A Daunting Project – at least for me!


Construction of the Classic 40 – a Direct Conversion

Receiver for the 7 MHz band.

Designed by Rick Campbell, KK7B.

Published in ‘QST’, August 1992.


000 DC a small


8th January 2016. As mentioned on the adjacent web-page, since Christmas we have been engaged on a 7 MHz CW receiver project. It started innocently enough, with the wish to knock up a simple Direct Conversion receiver. We had made a couple in the late 1980s, and they were great fun. I don’t think we made more than a couple of contacts using them, though; because if you operated even a 2 or 3 Watt CW transmitter near them, being very primitive, they rolled up into a small ball in the corner of the shack, gibbering.


The Direct Conversion receiver – also (very occasionally) known as the Homodyne – is a very attractive thing to many of us. The principle is very simple indeed. An aerial brings signals into our shack. Hundreds of them mixed together, actually! However, our aerials are usually connected to an Aerial Tuning Unit (Aerial Matching Unit is the more precise term), so many of the unwanted signals are greatly reduced & got rid of. We are currently interested in the 7 MHz Amateur band, so if we have our ATU tuned to that band, the stuff getting to the receiver will be mostly around that frequency, especially if we put in a tuned band-pass filter – more on that later.


All we do is to mix that incoming signal – say it’s 7.03 MHz – with another signal of nearly the same frequency, generated locally in our shack. Suppose we had set our shack generator (it’s called a Local Oscillator, or LO for short) to 7.0305 MHz. What will happen when it’s mixed with the incoming 7.03 MHz?


Well, when you mix two waves together, you get a whole series of products. The main ones are the sum of the two frequencies, and the difference between them. There are lots of other, weaker products, but that’s not important right now. Actually, it’s only the difference between our two frequencies is really of interest to us. What is it? It’s


7,030,500 minus 7,030,000


Answer: 500 Hz. Which is really great, because 500 Hz is an audible frequency; you can listen to the signal right away. 500 Hz is a whisker higher than B above middle C on the piano – but that’s not important right now, either. 8^)


What is important, is that there has been a Direct Conversion of the radio wave to audio. You can hear it straight away. Marvellous, eh?


Better still, because of the way our audio has been derived, it has a beautifully clear and pure sound. The sound has been produced by a classically simple and elegant process. That’s why a lot of people who are into audio like it so much.


Now we come to the inevitable downsides – for you can never have something for nothing.


Firstly, the audio signal we have obtained is very small indeed – probably just a few millionths of a Volt. It will need to be amplified an immense amount to listen to it on a loudspeaker, or even headphones. This degree of audio amplification is difficult to obtain without various problems: distortion, white noise (hiss) generated within the amplifying devices themselves, the interaction of spurious, undesired products &c.


Secondly, as you tune across our chosen band – which is 7.0 to 7.04 MHz – each station will come in at a high audio pitch, which will fall as your LO approaches the frequency the station is transmitting on. When your LO is exactly on the same frequency, you won’t hear anything. Because then, there isn’t any difference between the ham and your LO signal. But as you continue to tune, that station will reappear & gradually increase in pitch until it goes out of audio range. So every signal appears twice. To be sure, this is a problem, but it’s not too difficult to get used to; and having a good peak audio filter is a great help. (See the closely related page audiofilter.htm .)


The assorted ugly-construction boards at the top of this page have all been lashed up since Christmas as part of our tests & evaluations of circuits available on line. We’ve had great fun – plus a lot of frustration! – but eventually were able to receive some 7 MHz signals using mostly ramshackle stuff. Of course, we knew that a receiver will only work in an optimum fashion once it is rigidly installed in a metal box, and the LO properly screened &c., &c. But the performance of our lash-ups was so poor, it would certainly have remained dreadful no matter how nice a metal box we put it in!


A big, big re-think was necessary.


One circuit we had repeatedly come across during many hours of googling, was that of an advanced DC receiver designed by Rick Campbell, KK7B. It was published way back in the August 1992 ‘QST’ – the magazine of the ARRL, the national society of radio amateurs in the U.S.A. It looked quite complicated to us, so we sheered away from it – though not before borrowing the circuit of the simple tuned front-end bandpass filter for our project, whatever that was destined to be! But in time, we realised that if our project was to be worthwhile, we really had to bite the bullet & knock up something pretty good. The QST article is on-line as a .pdf, so we printed it out, read & studied it closely. You can do so yourself at:




It turned out that it wasn’t nearly as complicated as it looked from the circuit diagram. Which is NOT to minimise the years of work & study KK7B had put into it! Also, as the design was 25 years old, most of the components were quite normal analogue items that even I could understand. SO BE IT! We gradually bought in the components, taking care to find exactly the same ones as stipulated in the article. Happily, all were still available. Also, the circuit is laid out in logical modules, each of which could be prototyped and tested separately. A printed circuit board had been made available in 1992, and the resulting receiver was commendably small. But I have very big clumsy hands, so it was decided to employ ugly construction on rectangles of printed circuit board, just like at the top of this page; and though the receiver would come out quite large, I didn’t think that would much affect the performance. We shall see!


