Marconiphone 855 Silent on 49m

[cathoderay57]

Hello folks,
I've been working on a Marconiphone 855 using Trader Sheet 736. It was almost untouched although the electrolytics had been replaced a long time ago. Having mostly recapped it and realigned it, the radio works very nicely on all bands except 49m. It gets progressively more deaf towards the LF end of 41m and receives nothing at all, not even noise, on 49m. Ditto injecting a modulated signal from a signal generator. The tuning capacitor vanes are not shorting. I have changed the oscillator coupling capacitors C6 and C11, and the SW padder C7. The oscillator anode voltage measured very high (232v versus 178v on the Trader Sheet). I reduced this somewhat (to 195v) by changing R3 the anode feed resistor from 15k to 27k. The oscillator anode current on MW, no signal, is about 3.6mA. Now here's the thing - the frequency changer is an X63 which tests good on the VCM163 but replacing it with a 6K8G brings up 49m loud and clear. I also have a 6A8 which is a closer equivalent to X63 but I haven't tried it yet. Now, I have had trouble with X63 frequency changers in EMI sets before, notably a Marconiphone 851. On that set (which I no longer have), it was unstable on 16m band and also deaf on 49m. By juggling about with various components a sort of compromise could be reached but in the end I just ran it using a 6K8G. Anybody had similar issues and found a more elegant fix?
Cheers, Jerry

[Mike. Watterson]

Frequency changers are tricky to test on ANY valve tester. I've considered making a dedicated tester with plug in coils. Then measurements for conversion gain and leakage of LO out of RF in.
I've a 2V McMichael and it DOES work on the highest SW band with a DK91, octal adaptor and dropper resistor for 2V to 1.4V. I've two of the "proper" valves and they test fine and work on other bands, however very likely they don't have enough gain for the design at shortest band. No oscillation.

Unusual for the problem to be on 49m. It would seem something in the circuit, not the valve is wrong. I'd one set that switched increasing values of resistors in series with Osc grid at lower bands. Some part is out of wack and it's coincidence a different type of valve works.

[cathoderay57]

Thanks Mike. I should have said there is also some instability at the HF end of LW but I wasn't bothered by that as it works fine on 198kHz R4. I know failure to oscillate on LW sometimes indicates a weak frequency changer. I like to listen to 48/49m because there are sometimes pirate music stations popping up at weekend's, otherwise I would just live with it. I'll dig out some spare X63 and 6A8 and see if I get the same problem with more than one valve. To be honest apart from the oscillator coil there isn't much more to fiddle with. There is a loop inside the osc coil that you can bend to adjust LF alignment but as the frequency is spot on I didn't touch that. R3 grid stopper measures OK. I haven't checked R1 the Control Grid resistor yet. Cheers for now, Jerry

[cathoderay57]

Well, I tried another X63 and a 6A8G with same fault. I then tried a (probably new) X63GT and it worked on 49m for a few minutes then went silent again. The only valve that works reliably on 49m is the 6K8G. Comparing the voltages and currents between the X63 and 6K8G gave the following results (no aerial - Trader Sheet predicted values in brackets):
X63 Mixer Anode 262v (243v); Mixer Screen 78v (75v)
X63 Osc Anode 124v (178v) (non-oscillating; at 41m with valve oscillating Va=166v)
X63 Osc Anode current 4.8ma (4mA) (non-oscillating; at 41m with valve oscillating Ia=3.9mA)

6K8G Mixer Anode 265v (243v); Mixer Screen 82v (75v)
6K8G Osc Anode 118v (178v)
6K8G Osc Anode current 5ma (4mA)

Strangely the oscillator anode voltage has dropped dramatically from when I first measured it (see first post). Next I'll keep the X63GT in place and reinstate the osc anode feed resistor back from 27k to 15k.

[ms660]

Make sure the osc is on the correct side of the signal frequency, usually the high side in most cases.

Lawrence.

[G6Tanuki]

A most odd fault. One thing worth checking is the capacitor that decouples the oscillator HT-feed resistor to earth; I have in the past seen cases where this capacitor misbehaves and reduces the osc. feedback level which can cause some very bizarre effects like apparent 'gain' varying as you tune across a band, or in one case the oscillator would work if you switched the receiver on when it was tuned to the HF end and would then continue to work as you tuned it LF, but the oscillator would not start if you switched on with the receiver already tuned to the LF end.

