Help with Telefunken 655W last (?) touches

[Chuck-II]

Greetings,

First post. A lifelong ameteur/tinker, I'm hoping to complete my first antique radio restoration (cost-effectively) and declare success. But I'm only learning and I'm stuck, needing experience to evaluate my thinking, guide next steps, etc. Perhaps someone here is willing to help a neophyte?

The radio https://www.radiomuseum.org/r/telefunken_655w.html is a '57 and is French. It was not abused like many other flea market units I've seen; so nothing scavenged or obviously burnt, etc. and I grabbed it. The model seems rare as there are no surviving schematics I can find. I have the "D" schematic (German) but it uses different valves. I have a Polka diagram which uses the same valves. It is newer, lacks the EM34 Magic eye, and has other small variations, but may be the closest I will ever get. So, after re-stuffing the electrolytic "can," I replaced the rest and checked all the resistors. They seemed okay. With my test lamp -at first- I got uncontrollably horrifying audio (intermittently) with the volume knob at min. But, knocking on the 6AV6 AF valve with a chopstick, and using the phono input as source, it eventually ran well for hours --until power-cycled. So, I replaced that valve, linked all the chassis solder connects via wire and the howling stopped. Cleaning the band selector buttons, tuner, and pots with alcohol, I got MW to work, listened to a UK station with the internal ferrite antenna while watching football here in southern France. So, not too bad. But it seems weak, I get nothing on LW/SW and the EM34 eye never closes. It only fluctuates during cuts via the band-select buttons.

I've only a load lamp, bench-made cap discharger, 300V DTM, and an iron, so no tube tester, signal generator or scope - yet. And, though I've read many R/R hints/tips and found the best ones here (IF/RF Stages) I lack the skill to know what to do next. I'm looking for easy fixes, which may not be enough, but seem worth a shot. My diagnostic assumptions may need to be corrected:

Valves: since the audio works nicely and consistently now using phono input, I assume the main power rectifier and two last-stage audio valves are good, leaving the mixer/oscillator and remote cutoff valves as unknown. ??

Antenna: with some AM stations coming through from Spain, the UK and perhaps Northern Africa, only on clear nights, it may be that the internal ferrite antenna is just not good enough to deliver enough signal strength to hear well and also get the eye to close? My portable https://www.radiomuseum.org/r/majest...bl_rt_178.html with it's telescoping antenna receives nothing on MW/LW/SW, but it's my source for FM. If this is the case, perhaps the two unverified valves are fine and I should get a tuned-loop and/or drop a long wire out the window for SW & LW?

EM34 Resistor failure. Perhaps I just need to replace all the resistors at the base of the valve socket and see if that helps the eye to close?

I've "rocked" and returned the 4 (trimmers?) on the IF transformers back to where they were, to be sure they were not oxidized open. They did not seem "stuck." And, the alignment may have drifted from age and my recapping, etc. I have no spec sheet or tools to align anyway.

Perhaps there are other keen observations.

Thanks so much if you have any suggestions on how to proceed from here.

-Chuck

[agardiner]

Hello and welcome to the forum and the wonderful world of vintage radio!

OK, so it sounds like you have done everything important. A few observations...

1) The fact that you are getting MW stations including from abroad, suggests that the set is working quite well, especially as your portable gets nothing! You won't get anything noteworthy on SW without adding a wire antenna. Try connecting between a 3M and 6M long wire to the antenna input and I bet your set will burst into life.

2) Magic eye - this will only give an indication when the signal is strong enough, so I would not be concerned that it isn't working at this stage.

3) Although it is probably fine, it may be worth trying another ECH81 valve. I have restored many sets that use this inexpensive valve, and they do become tired. A fresh one may well add a worthwhile improvement.

However, do try the wire aerial first!

Regards,

Adrian.

[marceljack]

Hi Chuck-II,
Do you know the french site doctsf.com ?
The 655W is listed there with photographs but no diagram:
https://www.doctsf.com/telefunken-655-w/f41777/o=y
Thee is another model D655W which has the same set of valves and for this one a diagram is available but with the wrong set of valves (Rimlock).
However apart of the valve pinning the diagram should be very similar.
https://www.doctsf.com/telefunken-d-655-w/f20915/o=y
You might also get some help from the associated retrotechnique forum:
https://forum.retrotechnique.org/

[Chuck-II]

Gentlemen, thank you.

Adrian, I got 6M of lamp wire, split and tried it. No joy on any band. Then, strapped it to 12M and strung it across the apartment and also threw it out the window, but to no avail. I then made a ring/can-tenna from a 6" yogurt cup. It boosted the interference (LED, WiFi?) quite a bit, but no good radio signals. The magic eye lights, but still does not change.