Rick Campbell includes no particulars for a VFO (LO) circuit, saying candidly that those have already been dealt with by consummate experts such as Les Hayward K7ZOI, Doug DeMaw W1FB, and Roy Lewallan W7EL; and that he had – as yet – not been able to exceed their attainments in that field. (At least, he hadn’t in 1992.) 8^)


Very well then. The final batch of condensers – sorry – capacitors, arrived this morning. At last the Quest could begin!

circuit part 1

Here is the circuit of the first prototype board, plus the input filter. I’m sure Rick and the ARRL won’t mind me reproducing it bit by bit; after all, anyone can download the whole .pdf at the link above. I’ve actually been a bit clever and incorporated the Input bandpass filter from elsewhere in the article. (Hee hee.) It’s self-explanatory really. The aerial – no, I shan’t call it an antenna, so there! – goes into a tapped coil wound on a toroid, and goes via this symmetrical filter into the Magic Device contained in the Red Box. This is, straight away, is the very heart of the receiver. The SBL-1 has been made for many decades by Mini-Circuits™ of New York, and quite rightly so. It will take in any two signals between 0 Hz and 500 MHz, and mix them perfectly. While being the most expensive component in the receiver, its cost is surprisingly modest. Mini-Circuits™ themselves are happy to supply them at $9.50 apiece, in quantities of 10. . The LO also goes into the SBL-1, as shown. (We’ll get to the LO later.)


000 DC b small


The first rough prototype board itself. We haven’t shown the Input bandpass filter here, but the signal from it (the Radio Frequency or RF) goes into pin 1 of the SBL-1, as indicated. Of course, we needed to test this first board, in order to prevent creeping insanity. If it didn’t work, then there would be little point in making all the following stages, which even if they were OK, would not do their job, because this first board hadn’t. So we stuck the lead at bottom right into a nearby standard domestic audio amplifier, just to see. Or rather, hear. And behold: Morse signals, though very faint, could distinctly be heard. Oh joy! It was a pile-up in late afternoon. We are evidently on the right track. Here’s a sample. It’s alive – it’s ALIVE!


Right-click and open in new tab: First Test .


On to the next part of the circuit – the CW filter.


1000 Hz elliptical filter final

Values are also given in the article for a 3000 Hz bandwidth version for reception

of SSB – whatever that may be… 8^)


000 DC c small final


This is a ‘7th-order Elliptical Low Pass Filter’. I have no idea what that really means, apart from it being a darn good low pass filter. That’s why I am a retired jazz musician who seldom got to play the sort of stuff he really wanted to, while KK7M was already in 1992, an academic at Michigan Technological University. Still, this low pass filter is a vital component of Rick’s design. You’ll notice of course that all the condens – capacitors we’re using on these prototype boards all have very long leads, so they look gawky. The reason is, of course, that these boards are merely prototypes. If we snipped off the leads to be neat and tidy, that may hamper us when we come to build the finished version, in which we will use the same caps. It’s not that these are the only capacitors we have; far from it. We bought them in 10 at a time to get them a bit cheaper, and that’s fine, because we can make other receivers later on with them. I suppose it’s because were old, and therefore somewhat parsimonious. Still, that doesn’t matter here; we’re ‘only’ dealing with audio frequencies, so three inches (7.5cm) of leads on a capacitor will have approximately zero effect on the performance. Of course, if we were dealing with UHF or something esoteric like that, I’m sure such long leads would be catastrophic! 8^)


Anyhow, it works great! There’s no audio sample, as 7MHz activity was at a low point when we finished it.


9th January 2016. We thought that today we would receive the LM387s we ordered a few days ago, along with the TIP29C & TIP30C transistors for the power amp. But the postman doesn’t come until about 1100z, so in the meantime attention was turned to our 7 MHz VFO. Yes, even I know you’re not supposed to have a VFO running as high as 7 MHz; but while googling for inspiration, I had found a great little circuit on the G-QRP-Club website, in a DC receiver called the ‘Direx’ - a datasheet of a constructional article by the Rev. George Dobbs G3RJV. Don’t think I’m name-dropping, because I am. (I first met George at 0009z on 1st January 1982 – we were each other’s first QSO on the 10.1 MHz band, which had been released for amateur use nine minutes before. He kindly enclosed a copy of SPRAT, the G-QRP-Club magazine, with his QSL card, and incredible vistas of amateur radio gradually unfolded for us.) 




