Getting our wartime mobile carrier-telephony system TFb2 in a working order again.
Start of our Survey on 25 February 2017
Status: 6 April 2017
Approach B + C
The left-hand side our TFb2 apparatus under investigation
We would like to see how it looks on the rear side
Secondly, how it looks behind the front panel
We are looking at a typical Siemens way of telephony module construction (Bauweise)
From its wiring we can see that it concerns quite a complex schematic
For those interested in a better printable schematic, click at it
It has to be noticed though, that this schematic shows only the so-called: two-wire- in and two-wire- out situation. Whilst it also possesses the options for, either 2 - in, 4 - out or inverse 4 - in and 4 - out.
TFb2 is operating at a carrier frequency of 11 kHz.
Later we, Deo volente, will go deeper in to the technical aspects.
Viewing at it from a different perspective
After having switched on the 12 V supply, buzzing sound started and the control meter responded by erratically going a little bit up and down, but never reaching its full voltage, however, after a few seconds all stopped suddenly
My first thoughts went, the vibrator pack is striking
My second step was dismantling this device
But contacts weren't looking oxidised or being jammed
The concerning HT power module more closely
Before starting with dismantling it, I inserted in first the open vibrator pack again, so that I can touch the contact fingers
Then I measured the resistance between ground and +A, where a resistance existed of, say 41 Ω. A far too low value, and apparently overloading the power pack inside.
I also became aware, that the battery + is connected onto (system) ground, whereas the - pole being handled as the actual battery supply.
After removing 4 heavy screws (not meant ones keeping the power pack together), of which I have no idea yet how these being fit. It was found, that an electrolytic capacitor being apparently defect
But, inside the power pack, the construction is very crampy built and lacking room for by-passing another electrolytic capacitor. What also was found, is, that the components being numbered but not on the schematics at hand. I therefore decided, to disconnect the failing component on one side, and adding at the output contact, an additional 22 µF 450 V capacitor. At least, without load it should run smoothly.
Seemingly the power module being mounted insulated within the (main) frame. Consequently ground is determined at a single point. From the electrical point of view, this has an advantage
Closing for today, we notice:
Rel bk 61a
Rel SK XX D 3/3b
However, even operated without outside load, the voltage doesn't exceed 89 V. Hence, we first should consider this device first.
On 11 March 2017
Although, in the meantime some small experiments have taken place, these have been in average unsuccessful.
Yesterday, I took time to reapproach the annoying repeating defects again. First it was found that with increasing test duration the vibrator tended to stop and sometimes did strike entirely.
Visual inspection shows that this device must have been quite warm, because tar is being pressed outwards
Such phenomenon is quite common when dealing with tar-composed capacitors. What actually happens, is, that tar shrinks and the dielectric wax is hygroscopic and therefore ultimately constituting an inferior dielectric and insulator, when being confronted with high tensions; such capacitor warms itself up and finally breaking down.
Every experimental step was followed by a next nuisance
Although, measuring it on our General Radio Bridge type 1602 all seemingly was OK. Low loss and correct capacitance.
Please notice the wartime Siemens specs: - 40° to + 70° C
The only strange fact is its apparent maximum voltage range given for 15 V pp, whilst it should operate at 12- 13 V dc continuously.
However, I replaced the device and, it might be that the overall performance became better.
The nuisance, the clamping provisions have each time to be pushed to the left-hand side, which isn't always easy
In the back, invisible to us, are rectangular kinds of screw-nuts. Its shape prevents its rotation, but it is time and again tricky to move them out of the way.
Left of the metal strips are kept the rectangular nuts
Now a range of parallel set of measurements being commenced
The meter in front showing the HT current loadwhich is, say, 10 mA.
The meter on the left, measures the 12 V current consumption of the entire Tfb2 apparatus, which is 0.62 A; the right one shows the power-pack HT of 221 V (loaded condition)
Viewing it from a different perspective
It was also found, that the overall performance enhanced, when a stabilised power supply parallel onto the existing (genuine) power supplied is added; which might be caused by instability of the genuine Siemens wartime stabilised power supply
11Rel.besch 510/04a Bl 2 v 4.11.42
A quick internet search didn't bring a match.
Is there someone around who can provide additional information on this quite mysterious circuit?
Our apparatus had been once modified in Denmark (before I swopped it in the early 1980s), where at least the selenium rectifiers had been replaced, mounted on a rectangular PCB, by a bridge of silicon rectifiers.
The series L/C parallel onto the transformer designated Tr, might be like a regular series tuned at 50 Hz; or should it be 100 Hz? The transformer device designated VD might constitute a transformer having an 'air-gap'.
All bits and pieces of information on this topic can be very helpful for our understanding of this type regulating loop-system.
This additional stabilised power supply works fine.
But I have been too many times too optimistic.
Let us see what the next confrontation might bring us!
On 23/25 March 2017
Time has come to test the power supply over a longer operational period; and it was also decided to access the Siemens 12 V magnetically stabilised power supply inside.
A new approach has been initiated
The the battery controlled Tfb2 (left-hand side) runs now rather stable, and we would like to look why initially the Tfb2 power supply fail operating appropriately.
Bez. NA-Gerät für TFb
Zchg. Nr. 11 Rel.trgb. 17a
Spanng. 90 ...250 V
What I never expected, this device indeed operates below 110 V perfectly!
However, when temperature of the internal components reached a certain level, it even operated more reliable when the mains was reduced to 110 V or beyond.
The voltage swing from 220 (230)V down to 110 V didn't change the loaded output voltage (12,90V), within the range we measured; hence > 0.01 V.