The radio seems to be getting weaker. I can't test the ECH81 (mixer/oscillator) or 6BA6 (IF/RF remote cutoff) and can't find a good list of diagnostic "fail symptoms" for either. They both heat up well, rock in their sockets with no trouble, ping with a chopstick, and have no obvious defect signs (like white from vacuum loss). If it makes sense to buy replacements to see what happens, you think I should start with the ECH81?

marceljack, D is for Deutschland (different model), and the doctsf shows good images, like radiomuseum, but also no data. Thank you for the retrotechnique reference. Maybe they will know of reference data?

-Chuck

[agardiner]

Yes, start with the ECH81 as I often find them to be weak and always keep them in stock. Fortunately not particularly expensive.

[Chuck-II]

Adrian, bingo!

Found someone with a tester, and the ECH81 was less than 50% of it's target. Radio interference now drives the eye to closure, so the rest of it is antenna related.

Thank you again.

[agardiner]

Great result.

[Mr 1936]

Hi Chuck-II

I see you are in Nimes. The beautiful Roman Ampitheatre (which I have seen) has survived for thousands of years. Unfortunately the medium wave transmitters in France have survived less than 100 years ! I don't know if there is anything left on Long Wave or Medium Wave which is within daytime ground wave range of you.

If you listen after dark you will have a better chance of receiving distant AM stations via reflection from the E layer of the ionosphere

Shortwave reception should be possible at all times, but there are fewer stations than there used to be.

If you want to try a Medium Wave wire antenna, it's best if most of its length/height is spaced away from interference producers (most modern electronics) and the mains wiring they are connected to. The vertical part of the antenna is more useful than the horizontal part.

In addition, if you can put in a dedicated radio earth rod, it will be less noisy than just relying on the mains supply to provide an earth (via stray capacitance in the radio)

[Chuck-II]

Hi and thank you.

Yes, I never thought I would think nothing of walking by a more or less intact Roman Ampitheatre every day. Strange. Also disappointing to know after all that work there's not much out there to tune into, LW/SW/MW and I'm behind a large hill north of me trying to get the match in UK English. Too far away and bad weather... May try a passive DIY ring or run a sloper just to see if any of it works, and will look at the lightening->earth system here.

Or, is it worth it to dig your own hole?

Had to pitch the electrostatic tweeter and the philips piezo doesn't add much running off the main speaker, but I ordered an FM/BT converter to sneak in the back so I can get FM and try to learn some French!

[Chuck-II]

Not sure if I should continue my awkward journey here or start another thread?

The bug has bitten more deeply, so I am now even more dangerous. So I'm seeking some advice before proceeding. An active antenna brought in weak signals, but my SW bands were still dead. So, I read Marcus & Levy several times and various forums suggested searching for open coils. Under magnification I found one hair-like lead of the SW coil (under the piano keys, about right in the schematic for the similar POLKA set, so I think these are the RF coils?) disconnected. Soldered it back on and with my radiator earth ground and active antenna: bingo! I now have all the bands picking up something, so I want to do my first alignment to get it done right. But, I don't want to die doing this. Not on the first go, anyway.

I bought a 20 MHz FY6900 and an oscilloscope-on-a-board to become familiar with all it's functions and produced modulated RF test signals that work. Then, sadly, a loose dummy ant. wire hit the 3mm screwhead of the AC main to fry my sig-gen CH1. No more modulation, and the VCO-in appears to only use CH1. Or, it's fried too. But, CH2 and the counter still work. The wife screamed and perhaps a good lesson was given about respecting current.

Fear remains powerful. I do not have an isolation transformer (with no earth passthrough) for my DUT or for my generator. I can rig up my load lamp, risking the last incandescent bulb in France, but... Both my DUT chassis and the sig-gen are grounded, as I attached the earth via a new 3-wire (main AC) cord. I THINK this means they are operating at the same potential, but I have no training and I'm not sure. Also, I THINK it's not a "hot" chassis (I'm still here and there was no continuity between the frame, and hot or neutral when cold, but the switch was on) I. noticed (perhaps stray capacitance, low sparks) while dragging the ground against the frame, so I made it permanent.

Using the EM43 "eye" as a peak amplitude meter, I THINK I can inject 455kc unmodulated at g1 (Pin 2) of the ECH81 (through a .1 mF 400V cap) at a low amplitude (0.005v) to tune up the IF cans as prescribed by Marcus & Levy, but I've never done this before and don't want to lose what's left of my generator, or me, etc. on connecting the ground of the generator.

