Page initiated on 5 May 2017
Although, this project started on already January 2017
On 2 April 2018
+ C +D + K + L + M + N + O + P + R + S + Ta + Ua + V
This project should be regarded Hans’ project
Hans Goulooze was concentrating first on making the remote control to the Ebl 3F operational
This didn't commenced well in first instance, because the parts obtained via Ebay weren't of sound quality.
It soon proved that quite some was lacking, such a the driving motor within the right-hand section of our Ebl 3F
On the right-hand side the FBG2 rmote control unit, which includes an indication that inside the Ebl 3F the correct setting being accomplished.
On 20 April 2017
Hans Goulooze was busy with getting sufficient sensitive reception.
With some care the EBL 3F started to operate, albeit, in the beginning quite some work laid ahead
YouTube films demonstrating the FBG 2 remote control interacting with the Ebl 3F mechanism
On 8/10 June 2017
Hans Goulooze has made good progress in getting the combined EBL3F and EBL2 operational again.
The previous week he encountered all sorts of nuisances, but didn't yet grasped what the faults caused.
On the left-hand side the EBL 3F receiver, next to it the EBL 2 module, with removed screening plates. In the centre Hans Goulooze holds an open AFN 1 indicator in his hands
For this occasion we operate an AFN1 because this sample is at hand including cable connector as well as neon indicator lamp.
Getting a better vision on what Hans is actually doing
Please notice the device with connecting-pin-holes. It is a very versatile device by which means one can measure signals inside a cable interconnection.
Its type is P Mst1 (Prüfmeßstecker Typ 1)
The centre connector-section is feeding the connector pins 1 : 1 onto the according test (Büchsen) connections of the remote P Mst 1 module
The pin and cable numbers are maintained equal. Albeit, that in one occasion a resistor being implemented; its purpose is to allow measuring across the resistor as to get the actual current flowing.
For this occasion the SMS signal generator being used as to supply the modulated carrier at 32.0000 MHz, which equals EBL 3F channel number 21
As to get 1150 Hz necessary for simulating the long-distance landing-beacon-signal an external tone generator being used.
It finally was discovered that an incorrect contact number inside the EBL 2 connector had been once made
This easily can happen as all cables are of equal yellow colour.
However, the fault was determined and it all started to operate as may be expected in this experimental stage
The small circuit between the coaxial antenna input and the coaxial cable interconnecting the SMS has been implemented as to prevent damaging of the SMS signal-attenuator.
To be continued on 15/22+26 June 2017
Hans Goulooze continued with considering the means with which the EBL2 module can be interconnected onto the dipole matching box DAG1* ; the way this once had been done was accomplished by means of so-called twin-ax of bi-ax cable.
* DAG 1 stood for: Dipolanschlussgerät (type number) 1. It has to be noticed, that we have to make a - replica like - DAG device, albeit electrically of equal properties.
The EBL2 and DAG 1 combination being meant for receiving the 38 MHz landing beacon signals (700 Hz and 1700? Hz). Please consider in the next drawing the VEZ and HEZ signals. Such kind of beacon signals crossed the virtual landing path, informing the pilot that first he is 1000 m ahead of the the landing strip and with HEZ that only 300 m being left.
This drawing originates from my CHiDE presentation on Navigational Aids, of, when I remember well, 1997
It shows what the "Funk-Blind-Landefunkfeur" is about.
Before the war the so-called Lorenz Blind-Landing-System was used nearly everywhere in the civilised world.
Please notice, that when the pilot reaches the main warning signal 300 m in front of the landing strip his altitude (h) should have been reduced to 30 m above ground level.
Although, not yet dealt with in great detail, it is worth noticing what the system was about; because sooner or later we have to deal with its aspects
An aircraft approaching from distance will likely receive the airport beacon signal (some even from > 100 km), telling the pilot that he is whether, just on the exact approaching path - hearing in his earphones an uninterrupted tone of 1150 Hz; or right of the virtual path receiving dashes and left of it getting 1150 Hz dots in his earphones.
The trick was - that when both alternating antenna patterns coincide in its virtual centre, that the pilot receives both complementary signals which is then recognised as a constant tone. The course meter (AFN 1 or AFN 2) pointing (staying) in the centre between L and R. In case the aircraft tends to become off the virtual navigational track the signal is becoming modulated with dots or dashes; the way it sounds telling him whether being off to the left or right hand path side.
I bore in mind that about thirty years ago I obtained a bunch of German antenna cables. In our cable storage we indeed found where looking for, a special brown cable type. Its brown colour indicated that its application once was designated for the purposes where frequent cable-bending occurs.
For example, the early Würzburg type fit with its strange, soon obsolete, IFF facility. The two antenna-dipoles were mounted inside the parabolic mirror; left and right of the antenna arrangement. When the Würzburg antenna-mirror changed elevation cables need to cope with (some) bending.
