Type T 8

is starting to operate again


Page initiated on 31 October 2014

Status: 30 March 2015

2 + 3 + 4-5c! + 6d + 7e + 8f + 9g + 9gb + 10h + 11i + 12j + 13k + 14L + 15


We have, for what ever reason, lost more than a year in bringing our Magnetophon T8 apparatus to operate again.

As to create new momentum, I have decided to start this new approach by taking over the task myself.

The aim in first instance, is to get the system operate again. My personal impetus is to create sound cabling which is designed such that modules can be operated experimentally when these are placed on our small moveable lab-table. Hence, we need longer than usually interconnecting cables.

Luckily, we possess the original Tugel connector, both male and female.

When this experimental stage has been accomplished, Dick Zijlman's task will be re-styling the cables in a proper manner; as he has a penchant for.

However, what counts now is getting the Magnetophon system in a working order again.

Please bear in mind, that we obtained it in bits and pieces. Luckily, some documentation was accompanied, which makes life more easy. Lacking particularly technical documentation on the Telefunken T8 machine. Like, for example, how bearings should be lubricated. Who can help us?

However, its electrical (sound) recorder wiring is quite basic and could easily be understood. This will not mean that I fully understand the various, Magnetophon, motor controlling circuits.

For it I started some days ago with finding out the various wiring connected to the 'recorder head' assembly.



Filter unit W55 was already connected (wired) by the previous Belgian owner, who very kindly donated this system to our Foundation



Kop means recorder-head assembly

'oude verbinding' means the connections the previous owner had once wired onto it, I guess, many decades before.

Luckily, the connector numbers match one-to-one onto the 'recorder' pre-amp combined with unite RTI V47/b1.



Note-block page 3. Explaining the concerned interconnection at the output-side of the 'recorder module' V47/b1



Moveable lab.-table, viewing my first point of attention: the de-mounted recorder-head assembly

It is quite understandable, that we start our investigation survey in a reverse way; from the recorder head wiring towards both the play- and recorder amplifier modules V41 and V47/b1.



My second move was following the wiring between the Magnetophon and the electronic modules

The cables hanging on the left-hand side downwards are the recorder inter-connections. Whilst, the small metal box in front constitute the 'filter' module (HF low-pass filtering W55) between the 'play-head' (Widergabe-Kopf) and amplifier module V41.

Some are calling module V41 a microphone amplifier, however, on its genuine, maybe wartime or soon thereafter, RTI schematic is it described differently.

After having disconnected the electrolytic capacitors and putting instead modern types replacing them, I did not remove the electrolytic Cs from the chassis. I also investigated two block-capacitors both showing signs of typical high losses; which normally is widening-up their metal-cans. I kept these on place, but disconnected a wire at a single end. One solder-tag is used as to mount a modern type capacitor. Keeping the unit as original as possible; the previous situation is by this procedure easily to be restored again.



My first move was checking whether the voltages are about what is given on the schematic diagram. Which proved to be the case


The great moment of starting up the T8 Magnetophon again is due to come.

What is more appropriate than using a video camera, which is also recording movement and reproducing its accompanied sound; and placing them on YouTube?


YouTube films

Film 150:    Viewing first our moveable lab.-table with on it amplifier unit V41 on the right-hand side we see the recording-pre-amplifier V47/b1. Then I switch on the Magnetophon in the reproduction mode (Widergabe). We are using a small partial tape, which was accompanied with the machine when we collected the system in Belgium, in March 2013. The tape reproduces apparently a Belgium radio broadcast, but sounds over-modulated. It proved, however, that lowering the V41 sensitivity eliminated the sound distortion.

Film 151:    Viewing now a tape which we obtained last year at a NVHR rally. What was recorded on it is unknown. It proved to be a broadcast interview. Most likely recorded in stereo, but reproduced here in mono; as our machine is fit with a single track-head. It becomes quite soon apparent, that the tape was recorded with a machine having its head-assembly at the inside; whereas the older Telefunken and AEG machines used the outside of tapes. For it the tape has to be twisted (rotated) between feeding tape side and the first 'rolling-wheel'. I apologize for not knowing the appropriate expression for it. I am astonished by the fact that although we are listening to the 'wrong side' of the recorded tape, sound being reproduced is quite good.     Dick Zijlmans very kindly provided a small loudspeaker module, which is fit with an amplifier having a symmetric input channel.

Film 152:    Listening to the same tape, but the tape being twisted, as to get the recording layer facing outwards. We are reproducing a tape which we once bought at a really; with the objective having a tape for future experiments. Apparently, the pitch reproduced clearly sounds better.

Film 153:    Viewing in detail the way the tape being 'wound-up'. Although, being not an expert, I guess that it does it rather well.

Film 154:    Viewing the V41 pre-amplifier from inside. The newly implemented capacitor components are good visible. Our aim is not letting it look genuine, but to operate the system the best way possible. All the time bearing in mind, that by simply removing the added components the once genuine situation can be re-established (restored) fully.

Film 155:    Showing the component side of the V41 pre-amplifier of our system. Explaining why and how we accomplished it. Operating the genuine components would sooner or later causing troubles and, most likely, destroying other parts.



On 5/6/7 November 2014


Our machine is capable of replaying, time has come to switch on the recorder amplifier V47/b1



We purchased a simple DVD/CD player by which means we would like to test the recording mode of our Telefunken T8 Magnetophon apparatus


Let us first notice what the V47/b1 input condition are.


Transformer marked 1 of the far left constitutes the input signal channel

In studio technique it is custom to use symmetric cabling, where the induced hum-currents will be cancelled by common-mode currents against environment.

The impedance in German sound and broadcast practice was in the 1930s 1940s and maybe also in the 1950s 200  Ω, later higher impedances were used; maybe that was introduced by adopting American standards.




The only matter we would like to know, is finding out a simple means mixing the left (L) and right (R) channels of the CD player output-signal

The miniature transformer was found in a junk-box, and apparently once meant for coupling a low-power transistor push-pull stage (portable radio) onto a small loudspeaker. Its impedance transformation will be most likely being not optimally. As to prevent overloading of the player output channels I have implemented two small 100 Ω resistors; because I guess the player output impedance will be in the order of, I guess, ≈ 1 kΩ.



It worked instantly well

Although, with a small hum content in the background. Interconnecting the screen of the (symmetrical) recorder cable with the ground of the cinch connectors this downside was cancelled. Good visible is the interconnection between the recorder cable and ground of the CD player. Still a choice of junk-box transformers around. The advantage og using a push-pull transformer is the absence of an air-gap in the transformer core.



Maybe this photo is providing better information as to how we convert the two sound channels (L) and (R) into a single signal channel, and also converting it from an asymmetric signal against ground into a symmetric signal; which is necessary for the recording amplifier input channel 



During the course of today's experiments, it was discovered that the already noticed remaining hum was to find in a not connected earth contact at the input connector of V41. Apparently the open transformer input circuit is so sensitive that it picks up hum; as hum remained with an open input connector

It proved that hum existed even when the input-signal-plug (Tuchel Stecker) was disconnected. After having changed an earth connecting and loading the input transformer, hum is no longer bothering. Maybe that when we later will adapt a real reproduction system (with a good wideband loudspeaker system) that some hum might be noticed in the background.  



