Dave Richards AA7EE

September 19, 2012

Courtesy Of F5VJD – A Working G4JST/G3WPO DSB80 80M DSB Transceiver!

Filed under: Amateur Radio,Ham Radio,QRP,Uncategorized — AA7EE @ 5:03 pm
Tags: , , , , ,

As I’ve explained in earlier posts, my interest in the G4JST/G3WPO DSB80 rig goes back a long way – back to the mid 1980’s in England when I read the description of this little rig in Ham Radio Today and scratched together the money to buy the basic kit. An enclosure was also available but I didn’t have enough money for that as well, so built it into an aluminum enclosure I already owned. That little rig gave me much fun until one day in my apartment in Hollywood in the mid-1990’s when I accidentally connected the 12V power to it the wrong way round and it instantly stopped working. I didn’t have the patience, time or experience to get it working again, so (and it pains me to say this) I tossed it into the trash. There have been several times since then that I have regretted doing that, and since then, the inability to see my DSB80 only served to further memorialize that rig in my mind.

Recently I dug up the reprint of the article that came with the DSB80 kit (I did keep that) and tried to rebuild it Manhattan style – here and here.  The receiver worked as well as I remembered the original working, but I experienced a problem with the transmission of a residual carrier on transmit. After a few weak attempts to cure that issue, I shelved it in favor of other construction projects. It’s a weakness of mine – if something doesn’t work well the first time I switch it on, unless the fix is relatively simple, I don’t always pursue it. I didn’t consider it a failure, as the receiver worked well – and it was wonderful to once again listen to a direct conversion receiver that utilized a diode ring DBM . The TX seemed to work up to and including the final,  so there was really not that much to troubleshoot if I ever wanted to give it another go.

Then a very exciting comment was left by Richard F5VJD on the blog post about my DSB80 build. It read as follows -

“Hello Dave

I also built one of these in 1983 and I still have it… but like you I fried it during a fit of nostalgia.

I do still have the original instruction sheets and the original Ham Radio Today articles which I could copy and send you if you like… and as I can’t really see myself replacing all the fused components in the RX, I would also be happy to send you the radio if you promise to repair and use it. It has some nostalgia value for me because it was the first kit I ever built and I had my first ever QRP QSO using it, so it would be nice if it was back on the air again after all these years.

Let me know what you want to do and I’ll get things started my end.

73

Richard”

Well, how do you think I replied? This was a chance to see that little rig again – and to see how another ham had built it too.  Richard told me that he had used the original case suggested by Frank and Tony, so I was curious to see a more “original” version of the DSB80 than mine was.  Richard packed up the mighty little transceiver and even took a picture of him with it outside his home in Northern France to mark the beginning of the journey to it’s new home in California:

About a week later, it turned up on my doorstep in California:

Wow – talk about well-packed! The DSB80 was very well protected with much packing material, and Richard had written his address inside each layer while packing, just in case it became partially unwrapped in transit.  We needn’t have worried, as it made it’s way to me in one piece. It’s just as well that I’m patient, as it must have taken the best part of 10 minutes before I finally saw the object of my desire. What a great looking transceiver! As well as the several address labels in the package (including one taped to the rig itself), there was an envelope inside the enclosure, with a note from Richard and some extra J310’s and BC182’s. The BC182’s are near-equivalents to the BC238’s and BC239’s used in mic amp, AF pre-amp and RF driver.

I could barely wait to open it up and see that circuit board once again:

I could hardly contain myself.  It was so great to see that circuit board that I remembered so well from my youth.  Notice Richard’s ingenious use of Meccano to help support the mounting bracket for the tuning capacitor. I think that Meccano sets were called Erector sets in the US:

Here’s a view from the other end of the rig, showing the input/output bandpass filter, with the 2 inductors wound on T68-2 cores. Fixed to the back of the case, that board with the relay gives full break-in operation:

In case you’re not familiar with the architecture of this rig, or don’t remember my previous posts on the subject, it is a CW and DSB transceiver based around a Mini Circuits SBL-1 diode ring double balanced mixer package.  The SBL-1 was rated down to 100KHz so gives excellent out of band rejection at the design frequency of 3.5MHz. My earlier Manhattan version used an ADE-1 which is only rated down to 1MHZ, but there was not a trace of any out-of-band breakthrough.  Here’s a block diagram and device breakdown:

Block diagram and device breakdown of the G4JST/G3WPO DSB80 80M CW/DSB transceiver, details of which were first published in the March 1983 edition of the UK magazine “Ham Radio Today”.