Preparing the cable is more or less done in the usual manner
The main dielectric is most likely of the low-loss Opanol type, well suitable for its purpose. It isn't really solid but it behaves like foam. Foam implies air-inclusions, which is increasing the so-called velocity-factor of the RF cable parameters; I guess ours ≥ 0.7 and also improving cable flexibility. Regular coaxial cables have a velocities of, say, 0.66
A brief tour through some aspects of speed of light (c) versus the signal velocity in systems consisting of dielectrics
More technical approach: v/c wave velocity through a medium versus speed of light; and v = 1/√ε
However, copper antenna-wires hanging in free air are still being considered having 5 % velocity loss. An example, designing a dipole antenna for the 20 m amateur band (14 MHz); its λ = 21.428 m.
This antenna should therefore have a mechanical dipole length 21.428 : 2= 10.714 m
Considering 5% wire velocity loss is reducing the antenna wire length to: 10.714 - 0.5357 = 10.178 m (considering that the central feeding points are very near to one another).
Let us now consider we would like to use our brown cable as to constitute a ¼ λ stub. I considered our cable type is having a v = 0.7. The length of our imaginary stub would become then: (21.428 : 4) ∙ 0.7 = 3.749 m.
The consequences of all this, is, that signals travelling in a cable or transmission system is doing so never with the speed of light - but with a considerable lower velocity. The less dielectric constant (ε) involved the more a transmission line reaches the speed of light; but will never attain it. How light behaves in fibre optics I don't know, but it is most likely that these also have to deal with velocity reduction owing to dielectrics.
Maybe recognisable: around the two conductors being wounded a 'silver like' tape (the phenomenon is not well visible on our photos)
This silver-like shining isn't like that but might originate from 'surface light refraction', as this tape proves to be fully transparent.
Maybe this photo helps you to understand what it all is about
It is astonishing that this cable remains over more than 70 rather flexible. German coaxial or related cables where of a rather high quality standard. Maybe, rather expensive too, as, for example, the copper screening breath is rather heavy; hence, production consuming quite some Cu quantities.
It has to be noticed though, that in those days they may not have operated this more expensive (brown) cable type between the EBL 2 and the AAG1, but the less flexible 'blue colour' cable type. However, we only possess of it a short sample of not yet, say, 70 cm length; hence, for our display purpose too short.
I suppose: that the application of tape surrounding both copper conductors is to create flexibility between the foam-like cable dielectric and the (two) cable conductors.
Maybe this presentation is more recognisable
Just before soldering
Time and again it proves, that decisions made some 30 to 40 years ago are later often decisive for (new) projects. Gathering bits and piece, where the direct purpose wasn't always understood, but the 'trust' that it might later fulfils a purpose.
Hans Goulooze has started considering as to how it later can be mounted on display, of course, fitting within the designated display space
Viewing it slightly different
Hans Goulooze's first idea is constructing a light wooden frame so that we can learn how it all fits together well; with the aim later creating metal mountings.
The space where the FuBL consisting of an EBl3-F and EBl2 + U8 is planned to be positioned is just between the border of the glass-window and the (white) wall below the two blue-window- frames
It is evident that space is quite restricted, but with some care it should be makeable.
The 'Funktisch' has been pulled towards us, as to show what it actually is about; but regularly it is to be noticed through the glass-window.
The FBG2 (the device with the number scale) is to be mounted within the previous shown genuine 'Funktisch' left of the Morse-key; this was its wartime placed within the Siebel type 204 aircraft
This aircraft type was used for carrying passengers on quite long distances (I guess up to 8, excluding the crew); and it flew still in post war days, in several countries; I know in Holland and Switzerland, but likely in more countries.
The FBG2 is the electrical remote-control of the blind-landing receiver type EBL 3-F.
For those interested in the FuBL system its genuine manual might be helpful. Also increasing understanding of what we actually are doing.
D.(Luft) T.4058: Funklandegerät Fu Bl 2, Geräte Handbuch Februar 1943 (please notice that its data content surpasses 12 MB) This data has kindly been made available in digital format with courtesy of Ernst Wagner, Kemnath, Germany.
On 8 September 2017
Summertime has apparently passed, albeit not yet according the regular calendar.
In the meantime Hans Goulooze has done a superb job, he created a wonderful wooden frame which just fits within the board-display-space available.
Please notice first the second foregoing photo.
Our consideration first was how to integrate the FuBl 3F - Ebl2 installation in to the already existing display space
The locks (Schlösser) should arrive, Deo volente, Saturday 16th, this month; Hans is desperately waiting for them.