We have reshuffled the space below the Magnetophon, as to create space for both amplifier modules V 41 and V 47/b1 (bottom and top respectively)

Right of the electronic modules we notice the loudspeaker/amplifier unit. Its specific advantage is that it has a symmetric line-input. This signal comes from the output stage of the V 41 (pre-amp) module.



Whatever we do, some kind of 'cable spaghetti' is always remaining; in this case mainly  because we kept cable-lengths longer than strict necessary, as to allow operating modules from some distance

The way we have currently fixed both modules is preliminary. We would like later to fit them in a serviceable mounting frame.   


YouTube films series II

Film 156:    Viewing and listening to the first recordings we made with our T8 Magnetophon. For it we obtained a very basic CD/DVD player and using it as a signal source. We are listening to music which I favour very much.

Film 157:    Viewing first the loudspeaker then the recording setup, then the tape deck with in the background the just running (playing) CD/DVD player.

Film 158:    Viewing the very simple means by which we convert (mixing) the left (L) and right (R) output channels from the CD/DVD player; the miniature transformer secondary output is providing for this occasion a symmetric signal fed onto the V47/b1 unit.

Film 159:    Viewing again recording the recording setup with the loudspeaker still standing on the ground.

Film 160:    Noticing the Telefunken Magnetophon type T 8 currently recording. When you look on the left to the tape, it is clear that both the Al plate and the once wound-up tape are not flat (plane) and that this tape previously was rewound not in a proper manner. On the other hand, winding up is doing rather well and this tape does not show any sign of wobbling, although it virtually is 'hanging' unsupported.


After having rewound the tape several times completely, the pitch constancy has improved rather well.


On 20/21/22 November 2014


We have experimented with recording sound picked up by our just made operational 'Neumann-Kondensator-Flasche' microphones.



When I switched on the Magnetophon and put the Magnetophon in the recording-mode nothing happened, only listening to what once was recorded on the tape previously


My first conclusion was, that even when the chain from microphone up to the recording input of the Magnetophon would fail, that at least the erasing function - of what is recorded on the tape should have been wiped-out. Which it did not!

Thus, the V47/b module should be accessed firstly.



All those acquainted to fault searching will recognise this setting

It took me more than expected time to determine what the actual fault did cause.

It finally proved to be inside an old, but genuine rubber cable running from an output connector of module type V47/b1  towards the switch mechanically connected onto the recorder-mode-selector part of the tape-deck. (meant is the hanging in the air connection on the left-hand side of this photo)  



Please notice the connection on the right-hand side at pin 7, which is to be interconnected onto connector pin 5 (280 - 290 volt) in the recording-mode. Though, when the connection being interrupted, thus in the stop or play mode, at pin 7 stands stays then about 400 V dc. The 400 volt apparently being a too high value for the just described rubber insulation of the about 65 years old cable. After having cut the wires, but still leaving it for the time being in place, now being by-passed by to new wires; since it all operates again.



Investigating the V47/b1 module at our moveable lab table


This table is a versatile asset, as it carries a wide range of screw-drivers and other means, as well as, an oscilloscope and various test gear.



When you look carefully, you can recognise the by-passed cable which are bound by small blue thigh-raps (binders)


All does function again, also the reproduced sound of the Neumann-condenser-microphones.


We also intend to implement a new webpage, dedicated to bring my old amplifier interface operating again. Something has to be done as to facilitate the adaption of the two microphones which we got from Belgium. The previous owner has removed the original RE084 valves and replaced them by EF86 valve type. But, our own microphone should, which isn't changed, should be operational as well (4 volt dc filament).  



On 23/24/26/27/28 December 2014

On 23rd I received a quasi instant response from Friedrich Engel, in which he adds worthwhile information. He mentioned that - Christmas day did start today, as he was looking for so long, in vain, for such a photo. His and my remarks being marked in colour.


Working on the preparation of BIOS 28 in regard to the Voith-Schneider propeller, I knew from memory that somewhere in the wartime periodical Signal I once have noticed (maybe 45 or 55 years ago) photos on this revolutionary ships propeller type. Although, browsing through all we have (running from 15 May 1940 up to May 1945) I could not trace it. However, I came across the next photo which fits rather well into this webpage.


Please notice the three control handles* on the left-hand side of the lady's (Fräulein) left hand! Are these the forerunners of what later became known as 'fader'† Source Signaal 1st June 1943 issue No, 11 (Dutch language)

Mr Engels received a better quality reproduction, which for this occasion has to be prevented on the web.

Not quite unexpected Jan ten Have responded (26th):   In Engels is het:  “fader”  Van; (doen) verwelken, verbleken, geleidelijk verdwijnen enz. (Old soldiers never die, they just fade away…..)Wetenswaardigheidje:In Duits sprak (spreekt) men eenvoudig van “regler” toen er nog sprake was van de bouwvorm die meer overeenkwam met de bekende, ronde, uitvoering van een potmeter. Die regelaars waren ook toen al uitgevoerd met talloze draadgewonden weerstanden met evenzovele contacten.Toer er later, vanwege het gebruikersgemak, een andere bouwvorm kwam, koos men, naar de vorm, voor de term: “Flachbahnregler.” Met de ronde uitvoering ben ik mijn studio begonnen, overgenomen van de pianist André de Raaf van het toenmalig bekende pianoduo Andre de Raaf en Jacques Schutte. (Jacques was destijds bekend als begeleider van de ochtendgymnastiek van Ab Goubitz ’s morgens om 7 uur, directe uitzending!!) De vlakbaan regelaars heb ik destijds overgenomen van de KRO. Hence, I should designate it: fader instead of feeder!

* Known in Germany as 'Flachbahnregler' in the block diagram below indicated as: W 24a. According my friend Jan ten Have, who told me today, the he used 6 or 7 of these types. The last one he donated it to a British friend. Oh, would I have known this before!

On 26th December, thanks to Gerd Kuper, we can have now a quite comprehensive idea of what such a sound-level-control (W24 / W24a) is about!

We can also notice, that the way the (RRG) operated their Magnetophon machines was practiced at least up to the  1990s; of course, using modern machine types! 

It might be, that this wartime Magnetophon machine (R22a - R122a) operated the same type of capstan motor (Tonwellen-Motor) than does have our T 8 Magnetophon apparatus.

The caption to this photo is only mentioning (brief translation): that Miss (Fräulein) Käthe St. volunteered for a technical profession, and she was trained for a job in a broadcast studio, operating Magnetophon systems.     This was standard gear in the RRG*, but unknown to Allied broadcast services; until they captured these standard apparatus first in Luxemburg and later in all major German broadcast centres.

* Reichsrundfunkgesellschaft


The three sound-controlling 'faders' W 24a are clearly visible (like three are visible next to Mss Käthe). I don't know whether this technique was known generally before the war  (Courtesy of Mr Friedrich Engel)

As to obtain it in PDF format please click at the above drawing

Not long thereafter, I received a response from Friedrich Engel, but more importantly a kind e-mail from Mr Gerd Kuper.

It is sometimes unbelievable how internet-contacts can speed up a progress.


Gerd Kuper first did send me wonderful photos which he had made some years ago The sound level Regler type W24a (maybe it actually carries a different type number, but I guess that they differ more in impedance concerned than mechanical appearance). 

As to get a better understanding of what these kind of sound-level-control devices electrically work, we have to digest its schematic first.