This DSB80 of Richard’s had been inadvertently connected to 12V with reverse polarity, just as I had done with mine. The first order of the day was to remove the board from the case and, well, make it work again! Richard mentioned that one or two of the electrolytics had exploded, so I knew that at the very least, I’d need to change some of the electrolytics and one or more active devices.  On getting the board out of the case and onto the workbench, one of the big advantages of boards without plated-through holes became apparent – with the help of desoldering braid, it was very easy to remove old components and replace them with new ones. After replacing the big electrolytics that performed service as smoothing capacitors for the 12V supply (one of which had very obviously exploded) I realized that it would be very little trouble to go ahead and replace every single electrolytic. After all, the rig was about 25 years old, and those caps might be getting a bit dry by now.

After replacing all the electrolytic capacitors (even the small signal ones – including replacing the 1uF electrolytic in the diplexer with a polyester cap just for good measure), I connected wires for the volume pot, antenna and speaker, connected a 12V supply (the right way round this time!) turned the volume up and – as predicted – heard absolutely nothing. This was not a surprise as I had a feeling that amongst the devices well and truly cooked would have been the LM380.  This rig used the 14-pin version of the LM380 and I had planned in advance and purchased one. In it went (thank heavens for de-soldering braid!) On re-applying power, I heard hiss in the speaker, and on touching the input to the DBM, I heard atmospheric noise.  It couldn’t be this easy could it? Well yes, it could!  Further investigation revealed that the VFO and buffer were working and the receiver was fully operational.  Not only that, but the transmitter was working too – and without the issue of residual carrier emission that I had experienced with my Manhattan version. Woohoo!

This was the point at which I became a little uncertain what to do next. I still felt that it was not really my rig – it was Richard’s that he had sent to me.  I knew that he had given it to me, but still felt a reluctance to change it too much from the way in which he had built it, for fear I was being disrespectful to the history of this fabulous piece of home-brew. After all  it was the first transceiver he had ever built, with which he had conducted several hundred QRP QSO’s.  Richard put my mind at rest by saying that it had ceased being his rig once it left his house in that brown paper package. He was happy to know that it would continue having a life in California, and wanted me to do whatever I wanted in order to “make it mine”. What a helpful and wise sentiment. I wanted to retain something of the original rig other than (obviously) the circuit board. Richard suggested that perhaps I could fabricate an enclosure from copper-clad laminate and use the original top cover. What an excellent idea!

First of all, I wanted to modify it to cover the entire US band of 3500 – 4000KHz. The original version, being a UK design, covered the UK 80M band of 3500-3800KHz. I used a different tuning capacitor – a 365pF air-spaced variable with a built-in 8:1 reduction drive that I bought new from Midnight Science. I also re-wound the VFO toroid and changed a few of the other caps in the VFO circuit to achieve a coverage of 3485 – 4019KHz.  Even with the reduction ratio of 8:1, this still gives a tuning ratio of over 125KHz per turn of the tuning knob.  In order to help, I added a fine-tuning control consisting of a 1N4001 diode across the VFO tank circuit, acting as a varactor, and tuned with a 1K linear pot. The schematic of the modified VFO, with values, is as follows:

The other modification I needed to perform was to redesign the input/output bandpass filter in order to accomodate the wider US band.  Thank goodness for the software that comes with EMRFD – the program DTC.exe makes the business of designing a double-tuned filter quite straightforward.  The filter used in the original design of the DSB80 uses taps on the inductors for impedance matching into and out of the filter, while the circuit that the EMRFD software is based on uses a tap point at the connection of 2 series capacitors.  The filter I used keeps the output of the transmitter relatively constant over the 500KHz bandwidth of the US band:

Here’s what the board looked like with the mods finished.  The redesigned bandpass filter is at the left-hand side. Just to the right of it, the ferrite toroid wound with the green wire is the transformer that matches the low output impedance of the MOSFET final to the 50 ohms the bandpass filter needs to see. The VFO coil is on the far right of the board, just above the LM380.  The double-balanced mixer package (the heart of the rig) is the silver rectangular package just to the right of center:

Here you can see a closer view of the VFO toroid.  It’s not very easy to see, but very close to, and to the right of it is the 1N4001 diode that is used as a varactor to provide bandspread. The lead sticking out to the right is for the volume control and the lead leaving the board in the foreground is for the bandspread potentiometer:

Even though this particular board is 25 years old and has been worked over a few times by my soldering iron, I still think it’s pretty nice-looking:

I have only reproduced the schematics of the parts of the circuit that I altered. It didn’t seem appropriate to show the full schematic of the DSB80 here without permission. However, although the original article doesn’t seem to be available on the internet in good resolution, it is available from the right people if you know who to ask :-)

Now that the board was fully operational in both DSB and CW, the next task was to fashion a new enclosure.  I did want to retain some of the flavor of Richard’s version of this fine little rig and had wanted to keep the 2 small pieces of Meccano that he used in his tuning capacitor mounting bracket. After the decision was made to use an air-spaced variable instead of the polyvaricon that the original used, this didn’t seem possible.  I went with Richard’s suggestion to make a new enclosure from copper-clad laminate and use the original top-cover.

I cut 3 pieces of laminate for the bottom, back and front panels, and started marking the positions of the front-panel controls on a piece of paper.  When doing this, I mark the outline of the front panel life-size on a piece of paper and keep re-arranging the controls to find a suitable layout. This picture gives you the idea, though when this picture was taken, neither the paper layout nor the partially-drilled front panel were complete. To the right of the horizontal rectangular cutout for the frequency display are 4 holes that I drilled at what were to be the 4 corners of the vertical rectangular cutout for the modulation meter.  When making square or rectangular cut-outs, I drill a series of overlapping holes around the perimeter of the rectangle, remove the center piece, and use a small file to file the edges of the cutout smooth.  It’s slow work, but with care you can make some very acceptable cutouts:

Wherever possible, I mark the cutouts and hole-centers on the copper side of the board:

The finished front panel (before I realized I was going to have to enlarge the hole for the main tuning capacitor) -

The finished 3 panels with all cutouts and holes.  I think I might have cleaned and lacquered the boards at this point.  I scrub the boards clean with a Scotch-Brite pad, a little dish soap (washing up liquid for the Brits!) and plenty of elbow grease. This was the first time I decided to drill extra holes for those little “spigots” that help to keep switches and pots from rotating off-center -

 I use a hand punch to make holes wherever possible. It’s quicker and more convenient than drilling.  For larger holes, I start with a hole made with the punch, then enlarge it with a T-handled reaming tool. I got lucky and bought my punch from Harbor Freight for around $20.  They don’t carry this anymore, but similar punches are available online.  Mine came with a series of different-sized dies and looks like this:

I think it turned out pretty well:

Aw heck – I’m just dying to show you how the rig looks in it’s new case with it’s fancy digital display from N3ZI. The 2 knobs on the upper left hand side are the volume and bandspread respectively. Under the bandspread knob is a switch that will be used to switch between 2 different bandwidths when I fit a single-stage audio filter in between the AF pre-amp and the LM380 output stage. Under that switch is the headphone/speaker socket.  I did buy a nice speaker and grill cloth from Elecraft (a K2 replacement speaker) to fit in the top cover, but realized that drilling holes in the top cover might possibly damage the crinkle finish.  I didn’t want to risk that, so will probably live without an internal speaker.  I could always make a top cover from copper-clad laminate and drill holes in that for a speaker, but I like using the original top cover. Underneath the right-hand side of the frequency display is a toggle to switch the LCD backlighting on or off.  The meter is for monitoring modulation level in DSB transmit. The bank of switches on the far right are (top to bottom) TX/RX, DSB/CW and the on-off switch for the whole rig -

You’ll notice that I’ve used a hot glue gun to secure the VFO toroid to the board. This is also the first view of the 365pF air-spaced variable cap from Midnight Science. -

A few more views -

There are a few small issues that I’m working on. The TX does chirp a bit on CW transmit.  It didn’t do that on the bench, and I know that Richard’s version didn’t either. I’m pretty sure that a few more well-placed  RF bypass caps and/or feritte beads will cure the problem.  I think that I need to do a bit more work on how the board is grounded also.  Future projects will be to add a switchable bandwidth audio filter.  The later version of the DSB80, called the DSB2, had one, and I plan to use the same circuit.  It uses a single op-amp.  Richard’s version had a sidetone oscillator for CW and I’d definitely like to add one.  My antenna is rather inefficient on 80, so if I want much success on this band, CW will be the way to go.