Invisible on the rear side a common 'ground rail' being implemented (please notice the next photo the Al strip at the rear side of the wooden panel
Hans has used cable colours as far as possible according the genuine cable data, but where we couldn't match to this, he used other colours but consequently
He also has drawn a wonderful sound cable-plan accompanied with full details.
Please notice, just visible the Al guarding (ground) rail.
The switch-handle is for selecting the two possible power modes.
Hans provided, a facility with which we have a choice between operating the alternator U 8 (Umformer type 8) or allowing operating it via regular external power supplies. These supplies are to be interconnected onto the terminals: BA (24 V) and A (260 V)
The coloured wires are well visible
We are happy that we just have enough complete sets of plugs and according cable connectors; as these nowadays are most difficult to find.
These quite odd connector types genuinely originate from pre-war days; but have been used as long as design originated from pre-war or very early war period
It should be noticed that Blind-landing gear existed already since the later 1930s, and accordingly systems were implemented in GAF service.
However, technology improved and systems have been modified step-for-step. Therefore it was quite logic that when implementing single modules into an existing rig that, for simplicity, the original connectors types were still utilised.
On 15 September 2017
Hans Goulooze continued yesterday with 'his' FuBl2 - EBl3F project.
Hans' wooden - combined EBl 2 - EBl3F - being placed, temporarily, where it ultimately should be brought on display
When brought to a conclusion it should: constituting an integral part of our aircraft wireless rig.
Viewing for a bit different perspective
The remote control at the glass-window is finally to be implemented in the genuine Siebel 204 wireless/navigation table.
However, we encountered a problem concerning the output signal of the EBl.2 module; the signal often drops 10 or more dbs.
For this reason the EBl 2 module had to be taken out-of the wooden frame
Please notice that the so-called locks should, Deo volente, be collected on the 16th in Driebergen.
Hans's first action was to look for supply voltage faults
Knocking at various points on the EBl 2 chassis resulted in a wild response on the oscilloscope. Those familiar with those kinds of annoying faults, will not wonder that after all plates had been removed the failures couldn't be reproduced!
we also know - that this doesn't mean that the faults have been solved!
My proposal: let us wait until the next week
Thinking the encountered effects over again, it might be caused somewhere within the VEZ or HEZ signal filters. The latter consisting of an 1150 Hz modulated tone for this occasion.
Albeit it, that we haven't yet implemented the landing-beacon functions (@ 38 MHz), but instead of modulating the EBl3F signal with 1150 Hz. Because both systems join a common audio channel.
On 21/22 September 2017
We just received another Ebl 2 module:
After having demounted the front-cover-panel it was discovered that this device is in a rather sound shape and most likely - inside - untouched
We have been informed before we obtained it, that the so-called Kabelschänze being cut off, but substitutes are accompanied.
Photographed from a different perspective
It is noticed why the cables had been cut, as the cable insulation deteriorated; considering the forgoing photos this module might originate from the end of the 1930s or early 1940s. Whether this is the reason for insulation deterioration I cannot say. My suggestion to Hans Goulooze: let us operate it first in this genuine fashion.
What might occur is that some capacitors will fail in due course
On the far left-hand side the notice the 38 MHz VHF front-end section.
Schloss type Fl 28251 (Fl was a GAF part designation); 'opened'
Aren't these wonderful replica's?
He did a really wonderful job!
Preliminary checking how the Schlösser should be mounted
This lock had to turned as to ease lock-access
On 18 October 2017
Some progress has been made with respect to replacing the wiring of the two EBl 2 connection cables; one for the supply the second one for interconnecting signal cabling.
It is well visible that Hans does his work meticulously!
Viewing it from a slightly different perspective
On 19/20 October 2017
Hans Goulooze has accomplished renewing the so-called Kabelschwänze (the cables between the EBl 2 module onto the connectors.
After some queries, likely caused due to confusion, all started to respond appropriately
Viewing at it slightly from a different perspective
But Hans has built also a beautiful substitute antenna, which normally was built just inside the belly of the fuselage; as not to much causing drag, but still receiving the two beacon signals at 38 MHz un hampered.
Isn't it really wonderful?
It electrically should be fully operational; as the measures are copied from the genuine manual
Even the height above the metal plain is kept according the data in the manual.
We possess genuine German twinax-cable (bi-ax) and Hans has already calculated the values of the antenna matching substitute module. Both - his calculations - as well as the data given in the manual are about equal.
Our idea is, to build the matching unit into a box that should be mounted near to the antenna feeding point, but at the rear side of his antenna rig. Albeit, such that optimal tuning can be accomplished conveniently.
In my vision this is the optimal place to mount it at the wall (just up of the central heating tubes)
We have also discussed it being mounted on the sealing, but in my perception people have inconveniently looking straight upwards.