This is a very simple VFO, which George says was ‘surprisingly stable’, and indeed it is. Plenty good enough to act as the VFO module in our prototype lash-ups. A 2N3904 was used instead of a BC109 – remember those? 8^)  I have the impression that the stability might be something to do with the fact that the base is biased by two equal resistors (10 KΩ) and the collector & emitter resistors are also equal (470 Ω). I don’t, offhand, recall seeing transistors suspended so symmetrically between the voltage rail & deck, except in unity-gain circuits. Anyway, it had been working fine, so we tidied it up, provided some varicap diode tuning voltage from a 10-turn pot, and slung it into a die-cast box.

direx VFO


We had to adjust it a bit; it seemed surprised, even embarrassed; for a single-transistor VFO to be given its very own Eddystone die-cast box must be a fairly rare event, but eventually it settled down. (We made the mistake of putting the lid on after we’d just soldered stuff in place. It took ages to cool down & get more or less stable!)


dc rx w vfo copy


Here it is: dear, sweet little VFO! By the way, the 25pF variable in the circuit above is actually a miniature air-spaced preset, rescued years ago from a piece of junk. It’s a beautiful little job, & we always knew it would come into its own eventually. You can see it next to the pot. Unfortunately, the postman did not bring us our LM387s. Grrh! So we made do with our existing LM386, which you see on the top right of the shot. There’s a contest on at the moment, and although it was late in the day, we picked up the following sample as the band closed. The DC rx is getting better, though we are far from the finishing point.


Right click & open in new tab: VFO in box with LM386.


UC2K is a Club station in Kaliningrad, the Russian exclave on the Baltic Sea – about 900 miles from my QTH. The hum is due to everything (except the VFO, which did have its lid on) just lying on the bench, as you see above. The hiss is partly due to the modest LM386 audio chip.


Onward: ever onward!


Sunday 10th January. A cool, calm & sunny morning. We should really go out for a walk, instead of brooding over the continuing absence of the LM387s. The LM387 is not, as we imagined, simply an improved, low-noise version of the LM386. Well yes it is that, but it also contains two amplifiers instead of one. Here’s the next section of the receiver:

LM387 circuit

The components shown dotted (C21/R15 and C25/R22) are optional, and will reduce residual hiss, if this proves a problem. The 2N5457 FET in between the two halves of the LM387 will mute the receiver, when grounded, while you are transmitting. It is our intention to build this receiver exactly as it appeared in 1992, but other chips have doubtless appeared in the meantime that are better than the LM387. There are things we’ve come across like the OPA2134, used in hi-fi audio applications; we have ourselves used the NE5532 in a phono. pre-amp (see phonopreamplifier.htm ).


There was, however, nothing to stop us carrying on with the construction of the above stage in the absence of an LM387. We had intended to use a socket anyway. Accordingly, after a nice morning walk in local park




grove park small


For a very long time, we have been meaning to try feeding ducks, not with bread, but peas. Though bread has been traditionally used to feed ducks for many centuries, it’s not really appropriate, as it’s largely carbohydrate; apparently that is not normally a major constituent of the diet of ducks. Peas were advised as better, on a website I happened to visit. So I finally remembered to take a small plastic bag containing frozen peas to a local park. At first, the ducks didn’t seem to be too bothered. Also, because it was only about 4°C, the peas were still mostly frozen. But after a while, they got the idea, as did the black headed gulls, not to mention the many Canada Geese – though there are none of those in the shot above. The peas ended up being eaten greedily. Then we took our little walk, hoping that the rest of the peas would thaw by the time we got back. This was not the case; and worse still, somebody had emptied two carrier bags full of bread into the pond in the meantime, so all the birds were glutted. The experiment will just have to be repeated at a future date.

End of Interlude.


On our return home, work began in earnest.


filter - LM387 annotated small


Here is the stage completed. We always hold down our boards with insulating tape; there’s nothing so frustrating as chasing a board across the bench while trying to solder in a tiny metal film resistor. To solder from another angle, you just pull the board up & stick it back down at the angle required. You’ll see that we’ve re-wired the audio filter ‘properly’, to make more room on the board; also, we’re too clumsy – and impatient – to fix in ICs properly, so have just soldered an 8-pin holder to a piece of Veroboard. 8 short lengths of bare wire were soldered to the strips, and those in turn fixed to two strips of PCB each with 3 saw-cuts made with a razor saw, which gives 4 insulated pads.


filter - LM387 1 small


The same board from another angle. Of course we won’t be able to test it until those pesky LM387s arrive!