It actually is showing type W 24 of 1936

(with courtesy of Mr. Gerd Kuper and the AST group of the NDR Hamburg)

It is apparent, that when the selector is set at 0 Np signal level, that the in- and output lines are interconnected without any interruption. Hence no attenuation will occur.

By the way, some may wonder, that the above circuitry is asymmetric against ground. The regular interconnecting cabling is symmetric; reducing the pick up of environmental hum. The hum field induces a field in both symmetric wires and cannot be converted into a line-signal carrying hum. Normally, in an ac system, the current fluctuates in each wire alternatively in either direction. This is its typical appearance; though, when current flows in both wires into the same direction at instant, these cannot pass transformers, because there does not flow a current in the concerned transformer windings. Technically known as: common-mode rejection

My guess, therefore, and in the meantime confirmed by Gerd Kuper, is that on either side (in- and output) they must have implemented a transformer. Likely having a transformer ratio of 1 : 1, as their only purpose is getting rid off asymmetry.

Asymmetry is a curious matter. In classic telephone systems the telephone-line-wires are symmetric against environment. Even when you touch with a screwdriver at a single line-connection, most likely you will hear a humming sound through the telephone! Asymmetry also incorporate the line signal-phase in both wires is influencing this phenomenon.  



Viewing the Profilregler type W 24 from outside

(with courtesy of Mr. Gerd Kuper, and the AST group of the NDR Hamburg)


The German - as well as the Dutch PTT, were, in their postal services, technically counting, when signal levels being concerned, in  Neper, a logarithmic measure, which in contrast to dB is being based upon e or ln. (also known as: natural logarithm)



It is clear that 1 Np equals a signal strength difference (ratio) of 8.68 dB

(source: http://www.sengpielaudio.com/Rechner-neper.htm)

Np and dB are mostly concerned in a signal-strength-division ratio. Often derived from a voltage across a certain impedance. When we look back some drawings, at the RRG studio block diagram they use somewhere 1.55 V. This latter level might have been equal to 0 Np and 0 dB.

Very easy to know - is that in voltage-ratios in dB:

For example: 12 dB (6 dB + 6 dB) represents thus a ratio of 4 times (2 x 2); and 32 dB is 20 dB + 12 dB gives 10 x 4 = 40 times

Consequently: considering dB: a multiplication in value - is a summation or a subtraction of dB values.   


However, time has come to look at the pictures Gerd Kuper once took when he visited the AST group at NDR in Hamburg



The sound-level control W24a, which is just visible on the far left-hand side of this nice picture

(with courtesy of Mr. Gerd Kuper, and the AST group of the NDR Hamburg)



I guess that the brass screws are part of resistive coils (bobbins) as wires run towards the small contact strips

(with courtesy of Mr. Gerd Kuper, and the AST group of the NDR Hamburg)

These latter strips are usually mounted quite near to one another, as to secure that the moveable contact-strip touches several of these at once and by this means providing a smooth uninterrupted level regulation.

Doing so, also omitting the downside of carbon-potentiometers, where a governor is scratching over the quite sensitive carbon surface.



Viewing finally the rear side of the W24a Profilregler (Flachbandregler) In my perception one may also designate it being a: Flachbahnschieberegler Please notice: Aus means off and the signal passage being blocked entirely (interrupted)

(with courtesy of Mr. Gerd Kuper, and the AST group of the NDR Hamburg)


It is rather astonishing, that a simple hyperlink sent to someone can trigger a quantum leap in understanding what the sound-level-control left of the blond lady actually is!

Thanks to Mr Friedrich Engel and Mr Gerd Kuper and the internet, of course!



On 28 December 2014

I would like to add another interesting e-mail in which Dick Lucas, who was for a long time engaged in 'sound related' means, is explaining what he from his former profession remembers on faders (Flachbahnregler) . For those interested:

For those interested in a translation, please try it via Google translation.

Over faders. Die dingen op de foto waren tot lang standaard in de studio wereld. Kijk maar naar foto's van de Beatles bij hun eerste opnames!

Grappig is dat de beperkingen van de techniek, nl allemaal weerstandsverzwakker schakelingen weer terug kwam bij de eerste digitale faders. Met ook hetzelfde nadeel: je kon, wel erg zwak, maar toch hoorbaar schakelklikjes horen. Wat -voor zover ik weet- een flachbahn fader werd genoemd was wat anders, ik denk een uitvinding van Telefunken, De techniek van een laagje koolstof op pertinax was wel zonder stapjes maar te snel versleten bij doorlopend gebruik zoals in een radiostudio. Telefunken kwam met de oplossing: net boven de koolbaan werd een folie gespannen en de fader was een loper met een cilindervormig rollertje dat zorgde voor plaatselijk kontakt met de koolbaan. Er was zo geen slijtage van het koolstof meer! Zo'n soort fader zat in mijn eerste Telefunken M5 ingebouwd. Dat waren losse modules, zeg maar ter grootte van een pak koffie. Die zaten in bijna alle pro mengtafels van die tijd. Telefunken moet er vele duizenden van gemaakt hebben. Hilversum was er mee bestraat zo ongeveer.

Die zijn de norm gebleven, tot in Engeland Penny&Giles het conductive plastic uitvond, dat was nog beter.

Het soort fader van jouw foto heeft het ook zo lang volgehouden, tot in de jaren '70 in studio's voor filmgeluid, vanwege het feit dat ze makkelijk goed stereo gemaakt konden worden. Bijna alle andere technieken hadden zo'n variatie dat faden een links-rechts zwabberen veroorzaakte.



On 14/15 January 2015

The last weekend we obtained a 'light-spot-meter' accompanied by a special interface module type U 70a.

Nothing was known about it. Of course, Google and the internet was of great help to us, among it some other supporters too, like Frank Bell and Bernd Fischer



Siemens & Halske (S&H) Type Rel Bv 663K3010. Siemens Serial number F: Nr. 5/504579


The other device we got:


U70a Serial Nr. 115

Thanks to Frank Bell, we know from the Braun Buch (to quite some the holy grail). The Braun Buch is an ongoing series of internal publications meant for the Official German Broadcast Organisations and later also their TV services. These consist of data sheets, which can be add or taken out. Someone pointed recently, that he did find somewhere that IRT mentioned U70a in connection to the year 1975. IRT stands for Insitut für Rundfunktechnik. In the early years of German broadcasting representing the RRG (= Reichsrundfunkgesellschaft) and, shortly thereafter, also known as: RTI; just inversing the word sequence. Standardised modules were built equally by several companies, like Tonographie in Wupperthal, but also by other companies.

By the way, the latter company sometimes is also connected onto the name of Willesen. He  together with Mr Erbsloh were the founding fathers of the pre-war company GEMA, the first company in Germany producing radar before the start of the Second World War (Freya - Seetakt and other systems). Today, GEMA is name of the German: music branch copyright organisation. However, to what I remember from Harry von Kroge's book on the Gema Company, Mr Willesen died somewhere in the 1950s. Erbsloh passed away in the early 1990s, I did have some lively telephone conversations with him; who was by then a rather old man!


Aussteuerungsmesser U 70 (Braunbuch-Beschreibung)

It is quite evident, that this catalogue equals the previously shown photo. Apparently it was introduced in 1957 (courtesy of Frank Bell and the web)

What also is apparent: such units can drive various light-spot meter types; apparently the scale and current rates must have been standardised, likely according a DIN standard.

