Miles J talking about triodes

[radionortheast]

couldn’t believe I actually found this again, video were turned the triodes from the energy 106 transmitter into two lamps, make good ear rings/lamps, the part when he said he’d broadcast to belfast on 95.8fm sounded like it was direct but the coolest thing. Wonder if that was different spelling of capital fm, i did find a station with a simliar name i was amused by the spelling.

When he starts making the jingles for intelligence transfer squencer it is not something to be missed. It was upgraded the triode in the energy transmitter to 6 or 8kw, often run at half power was clever they lasted longer, (never though they did this at big stations too) the ones in kiss, 1.5 ones were run at full pelt, so it could really take off achieve full thunderpants operation. Its interesting he said the last one was taken away he managed to get it back, do wonder if energy had become to powerful 800kw erp, sounded like something happened prior to the big raid in may 2005, maybe someone official had taken something the valve away that was fundamental running of it?

I always wondered about the energy exciter, weather it was a special type, it always sounded clear even when it was weak, the modulation index was kept low, sounds like it was all to do with the processing not the actual transmitter at all, what modulation index is I don't know. I did wonder if it might turn up on ebay, if its out there some were, maybe at some community station.

[radionortheast]

The valves contain same stuff as in a magnetron you wouldn’t go sawing them apart. Miles is wiring them up to transformers making them into lamps, they are simliar to light bulbs, likely rated to work for so many hours.
Nice to look at, electromagnetic waves emitted in the visible light wasted energy though, looks like theres noway to get high power without the warm glow of a nice valve.

https://en.wikipedia.org/wiki/Triode

[Albert H]

The exciters Miles used were the NRG Pro II and Pro III. He had tried others, including the Broadcast Warehouse offerings, and most of them were wildly unstable with all the RF flying about! I pointed out to him that the NRG exciter oscillators were at half frequency, so wouldn't suffer from RF feedback problems. Stephen provided Miles with a couple of them, and they were installed successfully - completely stable, with nice, clean, low noise carriers.

The processing they used on the audio was pretty aggressive, but kept them at a consistently high modulation level, without over-deviation. It always sounded LOUD, bt in common with most highly processed audio, it becomes tiring to listen to. I remember some discussion about whether the high amount of compression could actually be losing them listeners.

All things considered, Miles probably ran the most powerful FM pirates ever - anywhere! I can remember hearing Energy on the west coast of Scotland on a pretty average car radio, in fully noise-quietening stereo for the entire journey! The stations could also be heard elsewhere (as you know), and my brother could hear one of them in Brittany - effectively off the back of the antenna array!

It was a remarkable piece of engineering!

[radionortheast]

I might of had a pro lll, (the pro II looks to have trimmers I never had) shame this stuff is long gone, along with the radio stations too, I miss energy everyday even though it was over 20 years ago since I last heard it. Years ago remember reading Lawrence John had been on riveria, some station in the south of france or Italy I don’t know.
Albert is there no small box that would do audio compression, I tried ‘really good audio compressor’ not so long ago, it sounded not very good at all, didn’t seem to doing much either. Is a behringer any good? sounds like someone should not be trying to get the sound of energy 106.

Maybe abit over kill to get something to a nearby radio, I do miss it, I suppose the choice would be the behringer, software, would be better to make your own, I can only deal with through hole components here. I do feel like i'm the only person on these forums to of liked energy, I think it was not underground enough I don't know.

[Albert H]

The Behringer stuff is cheap as chips and very cheaply made. It works reasonably well, but is aimed at recording studio music production uses - they're not really suited for broadcasting.

If you want a basic audio limiter for the last point in your airchain before you go into your stereo coder or link rig, the essential specifications are that it provides an absolute brickwall to levels, without overshoot. You need fast attack and slow decay, and - ideally - you want split-band processing. The advantage of band splitting is that it prevents heavy bass sounds from causing the mid and top to "dip". The disadvantage - of course - is increased complexity.

Many stations back in the 80s used a simple limiter circuit, based on the Philips Databook / "Compandor Cookbook" for their NE570 series ICs, The simplest circuit used just two ICs - the NE570 (or 571) and an LM339 comparator. The '570 has two identical channels, so is stereo capable, and has rectifier circuits, voltage-controlled attenuator circuits and Op-amps inside a 16-pin package.

The basic circuitworks astonishingly well, and prevented many stations from overmodulation. Broadcast Warehouse sold a cheap 'n' cheerful version of the circuit as a kit, but cocked it up by trying to add a clipper (which led to gross distortion!).

