That being so, what does a Contemporary Hombrew Comms Rx Look like???

[Bazz4CQJ]

There have been some very interesting threads going on in recent times discussing (among other things) the evolution of Comms RX since WWII, and Valve Questions and FET Questions. It seems to me that it would be extremely useful to try bring these topics together and see if there is a consensus on what a 'homebrew comms receiver' would consist of now, within certain limitations.

The forum is focused on Vintage Radio, and I think it is true that the majority of members are amateurs/hobbyists rather than professional engineers. So what I have in mind here is probably something that will be using components that were around in let's say 1980? It won't have any software in it and it won't need a VNA and deep understanding of Smith diagrams to set it up! As a stake in the ground, I guess what I have in mind is a home-built receiver that would exceed or at least meet every parameter of an a AR88, except that it should weigh a great deal less. Is that too ambitious?

Of course, in times gone by, RadCom published design classics like the G2DAF; I have not kept up with RadCom or the late SWM to know if there has been anything since ~1980 which would fullfil this concept. But in any case, it would be interesting to draw on the recent discussions on the forum.

Who'd like to go first; are there any valves in there?

B

[Radio Wrangler]

1980 parts?

Well the SD5000/Si8901 were around so the H-mode mixer was possible. Dynamic range close to that league was achievable from diode rings, but with rather fancy drive (write up came out in HP journal April 1982) Quadrature hybrids existed long before then and image-rejecting mixers. THe idea of a quadrature pair of IF filters was known.

I'd been experimenting with ladder filters using crystals on overtones to make narrow (ssb) roofing filters at 60 odd MHz. At work we had superb filters for SSB channels done at lower frequencies in LC pot core technology.....

You see where this is going? The IC7800 minus the DSP!

I'd built a motor controlled preselector (driven by a look up table from a frequency counter o an LO of a receiver)

We didn't have inverted-mesa VHF fundamental crystals, we didn't have DSP. We could get very good performance with all-semicondutor technology, though there probably would be a valve... a CRT for display.

Could these things have been home brewed with about the same amount of dedication G2DAF showed... Certainly!

David

[dave walsh]

This is interesting. I haven't seen anyone else but myself initiate the G2DAF theme but that's probaly untrue. Generally speaking I don't have much more technical undersatanding than I did at 14 when his design came out. I'd forgotten, as was quickly pointed out, that you even had to build your own filters in the initial version. Not fully understanding the technicality doesn't mean you can't see the significance though [I think] so it's great to see Bazz initiating this discussion.

A modern version would be very sophisticated though and I can see that. The question is, who would be listening? On the other hand, the major Broadcasters can junk their transmitters as outdated and save a fortune but [as in the early days] amateurs are free to experiment/transmit more cheaply-so far! I saw a comment on here today that 2 metres is just a lot of silence. Well CB didn't help. I'm not sure that vintage radio is just
destined to be a nerdy retro hobby. It may well have a revival eventually as something alternative.. Cell Phones and the Internet are great but not very independent,. If the system goes down they are lost [ unlike SW Comms] but maybe I've got that wrong too?l?

Dave W

[G6Tanuki]

I'd be thinking of something along the lines of a modernised G4CLF IF-strip (Plessey SL600-series chips, 9MHz filters, double-balanced diode-quad mixer) and then stealing a few ideas from the likes of the Clansman PRC320 for the local-oscillator generation.

Though with an optical-shaft-encoder (available from a computer printer of the era) for ease-of-tuning rather than the PRC320's "Safecracker" decade-dials.

A Kbyte or two of battery-backed CMOS RAM would give you, say, 100 memories. Scanning would be reasonably easy to implement too - I did that a few decades ago when converting an old Pye 290 radiotelephone to scan the entire 2-metre band in 12.5KHz increments, using CMOS 4017 chips and a 2716 EPROM as address-generator/decoder.

Some address-decoders (7447 TTL) and 7-segment LEDs (or perhaps even using TTL 74141 and Nixies?) to actually display the frequency: Job Done, using junk from the 1980s.

Remember that even in the very early 1980s we had access tyo cheap FPGAs and ULAs so no doubt the synth/memory/scanning stuff could be made a lot simpler than using boards of TTL and CMOS.