Please notice also the text header



Let us compare its construction. It is quite evident that the difference is: the U70 is using valve circuits, whilst our U70a is based on transistor techniques

The vertically mounted boards are pluggable, and can be removed for service purposes. 



This data sheet was somewhere found on the web

My first move was getting the 'light-spot sound-level-meter' operational. On the web the only information I could trace is the above shown data sheet, accompanied information telling us: that it concerns a rare device and that its weight is approx. 3.5 kg and, it has an integrated transformer for feeding the projection lamp. This is all I did find.



Viewing what is inside the meter housing

The transformer is in front of us. Just left of the painted number 2951 we notice the mounting of the moving-coil-instrument. Such kind of meter is not fit with a pointer but is using a quite small circular mirror instead.

My first attempt getting some sign of a vertical light-bar did not work out well. Then I remembered that in the circuit descriptions they mentions at what db scale level a certain current should be supplied. The current region is a bit over 1 mA.

I therefore took a small power supply, and put in series with one of the wires a resistor of 10 kΩ. The advantage is that when we supply 10 V dc that this will cause a current flow of 1 mA (we neglect the resistance of the moving coil). Hence, 26 V represents 2.6 mA. After changing polarity the meter did start to respond.



Light-slit simulating an audio-signal-level of just under 10 db



Just about - 20db



Simulating a sound level of + 2.5 db


Neglecting the Ri   of the moving coil

This is what I measured, and the values do not claim to be very accurate, only indicating the approximate value which can be expect for some of the meter readings. Please notice its inverse value response, thus when there flows no current the spot- or light-bar is outside the meter deflection range (screen). When increasing its value to about 1 mA the light-slit will appear from the far right-hand side towards the left-hand side of the frosted glass window.


It proved to be quite simple to reconstruct the true meter connections

Considering that this is all we know of the two module types, which's outside appearance are identical including the calibration provisions. Isn't logical to estimate, that U70a should replace the older module type U70, without changing other means?

Of course it does.


I therefore considered the outside connections provided on the right-hand side of the schematic, found on the web as well as with courtesy of Frank Bell


The + voltage line at connector pin 3b of U70a is via the + of light-spot-meter-coil connector pin 2a,  interconnected to pin 4a, and from there flowing via a resistor and both ECC 81 anodes toward ground. The latter valve circuit acting as a variable current source.  The 220 acV mains is connected onto the pins 6a + 6b.

As to be secure that no dangerous current is flowing, I used first inserted an external meter. It proved that all is responding well. Next, the Siemens meter being connected and all work as might be hoped for.

An essential feature of such electronic controlled system is, that the meter responds virtually instant for signal-levels moving upwards, but the light spot is delayed responding (≈ 200 ms) in its movement towards lower reading levels.



The projection lamp being pulled-out of the meter housing

It seemingly is a 10 W 12 V lamp (made by Philips).

I am not (yet) impressed by the brightness of the slit-light, but there is no means for inserting a more powerful lamp in the holder, as it cannot be inserted into its meter housing. I tried to rub the brass contacts, but this brought no visible improvement.

Considering this latter disadvantage - I would like to fit on an equivalent base-mounting (bajonet) a small 12 V halogen-lamp, which might act as a stronger light source, therefore providing a brighter light-bar. 



On 17/18/19 January 2015

I commenced some experiments with interconnecting the light-bar meter (Lichtzeiger-Instrument) onto the main signal input of V47b1

It soon became apparent, that for linking additional gear onto the purely symmetrical line system does cause sometimes trouble. We simply need 1 : 1 coupling transformers. These allow interconnections be it: symmetric onto symmetric system; or symmetric onto asymmetric system (v.v.); neglecting asymmetric to asymmetric.

My first attempt was linking the U70a onto the CD- player output. After having chosen the right ground connection hum virtually disappeared.

A second experiment, I did feed the asymmetric tone-generator onto the T8 recorder amplifier V47b1. Here we encounter already the disadvantage of working with testing gear lacking symmetric insulated outputs (free of ground).

We have to wait now for a couple of 1 : 1 transformers, as just have been described.

Yesterday (Friday), I encountered a nuisance, when recording a tone of 1000 Hz I got the strong impression that the tape was running irregular. The effect did appear as a kind of motor-boating but now caused by the left-hand tape motor (WML). The latter is providing a strain contrary to the tape movement. When I pushed the brake lightly against the flee-wheel that seemingly this 'Pendeln' vanished.



The schematic of a type T8 Magnetophon (Courtesy Friedrich Engel)

Please notice the red encircled two resistors.

Their function is, to provide, when we record- or operate the play-mode, that the left-hand tape is being kept a bit at a strain; not too much, of course. The motor gets virtually a higher inner resistance. This effect can be noticed, when the tape runs-out the tape-reel that it starts running into the opposite rotation direction; this rotation movement can be stopped manually.

It proved, however, that in our case the power resistors being positioned and wired a bit differently, than is shown in the above schematic. Hampering is, that when an ohm meter being connected onto the left-hand tape motor connection U, that we encounter first a 90 µH coil in series. Though, where can we find this coil, as seemingly there exist some of those; all wiring is coloured grey?

After quite long trials I might have found the two resistors. Then something curious occurred. First I measured a certain resistance (329 Ω), though, at some point it apparently did fail to have a resistance. After some careful considerations, I found that its very tin wound resistor-end has been broken-off. For convenience, I disconnected it totally and replaced it by 3 x 1 kΩ resistors (as to manage the power rate). Leaving open a final repair for a later date.

Now the tape test was repeated. When we record at the T8 Magnetophon, as is the case with most commercial recorders, the recording can be monitored life. Thus, what you record, after some milli-seconds is being reproduced. A guess, tape speed is 38.1 cm/s. Let us estimate that the recording- and the play-head is having a space of, say, 4 cm the time delay will be then: 4 : 38 roughly 0.1 second.  Please notice in section 13k that the actual space is approx. 7 cm and that, consequently the delay is about 0.18 seconds.

However, there are clear signs that after skipping the likely defect power resistor of 300 Ω, our Magnetophon might run better than it did yesterday.

But how checking that this is true?



This is the way I would like to check the inconstancy of the Magnetophon's tape speed

By the way, the transformers are of the type I just previously have discussed; having a transformer ratio of 1 : 1. These operate at 600 Ω systems and partly at 200 Ω systems. As long as both sides use a certain impedance this will be transformed onto the opposite side; as it is theoretically in some way determined by the transformer windings ratio. I do not care much whether we deal with 200 Ω or with 600 Ω systems.


Whatever the phase difference will occur, this only will influence the rotation of the painted ellipsoid or in a particular case being a circle (90° phase difference between the X and Y channel and both amplitudes are equally matched onto the screen reading, or, when there is a phase difference of 45° creating a straight line at that angle) Mathematically expressed: drawing a first order Lissajous . But we may expect, that minor frequency deviations will occur between the tone-generator signal and the one returning via the play-head. I dare about phase shift caused within the transformers, as for a certain frequency their nature will be constant, as long as the transformers not being overloaded. But that will not occur in our experimental setup. The tone generator does not provide signals much over about 0 db which is equal to 1.55 Veff.   

The main problem might arise, judging (interpreting) what will appear on the scope screen, as changes might move quite fast; not only in phase shifts but also in amplitudes.