PM me, and I can give you the details of a simple-to-build stereo limiter using this approach, and of a more complicated (but better sounding) split-band limiter based on the same circuit. The basic version uses just two ICs, 18 resistors, one transistor, 8 small electrolytic capacitors, two ceramic capacitors, and a piece of Vero stripboard to build it on. The level detection is done for each channel individually, and the highest of the two levels is used to control the voltage-controlled attenuators. The basic version really works well, considering its simplicity.

[radionortheast]

Hi, I will do that, it will have to completely numbty stuff, I don’t mind if it users modern components, don’t want to be having match fets whatever it was. I guess the rather complex circuit limiter you were mentioning reduces the bass frequencies individually, seems fm has a problem producing them, easy to get the bass to make an fm signal sound really naff, or receivers have got more narrow suspect they had a hard time producing them, suppose a super wide fm receiver would produce perfect bass but be no use to anyone.

[Albert H]

I've done extensive work on getting audio limiters to "sound" right. It's very easy to make a complete mess of things - even with the most sophisticated multi-band limiters. If they're wrongly adjusted, they may stop overmodulation, but will sound horrible.

The designs I favour make use of PWM level control - this eliminates the need to match active components channel-to-channel, and gives the least distortion of any kind of active dynamic attenuator.

Part of the trick of effective processing is to avoid too much compression. This is for two reasons: Compression tends to amplify low level sounds unnaturally, increasing the noise content, and too much compression eliminates the dynamic range of the material you're playing - a single piccolo will sound as loud as the whole band!

That said, some gentle, slow-acting compression, with low-level gating to cut off the background noises during "silent" periods, can be very beneficial. It allows the average level of your station to be roughly the same irrespective of the level of the source material. This is especially useful if your stations take programmes from several sources - level errors between sources are gently ironed out, giving the main fast limiter a fairly consistent level to work on. The slow-acting compression (called "AGC" in my designs) only has very a low compression ratio - typically 2 : 1.

The limiter has to have a very high ratio (at least 20 : 1), very fast attack, slower decay, low distortion, accurate channel-to-channel tracking, accurate threshold setting, and - ideally - at least two bands to prevent the high energy, low frequency sounds from "dipping" the mid-range and treble sounds. Done right, limiting is unobtrusive, done wrong, it sounds horrible!

Unfortunately, basic limiting cannot be perfect, even using the split-band approach. Really fast transients can get past the limiter, because even the fastest practical attack time will be too slow to catch them. There are two options to prevent transient overmodulation at this point - either eliminating the attack time (which I'll explain shortly) or Clipping.

Fortunately, really brief periods of clipping are virtually inaudible, so adjusting the levels into and out from your clipper can minimise its (adverse) effect. You can also use "dirtortion cancelling" clippers which do reduce distortion somewhat, but not perfectly. The Orban "Optimod™" tries to minimise the distortion it (necessarily) introduces by using lots of very small audio bands, and having a separate clipper for each. It (sort of) works but I can always hear its operation! I really don't like clippers, but sometimes they're the only really viable solution.

The other way of eliminating transient issues is to introduce delay into the audio. The delay required is pretty short, because all you're trying to do is to eliminate the attack time of your limiter(s). The old way to achieve this was to use a whole heap of "all pass" filters, each intoducing phase delays over a small band of frequencies. Pye made a limiter in the 70s that used around 40 pot-core inductors and a whole pile of very high accuracy capacitors to make their delay line. The results from these limiters were superb - for mono - but pairing them for stereo was always problematic, and they really couldn't handle the whole audio bandwidth needed for FM stereo.

The principle of delay-line limiting is that the level-sensing is done at the input to the delay line, and the attenuation is done at the end of the delay, giving the sidechain time to react to transients and pull the gain down - to effectively "see the peaks coming"!

Back in the 80s, I designed and built a delay-line limiter for microphone channels. It had vanishingly small distortion, an inaudibly short delay time, no overshoot whatsoever, and very low noise. The drawback with the circuit was that it used around 60 high quality op-amps and plenty of expensive, high accuracy capacitors and 1% resistors!

My next delay circuit made use of the Panasonic (or Reticon) "bucket-brigade" analogue delay ICs. These worked really well, but added a fair amount of complexity and expense to the limiter. The noise and distortion specifications of these ICs weren't spectacular, but careful design could mitigate the worst of the nasty artefacts. Unfortunately, these ICs are virtually unobtainable these days (they haven't been manufactured since the 1990s), and the few that remain available are rather poorer quality than the ones we used all those years ago. The modern reproductions of them are pretty poor, too.

The modern approach - of course - is to use a digital delay, but if you're going to the expense and distortion of conversion from analogue to digital and back again, you may as well do the whole thing in the digital domain! Modern processors usually go for this approach.....

Audio processing is one of the "black arts", and even the very best designers have difficulty in getting the idealised, unobtrusive processing that we all want!