AM 'Pantry' Transmitters

[Kalee20]

Doesn't applying modulation in series with the oscillator also produce a lot of FMing?

That's what I found, using a Colpitts oscillator based around an NPN transistor.

The 15 transistor jobbie that Yorkie kindly referred to above (which can be trimmed down to 9 transistors if you dump the overmodulation indicator and the negative-feedback circuitry), used a dedicated stabilised supply to the oscillator, for good frequency stability. Mind, careful attention to the oscillator could well give decent immunity to frequency varying with supply voltage. Alternatively, use a crystal rather than an LC tuned circuit (if you can get one of the right frequency). Then you can hopefully amplitude modulate the oscillator to your hearts content.

The frequency varies with supply voltage because the active device's characteristics, junction capacitances etc, change with voltage, and this naturally has an effect on frequency. Keeping the active device loosely coupled to the tuned circuit would help. Gut feeling also is that valves or FETs would be better than bipolar transistors.

FM on the carrier manifests itself as an ill-defined tuning point - depending on the selectivity of the receiver. It's because having the receiver off-tune allows it to slope-detect, so the audio actually gets louder. Naturally, there can be distortion too, with the two lots of modulation hitting the detector, giving either partial aiding or cancelling.

[Radio Fixer]

The friend with the SSTRAN said the same thing about FM being produced but as long as we found a quite part of the MW band it still sounded as good apart from slightly less top (not a lot). In blind testing I wouldn't have known which was which.

The radio being used was an HMV 801 which was a quality item from 1937 and here is a report on it.

"So far as the reproduction is concerned we have no hesitation in declaring the 801 to be easily the finest self-contained radio-gramophone that has ever been in the London Office of THE GRAMOPHONE”

"If manufacturers were to reduce the number of H.F. stages, compatible with reasonable selectivity, and concentrate a little more on the design of the L.F. amplifier, the pick-up and the loud-speaker, and thus provide ' de-luxe' quality also, then not only would the radio enthusiast enjoy his home and foreign transmissions more, but the gramophone lover, whose tastes have hitherto been somewhat neglected, would take a real interest in electrical reproduction. Distinguish the de-luxe class by quality and not by quantity of valves or of noise."

We use a radio tuned to the Mate most mornings and it sounds OK even on ones with quite narrow BW down to 5kc's say. I will have a better listen on sets with switchable BW (10 or 5 kc's). It isnt HI-Fi maybe but of course neither are the radios. The audio BW of most isn't great, rolling off at around 4 kc's even for quality sets.

It will be interesting to do a comparison with the Mini Mod when I get them built.

Gary

[Ian - g4jqt]

Hi Garry,

Being quite basic, the MiniMod audio drops to half at about 50 Hz and 12.5 kHz, but is almost flat between 120 Hz and 5 kHz, and the upper roll off is probably due to the bandwidth of the antenna tuned circuit.

Ideally there should be some pre-emphasis at about 3 to 4 kHz to compensate for IF bandwidth restrictions, and anything above about 7 kHz is normally well outside the IF bandwidth of a normal domestic vintage radio. But even some cheap MP3 players have useful tone correction that can be selected.

Of course 'proper' AM broadcasters must restrict their transmitted bandwidth to not more than about +/- 4.5 kHz so as not to spread excessively onto adjacent 9 kHz channels, although in practice during daylight you wouldn't have two stations nearer than about three channels away in any given service area. And after talking to the Engineer in Charge at Drioitwich (so that was a long time ago!) she told me that AM music stations are allowed a bit more bandwidth than speech stations. All and they have some pre-emphasis just before the higher modulation frequencies are cut.

Then there's the whole complex area of audio compression to make it sound 'louder' and overcome noise, particularly at the edges of a service area. But the whole point of our pantry transmitters is to have a very good signal with maximum audio bandwidth. Luckily that's quite easy to do...

Regards,

Ian

[Radio Fixer]

I'm not sure where the Mate rolls off, its certainly flat to 6 kHz and I have used it for measuring receiver bandwidths with a sig gen connected.

Spent an hour at lunch time listening to Classic FM via it on an HMV 650. Really couldnt tell any difference to the quality with the radio on narrow or wide band, just a drop in treble as expected. So we listened to it on wideband and it sounds remarkably good for a 2 valve 40 year old design. Better than any actual MW station in my opinion. Its a strong signal with the radio in the same room which no doubt helps.

With the modulation input switched off then what I call intermod hum isn't really. Its certainly not 50Hz or 100Hz rather like a gritty noise. Its inaudible sitting in an armchair, at decent listening volume, and 12 feet away even in the quite passages.

Be interesting when I get a Mini Mod to see how they compare.

In my previous post I said it had been playing 24/7 for 5 years but I checked up and its about 8: how time flies !

Should anyone fancy building a Mate then if e-mailed I can give some advice on construction to achieve low or no hum.

Gary

[Ian - g4jqt]

With the modulation input switched off then what I call intermod hum isn't really. Its certainly not 50Hz or 100Hz rather like a gritty noise. Its inaudible sitting in an armchair, at decent listening volume, and 12 feet away even in the quite passages.

I had the same problem with a mains powered mini tx on MW. It was modulation hum. It came and went. Turned out to depend on what was plugged into the mains that had a semiconductor somewhere. Dimmer switch, electric blanket, heating pump - neighbours' stuff too. When some or a certain combination of these came on, the TX signal had hum.

I may be wrong, but I believe it's RF getting into the mains and getting roughly modulated with 50 Hz in the power supply. I significantly reduced it by careful decoupling all diodes in the power supply of both tx and radios.

It's much less of a problem with battery-operated gear, both tx and rx, but still possible on rare occasions.

Ian

[Radio Fixer]

Thanks Ian.

I too decoupled the diodes (used semis rather than a valve) and went to quite a lot of trouble with the construction. Used a fully shielded mains Xformer, I had it any way and an IEC mains socket with RF components. Then screened wires on all audio leads and so on. I could even build it better now but probably never will.

For what it is, its excellent and for value for money beats an SSTRAN hands down. I looked up my notes from when that listening test was made and the notes say the "... SS had more presence perhaps a little more treble". But maybe it was the pre-emphasis you wrote about.

The friend with the SS was an audio buff having Thorens, Willi's and huge Tanoy speakers + an EMG Horn gramophone (Boy! one of those set up right, and his was, sound fantastic on good 78's). So even if my ears are not that discerning his were (sadly now he's gone).

cheers Gary

[boiss]

Tut tut all these 3 legged fuse designs. A nice PL36 with the heater volts doubled for HT, a crystal (plenty of 1MHz Xtals around/300metres on MW band), half a dozen components and Bob's your uncle, a stable couple of mW on MW plus eye appeal. (Mod around 30% and sound is good)
I have found that a few smaller caps across the main decoupling electrolitic helps a lot along with 10nF caps across the diodes or bridge connections. However there is so much rubbish coming from computers, games, TVs, modems and telephone and power lines that it is is very difficult to have a completely clean band in AM. Here there are still MW and LW transmissions but actual location of the radio is very important in improving S/N ratios in fairly quiet bits.
Mike