Nachtfee MLK-lab-survey


Status: 13 December 2011


After the quartz and signal processing modules have been separated from the Nachtfee's main chassis, the devices (according survey page 2) has been brought to our MLK premises, where we have a well equipped laboratory facility.


Our first aim is getting the 10 oscillator channels in working order

This proved to be far more difficult than expected.

As to get one of the oscillators modules working, I took module number five ('5') and disconnected two components in such a way that we unsoldered them at a single end; as to guarantee that the genuine condition could be restored at any time. Of course, bridging them with an appropriate substitution. In one respect choosing module number 5 was a missing chance as this module tended to be not very active. A fact, as we will notice further down, the Germans might have faced themselves during their conception of this particular Nachtfee apparatus (maybe prototype serial number 5 or V).

Straight from the beginning of our survey I keep a survey-report which is reflecting all my thoughts and observations. This report though, is written mainly in Dutch language, but the hand written information is, nevertheless, a useful document.


The module in front is having a double size, which function is still not well understood. My main attention concentrated firstly upon to the channel oscillator module number 5 (15,000 Hz)

Its wiring was figured out briefly and is done more seriously later. When it was tested laying on the lab. bench, it showed a rather strange kind of signal; which proved to be of spurious origin. This phenomenon stopped when the module was mounted as is shown in the last photo. However, we still couldn't get now any sort of oscillations at all!

Changing the position of the module at another channel place (thus using a different quartz vibrator) didn't brought any result. Are the longitudinal quartz vibrators all defect?

For this occasion I fed at one side a signal generator which was also linked with a frequency counter and at the other quartz-end (X-axis-vibrator) I connected a scope. It proved that all channels vibrated in series resonance, hence, constituting a rather low resistance (in this case representing a pure ohmic value).

However, why couldn't we get any sign of oscillation? Like so often with long enduring trials, at a certain point I saw only for flash that the time-base-line tended to widen up in the 'Y' direction. I still didn't grasped the clue yet. Finally it proved that this module demanded an additional capacitive loading at its anode; like is also the case with some Pierce type oscillators although, this circuit in not Pierce type alike. The optimal capacity is 150 pF. From now on the circuit is working, however, it takes some time before the oscillation is starting-up. This may be owing to the fact that the Q-factor of these kinds of quartz-vibrators is very high, and may reach >10,000. Modifying the two other modules (number 4 and 7, only bridging a 0.1 F and 25 F capacitor) brought to the lab., they responded instantly, without additional modification. Nevertheless, also demanding some time before the oscillation is starting.


Let us go back to the situation where oscillator module 5 didn't respond at all. What the meaning of 'Gei' is I do not know, maybe the name of the guy who prepared it. Please notice that the transformer core screws are locked by means of soldering; presumably as to keep its effective -factor constant

What was eventually causing detuning? Inside I found this transformer. Please notice my survey-investigation-report at page 7. Bearing this in mind, we might have to consider that at least some other Al. cans in the Nachtfee main chassis may contain similar kinds of transformers.


What is happening at their common-output-line (all channels output is connected in parallel) when three channels is being switched on?

Oscillations at 14940 Hz, 15000 Hz and 15120 Hz. Our scope time-base might find it difficult triggering upon it correctly. My guess is, that a single channel should actually be operated. Although, as everything is different from what it looks like, this guess might also be wrong.


The three modules being plugged-in and according the signal lamps channel 4 and 7 is activated. As we have noticed before that starting up oscillations take some time, they are kept constantly running and switching on- or-off signal is only done by means of disconnecting the output-line and also switching the indicator lamp accordingly


Now the combined (double sized) module is being attached and the signal at g1 of R 16 is monitored


At the anode of R 16 we notice that two sorts of signal is available. A lower frequency sinusoidal component and the component originating from the quartz controlled channels, although, not entirely

The behaviour of R 18 is so strange that a circuit fault might occur, which is demanding a further circuit determination. Hampered, though, by the fact that all wires are extremely thin and most having the same colour. As the wiring has been done in a proper PTT like stile, following their actual route is extremely difficult and confusing.