Looking again at this beautifully made construction It looks like that this FuBl 2 test set isn't operational, but it fully is.
It has been opened because the antenna current meter isn't responding, but likely the meter or the thermocouple is defect
This test set generates most relevant landing beacon signals including the VEZ and HEZ signals; please notice the yellow text "Modulation".
On 16/21 November 2017
In the meantime Hans has, with great ingenuity, adapted the sockets of neon indicator lamps (glass bulbs), as the one used first wasn't suitable for an application within the EBL2 marker-beacon circuit indicator (38 MHz marker signals) (mounted within the AFN 1 or AFN 2 instrument).
I wasn't aware that it is, in many cases, possible to separate a glass-envelope from a lamp base by means of simply using a hot-air-blower set for at at least 300 C°.
With some care and suitable tools, as well as precautions concerning the high temperatures involved, it isn't a too difficult job.
A range of types have been tested, one was found more or less suitable, but it just not yet performed perfectly.
The reason was found on the web, where special neon indicators were on offer. These wore a red colour ring on its glass surface. Hans could order, luckily, a few samples. The GAF Fl. stock-number clearly indicates that it concerned a special performing indicator type.
Our commitments and results on these aspects were nevertheless worth it.
In the course of our neon indicator survey we experimented with matching appropriate bias voltage, where ignition should take place only as a result of the demodulated marker beacon signal content
Please notice the neon indicator bulb a bit up during experiments.
Also was noticed, confirming an old GE publication, that neon lamp cathodes are sensible to external light approaching the cathodes.
When we have to observe the behaviour of Marker beacon signals, at least we should possess a test generator tuned at 38 MHz and modulated with the appropriate beacon-tone signal.
But after seriously testing, it we found some faults, among it a defect antenna-current-meter (0-50 mA)
Secondly, a more serious problem, the 35.2 MHz quartz oscillator stage isn't functioning.
We went back to the LMK lab and tested it on our Saunders rig. Its R1 was about 50 Ω. Please notice these quartz types are oscillating at their fundamental mode!
Only Carl Zeiss Jena, the famous optical and fine mechanical company could manage the task!
The implications have to be investigated in due course.
A disadvantage, we possess the schematic of a PSU-A but our apparatus carries type PSU-B. We already encountered quite some differences.
Hans did a search on internet, and time and again he landed on our website!
Also in the 35.2 MHz oscillator stage.
Which should become a special project or call it survey to be reported upon.
Hans has already dismantled it, as he removed the moving-coil meter section, because apparently the thermocouple is defect, as is so often the case.
The thermocouple device is just within the 'Mipolam' frame construction
In my perception a bit crude.
Our test generator was designed by Lorenz, but manufactured at a Philips factory somewhere in Eindhoven.
Formerly we couldn't see where the thermocouple had its actual defect, but now it appears that just at the thermocouple junction one thermocouple wire is detached (broken off)
Maximum meter deflection
Getting a feeling what the actual current sensitivity of the moving-coil meter is.
Not yet minimum scale deflection
We do not possess a thermocouple for 50 mA HF, but Hans possesses a Philips TH 2 type fit for 15 mA. For our preliminary experiments it should do.
Maybe a bit bulky compared to the 'Mipolam' type frame. But we have found room outside the moving-coil meter.
What counts first is short wiring onto the HF source, onto the meter is dc current fed only.
Our meter is now nearly ready for mounting it into its meter-housing again
We tested it by means of a dc current and 40 mA reading equals the TH2 thermocouple maximum current load.
On 30/11 and 3/12 2017 we continued
Please notice the neon indicator on top of the AFN 1 Blind-landing moving-coil indicator
EBl 2 marker beacon receiver input being fed by means of a: BALUN (Balance- unbalance) device.
We encountered a curious phenomenon, when the neon indicator ignites the sound pitch changes considerable accompanied with strong sound distortion
The tone just represents equals what being generated via the signal generator.
Please watch the additional distortion
In our perception: The neon-indicator is loading the circuitry, but by its own means it is starting to load-and-unloading; causing an additional signal frequency. This is what being noticed.
When we look a bit down at the CRT screen, we still recognise a quite sound sine wave, but with a strange signal form.
In earlier days, there existed tone generators relying just on this technology (effect), providing audio signals; but not sinusoidal like.
Having just changed the setting and feeding at a lower signal level onto the regenerative 38 MHz landing beacon receiver, within the Ebl 2
Hans Goulooze could obtain some of the special indicator bulbs
The red ring indicating that it concerns a special selected type, just fitting for its purpose. The criteria: igniting voltage - cutting-off behaviour; a red dot (not visible) on the bajonet socket indicates the way it should be mounted. As it does matter onto which electrode is fed plus or the minus voltage.