Could be simpler still if we were allowed to use something like an Arduino!

[Biggles]

I built a dual conversion multimode IF strip from Radcom circa 1980's but never got round to building a decent front end and VFO for it. That uses the Plessey series of chips and is pretty straightforward. Cost me a fortune to build as well as there are three seperate filters in it. Every few years I look at it and try to come up with ideas but it seems to be a lengthy task so back in the box it goes....
Alan.

[Bazz4CQJ]

Mmmmm... I think I may have 'mis-underestimated' the sophistication of the the state-of-the-art components available in 1980 that Wrangler and Tanuki are so familiar with. I do wonder how easy it would be to find those components today. Additionally, I'm not sure just how many of the hobbyist group would enthuse about the ideas put up thus far? I'd guess that we'd have to come back down the performance - v- complexity curve to pick up more enthusiasts, but perhaps it's just me who's Chicken . What a pity I didn't manage to acquire a few of those Plessey chips when I worked for them in 1980!

[Radio Wrangler]

The HP mixer diodes we used are still available, though now in SOT23 packages so that's easy. The microeave bipolar transistors that drove the LO are now mundane. BFR93 will do the job.

The SD5000 for the original H-mode mixer is now mounted in an SMT package and called SD5400 but they might have changed the name but we all know it's the same beastie.

Crystals used for their overtones in 60MHz region.... available. surplus ones are interesting.

Quadrature hybrids... bit of enamelled wire, a small ferrite toroid and a couple of capacitors... easy!

High dynamic range bipolars... still very easy.

Quite within the reach of an interested amateur feeling a bit adventurous, but awfully easy to categorise as 'too hard, I won't bother'

So it's not actually difficult, it's not a lot of effort, but few will do it. Those that wnat to, can.

It's quite good fun leading someone into terrifying lands peopled with head-shrinkers, dragons and the ravenous bugblatter beast of traal, and watching them discover that they can handle it after all, and that there are interesting things to do.

David

[G6Tanuki]

Birkett's used to have a range of "out-of-military-spec" SL600-series chips available at bargain prices. Truth is, it would be possible to replace the SL600-type balanced mixers with off-the-shelf cheap parts like the NE602 at a significant cost-reduction.

The MD108-style diode-quad mixers are still available; 2N5109 transistors, run with 25mA of standing current from an 18V supply [don't forget the heatsink!] are great for broadband stages to lift a VFO/synth's output to the level needed to drive diode mixers.

If I was building a "next-generation G2DAF-type" comms receiver I'd consider the biggest cost/problem would be in sourcing a pair of decent USB/LSB filters whose performance [shape-factor, ultimate sideband rejection, ability not to go non-linear in the presence of strong signals, consistent group-delay] would not degrade the performance of the other components. Homebrew "Ladder" xtal filters are OK, but hard to produce repeatably/cheaply without access to good test-gear.

It's also worth pondering the issue of frequency-stability: military/professional HF-types expect to be able to dial-up a receiver to a given frequency in the absence of a signal and for the signal - when it appears - to be ~clean~ without needing to retune or fiddle with 'clarifiers' to resolve it. Amateurs are more interested in short-term frequency-stabilty [how long does a typical QSO last?] and don't fret about long-term stability/resetability.

The upshot of which is that an amateur HF comms RX can *perhaps* make do with a free-running VFO and xtal-controlled calibrator, whereas a professional RX needs a reliable, resettable master-oscillator [accurate to a few parts-per-million up to 30MHz] which mandates some kind of synthesizer.

[turretslug]

We (as in on the forum) have at least two distinct advantages- we can peruse developments over the decades and chose what was genuinely good (as opposed to trendy or expedient at the time) and we're not tied to particular suppliers for commercial reasons, patents, politics or "NIH".

Sadly, I can't think of water-tight excuses for involving thermionics anywhere (other than VFD!) but I'll watch the thread with interest.

[G6Tanuki]

Originally Posted by turretslug

Sadly, I can't think of water-tight excuses for involving thermionics anywhere (other than VFD!) but I'll watch the thread with interest.

I did suggest Nixies for frequency-display. OK they're cold-cathode, but still free-electron devices.