That one had me foxed for a day!

[cathoderay57]

Thanks for the inputs. I had another go this afternoon. This radio looks nice externally but is a bit of a pig to work on. There is an access panel underneath the cabinet but you can't access many of the components. When you get the chassis out the underneath of the frequency changer is inside a screened coil box. To get the cover off you have to release the push-button box from the main chassis. When you remove the cover of the coil box and do the alignment, the RF alignment shifts when you refit the screened cover. Not one of EMI's best IMHO. Anyhow to cut a long story short I fitted 2 RF bypass capacitors (0.1uF) one in parallel with the (new) 4.7uF HT decoupler for the anode feed (thanks G6Tanuki) and the other on the screen pin of the frequency converter. Then tweaking the bendable coil turn inside the oscillator coil and ditto the aerial coil achieved stability and some reception on 49m. I'll check reception when it gets dark. If it works I will quit while I'm ahead! Cheers, Jerry

[cathoderay57]

Well, tried it tonight and guess what - no signals on 49m until I put the 6K8G back in. There must be a fault somewhere, as I can't believe it was originally sold with this sort of performance. The design isn't great for stability as the last oscillator anode feed resistor is on the main chassis and there is a wire about 4 inches long going to the wavechange switch and thence the oscillator coils. The only thing that is significantly changed from the original layout is that all 4 HT smoothers were in a waxed box on top of the chassis, long since removed by a previous reparer. Presumably they were all earthed at the same point, whereas now there are several different chassis connections for the individual can electrolytics. I might have a play around with that but otherwise I'm stumped. Cheers, Jerry

[Hybrid tellies]

There is not an awful lot on the 49m band these days. I use a sensitive portable communications receiver and have noticed that quite a few times during the evenings there are only a few weaker signals to be heard. Just at certain times of the evening you can hear the stronger signals from China Radio International and Radio Romania International but these broadcasts only last for 30-60 minutes. So it could be that you are listening during quieter moments.
What type of aerial are you using for 49m band reception?

[cathoderay57]

Hi Simon, normally I use a 25ft longwire in the garden but I also have a homebrew active loop in the loft. I know what you mean as the reception of actual signals, even background noise level, seems very variable on 49m. However, I know that the oscillator was stopping as I used the sig genny to test it. Also swapping the X63 for a 6K8G immediately brought up both noise and actual signals. I have a feeling that G6Tanuki is close to the source of this fault so I will play around disconnecting one or more of the 4uF (now 4.7uF) decoupling capacitors. I reckon the designers might have experienced difficulties "taming" this set on shortwave because the oscillator supply is taken directly from the rectifier cathode and then smoothed with 2 dropper resistors and 2 separate 4uF decoupling capacitors. The remainder of the HT supplies come from the remote end of the field coil. Cheers, Jerry

[G6Tanuki]

Do you have another radio [preferably a 'communications' type with a BFO] that you can tune to the 49M-band plus/minus your receiver's intermediate-frequency?

If so, tune it to the 49M-band-plus-your-radio's-IF, with a few feet of wire plugged into its antenna-socket.

Then tune your problematic radio across the 49M band. If your radio's local-oscillator is actually oscillating you should at some point hear a whistle/screech [if it's a communications-receiver-with-the-BFO-working], or if it's an ordinary broadcast-radio the problem-radio's local-oscillator will produce a dip in the background-noise.

[turretslug]

Taking the oscillator supply from the rectifier cathode may simply have been to ensure the minimum possible change of supply voltage with changes of HT voltage with AGC action- field-coil smoothing is often of both high resistance and high inductance and a signal subject to fading may result in series modulation of the oscillator supply as AGC changes the current drawn by the signal valves. That can cause an unpleasant fluttery effect as the oscillator changes frequency slightly, the IF amp output decreases, AGC decreases, and an LF oscillation loop is set up.

