the electrical circuit being disconnected from electricity (please see for it Film 175)

When matters goes normal, thus electricity is available, the clock mechanism relies upon gravity and the drum rotates a bit anti-clockwise. When the seconds-hand passes a few seconds '0 minute' the drum movement has forced the nickel arrow to de-block the lever-arm and electricity just can provide two pulses and the nickel arrow again moves the lever-arm clockwise. Please notice Film 175 at the end of this webpage. Rob Timmermans told me today also, that the first pulse is where the 'seconds hand' (pointer) is being aligned upon at the beginning of a new minute (0 degree or facing entirely upwards); these two pulses being conveyed onto the slave clock (Tochter-Uhr) as to keep it in concert to the master clock (Hauptuhr).

Hence, when - a new minute starts - two winding-up pulses occur then the electrical circuit being disconnected and the drum being driven by gravity anti-clockwise (again).  


Maybe not well visible, but behind the vertical mounted electrical contact there is a small kind of flee-wheel, which movement being triggered by means of a relay. A vertical movement being converted into a rotation, which by its very means a wire is wound-up at the clock drum.



Viewing an empty drum. Nylon wire not yet attached

At the front the lever-mechanism that has such essential function. The nickel arrow is just before pushing the lever anti-clockwise and by this means forcing electrically winding-up of the drum-wire and stopping the pendulum movement. Failing electricity will then (eventually) causing a clock stand still.



This happens when apparently there exists a mismatch between both extreme winding situations. The winding up is acting not in concert to the synchronised blocking mechanism

When no action is undertaken the nylon wire will be stretched up to is limits and ruining our clock will be the result.


YouTube films explaining the various aspects

Film 171:    Viewing the Siemens & Halske master clock type HU 20/48 The clicks you hear in the background are due to the electrical winding-up system. Film was taken during earlier experiments, trying to replace the former 0.6 wire by means of a 0.5 nylon wire. Which caused curious derailing of the wound-up-drum-wire from its drum. The clock load (weight) is about its lowest position possible. Far down on the right-hand side, we see an electrical magnet which should interact with a small metal bar attached at the lower pendulum. By this means once the clock could be synchronised onto the timing signals originating from the former Dr. Neher Laboratory in Leidsendam.

Film 172:    Shown the electro-mechanical winding-up of the clock. In this mode, from every start of a new minute (0 second) 17 pulses being generated causing a slight winding-up of the drum. Total winding-up time is about  2 to 2 hours. The clock weight is at about its lowest level.

Film 173:    Viewing the clock winding up and the related trigger mechanism. Please also notice, further on, the revolving wheel in the background, being activated by means of relay pulses.

Film 174:    Viewing the winding-up mechanism in detail. A vertical relay movement being converted into wheel rotation. The wheel movement being coupled by means of gears onto the winding-up drum mechanism.

Film 175:    Viewing the ultimate upper winding limit and the repeating - winding-up and blocking sequence; and its operational bouncing between gravity- and electrical winding-up period for two pulses and relying the rest of the minute on gravity operation again. This should become the regular alternating clock status, as long as uninterrupted electrical power is available.


What remains firstly - is checking whether the wire winding-up does function repeatedly. Especially worrying is the proper way of winding up the wire at the drum perimeter next to one another soundly.

When this has been commenced successfully, we can mount the clock- face or dial and attaching the clock hands again.



On 7 April 2015

Today we found the clock load at its lowest ultimate level, which is what should be the hoped for status.


Clock-load being wound-up at its highest level

Down on the right-hand side the electromagnetic coil, which was being activated temporarily during the sequence of synchronising upon the time-signals originating from the Dr. Neher Laboratory in Leidsendam. The switch in the middle on the far left-hand side was activated only during synchronisation, when this was commenced, it was put in its off-mode. We would like to lock our clock upon the wireless time-signals originating from DCF 77. This can only work in the region of 'parts of a second' deviation. In my perception such radio controlled signals might also causing nuisance! For instance, how does the synchronising system know what the status of the clock is - as well as that of the pendulum?



On 8/9 April 2015


Yesterday (7 April) I did disconnect the power onto de clock before leaving.

The clock mechanism had stopped somewhere yesterday and the next morning the power supply (60 V dc) being switched on again. Nothing was happening. The day before, by pure coincident, I discovered that when one of the switching contacts was being shorted by a screwdriver the electrical winding up did start functioning again. I repeated this several times, and it was found that sometimes it does start-up from it selves, sometimes it did not.

Let us go into the problem a bit.



Viewing the essential electrical section of the clock

On the left-hand side, what I designate, an eccentric arrangement. The contacts being activated by the cam-wheel is for generating the pulses meant onto the slave clock.



The two contacts within the red ellipse is constituting an interrupting switch, which should block winding-up and synchronisations pulses during the commuting sequence of two pulses; which sequence being interrupted after about 33 seconds has passed, lasting until the start of a new minute

Considering all the facts, it became apparent that between 33 and 60 seconds there never can be an electrical action. There must thus exists a means to bridge these contacts temporarily.

But how?

The only means possible is that these contacts being fed onto outside somewhere.



My considerations proved to be correct, the resistance between the two contacts on top and the switching contacts was 0.6 ohm. The left one being wired onto outside contact 1 and the right-hand switching contact onto outside contact 2; mounted on top of the HU 20/48 Hauptuhr (master clock)

Please notice also the contacts 7 and 8 which accompanied text tells us that on these two point the synchronising signal being fed onto. Inside the clock housing there is a switch which can further interrupt the synchronising signal, in the far past originating from the Dr. Neher Laboratory in Leidsendam. Regularly this synchronising signal was only acting temporarily. 



This is what actually is interconnected onto the previous contact strip, in combination with the encircled pair of contacts. These contacts being open during the sequence of 33 up to 60 seconds (the latter figure meaning the start of a new minute)

Parallel onto this pair of contacts is a resistor of 26 kΩ ; why is only a guess: it might assure that in the load foregoing the circuit, such as an inductance, that some current is flowing all the time.

The activating provision should only be operated when the moving energy stored in the drum becomes zero (mechanical blocking of the drum). It needs only to be activated for a few seconds, maximally 27 seconds, though I estimate in practice for a few 'clicking sequences' only. 


Experiments proved also that 48 volt power supply might be a bit too low and we therefore chose for continuation with 60 volts instead.



Having placed the dial, the clock stopped running after half hour. I did remove everything - all runs fine again

Eventually I discovered that I did not push the 'hour-hand' far enough upon the shaft and it coincided (friction) with the movements of the minutes-hand.

My fault, of course, but that is all in the game. A necessary learning process.


Today I was picking up some parts in the Klooster and found that the pendulum magnitude has increased. On the right-hand side, now reaching 70 degrees and towards the left it reached 80 degrees. Some times previously, it was already discovered that the pendulum magnitude can increase over time.

The clock operates now stable.


To be continued in due course



By Arthur O. Bauer