[mole42uk]

Well, I'd go for a no-frills manually tuned rock solid highly sensitive receiver.

My biggest gripe with most receivers is their inability to keep the LO frequency stable, my second gripe is that many require a huge aerial which sometimes just isn't feasible.

I'd use valves too, but only because I understand thermionic emission much better than I do semiconductors!

Richard

[kellys_eye]

I'm not so sure we should be looking for the ultimate-anything since the greatest delight in amateur radio was in achieving the best possible results with what you had - no matter how good or bad it was.

Making a receiver that is the be-all and end-all takes all the work out of a challenge (all other things being equal).

My recent return to amateur radio (old c/s was G4GQB) has had me searching for a receiver design that just works, isn't overly complicated, would be a challenge to use to it's full extent, isn't impossible to source parts for and, ideally, recreates some of the more well-known and respected designs (G2DAF is a good example).

I've gone from 4-tube CW receiver (hard to source the Miller coils for) to a cut-down Eddystone EA12, to an all-germanium superhet (using Denco coils), to a 4-tube CW transceiver, an Elektor 40m CW receiver (BF900) and the YU1LM SDR series! And I'm STILL not settled on what I want!

All of the above have their own challenges, pro's and con's but I think nostalgia will prevail and I'll (eventually) settle for a tube design.

[Bazz4CQJ]

Something that has gone without comment is the original use of the AR88 as a baseline. I think I'm working with the belief that a receiver with AR88 capability, together with a modern SSB detector, and a mass reduction of about 80% , would satisfy the requirements of many/most amateurs as a receiver. The fact is that there are still lots of fairly standard AR88's, HRO's, Hammurland's etc which are good enough to mean that they are in practical use rather than just being items of nostalgic value.

I certainly think that we can draw a broad line between a professional RX and and an amateur one.

As for exploring " terrifying lands peopled with head-shrinkers, dragons and the ravenous bugblatter beast of traal", that man left Yorkshire and went North whereas I left Yorkshire and came South; n'uff said I think

[G6Tanuki]

Re post #13: remember that the AR88 was designed to fulfil a WWII professional/military specification - so the original conjecture having specified "AR88-equivalent" performance my design-comments have been based on extrapolating this forward to fulfil the same niche but using 1980s technology and receiving the transmissions [essentially SSB] predominant in the same time-frame, and with the same [inflation-adjusted] budget as the AR88's designers were given to work within.

Consider: in the late-1940s the production-cost of an AR88 would have been equivalent to something like the price of a new 10HP Austin/Morris car of the era - nigh-on a year's wages for a 'white collar' office-clerk. That such receivers were sold off at spectacularly-discounted war-surplus rates only shows how lucky the hams of the time were.

[Bazz4CQJ]

For amateur use, I hadn't intended to "inflate" the AR88 specification, but simple provide what we used call a "mid-life update". It's interesting that for a hobbyist like myself, we probably just have no idea or concept about what the spec. of a professional receiver of 2014 looks like or how much it costs. Come to think of it, I've never looked at a tabulated comparison of how an AR88 would look against whatever amateur stuff comes from the Far East these days. But just noting kellys-eye post, using the modern equipment (Kenwood, Yaesu, etc etc) just bores the pants off me.

[Radio Wrangler]

They still let me back in for visits.

I grew up with an AR88 atop a hill above Huddersfield during a glorious sunspot maximum long before most QRM generators were invented.

I look back on the days with the AR88 and smile. They were wonderful. Nothing since has come close. They don't make reminiscences like they used to.

I've used:

Icom R9000,
IC-765 and some of the newer stuff.
Racal RA117,
RA1217,
RA-1792
FT107
FT1000D
Eddystone EA12
Marconi H2900 (very rare and an absolute beast)
HRO MX
R390
S-line
AR8516L
and numerous homebrew receivers.

This is a range encompassing some of the sublime and some of the ridiculous. They weren't all mine. None of them captured the magic of the AR88. It was probably a mixture of my youth and the marvellous conditions. But for me, it can't be equalled. What's more, the amateurs on the bands were talking about interesting things, not just hernia operations and contests.

OK, does the AR88 have something going for it other than my personal circumstances at the time. Most definitely yes.