The final part of the circuit is a power amplifier to drive a loudspeaker:


power amplifier


Wednesday 13th January 2016. Yesterday we reached the nadir of the project. By which I mean, the ruddy LM387s have still not arrived. So we checked, and found to our dismay they had been scheduled to arrive over a week ago. We have emailed the seller and asked them to check it out; also, we looked for another source of 387s. They really are pretty elusive these days. There are plenty on ebay, but the local ones are quite expensive. Then we found a cheap mail order source, believe it or not, in our own city, Birmingham! We’d missed it before, because they had used a generic image of a 14-pin IC, so we assumed it was something else. The supplier is, however, mail order only. We anxiously await their delivery.


So in the meantime, we carried on with the power amplifier. We managed to shoehorn it onto the same board, though we had no real reason to make it small. Just thought we’d have a go. It got quite cramped & fiddly towards the end, and expected it wouldn’t work, but it did – the power amp, that is. Audio signals of about 50 mV were injected & sounded fine from the loudspeaker. Whew!


final board small


Well, there we are, then; half-satisfied, but half-frustrated. Does the LM387 stage work? I suppose we might as well put it in the nice box we have waiting for it. It seems somehow wrong to drill & spray the box in our ‘house colours’ of red & black before the darn thing works, but what can one do?  >8^(


99 main rx in box small


99 LO & rx small

LO & mixer/audio section all present and rarin’ to go – except for an LM387.


Thursday 14th January 2016. The LM387s did not arrive, nor have I yet heard back from the company I first ordered them from. Though we hoped we might get the other 5 we ordered locally, they didn’t come either.


Friday 15th January 2016. There is now no doubt in my mind that a Definite Conspiracy exists, probably involving Satan, Goblins, Ghosts, Ghouls, Malignant Sprites, Komodo Dragons, Fruit-bats &c.: the whole ruddy shebang! Not only is there still no word about my original order for two LM387s, but my second order for five, made to the local supplier, has been sent back with an apologetic refund; they obviously haven’t got any. Now I’m not getting paranoid, please believe me; but I am very nearly certain that the few thousand remaining LM387s have been carefully gathered together from all over the world, and locked away in a Sealed Storage Facility, probably in the middle of the Gobi Desert, just so’s I can’t get one. Yes, that’s what it looks like to me. Definitely. Yes, siree. Very well then: my Vengeance will be unleashed, and they all jolly well better LOOK OUT, that’s all I can say! In the early afternoon, we found a place in London that offered LM387s at great expense. We phoned them, diplomatically establishing that they did actually have some (three actually), and so ordered one. I would have had all three, but couldn’t afford that amount of bread. However, the guy was affable & efficient, so we made the order.


Saturday 16th January 2016. Hurrah! The LM387 arrived. It just looked like any other 8-pin IC, but we danced around the room, cradling it, singing to it, placating & extolling it in a crazy ritual of fulfilment. But would the receiver work? Trembling, we carefully bent the pins so that it would fit the socket. (Why do you always have to do this? Why don’t they come ready to plug in?) Eventually, in it went. All was set up, but a profound silence issued from the loudspeaker. Oh dear. A fresh mug of tea was prepared, and the Crucial Question ‘WHY?’ was asked. Since the entire circuit apart from the SBL-1 is at audio frequency, we injected an AF signal to the final PA. That was OK. The signal was injected into the LM387. That was OK. So was the 1000 Hz filter. So was the initial board; injecting the AF signal at the output of the SBL-1 worked fine, too. So why weren’t there any CW signals? After another mug of tea, we checked the frequency of the LO, and all became clear. Because it was a lash-up VFO, it had no buffer; and when we boxed it & set it up to tune 7.00 to 7.04 MHz, it wasn’t connected to anything. On connecting it to pin 8 of the SBL-1, it was pulled it down about 150 KHz to around 6.85 MHz. And there were no signals there at the time – hence the silence. The VFO was tweaked up to begin at ~7 MHz, and Lo! there were the CW signals, as clear as a bell. Whoopee! We were physically & mentally exhausted by this time, but came back later and recorded this little extract from a contest that was going on. Oddly, it doesn’t sound like a busy contest, though it was, because you can only hear about three stations; on an ordinary DC receiver, you can hear eight or ten; this, of course, is the Ultimate Tribute to Rick Campbell’s superb design.


Click here for example.


This, at last, concludes our Christmas Project. If anyone is still reading this guff, I commend you for your persistence, and thank you for your patience. I think I might even make a YouTube video of it, using some of the stills. But now for a day or two off. Then, we shall have to use this receiver to make some contacts…





Page finished 16th January 2016.



























Page in progress, 10th January 2016.