Film 249 (267) Showing the AFN 1 neon indicator driven by the Ebl 2. Viewing and listening to the mean beacon signal (1700 Hz); warning that the landing=strip is 300 metres ahead. Notice, that in practice the 1700 Hz signal being interrupted at a high rate; as to enhance the attention of the pilot.
Film 250 (268) Listening and showing AFN 1 pre-warning the pilot that the landing-strip is 3000 m ahead. Signal interrupted less frequently.
On 27 December 2017
Hans Goulooze has started with provisions to install the 38 MHz, landing beacon antenna reconstruction.
For it, it is necessary to create an electrical substitute for the antenna matching unit. Also providing two small Al mounting frames for fitting it onto a wall.
Hans works always meticulously
The frame turned upside-down.
Viewing the 38 MHz antenna-tuner substitute, because a genuine one is lacking
The tuner in- and outputs are both symmetrical.
Luckily we possess genuine German screened symmetrical cable; sometimes known as: biax or twinax.
Fortunately, the manual is providing the coil winding measures.
It looks sound, isn't it?
Hans is removing the cable insulation
Its construction is most elaborate. Its brown colour indicates, that this cable type being meant for flexible operations.
Notice, that the 'Opanol' insulation (dielectric) should not stick together with the (quite heavy) copper core due to adherence; as the cable structure could otherwise become faulty.
For it, double tapes being wound in counter-spin, around each of the two copper cores. Due to this technique, there exist flexibility between the quite heavy cable dielectric (Opanol) and the Cu cores; allowing some amount of bending.
When you look carefully, you might recognise the two 'counter-wound' tapes structures
The Al-shining is resulting from light refraction, because separated tape strips look quite transparent.
Cu might have been a strategical material though, the Germans allowed lavish application of Cu in cable industry.
On 4/8 January 2018
Hans continued with approaching the faulty trimmers and thereafter with getting his Blind-Landing-Beacon antenna mounted onto the wall.
Hans, from his professional background in 'process quality-management', possesses a sound understanding of the behaviour of various materials.
Let us follow his line of approaching the repair of our defect trimmers.
When I arrived in our Klooster premises last Thursday morning, Hans had already removed the oxide at the silvered trimmer disk; by means of a heavy version of "Scotch Brite", as to allow proper soldering
Please bear always in mind: that one should never approach these kinds of (silver deposited) defect trimmers with a solder type containing lead (Pb)!
The proper way is using a mixture of silver-tin.
We luckily possess a bobbin of silver-tin solder; albeit purchased about 45 years ago.
I have repaired trimmers with this solder decades ago, these weren't always looking nice, but fulfilled my requirements.
Hans came up with a crucial suggestion: that during soldering 'the trimmer body' should be brought up to a temperature level slightly lower than the solder-melting-point.
These kind of "hot-air pistols" can be bought nowadays in nearly every tool shop or market.
However, for it, you need at least some assistance.
Before approaching the delicate trimmers, we used another sample, showing a similar defect
Please look carefully, you might notice what our current problem is
Just the failing solder contact between the silvered trimmer-disk and the contact-screw.
The actual defect shown more in detail
It is clear that we have created a sound contact between the trimmer disk and the spindle-screw
Don't worry about the brown residue - because this originates from the necessary solder-flux one has to employ. Which can simply be removed by means of, for example, alcohol.
Our advise, be careful, that the flux residue does not enter the space between the trimmer disk and the fixed trimmer body.
Therefore, we turned the trimmer such that the flux residue washes easily away. Even renewing the alcohol liquid ones or twice, is advisable.
Finally we approached the defect trimmer belonging to our Schwabenland receiver survey project
Again: don't worry the brown residue, like the foregoing trimmer, originates from the necessary solder-flux.
Our silver-tin solder does not possess a flux-core, therefore a substitute flux has to be attached before starting the soldering process.
Later on we built the latter device into its Schwabenland coil box again and it responded as what may be hoped for.
The first trimmer we did repair, belonged to the antenna matching box to the 38 MHz main Blind-Landing-Beacon system.
Now time is right for attaching Hans' 38 MHz beacon antenna arrangement onto the wall. Our criterion: best placed for demonstrations.
First, Hans had to measure where the screw-holes should placed
Hans really did a great job
The construction allows still tuning of the antenna matching unit; mounted at the rear side of the antenna substitute.
We should be careful with mounting the biax-cable onto the wall
In particular in regard to the central heating tubes, of which one might become, sometimes, quite hot.
On 11 January 2018
Hans encountered strange instable EBl2 responses on 38 MHz.
It appeared that the concerning NF 2 valves suffered from defect metal screening
The date concerning week 34 of the year 1938, noticed on the metal valve-screen might also indicated the age of our EBl2 device, which internally proves to be virtually untouched.