Has the radio been unused for a long time?- the SW oscillator former may be a type that absorbs water vapour, decreasing Q and aggravating the effect of unfavourable L/C ratio at the LF end of the band. It maybe that the drying-out effect of continued use helps, though the former or any impregnating wax, varnish etc. may be permanently degraded.

[cathoderay57]

Hi G6Tanuki, I tried your test using my Panasonic DR49. I set the DR49 frequency to 6465kHz and tuned the Marconiphone 855 across the 49m band. Not a peep of sound. I'm now certain the oscillator stops running just south of the 41m band; it's like turning a switch as you tune down in frequency, the background noise suddenly disappears and the set goes totally silent. Sometimes if I hover over the break point there is a squawking noise. Tuning back up in frequency normal operation resumes around 41m.
In answer to Turretslug, the radio definitely was unused probably for many years. The speaker field coil was open circuit which is maybe why it was abandoned but I managed to find a replacement. Inside it was very dusty and no evidence of repairs since the electrolytics were replaced maybe 20 years ago judging by the types used. You may be right about the Q of the oscillator coil - it would certainly explain the symptom. I've been struggling to compare the characteristics of an X63 with those of a 6K8G (which still works on all bands). I'm aware the 6K8G is a triode hexode whereas the X63 is a heptode. However, the pin-outs are compatible because the X63 grid1 serves as oscillator grid and grid2 as oscillator anode. According to A.M. Ball the 6K8G peak oscillator voltage is 7.5 whereas the X63 is 25v. The X63 also has a much higher grid-anode capacitance. Other than that the parameters are pretty close. My knowledge of valve technology is reasonable but I couldn't explain why one type works in this case whereas the other doesn't. Answers on a postcard? Unless anyone comes up with new ideas I will try relocating the final oscillator anode load resistor into the coil box; that would dramatically shorten the length of wire between the non-decoupled end of the anode feed resistor and the oscillator coils. Cheers, Jerry

[ms660]

I would double check the oscillator frequency with reference to the signal frequency when it's working just in case.

Lawrence.

[cathoderay57]

Hi Lawrence, OK I set my signal generator to 6MHz and tuned in the Marconi 855 to the tone modulated signal. I turned off the sig genny and switched on my Panasonic DR49 (which has a digital freq read-out) and tuned around for the heterodyne whistle from the other radio. Picked it up at 6.465MHz which means the Marconi 855 osc is running 465kHz above sig freq. I think that's correct. Cheers, Jerry

[cathoderay57]

And for the viewers at home here's a picture of the radio. By the way nothing I have done can make it work properly with an X63 frequency changer to I've left the 6K8G in!

[ms660]

If the osc is running at 6.465MHz when the receiver is tuned to 6MHz then the osc must be working but that implies different to what was said in post 13 unless I've got mixed up.

I'd be checking it with a 'scope to see what's what.

Lawrence.

[cathoderay57]

Hi Lawrence. In post 13 the Marconi 855 was, I'm certain, not oscillating and was therefore receiving nothing anywhere on the 49m band, and so it did not pick up any signals from the DR49 oscillator. I had already done a systematic test anyway with my signal generator injecting tone modulated signals at 41m and then slowly decreasing the Sig gen frequency and following the signal on the 855 by adjusting the tuning knob. As it approached 49m everything went silent. That was done with an X63 frequency changer fitted whereas in post 15, I had a 6K8G fitted. I think that was fair game as merely swapping the 2 valves could not have shifted the alignmen by 2x465kHz. I do have a basic scope but no RF probes so I have never had any success connecting it to old radio oscillator circuits as the act of connecting the scope leads (presumably capacitive effects) prevents a meaningful measurement. Cheers, Jerry

[cathoderay57]

However, you're right - in post 13 I should have set the DR49 to 5.535 MHz to radiate an osc signal at 6MHz

[ms660]

Re post 18, ideally you need a min. 30MHz 'scope and an X10 probe, the capacitance of the probe will be small compared to the value of a fully engaged tuning capacitor.

The reason I originally mentioned the high side osc check is that sometimes it's possible to tune to the low side by mistake on shortwave if the trimmer has sufficient range, which can mean that tuning down the bottom end of the band the L/C ratio can become unfavourable and sometimes the circuit won't oscillate.

Lawrence.