The designers of the oriental big three thought that once you'd got through the main crystal filters no care was necessary. They foisted off on us nasty distorting IF amps, demods and audio stages. They sounded awful. Try an AR88 or an HRO and they sound so much better. Easier on the ear, less fatiguing.

I'm not pro vintage stuff, I'm not pro modern stuff. I'm happy to jump around erratically in time, picking up things which are excellent and dodging the rubbish. I don't see a point in preserving something just because it's old. If being old justified preservation, then being older establishes a precedent, so we ought to tear down all listed buildings and reinstate the even earlier original forest.

I'd preserve things because they were good, At a pinch I'd preserve some bad things just to show people how awful they were. Man's past includes medicines from rhino horns and tigers, stenode reception, square steering wheels, 'fast bass'. We have a lot of things best forgotten, or perhaps better used as examples to aid avoidance.

On the other hand we have Murphy Baffles, HROs, original minis, Nagras, the end of apartheid, Dudley Harwood's speakers, Timothy Taylor's 'Landlord' We have things to be proud of, things to go in the trophy cupboard, things to enjoy.

I was going to say I have a perfectly fine motor car but it's back in the garage again, but some of the time I do have one... and yet, for enjoyment, I choose to be on a horse.

Does that qualify as a liking of old methods? Or is the important factor not the age of the technology? Motoring is boring, too easy. But I can't say the same of modern radios. So many of them actually aren't much good at what they are supposed to do. And as G6Tanuki says, they're boring as well.

David

[Synchrodyne]

From a different perspective:

Picking up on the AR88-as-baseline theme, bearing in mind that the goal is not a home-made RA1772, and at the same time looking at keeping things as simple as possible, a question that comes to mind is whether AR88 performance could be matched using the same topology with small-signal solid-state devices. That is, single-conversion, 455 kHz IF, two RF stages and say three IF stages with distributed selectivity. Use of dual-gate mosfets for the RFs, mixer and IFs comes to mind, possibly though with push-pull jfets or mosfets for the mixer. Large signal handling capability might be a possible stumbling block, but then in TV and FM front ends, mosfets were able to match or better valve performance, so the same may be true for HF applications. But with solid state circuitry come some advantages, such as the relatively easy application of wideband RF agc separate from the main agc system, and also the use of PIN diode attenuators as part of that agc system. A good solid-state oscillator would I think if anything be more stable than a valve oscillator.

Good SSB performance probably points to the use of lumped selectivity at the front of the IF amplifier chain, and if done for SSB, then it would be logical to do it for all modes. And once done, it opens up the possibility of using bipolar ICs as IF amplifiers. Maybe a chain of CA3028s or similar, but possibly a consumer IC such as the MC1349 (which has been used in professional HF equipment). Naturally a good product demodulator would be included, for example based upon an MC1496 or similar. But if one were after the simplicity of minimum IC types, then maybe both halves of a CA3026 for the product demodulator, with half-CA3026s used for each IF stage. I imagine that stock IFTs and filters would have been available for this kind of circuit; a four-gang tuning capacitor might have been a stretch by the 1970s; ganged two-gang units might have been needed.

Much better frequency readout precision than available from the AR88 and its ilk would no doubt be high on the desiderata list, so a digital readout could be used; I recall that Ambit were offering consumer level assemblies in the late 1970s (Catalogue #3, I think), so that should not be too difficult or complex.

Of course, there were HF receivers in the 1970s and beyond that looked a bit like this, and the Eddystone 1000 series (analogue frequency readout) and 1570/1590 (digital readout) come to mind. These I think were aimed mostly at professional applications which required professional construction but moderate overall performance (at least as compared with the top professional receivers). For example they had only three-gang front ends. The 1590 though, had a pair of quite tight asymmetric filters for USB and LSB, suggesting that the concept had credible SSB performance. How these receivers performed in practice I don’t have a clue. But any limitations that arose from the price points they were designed to meet rather than the boundaries of their topology would not necessarily apply to home construction.

If replicating the AR88 topology in solid-state is not going to work for one reason or another, then the next question is how far up the hierarchy of topologies would be it be necessary to go. Perhaps dual-conversion with a crystal-controlled 1st oscillator. If so, a la Collins with a tunable 1st IF, or following the Drake pre-mixer concept? Would the latter point in the direction of something like a “souped-up” SPR-4?