On 12/13 + 14/17 February 2018
The interruption of about a month, has been caused mainly due to quite annoying troubles encountered within concept of our PSU0-B test set; meant for the EBL2 and EBl3F beacon receivers.
We encountered:- that the 35.2 MHz quartz oscillator stage was generating, but not at its appropriate frequency; often between 130 and 160 MHz. Even not directly related to an according overtone mode.
The difficulty encountered, was that the genuine REN 904 valve, actually operating far beyond specs, oscillated spontaneously. Even in the case of heavy capacitive loading of the valve anode.
The actual circuitry differed from what being mentioned in the available manual.
This circuit we found actually existing in our test set, apparently genuinely manufactured this way
This schematic differs fundamentally from what is shown in the genuine PSU0-A manual.
This circuit oscillates between, say, 130 and 160 MHz; or it doesn't oscillate at all.
However, during our long lasting experiments, I understood that apparently the function of the trimmer, was not adjusting frequency, but was more-or-less adjusting the signal feedback.
In an attempt to restore the circumstances as shown in the real manual, we reconstructed it this way; equalling what the genuine manual provided
Do not wonder, that nothing fundamentally has changed in the response of our oscillator circuitry.
A range of signals discovered, but in no way at the quartz frequency spectrum of 35.2 MHz.
Could it be that our quartz crystal is defect? I don't believe so, because some time ago we tested it on a professional quartz parameter test set.
This circuit didn't change the mall-functioning of the circuit
In a desperate mood I decided to continue experiments in our MLK lab.
It is evident, that the wartime Carl Zeiss Jena quartz still resonates in its 'series mode' vibration on 35.203877 MHz
Measuring its so-called R1 value of 43.5 Ω
Constituting the series resonance resistance; a tuned circuit in resonance responses (obeys), principally pure ohmic.
Interesting, the optimal vibration frequency can be determined by watching the R1 value versus tuned frequency. There will be find a minimal resistance value somewhere on the digital reading; in our case, this being 43.5 Ω.
I suppose not too bad, for a fundamental quartz resonating on 35.2 MHz. Even nowadays, fundamental resonance at 35 MHz is a great exception. One would use an overtone circuitry.
I first used the same setup as being used in the Klooster premises, thus experimenting with a REN 904 valve again.
The annoying results being equal as was encountered in the Klooster previously.
My first thoughts went in the direction of phase-delay due to this odd valve type. After all, it most likely originated from the long metal rods (Stege) carrying the electrode system, within the valve envelope; because oscillation continued even when the REN 904 anode was blocked with capacitors of 15 nF and beyond!
Maybe not in accordance with wartime technology, I took ultimately an ECC 81 valve. Fortunately, in the valve box was also found an according noval socket.
The circuit responded well, but again at about 129 MHz and beyond.
My next move was restoring what was given in the genuine manual
Before I started the circuit experiments with the ECC 81, instead of the quartz crystal, I injected a 35.2 MHz signal from our R&S SMH generator and tuned the output trimmer for maximum signal output.
HT was again provided, and it started to run at the so desperately wanted 35.2 MHz
The grey valve envelope belongs to the REN 904 valve used in the forgoing (failing) experiments.
It is clear that the circuit operates
As to save the counter input circuit, I use the 10 x probe attenuator cable for the counter
Reading off: 35.21150 MHz.
A Pierce oscillator type does necessitate a complex tuning (loading) in its anode (collector/drain) circuitry.
Let us notice:- a Pierce oscillator circuit operates in the so-called parallel resonance mode. Parallel resonance-mode lays always a tiny bit higher in frequency. Quartz in parallel resonance allows some frequency adjustment, in contrast to series-resonance mode.
Don't tune the circuit at maximum output, it will next time not starting up automatically. What should be done:- is finding a proper point on the slope of the tuning curve. I luckily remembered this aspect from my commercial production of VHF/UHF RX in the 1970s. Albeit using transistors, but the principles do obey equally.
Only a scenery
Wednesday all should be cleaned and then preparing for transport
I have discussed with Hans Goulooze, this afternoon, that we should try to implement an acorn valve type 4675 made by Philips. Its advantage, it necessitates 4 V filament which equals the provision for the availabl for the REN 904.
The application of an ECC 81 would mean, that a voltage doubler arrangement has to be implemented as to make 4 V → (about) 8 V
However, its slope factor of 2 mA/V is quite lower than 5 mA/V for an ECC 81.
An advantage of adopting an ECC 81, is, that the output level (second triode section) can favourably match onto the "loading" conditions of the REN 904 mixer stage.
We will see whether it can be made operating.
My temporair opinion:-
It might once have been, that the criteria were:- a particular REN 904 valve sample and a selected 35.2 MHz quartz module allowed proper operation, otherwise, the adopted circuitry could never have operated reliable.