I guess, therefore, that an analogue moving-coil-meter instrument is to be favoured, as when the circuitry allows, we can read an average value.

We have to wait until a set of transformers do arrive.

Another major concern, is, to look after the state of grease of the concerned electrical motor bearings.

I estimate, that the 'capstan' driving motor does have the most influence on the actual tape-speed fluctuations.


The brown coloured capstan-motor, on the left-hand side, is the one just underneath the tape-heads needs our attention first.

According a chapter of one of the 'Braun Buch' subjects, is, that when the capstan-motor (Ton-Motor) does lack sufficient greasing, that the reproduced sound will be 'rough'. For it we have to separate it from the main tape deck. Which should not being too difficult, I guess.



On 22 January 2015

After a period of reflection in a hospital, I would like to start thinking on a way of designing a small sound-control-panel.


What should it accomplish?

It should allow controlling sound levels, by means of the previously described light-spot-meter (Lichtzeiger-Instrument) and faders.

It should (also) allow mixing at least two input channels, such as from a microphone and another signal source. Using in someway or another historical sound faders (Flachbahnregler). And also controlling the output sound level of the amplifier unit V47b1.

How should this be optimally fit onto the existing T8 deck?



 Is there a way exploiting what we have at hand?


Yes, it does.



It might give a confusing impression, but it certainly it is not


Both clamps at the right-hand side should be fit or attached onto the two vertical tubes on either side of the main mounting frame. Thus, allowing to use it on the right- or left-hand side of the T8 main deck. For it the wooden plateau is being kept removable.  Hence, when one chose to operate it at the other side of the T8 mounting frame, then one takes the wooden plateau out and keeping in the same design manner (direction) just at the other side. As to prevent that the external interconnections will be mounted at the impropriate place, my suggestion is to mount it at the outside left just in the middle (outside). Thus, when the frame will be turned 180 degrees as to hang or mount it at another side, the interconnections stay outside in the middle. The electrical interconnections can de arranged by means of a Tuchel-Siemens connector, quite commonly used up to the late 1950, maybe even 1960s.

The clamps can be rather simple made and though, still being highly rigid and stable.

Please bear also in mind the details shown in the drawing above.

Imagine, that roughly the stainless steel tube frame is made of 20 mm tubes. Take a block of metal having a, say, 35 mm depth and a broadness of 60 mm and a height of 60 mm too. It is on decision of the person to chose material dimensions at their disposal.

Drill now plane of 35 x 60 mm a hole of just 20 mm over its full length of 60 mm. Divide the drilled hole plane by sawing it through over its full length. We get now two halves when we drill horizontally two, maybe four, heavy screw holes (M 5 or M 6), we can clamp this arrangement easily by screws around the vertical frame tubes.

I guess, that this will establish a fairly solid mounting. The drawing will provide the way the control panel might be attached onto the two clamps.

The wooden plateau is kept in a frame, and can easily be lifted out of this frame. When necessary, a few screws can fix it temporarily.

The very day that returned from hospital, still a bit weak, I received a package from Bernd Fischer incorporating 2 Profilregler type 24a.



This photo I did copy from the already existing photo on this website (ours is more neatly)


This type is from wartime days, maybe shortly thereafter too.


For the time being, I would like to adopt a more modern type equally type W 66, though, this artefact seemingly originate from DDR production.



What I would like to obtain are two additional types W44, like this one (Courtesy Bernd Fischer)

However, I am currently indeed very happy with these two faders type W24a!


Isn't it strange, that about a month ago, I did know nothing on these kinds of controls, and a month later we possess most kindly two samples. We did not know a month a go much on studio light spot meters (Lichtzeiger-Instrumente), now we have a working system.

I only would like, herewith, to express, that having friends and connections, as well as a great deal of exceptional luck is indeed most essential in human life.




On 27 January 2015


In the meantime, we got 4 audio transformer having a conversion ratio of 1 : 1. When I measured their individual winding resistance I wondered the one is, say, about 91 ohms, though the second winding is having resistance of 151 ohms. Hè 1 : 1 ? I instantly contacted Marc, the kind donator but the told me, that these indeed are having a ratio of 1 : 1, but, one winding section is would just around the core but the second winding wound around the first winding section and, therefore, having a larger diameter.

However, our first aim is, to separate galvanic both core sections, as to allow asymmetric onto symmetric.


This transformer being wired parallel onto the V41 play amplifier. Its asymmetric output is linked onto a probe interconnected onto oscilloscope channel B. Channel A of the scope is directly fed from the tone-generator output, but this very output is also fed onto recorder input of V47b1




Viewing the the way the Tone-generator output is also fed onto the recorder amplifier input


Let us repeat the formerly used drawing


Please notice this block diagram and the previous two transformer, which are just the two being in use for this experiment.


The only way showing the results of this kind of experiments is combining pictures including sound.





Youtube films


Film 161    Showing the actual way the external phase difference setup is operating. Below we see the insulating transformer having a ratio of 1 : 1. The what the Germans call Schwebung in the sound pitch. The tape being recorded and replayed about 0.1 seconds thereafter. The A-scope channel being fed from the tone-generator set at about 1000 Hz, this signal being fed, of course, also onto the input channel of recorder module V47b1 . Scope-Channel-B being fed from the play-amplifier output, via an 'insulating' transformer. 


Film 162    We are first viewing what kinds of first order Lissajous figures is being pained onto the video screen. Checking the inconstancy of the tape mechanism. I have to admit, that this field is new to me, I cannot judge phenomenon from long time experience, as, in some respect, was possible during the Nachtfee project. I always have been "a sound consumer, and not a sound maker".


Film 163    Viewing now the running capstan motor. I am incorrectly suggesting that the tape is running at 76 cm/s speed; Because this would implicate 0.76 m/s. Considering the visible tape speed this is not the actual fact.


Film 164    Viewing the newly obtained W24a. These are beautiful devices. I was told recently by a sound technician that there exist photos where during the Beattles first record recordings they operated these kinds of faders.    


Although, I took the vision that the wobbling sound pitch is being caused by a mechanical reason.

My technical background, also encounters some doubts. Couldn't it having an electrical reason as well?

I can imagine, that the most likely section will be in recording amplifier V47b1.

Let us consider its schematic again.


The most likely stage where it might going wrong, is about valve 33 (EDD11)

I will not say that this actually is the case, but when something goes wrong that this stage might having a strange effect. Don't forget, that the left-hand section of V 33 is constituting an oscillator at 50 kHz which is used for the HF bias of the recorder head. Whilst, the two halves of V 33 together constituting a doubler stage. Its output frequency being 100 kHz.

In the second half of this survey I discovered, that my perception was totally incorrect! I did not read the left-hand cadre correctly!

It is just the other way around. The left-hand section of valve 33 acts as a 80 kHz oscillator and the push-pull stage being locked upon half that frequency (40 kHz). I measured respectively 82 kHz and 41 kHz. Although, position number '22' is providing a tuneable dust core, I could not tune it lower without damaging it. In my perception these frequency deviations do not matter much.



It might making sense, to check the behaviour of this stage first. Inside this module, both head-bias values being noticed.

Maybe it does make sense to implement the filter W 56, which is especially meant for being put in series to the V47b1 output connector 56. The only matter that is implementing it instantly, is that it contains. 