If it works without connecting the sig-gen ground clip to the chassis, I won't. But, if it doesn't, I will use another cap with the black clip and I could connect up my meter, or try a light bulb to see what may happen on contact. I do not want to be the cause of any more vaporizations

Also, Marcus & Levy's advice was for connecting test equipment from 1947. So, I do not know if their suggestions and values are valid for today's DDS, but there is only one channel left on my generator. And, I like living.

Sorry for the long post, but if you can confirm or correct any of my ameteur thinking, before I do this, my first alignment will go more safely. TIA!

[Richardgr]

This is an interesting alternative approach for exercising the circuitry of your old radio, but picking up online radio (so UK radio would be tunable) ...
UK Vintage Radio - DAB Conversion ...

[Chuck-II]

Thank you. Just want to be safe hooking up unfamiliar test equipment for the first time. I get the UK on SDR 7x24 & AUX it in via BT

[Mr 1936]

Hi again

I am 95% certain that your 'AC only' radio uses a mains transformer for HT and LT, and that the chassis is not directly connected to one side of the mains. If the chassis is left floating, there may be some stray capacitance to the mains which will make it sit at an AC voltage, but this is effectively in series with a high impedance so it is not capable of delivering much current.

If you connect the chassis to a reliable earth, the chassis voltage will immediately drop to the same voltage as this earth. I don't know what the detailed electrical rules are in France for house wiring, but ideally the heating radiators and pipes will be bonded somewhere to the property's main earth rod. Alternatively you could use a 3 core mains lead and the type of 3 pole mains plug that has an earth connection.

Your signal generator probably has its own isolated mains power supply, and the signal generator output socket earth is likely to be connected to earth via its 3 pole mains plug. It should be OK to connect this directly to the earthed radio chassis, BUT if you want to play safe use a series capacitor in the earth link which is low in impedance at RF but high in impedance at 50 Hz. I suggest 0,01 microfarads (10 nF) with a rating of 400 Volts DC or higher. This will have an impedance of 35 ohms at 455 kHz, rising to 318,000 ohms at 50 Hz.

The same sort of capacitor would suit the 'hot' RF connection from the generator, but you could try a much smaller value (100 pF ?) so as to reduce any loading on the radio circuit you connect to. The small value will reduce the injected signal level, but you should have plenty of level available so this is not a problem. If anything, these DDS generators produce too much level for radio work even at their minimum setting, so a small capacitor here will help.

I also suggest that if you use spring clips for connections, mechanically anchor any wires (for example with cable ties) before you apply power. This reduces the risk of anything getting pulled out of position.

Any alignment adjustments will normally be slight (just to account for aging) and should be carried out at the smallest signal level possible.

Bon Chance !

[Chuck-II]

Thank you, and very much. After frying CH1 of my sig-gen (unanchored wire) I want no more sparks or smoke.

Hot and Neutral are both on the transformer, but there's a cap to the chassis. The sig-gen also has a third mains wire, but it's Chinese. Your December suggestion for real-earth grounding improved my reception dramatically. So, I'll add the radiator as earth to both before hookup or power-on. I also appreciate the suggestion for cap placement/values. I can get to .05 mF and 110 (or 220) pF via series / parallel hookups. I left all the mica "RF/tuner section?" and rod-style caps in place, but all the rest in the set are new.

I made up a tuning wand, a screwdriver, and de-waxed/freed-up all the cores with no issues. Though the caps just arrived, I'll wait for the insulated pistol loop-grips to be more careful this time. Gives time to plan and clarify adjustment points, as several are unclear or don't exist / correspond with Marcus & Levy (p. 258-62 & ch. 22) multiband/general alignment. Could be me though.

I've only a "close" schematic. But, no alignment instructions or location map. The IF frequency of 455 kc is stamped on the can and adjustment points are clear, so I'll start with the second can. The oscillator and antenna trimmers are visible on top of the cut-blade tuner gang, but must be common (?) to all the bands. The lone coil up on the ferrite antenna is most likely the wave trap (shared by MW &LW)? And, the only other adjustable coils I can see are under the piano keys: SW has 2, LW & MW each have 1 and the SW spread-band has no coils, just caps. Once the IF is done, maybe it will be enough to get the dial to read "close enough" across the bands and go no further? I hope, but we shall see.