It has to be noticed though, our current quartz crystal is not the one once matched onto the moment of its acceptance.
Donald Prins, once employed at the Philips Quartz Crystal plant, had it laying at his desk (he wanted testing its parameters). He reported - that due to its cylindrical shape it rolled off his desk and smashed on the ground, the remains weren't useable anymore.
Luckily, at a meeting (Dresden 2012), I saw our current quartz device inserted in an old 1920s set, where it apparently wasn't meant for. I told the owner my particular concern and he very kind gave it to me.
A question should be razed, why have they adopted such delicate circuity?
In my perception, they only gained the lacking coil-body.
Though, was this all worth it?
I tend to understand - it was not worth it!
Gaining perhaps 1 RM lower material/production cost versus likely much higher costing efforts as to match (unreliably) valve and according quartz crystal devices.
On the 14th the MLK lab table has been reshuffled, the necessary 35.2 MHz modified oscillator circuit and other bits and pieces being prepared for transport to the Klooster premises again.
Quite a different situation than the foregoing scene!
Don't you agree?
On 1/5 March 2018
Continuing the foregoing experiments - it first seemed to perform well, but at a sudden moment the 35.2MHz quartz failed resonating entirely.
Is it really defect?
Therefore, I went back to our MLK lab; tested it, and found that the equivalent series-resonance R1 parameter being 40.3 Ω. Which proves, that it is still functioning.
After some consideration, I drew the conclusion, that mechanical overloading could have caused this nuisance.
Why not adopting a transistor stage instead?
A long time ago, I designed and manufactured VHF and UHF receivers (1970 early 1980s). Therefore we used for some circuits BF494 and BF495 transistors. A brief check showed, that some are still in stock in our MLK lab. I know - not the most up-to-date types, but according their specs these must do well on 35 MHz.
First I considered the same type of oscillator as was used within the valve driven oscillator; a Miller type circuit.
This didn't work.
My next approach, wiring it like in a Colpitt-oscillator circuit; where the crystal (resonating in parallel resonance) being wired between base and collector (in valve terms g1 and anode).
This worked instantly fine.
It proved, unexpedly, that some reactive component in the collector load should exist. Hence, tuning the tuned circuit a bit on the slope instead of at maximum amplitude. It proved to be necessary to adopt a symmetric load against ground. Always bearing in mind, that stable starting-up and operation is our main aim.
This became true, when I switched it on on the next day (2 March) no signal could be detected. A slight change in setting the tuning trimmer (the big ceramic disk) and it did start operating again; but just at a tiny bit lower signal level; caused by the slight detuning. When aligning the circuit I apparently forgot this very fact.
Quite basic isn't it?
I operated it first at 12 V dc, but the current was considered too high. Watching the current meter - I reduced the supply voltage and found 8 V a still sound functioning oscillator and measured 17 mA; implying a power consumption of 136 mW. Say, about 50% of wherefore the transistor had been designed.
After tests for several days, the open wire construction was put carefully in a box and brought to our Klooster premises.
My first action was to implement it into the existing facilities in the PSU0-B test set.
This photo has been taken afterwards, thus when it all pointed into the way it should be accomplished (the red wire supplies the +8V)
Just underneath the quartz socket we just see the: flying-wire technique mounted BF494.
I adopted the already existing tuned circuit, including using the genuine ceramic trimmer (the big circular disk)
After quite confusing experiments, it eventually was found that an ECC 81 can perform as was expected (but not more than that) and it was checked whether the genuine REN 904F valve, originally used, would perform too.
And it astonishingly did perform about equally.
Albeit, I kept a cathode resistors as was already adopted for the ECC 81.
Because I still used in the foregoing experiments the genuine valve base of the REN 904F; it being simple pulling out the ECC 81 "Ersatz" valve and inserting the genuine valve REN 904 instead.
Because it proved favourable to operate with a self-contained bias I implemented a 470 Ω cathode resistor blocked by means 15 nF; I measured across it, say, 2 V.
I kept this provision for the REN 904. Short circuiting the provision didn't influence the measure output amplitude much, but is found better for the life of the >75 years old valve sample.
Keeping the genuine situation as much as possible genuine.
The red clamp is supplying the 8 V for the transistor stage
It proved, however, necessary to the implement an additional tuned circuit in the anode of the driver-buffer stage; fed from the existing HT within the PSU0-B set.
By this means it became possible to separate completely the the mixer-valve grid from the buffer stage.
Quite a strange experience, tuning the buffer anode circuit, had to be accomplished by looking just for the point where the sine-wave shown on the oscilloscope screen is looking sound
Up or down this point the projected sine-wave looked a bit distorted, proving the appropriate functioning of the tuned anode circuit.