It is quite evident that the two 10,000 pF capacitors will be loaded parallel onto the 220 V mains

We keep it really at 220 V, as the entire system being operated via a variable mains transformer (keeping insulated from the mains lines).




In the afternoon of the 27th  I went back to the Klooster premises.


My intension to implement filter W56 in series to the output connector proved to be invalid. Because, the wiring is not complementary. It apparently should have belonged to another module. The previous owner did implement it in series to the output of V41. It should, however, be possible to adapt it onto the V41 output connector. Whether it will make much difference stays open.

I also investigated the recorder module V47b1 (V47/2)

I implemented first an additional smoothing electrolytic capacitor, which proved to have a too low operational voltage of 450 V. I therefore did put an equal one in series (making this arrangement for up to 900 V). 20 µF / 2 making 10 µF, which should be still sufficient, as the schematic gives for it a 4 µF capacitance. Then I heard some flashes, but could not yet trace their origin; it proved finally to be the connection onto the old electrolytic capacitor that I did have bridged electrically (using for practical reasons different mounting points). Whether the wire attached onto it just made contact and/or my mechanical intervention caused this, cannot be determined.

It still is apparent, that everywhere in the V47/2 (V47b1) at almost every contact-point, we encounter more or less traces of the oscillator frequencies. Some having quite high levels.

But, our first aim is, to deduce the origin of the phase fluctuations we hear and can measure. Therefore, YouTube is an excellent facility again.


Firstly, returning to the power section of the recorder module.

The so-called 'Leerlauf' voltage seems to be 465 V, this is the case, when the recorder is not being operated in the recorder mode; otherwise the loaded voltage is about 325 V. But, we have all the time to count for the worst case, that we do not record, which is happening most of the time (quite obvious).

I also found out (blaming myself) because it is clearly written on the left-hand cadre of the schematic to V47b1

Erasing frequency is 40 kHz and the HF bias frequency is 80 kHz. In our case it respectively is: 41 kHz and 82 kHz. Which I consider being workable.


After some experiments, I repeated YouTube recordings, as I would like to show that the amplitude fluctuations heard do not change its value, but it is the tape speed that deviate. Using scope channel B, where in the Lissajous mode, the play-mode output being fed onto the same channel B. The only difference being, that when I press the channel selectors A + B at the same time, the scope is then painting a Lissajous. We see the T8 system loop, where the tone-generator output and what is being reproduced by the play-section      



Film 165    We see the T8 system loop, where the tone-generator out and what being reproduced by the play-section of the Magnetophon do draw a first order Lissajous.

Film 166    The scope channel B being selected solely and is showing the output signal provided by the recorder. It is quite evident, that the fluctuations we still hear are not originating from an amplitude nature. The painted line being more heavy than it actually is, due to the automatic picture control of the camera recorder.

Film 167    Using all the time the same pre-conditions, but now pushing channel switches A + B together we are viewing the oscilloscope writing a first order Lissajous again.

All three films together proving to us, that the audible fluctuations are having a signal phase error, that only can be caused by tape speed deviations. But, where does these fluctuations originate from? What might cause the about 2 Hz (my estimation) phase fluctuations? And, in most cases over the tape-running-time having a more or less error constancy.

The capstan-motor (Tonmotor)?

Nothing in the system does have a, say,  2 Hz nature.

Thinking it all over again, my first move should be dedicated to the rewind motor (WML), as this one shows apparently some axis wobbling.


Those having an advise or answer, please contact us at:

Please, type-in what you read


On 1 February 2015


Waiting for some parts from Germany, I used the opportunity to measure some deviations by means of a 'micro-meter'. The main problem being - how to fix the device onto the Magnetophon frame?



Let us estimate that the pointer is indicating -8

According the scale definition each division represents 0.002 mm, which equals 2 µm.


Now we rotate the flee-wheel of the capstan motor a bit


This flee-wheel is mounted just at the lowest side of the capstan motor


Let us now select one of the deviation maxima


The pointer indicates now: 19 which, according the resolution indicated, means a deviation of: (8 + 19) 2 µm  = 54 µm. By this means it cannot yet be judged whether these deviations are having a local (limited) mature

At this point I cannot yet judge whether this is causing our problems. I don't think so.


In an e-mail directed onto Gerd Kuper, I also calculated briefly other dimensions.

Let us follow briefly the outline of my e-mail:

The diameter of the capstan (Tonwelle) is 9.81 mm consequently, its perimeter is 30.81 mm. We know that the tape-speed is 38.1 cm/s (380.1 mm/s) this implies that the capstan rotates 12,36 times per second. Per minute we get: 742 rotations per minute. This is quite near to what newly is discovered. The previous owner mentioned in his personal notes, that the motor is having 8 poles. On 20 February, Jaap Keijzer, correctly noticed, that a single (motor)pole consists of two coils (50 x 60 = 3000 rpm). 4 coils is thus a motor having 2 poles and, consequently, is having 1500 rpm. In our case we should count with 3000 : 4 = 750 rpm. Whether the actual capstan-loading is causing a 'rotation slip' of 8 rpm? I was already told, that some kind of motor-response might be observed. Whether this could be: (8 : 750) x 100 = 1.06 percent? In the meantime at 13k we have removed the bearings from the capstan-motor (Tonmotor); what we actually will encounter after the motor-revision is not yet known. My calculation was not relying upon capstan rotation-speed measurements, but was based theoretically on the diameter of the capstan shaft versus the expected tape-speed of 38.1 cm/s. On the other hand, from the T8 electrical schematic we know that a series-resistor might be involved, and that this might also influence it over-all behaviour.   


I, therefore, estimate, that nowhere in these values we find a dimension that correlates with 2 Hz deviations. Please notice, that my 2 Hz is a rough estimation, only derived from what the YouTube films show to us.

A second size was measured, the diameter of the rubber-roll which presses the tape against the capstan.

I measured: 45.51 mm multiplied with π (pi) this gives a perimeter of 14.29 cm.

Tape-speed being 38.1 cm/s. Calculating it all in mm we get 380 / 142.9 = 2.65 rounds per second (the 0.1 is skipped as its influence is neglectible).

Which implies, that this figure has something in common to the tape-speed deviations. At least we should try to replace it, albeit, temporarily. It does not yet tell us whether we have found the origin of our technical problems.


A last attempt was made to measure the deviations of the WML perimeter. Here I know, that a mounting screw of the WML motor is missing.



 I measure at some sectors quite some deviation, whether this is bothering I cannot say yet. But, the 'tape-support' shows also some irregularities in the horizontal plain (wobbling slightly).



A problem was, to mound the micrometer onto the Magnetophon frame. For it I used a so-called big 'Kocher' (Kogger??). The small one used in hospitals to block or cut-off blood or other liquid flow; this one bought at a big rally, is most convenient. 

By the way, a wonderful versatile instrument!



On 12 February 2015

I continued with experiments after we received, very kind, yesterday a package containing two 'fader controls' (Flachbahnreglers), one being S37 and the second one type W44.

Among it were also two rubber-rollers, which do press the running tape onto the capstan (Tonwelle).

Did it change the behaviour of our T8 Magnetophon machine? No. it did not much.

 Next I focussed my attention onto the capstan motor. With a screw driver pressed at an ear and on the other side touching the two motor unit ends, a rumble was apparent with some repetition like the discontinuity painted at the Lissajous figures.