If not, after IF comes SW-hi, SW-lo MW, then LW. The coils under the piano keys may be called "RF", as all the other points have been identified? Still, the mixer is the first tube; I'm not sure. And, 2 separate coils for SW perhaps correspond to hi and low frequency tunings? This also remains a mystery. I'm hoping there will be no need to slide coils on the ferrite rods. And with no modulation, I'll need to leave the wave trap for another day. So, as a learning process, I think it's coming along. TBD for the radio!

Thank you again for the help!

[Chuck-II]

Well, an IF alignment can indeed be done with an unmodulated sig-gen and a magic eye. But other methods are likely easier and therefore better. Still, I'm glad to have survived, learned something and done good work. The dial tracking part failed though, as I could not get a MW carrier signal through.

The last known-good MW state was before I reconnected that detached SW coil wire. I thought I had thrown the IF way off making a screwdriver and playing around trying to ID primary / secondary coils. Now that I've done the work, it turns out not to matter, and it was probably not that far off.

MW returned after disconnection so the IF stage is working very nicely and the pointer is only off by 10 meters or so. I'll finish the MW alignment tomorrow and explore other attachment options. Several connection points are within reach of this lead, so I hope one will be the correct one. I'm not clever enough to know, so more trial and error. At least it's not high tension.

Thanks again for the help and support!

[Chuck-II]

Got the hookup right and all bands work now. It's much louder too. And, I'm hearing stations on all bands (even LW!) though the dial is a bit off. But, before I start twisting any slugs, it's probably good to ask for a little more advice. All I have is an old book and better websites, like this one, but I think I can see the finish line.

What I believe/think I know:

Any adjustments I make will carry across all the bands, and be most noticeable on the lowest frequency band. I think this means start on SW carefully and finish up on LW hoping to be close; expect a few cycles to get the best overall result, sort of like IF tuning.

The tuner has two identical-looking gangs, but I've identified the oscillator from the antenna via personal capacitance. Each has a slotted screw up top (mica based capacitor "trimmer"?) and the oscillator's trimmer in this location is for high frequency adjustment. There is a (or are some) corresponding coil(s) elsewhere for low frequency adjustment. This/these may well be the coil(s) under the various piano keys, but this is where my fog develops. How to tell an RF/Ant. from an oscillator coil? Dunno, but there are only four adjustable coils. None physically under SW-spread, 2 under SW-full, and 1 each under MW and LW. They respond as follows when I insert my tuning wand:

SW-full-band position There are two coils. One squeals; the other does too, but it's less pronounced. Wanding all the other coils on all the other keys in the SW-full position produces nothing.

SW-spread The coil that squeals above squeals here, but wanding the second coil under this key does nothing. The others do nothing.

MW-position. The single coil under this key squeals. Others do nothing.

LW: Wanding this single coil squeals. Others do nothing.

Dial tracking fed/read (meters):

SW: Fed signal at 17m (reads out at ~16.15m) | 50 (50)
SW-spread: 48m (47.2) | 50.5m (49.8)
MW: 250 (260) | 500 (512)
LW: 1000 (1050) | 1700 (1880)

Except for the spread band, it seems to read shorter at higher frequencies, then progressively longer. Also, is there usually only one common antenna trimmer on the tuner, which is common to all bands? Bit of a mystery. Any comments on how best to proceed will reduce the error portion of my trial and error, and would be greatly appreciated.

Many thanks,

--Chuck

[Mr 1936]

Hi again Chuck

I do not have the schematic for your set, so I can only offer general guidance for alignment. Remember that this is a domestic set, not a laboratory instrument, so small frequency errors in the dial are normal.

The IF adjustment is done first, as it applies to all bands. Normally 4 tuning slugs, sometimes also a wavetrap to minimise IF breakthrough from the antenna.

Now check that the dial pointer travel is correct. Sometimes there is a small mark at one end of the dial to show where the pointer sits when the tuning capacitor is fully meshed or fully unmeshed. If necessary, slide the pointer on its string.

Next do the highest frequency shortwave band. The oscillator trimmer adjusts the frequency of reception and has most effect at the high frequency (low wavelength) end of the band. The oscillator inductor has most effect at the low frequency end. These two adjustments interact a bit, so you may need to do them 2 or 3 times. The RF trimmer and inductor affect the sensitivity, but do not change the frequency of reception (maybe slightly on shortwave) The trimmers for the highest frequency band may be the ones which are part of the main tuning capacitor. (lower bands have separate trimmers)

Next do the lower frequency shortwave bands, if the set has them.

Next do medium wave. The medium wave RF inductor may be the ferrite rod antenna (if used), and is adjusted by sliding the coil along the rod.