Or might it be necessary to go right up to a synthesized oscillator, upconversion and a high-level mixer to obtain the desired performance level?

Cheers,

Originally Posted by Radio Wrangler

long before most QRM generators were invented

Ahhhh... bliss.

My 'best' radio is a Perseus SDR (software defined radio) brilliant performer with all the bells etc. you could want, totaly impersonal though. You still can't beat a mechanically tuned superhet for the sheer feeling of radio. All my engineering background disappears when I have a large, well weighted, tuning knob to twiddle. (nearly missed out tuning)

[Bazz4CQJ]

Originally Posted by Synchrodyne

Picking up on the AR88-as-baseline theme, bearing in mind that the goal is not a home-made RA1772, and at the same time looking at keeping things as simple as possible, a question that comes to mind is whether AR88 performance could be matched using the same topology with small-signal solid-state devices. That is, single-conversion, 455 kHz IF, two RF stages and say three IF stages with distributed selectivity. Use of dual-gate mosfets for the RFs, mixer and IFs comes to mind, possibly though with push-pull jfets or mosfets for the mixer.

Yes, this is getting closer to where my own thoughts lie, though I had wondered if there might be proposals to use valves at the front, but none so far. There are still a lot of 455kHz components around if you look for them, including the Toko and Mutata multi-pole filters. Those components may be far, far from state-of-the-art, but may well be very adequate in making a "1980's AR88". Just to add a little inconsistency to the plot, I think that we give digital frequency meters a concession and let them in. The ones available in 1980 were quite expensive, but you can get a good one now to bolt or your AR88/HRO for ~£20, so that's a must-have .

[Radio Wrangler]

Hmm designing an AR88 for the 21st century....

There is a 3-pole tuneable filter design to use a 3-gang capacitor. It uses bottom-L coupling and the centre inductor is implemented as two in parallel (except at the bottoms) to split current. It sometimes gets referred to as a 'Cohn' filter but Seymour Cohn was instrumental in so many bits of filter technology, hanging his name on one is probably wrong. He was interested not in topologies but in how to tune the things for minimum loss given a required selectivity.

I've built a few of these and they work extremely well. Bandswitching is needed to cover the range.

Valves with their high impedances are easy to tap into tuned circuits. If semiconductors are used and there is the need to keep within hailing distance of 50 Ohms then a properly terminated filter design is needed. So the bottom coupled job is a good candidate which can rival the AR88 front end for sharpness. Varactors are right out, mechanical capacitors are needed.

I got some stepper motors which had been in one-armed bandits. (did you know the electronic ones decided everything in software, the spun the reels to match. Even the final judder as they stop in a detent is faked and programmed) These motors are brilliant. Microstepping gives ultra fine resolution. A look-up table after a calibration session can hold calibration positions for interpolation. You could count the LO freq and move the preselector to track. True one knob tuning. *Very* AR88.

RF amplifier, look up Noiseless Feedback and David Norton's patents. Jacob Makhinson did some for QEX in the early nineties.

Mixer. H-mode. It's just right for the job.

IF: There are superb variable gain RF/IF amps from Analog Devices (James, G4CLF works for them now, no longer Plessey... as does Barrie Gilbert) Distributed filtering is still possible

A really great multipole crystal filter can still be realised as a distributed structure with pairs of poles between amplifier stages, a la G2DAF.

Switching bandwidth would be a nightmare, but the IFs would be small. Build 5 full IFs for 5 different bandwidths and switch the whole lot. Relay switches for linearity.

Another H mode for product detector. Wind your own toroid transformers and they're cheap.

Nice beefy audio, low distortion and plenty of power.

The local oscillator could be a free running LC with either bandswitching, or else a switched binary divider series. Use a reclaimed BC221 drive and capacitor. THe counter feeds the microcomputer which operates the preselector.

A mixture of old and new. The 455 kHz IF single conversion structure implemented with very high performance modules. True single-knob tuning.

Two of the three pole preselectors could be combined to make an amazingly selective front end.

Does that hit any of the right targets?

David