This motor should now be accessed!



The motor being detached from the tape-deck



Viewing the same motor from a different perspective



The lower plate being removed which apparently acts as a centre bearing

The wheel inside is also acting as a flee-wheel



The only identification found: M1C1/1s

At the top section of this capstan motor we find M1C1/4s

The code following the forward-slash is likely telling us what motor part we deal with.

 Whether M1C1 is a motor code we don't know.


I opened the top bearing-cover and recognised a ball-bearing. What is curious, is the fact that the motor-shaft can be moved freely through the inner-ball-bearing space.


I discussed my concern with Jaap Keijzer, and he kindly indicated that he would like to investigate the motor troubles and curing it as well.

He is the same person who helped us recently with modifying the TX tuning of the Urechse transmitter of our Würzburg system. He also very kindly, solved the problem with our Nachtfee project, where the front-side-mounting of the dual-trace CRT is lacking. No one is noticing that what he beautifully made is his conception and not genuine.


Our first move is having patience and trusting fully Jaap Keijzer's workmanship.



On 17 February 2015


Time is spent to investigate some of the newly obtained sound faders (Profil-Regler)


My attention was first focused upon type W44



Viewing the front-panel of fader type W 44 serial number 760

This model was already operating in db steps, whilst the forerunner type W24a was using Neper (Np). The latter being based on log e (natural logarithm)  whereas db scale is relying on 'log 10'. However, it is quite simple to operate both, as 1 Np is equal to 8.65 db.



Viewing the rear-side of the fader type W44

My guess, the designation 44 might point at the year of its conception, as wartime matters often were re-designed as to obtain a cheaper version of a device. Accepting that its performance is a bit less sophisticated; but on the other hand demanding less materials and easier to manufacture.



Viewing sound fader type W44 Fader set in its 'off position'

(Profil-Regler also known as Flachbahnregler)

Please notice the many small bobbins (coils wound with resistance-wire) mounted just behind the front-panel and the switching contacts.


A curious discovery is type S37



Viewing the front section of Flachbanregler type S 37

Apparently, its division scale is running from 0 up to 11. Unlike the regular faders there does not exist a provision for switching off the signal-path.



Viewing the rear inter-connections of sound fader type S37



I never expected that this fader does not possess resistors inside its housing!

All connections are being interconnected onto the two Siemens-Tuchel connectors (2 x 6 contacts each). Hence, at least an additional unit or box must have been engaged.

If someone does have knowledge of this kind of device, please come forward and contact us at:

Please, type-in what you read



The most beautiful Profil-Regler (sound fader) type we currently possess, is type W24a. Already discussed previously on this webpage.



Viewing inventory label of: Rundfunk- und Fernsehtechnisches Zentralamt

Inv.-Nr. 204 067 (204067)

I guess, that this points on to a German wartime institution. In post war days, it were first the Allies who established execution powers. Under their role Britain handed over the actual daily affairs to the NWDR (later two institutions WDR and NDR saw daylight in the 1950s). The Americans established RIAS and like in the case of the NWDR they also had supervision over the: Bayrische Rundfunk. Considering the foregoing explanation, TV (Fernshen) was coming into being in the early 1950s. Though, TV was at a low level existent. 

However, yesterday evening I searched extra on Google, entering the full text of this label, and the information that came out of it, is, unexpectedly, that this very organisation was on behalf of the DDR broadcasting- and TV organisation.

This institution got the abbreviation: RFZ


Related links provide all sorts of nostalgic DDR information.

Quite some can be found on the web.



Viewing the main division section of this sound fader type W24a



Maybe not well visible, I encountered a strange behaviour of this W24a sample

The wire not covered within a green insulation-tube proved, after some tests, being disconnected. Careful viewing through an illuminated magnifier-glass showed finally where it apparently once was interconnect onto.



What also became apparent, is that although this schematic covers fader type W24 and we are dealing with type W24a, that their schematics do differ quite much

Maybe this schematic is only a brief version of a rather complex device.

Making comparison between the above schematic and our W42a sample is rather confusing and difficult. 



The two switching sections in the upper side of this photo being set in the disconnecting position (off mode)

It proved, however, that their silver-contacts were not electrically conducting.

Rubbing with some double folded paper sheets dripped in a 'contact oil' solved finally this failure. I dislike putting heavy friction (force) on silvered contacts!




The electrical contacts are being pushed in the 'off mode' position, like it is in the previous photo

In my perception, we are viewing now at the T-leg circuit side (cross section) of the sound fader. As long as both circuits does have enough equality (W24 versus W24a).



Shown is the interconnection lay-out

We  luckily possess also a second W24a within its genuine cover, and at its bottom we can read the designations as is reproduced above. Whether these designations point to DDR (RFZ) practice, or even wartime technology, is not known to us. It might originate from a different source, as it does not carry any kind of RFZ labelling.



On 20/23 February 2015


Last Friday, 20 February, Jaap Keijzer came along for dismantling the capstan motor (Tonmotor) of our T8 Magnetophon.

After having listened first to the rumbling noise originating from of the capstan-motor-bearings, he started with dismantling it. His first attention was focussed upon the top bearing. Evidence was found that the current roll-bearing is not what originally might have been employed; as he found provisions meant for lubrication by means of quite thin oil.



Jaap is investigating the quite small pivot-disk (taats-lager), which's centre is clearly showing signs of what he calls "jimmying". The latter is a tendency not to load at a particular centre region, but the single bearing-ball is drawing irregular movements (not entirely in a centre)

Just above the bottom-mounting-plate, we see the lower housing of the capstan motor (Tonmotor).



Motor being dismantled entirely

At most 'due north' in this photo we see the top bearing-cap of the T8 capstan motor.


I have selected three video films for YouTube, which is showing Jaap Keijzer's dedication and patience with which he is dismantling our capstan motor.

Film 168:    Jaap Keijzer is now warming-up the lower bearing housing of the T8 capstan motor. He is doing every new move with a great deal of patience. Apparently, at the end there was not much force necessary to separate the section from the centre motor unit.

Film 169:    Two quite long shafts are still hampering separation of the centre- and the bottom-motor-section.

Film 170:    Jaap Keijzer has to un-screw the long bold (trekstang), as it apparently is not releasing by itself.       

An additional discovery, was, that although, the kind Belgium donator of our T8 Magnetophon, mentioned in his personal notes, that the capstan-motor is having 8 poles, he might not have been aware that a single-pole-motor consists of two field-coils opposite to one another. We count 8 coils but each set consists of two coils; hence, our capstan motor is thus provided with 4 poles. A simple calculation gives for a single pole (two opposite field coils), at 50 Hz, a rotation of 50 x 60 = 3000 rpm. Hence, two poles providing 3000 : 2 is 1500 rpm; consequently, 4 poles, like in our case, is providing a rotation speed of 750 rounds per minutes. This is in accordance to my observation described on this webpage previously.