Richard and I have kept up a regular e-mail correspondence in the last couple of months and in one e-mail, he told me how his DSB80 started it’s life -

A bit of history for you:  the rig first squawked it’s mighty 2 watts in anger at 00h08 27 January 1986; signals received by G4BMR Derek (599 about 40 miles away from my then QTH Swindon) but it wasn’t really the first proper QSO as I was already in QSO with Derek on 2M FM and had been for the previous 4 hours.  He was helping me to try and solve the hum, drift and chirp problems and we were also in QSO on 80M CW so all I did was just change rigs when I thought that I had the problems licked – it was quite a thrill to at last hear a nice clean CW signal coming back to me via 2M.  As soon as that had been achieved and we had swapped signal reports, Derek went QRT – not surprising after 4 hours!

I then finished boxing the little chap up but at 01h03 with the adrenalin still running high, there was only one thing for me to do and that was to call CQ.  I received an immediate response from OK3CSA Juro in Myto who gave me a 569 report;  for me, this has to be the first real QSO with the rig and also my first QRP QSO and my first QSO with something I had put together myself.  (Paper logs do still have their uses – the one in question is in front of me as I write – which is why I still to this day log this way!)

Here is that log entry of which Richard speaks – the QSO with Derek G4BMR and the first “proper” QSO, with OK3CSA -

Richard’s note continued -

Enjoy your time with your visitors and I look forward to reading the next instalment of the saga of ‘The little rig who found a new home in America…’

Well, the next installment begins here! The big success of this rig for me is that it has connected me to my ham radio past and filled in the gap on a long-lost piece of radio nostalgia.  It all happened because Richard G0BCT/F5VJD shared with me his first ever home-brew rig and allowed me to let loose on it with my soldering iron.

For that, I thank you OM!

Note – although the wiring on my version may look a bit neater, my version chirps and Richard’s didn’t.  He found that it was important to keep the various leads to the board well-separated, and I think I will end up bringing my connections out to the top of the board like he did in order to experiment with them to find a layout that works best.  He told me that his choice was between pretty yet defective, or scruffy and functional.  I am going to have to make the same decision the next time I take the top cover off this rig!

The intention when taking this photo was to set the DSB80 to 3985 – the North American QRP SSB calling frequency. It wasn’t until later I noticed that I had set it to 3895. These senior moments worry me!

PS – I liked Richard’s original version of this rig so much that I have kept the case, tuning mechanism and associated components intact in case I wish to recreate it one day.  I do like the ability to see where I am on the band with the LCD frequency display but in my opinion, his version of this DSB80 looked more like a “real home-brew rig”. I have not had a QSO with my version of the rig yet, but have spent many hours listening to it already. I love listening to this DC receiver! There is still a slight chirp for me to get rid of before I use it on CW and truth be told, I am more of a CW person than a phone person.  I was 100% phone in the earlier days of my ham career and maybe the pendulum will swing back in that direction one day but for some reason, after a working life in which I earned my living by talking (DJ’ing and voice-overs for radio and TV commercials), I now can’t think of very much to say on a microphone!  Even on CW, I tend to be more of a listener than “talker”, but I do love to listen. Besides – listening is the perfect thing to do while I’m building , and this little rig sounds great :-)

Stop Press – I fixed the chirp.  It was a wiring error on my part and a simple fix. Will explain in a future post. I also had my first QSO with the rig this evening (Oct 4th) on 75M DSB! More info to follow.

June 3, 2012

The DSB80 – A Direct Conversion DSB Transceiver for 80M By G4JST and G3WPO – First Stage Of Building

I was 19 years old in March 1983 when the UK magazine Ham Radio Today published the article “A Low Cost DSB/CW Transceiver for 80M”. Being short of cash and wanting to get on the air, I sent away for the kit and soon after, was surfing the phone portion of the UK 80M band on my new DSB rig.  I didn’t get too many QSO’s due probably, to my poor 80M dipole, although G3UMV who lived just a mile down the road heard my home-brew signals and came over to see where they were coming from.  The receiver seemed to work very well, and I spent many hours listening to the chatting between 3.6 and 3.8MHz (80M only goes as high as 3.8MHz in the UK). The whole thing was enclosed in an aluminum case and tuned with a Philmore vernier attached to a polyvaricon.  I don’t remember any drift, so it must have been stable enough for sideband, and it didn’t have any noticeable microphonics either.  As it was my first DC receiver, I didn’t even know that this type of circuit often suffered from microphonics, as this one didn’t have any to speak of.