Finally do longwave. Sometimes not all the adjustments are provided for longwave. Again, the longwave RF inductor may be the ferrite rod.

Remember that the set was aligned once. In general, things do not drift with age very much. The normal reason for realignment is when a critical component has been changed, or when an "amateur" has been turning the screws

[Chuck-II]

Mr 1936,
I am that amateur! But, aspiring and trying not to make it FUBAR. But, both the set and I have come rather a long way since aquisition, with much kind assistance.

Others seeking info on the rare 655w may know the earlier AM/FM Jubilee Export (53-55) has full information online, and the Polka built a few years later (57-59) schematic is available, and was also a French version of the postwar arrangement. It has the same tube lineup, but lacks the eye, an electrostatic tweeter (sadly, mine was not repairable by me, but I did try), a tone control, and uses a single stationary ferrite bar rather than the rotatable dual-rod setup as in the 655w, but is otherwise almost identical. Both switch between the int. and ext. antenna at the knob, though the Polka seems more an economy version. None were lab-quality; that's quite true.

The IF work went slowly, but nicely, and safely (thanks again). The set is much louder now as I suspect more signal is passing through to the AF stage. Pointer placement is correct; I'm better mechanically than electronically. A "5th" coil (physically mounted beside the ferrite rods) has internal threading, but no slug in it. I am interpreting this as no wave trap, or no adjustment to be made there. 455 kc/KHz seems to go right through either antenna input. This helped me evaluate IF tuning success. I think the military used to transmit morse code at this frequency, and no longer do, making it less essential. Dunno.

The puropse and relationship between the oscillator trimmer (cap) and inductor (coil) (making an L-C or tank circuit?) was a bit foggy though. so your explanation helps a lot. In multiband sets, like this, it seems there is one shared trimmer on the tuner, and each band appears to have it's own inductor, as the 3 horizontally aligned coils (each under its own key) share a yellow wire --leading directly to the oscillator section of the tuning condenser. I gather the oscillator trimmer has to be averaged for the higher frequency section of all bands, while the 3 inductors (if I have this right) will provide some lower frequency adjustments, one for each band, and these are switched in and out. The two SW bands share a single coil (supported by wanding this coil in both positions with audible results) and will need to be averaged across both SW bands. I can do that.

Also very helpful to understand selectivity is the purpose of antenna/RF alignment procedure. For all the material they provided, M&L assumed the reader already knew these two concepts. There is only one "antenna" trimmer (cap) on the tuner gang, and (if I still have all this right) one coil remains unaccounted for under the keys. That coil happens to be horizontally aligned with a red wire --leading directly to the antenna section of the tuner gang, so I suspect it is RF/Antenna related. Also, it's under the SW key, reacting well to wanding in the SW-full band. Perhaps the ferrite coil-sliding adjustment is why there are no other coils in this same horizontal plane (e.g. under the MW/LW keys). The MW alignment instructions for the Jubilee direct the tech to do exactly this in the last steps before adjusting the wave trap. Does not prove anything, but builds some support for putting it all together. Maybe.

Hope anyone else reading all of this gets as much out of it as I have. I'm off to do it!

[Chuck-II]

Well, that went rather well. Good as new!

There is no "perfect" dial alignment, but it's close enough for government work and the cumulative effects of all this stumbling are pretty amazing. From a dead radio sold as "décor" at the flea market to 3 stations on LW (Algeria, BBC-4 and one other at 1750 m) more than a dozen on MW, including TalkSport, BBC-5, something in the Czech Republic, football in Spain, etc. And, half of the world on SW: China is an awesome investment opportunity --in English-- and who knew the Japanese were Rod Stewart fans? I didn't hear Moneypenny whispering any numbers to 007, but maybe on another night...

All coils on the "oscillator" row (away from the piano keys and toward the rear of the set) moved the lower frequency tune points of each band-key it was located under exactly as anticipated/suggested. The antenna trimmer on the tuner brought things home for the 2 lower bands and the coil on the "antenna" (front) row peaked a 30 meter SW signal nicely. I left the (shellac?) fixed paper coils on the rods alone, as reception was pretty good, but believe the red wire side is for MW, and the green wire side is LW. They resonate on their selected band touching with a finger, just like a tuning wand, if there's any doubt.

Wherever I searched for a Telefunken 655w schematic, there were others looking for one too, so perhaps this info will be helpful, if translated into French.

So, "Well done, Gromit!" I'm very grateful to the board and contributors, and really quite pleased with the results, for both the radio and myself. Now, there's got to be a way to kill off all that RFI, especially over the WiFi...