This is what I have substituted today onto the previous web-section (some details being here slightly adapted):  This is quite near (calculating the capstan-diameter versus the estimated tape speed of 38.1 cm/s, resulting in an actual rotation speed of 742 rpm), to what newly is discovered.    The previous owner mentioned in his personal notes, that the motor is having 8 poles. On 20 February, Jaap Keijzer, correctly noticed, that a single (motor) pole consists of two coils (50 x 60 = 3000 rpm). 4 coils is thus providing 2 poles and, consequently, is having 1500 rpm. In our case we should count with 3000 : 4 = 750 rpm. Whether the actual capstan-loading is causing a 'rotation slip' of 8 rpm? I was already told, that some kind of motor-response ('slip') might be encountered. Whether this could be: (8 : 750) x 100 = 1.06 percent? In the meantime we have removed the bearings from the capstan-motor (Tonmotor); what we actually will encounter after the motor-revision is not yet known. My brief calculation was not relying upon capstan rotation-motion measurements, but was based theoretically on the diameter of the capstan-shaft-diameter versus the expected tape-speed of 38.1 cm/s. On the other hand, from the T8 electrical schematic we know, that a series-resistor might be involved, and that this might also influence its over-all behaviour; hence, tape-speed.


Another detail: I measured the space between the centres of the record- and play-head which proved to be approx. 7 cm. My calculation in respect to the delay between a signal being recorded and being played (reproduced), has to be corrected: 7 : 38.1 cm/s = 0.18 seconds. My first guess was that the head-centres were separated 4 cm, providing in this case (incorrectly) a signal delay of, say, approx. 0.1 second.




  On 4 / 5 + 11 March 2015


I turned my attention today on the sound-fader (Flachbahnregler) type W 44 


Viewing the way this fader is being constructed

Contact strips are found on both side of this device. This may be in accordance to my guess that a so-called L-type sound-fader technique is involved.    

But, this is only a guess, which cannot be proved yet; as long as we do not have its genuine schematic or reference.     As quite usually, searching on the web, nothing but our website is found as, seemingly, the only source of information (skipping type-numbers of consumer products). This is touching a fundamental shortcoming of search-engines. Old knowledge is randomly found, because someone did put it once on the web. Old publications or data-sheets needs 'someone' who is doing the effort of reproduction; and quite much is still pending. Though, one needs to have access to sources of information, and this is just the bottleneck.     Another nuisance of search-engines, is the hampering effect of keywords related to famous persons or brand-names. In such case - skip searching, as it is not worth the efforts. I once encountered this during a search on someone (living in America about 1902-1903) in respect to Hülsmeyer (2004); who's name was: Kettler. Seemingly, the search-engine-provider is often not selective enough and sometimes reproducing references which critically viewed could have been reduced to, say, 5 or 6, instead of hundreds (as it deals with the same publication of a paper or that like).     On the other hand, we should also be grateful for the existence of search engines, as this is most often the only means of tracing facts and/or references.     The Germans call this contradiction a 'Zwickmühle'.      

However, when you look carefully, the bank of (tiny) resistor-bobbins being visible between the front panel (towards us) and the ruling mechanism. These resistors are wire-wound, though, as to prevent that these becoming frequency dependant, they are wound in a bifilar manner. This implies, that the incoming (bobbin) wire is kept parallel to the outgoing wire (bent half-way); their mutual inductances being cancelled (nullified).

Dick Zijlmans very kindly lend me some books on Magnetophon technology and Studio techniques generally.

In Johannes Weber's book on: Tonstudiotechnik, I found a chapter on sound-faders (Flachbahnregler).


However, how should our 'fader type W44' being interconnected?


After some experiments I concluded, that the most likely connections should be connector pins 1 - 2  and pins 7 - 8.


After some direct linking, I decided that implementing coupling transformers, which proved to be a very good means

The dotted line is only meant to indicate that both being regulated by means of a single control and that both being varied by it. This does not imply that both interact linear to one another. Their interaction is to get a logarithmic sound-fader response.

From Weber's book we know, that there do exist:



B - Regler


Considering the most simple means, I guess, that the L-type-fader technique is involved inside the W44 frame; as is shown in the foregoing principle drawing. The L-type is the most simple one, as it does have only two variables to count with, whilst most other have at least 3 some even more variables to count with. It is quite evident, that its input impedance at terminals a1 - b1 can be kept more or less constant, but the parallel resistor-chain can get quite low impedance values. I interchanged the connections, but there was hardly found a measurable difference. I believe, that the input of the following amplifier stage is less sensible to low loading than would be the output stage of the foregoing amplifier.  



This photo very kindly was made by Bernd Fischer

I received it after I had already found out that pins 1 - 2 and 7 - 8 are involved.




Viewing the experimental set up



I cannot measure below 5 mV, and maybe some external signals being induced


I came, however, to the conclusion, that the maximum attenuation I can measure is, say maximally - 50 db. Whether we actually can get 60 db deep?

The in- and output parameters have been changed between 50 ohm up to 600 ohms on either side separately. The general 'overall' results did not vary much.


On 4 / 5 March 2015

Dick Zijlsmans very kindly gave us superior class audio transformers formerly owned by the Dutch Broadcasting Organisation (NOB).

These transformers are really superior in their specifications. What to say flat from < 30 Hz up to 20 kHz, maybe even beyond?


I did use these transformers experimentally in conjunction with the vintage sound-fader type W 44 



The scope-scale set at 5 V/cm* , for this occasion being 0 db

* 5pp V = 1,76 Veff, effectively it is less, as the peak-voltage does not reach the vertical scale lines fully. 



The scope-scale set at 5 mV/cm, no trace of spurious


The old transformers (just visible north of the W 44 fader) performed quite inferior compared to former NOB devices.

I never expected that so much influence may be encountered.


On 11 March 2015

I had the opportunity measuring the real frequency response including the W 44 type fader.

I measured through - the high quality audio (balance) transformer - W 44 input - the same type transformer coupling the W 44 output onto the oscilloscope bridged with a 270 ohm resistor, having parallel onto it an ac voltmeter fit with a dB scale.

The following results were obtained:

0 dB = 10 Hz - 100 Hz - 1 kHz - 10 kHz

- 0.5 dB = 15 kHz

- 0.8 dB = 20 kHz

- 1 dB = 22 kHz

We may, in my perception, consider that this being the best possible frequency response.

Repeating the same range for a sound level of - 20 dB and the response is flat between 15 Hz up to 15 kHz. It has to be noticed, that the dB scale is not allowing high accuracy reading.    



On 30 March 2015

Yesterday I met Jaap Keijzer during a NVHR rally in Driebergen, he did take the capstan motor to this workshop. His first words were: I have two messages, a good one and and a sad one. He finished with: I will do my very best getting it, nevertheless, right.

He then told me that there apparently exists a 'faulty lining' between both the top- and the lower shaft-ends. Very disappointing. But, he hopes that he can fix the problems. He also told me, that the previous owner must have encountered problems and he therefore has removed the genuine top-bearing (maybe his Belgium technician did eventually cause the nuisance). Jaap would like to re-fit (modifying) both the top and lower bearings, replacing them by 'sintered-bronze-plain-bearings'*. Which do not need additional lubrication concern. 

* thanking Jan ten Have

The only option we currently have, is to be patience.

When it finally comes out that the capstan motor cannot be brought into a sound operational state again, we should consider adapting a so-called 'Papst Ausssenläufer' motor. Previously one of the best options possible.

Our problem, I have no experience with Papst motors at all; where we have to deal with 750 rpm - as well as an appropriate shaft diameter. We are bound to our pre-condition of a tape speed of 38.1 cm/s.


To be continued in due course.  


By Arthur O. Bauer

Did you already consider the previous Magnetophon starting page?