On 8 December 2011

We continued our MLK lab. investigation on the quartz-linked Nachtfee module.


The investigation steps are briefly, sometimes a bit more going into details, laid down in our Nachtfee-Survey-Report (latest version)


The double sized module under consideration

Photo taken before the failing tar-roll-capacitors and an electrolytic C had been bridged.


My first attention went to (valve) R 17. It apparently gets some signal from the quartz channel section, but its anode circuit responded a bit strange, as first there existed hardly any sign of influence whether the quartz stage was switched on or off. It proved to be very useful to solder all accessible (visible) solder-contacts, as apparently wire oxidation can causing a distorted signal transfer. I also discovered that R 18 in conjunction with transformer marked with NF5 is behaving as an oscillator, but resonating at a far lower frequency than originating from the quartz (channel) section. It proved also, that a 25 kΩ (notice please previous drawing remark) was not well soldered or what ever means inserted in the circuit I couldn't measure any resistance, after unsoldering all seems to be in order. But, mounted as necessary again an open circuit remained. I decided to bridge it by means of a 22 kΩ  resistor, and now all seems to function well.


Signal measured at the anode of R 17

It is clear that two signal components are existing. One originating from a channel quartz, and one originating from the oscillating stage around R 18.


Three signal components are existing, when two quartz channels being operated


Second photo from the same phenomenon


Measured at the same point, but now without a quartz channel being selected


Measuring at a different circuit point. Clear is that again we notice the two signal components (A and B)


Looking now at the output of the double sized module


The output locked upon a quartz-channel-signal


An un-locked output (no quartz signal provided onto this module)

The difference in locked and unlocked is about 250 s


We tried to measure accurately the signal output frequency versus locked or un-locked. Locked the output frequency is read-off our frequency counter about 964 Hz (never getting a stable reading, sometimes showing about 970 Hz) and un-locked, thus no quartz-channel-signal is provided, its output measured at the anode of R 18 at about 944 Hz. We have to be aware, that the indicated frequency might differ quite well from the actual output frequency, as the counter triggering, as well as often the scope triggering, is being hampered by the existence of various signal components.



Next scheduled tests in our MLK lab. is looking whether improved performance can be achieved with fine adjusting the 6000 pF capacitor parallel onto the primary windings of the anode dust-core-transformer of R 17. As I had, in first instance, replaced it by an 5600 pF C type, though the circuit performed better with the original 6000 pF C; which even might have a dielectric loss. Maybe using a variable tuning capacitor as to cover a wider range than is possible with substituting combinations of capacitors.


Quite hampering though, is to judge what we measure is correct or victim of an unnoticed technical aspect or fault.


On 12 December 2011

I continued with preparing a schematic sketch of the double sized module.


My hand-drawn schematic of the double sized module, which is fed by one out of the 10 selectable quartz channels. Its output is marked with '3'. The input is marked with '4'. That the numbering might look a bit confusing originates from the fact that these numbers had been pointed when my first investigation began. And we just have disconnected this section from the main Nachtfee frame (chassis)



The according screen shots have been noted in a Roman style numbering



Measured at point 'I'


Measured at point 'II'


Measured at 'III'


Measured at 'IV'


Measured at 'V'


Measured at 'VI'


Output measured at 'VII' also carrying '3' (module output)

Please notice, that the screen division have been adapted to the circumstances, in respect to amplitude as well in regard to time base. The locked output frequency is in a spectrum of about 970 Hz. The input signal (a bit lower in spectrum) is depending upon which quartz-channel selected. I must say, that I have not really measured the fine quartz signal steps being each 60 Hz versus the locked output frequency. However, there is a clear lower output frequency when non of the quartz channels is being selected (no locking possible), thus no signal provided at its input.


To be continued in due course



Please don't forget to use the handsome: Nachtfee Chronology page


And, the PowerPoint progress page (converted into PDF)


By: Arthur O. Bauer




Please notice also our: query page

By Arthur O. Bauer



Please go back to, or proceed with: Handbooks papers and product information