Also the counter showed stable performance
For it we even re-adopted the genuine Philips type anode resistor again
The circuit now used is in some respect on the lines of the genuine PSU0-A manual which we have to rely upon.
On the far left-hand side the flying small tubular capacitor is only 1.5 pF whilst in our test set genuinely a 100 pF capacitor had once been implemented.
But our test set type could never have functioned reliable!
In some way or another, once a tuned circuit coil must have been mounted.
A way should now be found to implement somewhere a miniature transformer 230 - 10 à 12V
The Philips type trimmer has been mounted a bit facing downwards, as to prevent it touching the metal case or box. Please notice it operationally carries about + 200 V against ground!
Its expected power consumption may be considered for, say, 30 à 40 mA due to the additional 7808 regulator consumption.
Final comment today: what looks promising, is the fact that the oscillator signal watched on the oscilloscope (output of its output) is steady, in contrast to the foregoing situation, where always a sudden irregular signal amplitude being noticed.
Further tests should proof whether this presumption is valid.
On 2 April 2018
Progress have been made in getting the PSU0-B test set operational again albeit, in a some modified fashion.
A sound tip I got from Raymond Domp-Frank, and Hans Goulooze obtained a sample for us.
Considering the size of the ½ W resistor (or the 0.1 inch separated holes) - this module is compact and and provides 15 V; even up to 5 W
Why using a big resistor?
The reason is quite pragmatic, we had no smaller type at hand.
The application of a 100 Ω resistor is that we can smooth the operational voltage a bit from the quite steep pulse slopes on the module output. One never should load pulsed power supplies directly with a 'too big' smoothing capacitor. Improvement was got using several capacitors types wired in parallel.
100 Ω has been chosen also as to ease reading off the voltage but mA. Hence, 1.5 V equals 15 mA current flow.
These two photos showing the two point at which the 230 V can be found, necessary for feeding our switched mini power-supply unit
The two connections in between (the middle) might once have been meant for 110 V operation.
Space was easily found to mount the small 15 V power supply
The new 35.2 MHz quartz oscillator stage being fed from this new power-supply module (just visible on the other side of the Al plate
Please do remember, that we had been forced to modify the 35.2 MHz quartz controlled oscillator stage for several reasons. In the concept once provided, it oscillated on roughly 140 MHz or 160 MHz, but never on 35.2 MHz!
My first modification was changing the circuitry as originally given within the PSU0-A manual (Pierce type cicuit).
Then creating an adapter as to adopt one side of an ECC81 valve, but using a base as genuinely used for the REN 904F valve. This operated well, but at a certain instant it stopped operating.
It proved - that apparently the 35.2 MHz quartz did strike due to overloading!
In a desperate move, I considered adopting a low power transistor controlled stage.
This proved to be a sound option. As to amplify signal and still using the genuine REN 904F valve, it finally started working reliably. Albeit, that additional modifications had to be implemented.
There are still enough PSU0-A/B sets around elsewhere, that this modification being worth implementing.
Looking at the antenna-output of the test set
Don't be surprised, the combined mixer and output stage is being supplied with three different signals:
The 35.2 MHz signal visible is e sine-wave ripple, the local oscillator given by the equation: Fx = Ft - 35.2 MHz.
Fx = the mixer signal frequency; Ft is the output or transmit signal; resulting from the mixing process.
For example when the test set should provide 33.33 MHz → 35.2 - 33.33 MHz = 1.87 MHz
For 30.00 MHz we get 35.2 - 30.00 MHz = 5.20 MHz
For 38 MHz, the landing beacon marker: 38 - 35.2 = 2.8 MHz
The Germans call such signal: Überlagerer, elsewhere known as local-oscillator (LO).
And, very significant, we get to notice the audio modulation: 1150 Hz; 700 (VEZ or 1700 Hz (HEZ). It concerns anode modulation of the mixer stage.
Being aware of a single-tuned anode coil, we may expect that all signals in some way will be measurable at the antenna base.
Is this hampering the purpose of our test set?
Not really, because the DUT, in casu: EBL3F (or EBL3H) is possessing selectivity from its own right. Maybe less within the EBl2 receiver front-end (audion).
Apparently another signal maybe 30.5 MHz
Maybe you can recognise the 35.2 MHz ripple super-imposed onto the summing output signal
Most likely the 38 MHz beacon signal und 700 Hz modulation
I suppose that that this signal concerns the 1700 HEZ* signal
* HEZ Haupteinflufzeichen, or main landing-beacon signal, received when the aircraft is 300 m in front of the landing strip.
Test set place in its genuine case
Please notice the 10x scope probe among the antenna 'substitute' antenna cable.
The arrows pointing at the fixing screws
To be continued in due course
By Arthur O. Bauer