Not bad isn't it?
The 'variac' being set at an output voltage of 110 V
Even beyond, I guess 90 V or even less, it wasn't possible to deviate the 12.90 (12.89) volt.
I never expected this incredible stability figure!
On the left-hand side the device designated Tr. corresponds with the one in the schematic below, on the right the device VD is the 'air gap' transformer
The device in front, designated ED is the series resonance choke tuned at 50 Hz
As to make resonate at 50 Hz, the foregoing showed choke ED apparently in series with this 8 µF capacitor is doing the job
The quality after some 75 years past may be regarded 100%.
Because of its hermetically seal-off technique; a most outstanding technique widely used in German wartime apparatus; particularly were critical circumstance occur.
As to understand the foregoing description better, its schematic being reproduced again
The capacitor parallel onto the meter has a value of 3 µF. Maybe this quite large value have been implemented as to prevent for 'switching-on annoyances'.
YouTube films showing some aspects previously dealt with
Thinking it over again, we might explain the circuitry as follows:
The transformer designated VD provides at its secondary side a proportion of the supply voltage needed for the ultimate 12 V dc output.
But, the VD device might have been fitted with a 'core gap' causing a less efficient energy transformation.
Transformer designated Tr is being loaded with the series tuned circuitry consisting of choke ED in series with a 8 µF capacitor. A series tuned circuit constitutes a low impedance at its resonance frequency, which must be considered here being 50 Hz; and is acting more or less in an ohmic manner (we may not neglect the loss within the choke device ED). But such circuit will cause that current will flow through the second winding section of this transformer. But its current isn't dependant upon circuit load, but purely on its series tuned circuit parameters.
Transformer VD will cause more energy transfer loss with increasing load current; hence, providing a lower voltage portion (all together). Somewhere there must exist an equilibrium, but this is given by the various interacting circuit parameters.
Now an estimation, significant is the tuned ohmic circuit at 50 Hz, which will cause a more or less constant current flow through its series tuned circuitry. Normally, loading is dependant on the load provided, but in this case its behaviour is only its ohmic value, but a constant loading.
In an catalogue yesterday (6 April '17) provided by Florian Eibensteiner from Austria, it was quoted that also vector summations took place. But the devices handled within it were pure AC power supplies. I cannot judge the amount of phase-shifts encountered between the output of Tr and the VD devices.
I must admit, that my explanation might not be fully correct, please come forward and explain it on the hand of existing literature.
I must also
admit, that the swinging choke following the rectifier (Graetz) bridge, is
curious. I can remember, however, that such a device
is provided with an air-gap
as well, it hasn't an
airgap causing - that with increasing current a decreasing self-inductance (core
saturation) will come into affect; as to compensate the normal (low-pass-filter)
voltage drop. Such a provision has to be designed for its particular task.
Interesting is to notice: that when there isn't a load existing at the output terminals, that the output voltage reached >> 20 V dc.
This is due to the lacking saturation effects caused within the VD and to a lesser extend by the swinging-choke loading the rectifier bridge.
Please, repeat reading this explanation, as to become more acquainted to the way this technique works.
I never expected that such a non-electronically controlled circuitry could operate so sound as does ours. Over an input swing between, say, 90 V and 230 V ac mains input.
On the other hand, when we started up this stabilised power supply first, I encountered a rather instable voltage output.
During operation in an opened condition the temperatures rose, over time, from room temperature, say 21° C up to:
55.4° C for the transformer core Tr; its windings became 47° C
32.8° C for the VD transformer core; windings about 32.8° C
42.3 for the ED core; whilst the winding temperature outside gave 44.4° C
The rest of the components, including the series capacitor of 8 µF as well as the parallel capacitor of 3 µF holds a temperature of 22.8° C
With increasing duration of operation, some slight increasing voltage deviations came to light. This could be cured fully by reducing the supply voltage down to 110 V ac. After a while, the voltage was being brought to 220 V; and all operated entirely stable again.
We might draw the conclusion, that the transformer and tuned choke circuit does cause this minor instability. Therefore we find at the front panel the rather unusual cooling grids.
The perforations are good visible and does make sense
On 6 April 2017
I proceeded with our Tfb2 (carrier-telephony) project.
First the second Tfb2 unit had to be taken out of its case, because apparently the anode supply voltage was far too low being 117 V.
It showed that blocking capacitors were having signs of blowing cases.
But a misfortune, some wires were broken off but where to connect these onto?
Maybe one point hasn't been cleared yet, but the filaments being supplied with about 12.6 V, as well as the HT is available.
In the meantime some time has been spend on getting at least two telephones operating appropriately. Problems have been encountered with the carbon microphones, for which a test facility is lacking. However, the two sets operating are performing perfectly well.
The Leclaché batteries being replaced by Duracells type D, within an appropriate battery holder . On the internet we could also purchase a good bargain of Industrial types.
The current test setup
After some gain adjustments all works remarkably sound.
Quite amazing that after say 73 years such a system can be operated again
It can be noticed that this set went into post war service again in 1948, and last revision had been commenced in 1953
According the text on the white plate it had been operated between Straubing and Deggendorf in Bavaria. This set constitutes post A, whilst our second set being operated in station B mode.
This set is being supplied by means of a provisional internal mains supply. A bit crude, but might do the job.
Viewing the speech signal on the carrier telephony interconnecting line.
Another speech example on the carrier interconnection
Hans Goulooze talking on the FT33
One of my tests
To be continued in due course
By A.O. Bauer