That little rig made it with me across the Atlantic and met it’s end one day in my apartment just a block from Hollywood Blvd. In a passing wave of nostalgia for my earlier radio days, I hooked it up to 12V DC to see if it still worked. It would have, had I not connected the 12V the wrong way round, and if I’d had the foresight as a kid to provide it with reverse polarity protection. I still don’t know why I didn’t just put it aside so that at a later date I could have replaced the damaged active devices. Unfortunately, I tossed it into the dumpster of my apartment building. What a shame – and it had an SBL1-8 mixer too!

From time to time either when moving or thoroughly tidying my apartment, I come across the copy of the original article that came with the kit. Trouble is, whenever I looked specifically for it, I could never find it, and the only copy of the article I was able to find on the internet is of too low a resolution to be of much use. I’ve been wanting to recreate this rig for a while and recently, when the desire became too strong to ignore, decided that I was going to find that article even if it took a day or two of searching. It did, but I did.

Pure nostalgia wasn’t the main reason for my wanting to build this rig again. A big reason is that I have always been drawn to simple receiver topologies such as regens and direct conversion receivers, yet not all designs are created equal. I remembered this one as working well and on top of that, it used something in the circuit that you don’t see too often in DC receiver designs these days – a passive diode ring mixer package (NE602 anyone?)  I wanted to build a DSB rig that used a diode ring mixer package, so this is why I am here.

The schematics for this rig aren’t that easy to come across.  I eventually a found a low-res version of the article online after some searching but it’s not really good enough to work from.,Ham Radio Today is no longer being published, and the company that sold the kit back in the 80’s, G3WPO Communications,  went out of business a long time ago. On top of that, Tony Bailey G3WPO is no longer an active radio amateur. On this page on his website he gives a link to a reprint of an article about another of his projects, the Minisynth VFO. Judging from this, and what he says in the whole of that 3rd paragraph, I don’t think he’d mind my publishing the schematics for the DSB80 here on my blog.  That is what I am hoping as I’m pretty sure that some readers will want to see the schematics, and there doesn’t seem to be any other way to get them. I’m having a bit of trouble with the VFO (more on that later), so by showing you photos of my construction and the schematics, I’m hoping someone may be able to help me.

The plan is to get this working well as a receiver first, after which I’ll add the transmitter stages.  So to kick things off, here’s a block diagram of the receiver, a pretty standard diagram of a direct conversion receiver:

I built the VFO first of all. It seemed to work OK and be reasonably stable, with drift of less than 80Hz/hour after warm-up. I know that’s not stellar, but a bit of temperature compensation could help that.  Somewhere in between adding the buffer and adding the rest of the receiver, I noticed that the VFO was drifting a bit more and FM’ing, which makes SSB sound pretty bad. However, if I can lower the drift on the VFO and get it to stop FM’ing, I think I’ll have a nice-sounding direct conversion receiver on my hands. There are virtually no microphonics – you have to turn the volume way up and really be listening in order to hear them. For all practical purposes, microphony is just not a problem; something I like very much in a DC receiver.

I’m getting ahead of myself. Here’s the original circuit for the 80M VFO in the DSB80. I did leave out a trimcap, 1N4148 diode and associated components that were used to switch in a CW offset, as I won’t be using this rig for CW:

Important – please note that I experienced instability with the buffer transistor Q2 (it wasn’t doing a lot of buffering). I don’t understand why Q1 was coupled to Q2 with a 100 ohm resistor instead of a capacitor, but at the suggestion of K4AHO, I replaced it with a coupling capacitor (I used a 39pF NPO) and put a 100K resistor from the gate of Q2 to ground. My problems with the buffer cleared up and the receiver sounded really great. I’ll publish the schematic of the entire receiver section of this rig in a future post.

The oscillator is a Colpitt’s configuration and the 260pF variable was, in the original design, a polyvaricon.  I wanted to modify the VFO for varactor tuning (at what point did we stop calling them by the more descriptive term varicaps and start using the name that makes them sound like a prehistoric bird?) and also figured that a 78L08 regulator in place of the 8.2V zener diode could only help. This is what I came up with:

All the caps marked “poly” are polystyrene. I changed the values of the 2 x 1000pF caps to 1200pF simply because that’s what was available.

Projects always look pretty when I first start them, before I’ve had a chance to mess them up -

Here’s the VFO, although I have yet to add the varactor at this point – it will be located at the far right end of the board -

One more view just for the heck of it -

Somehow, by the time I got around to adding the mixer, AF preamp, AF amp and bandpass filter to complete the receiver, the whole thing started to look just a little bit messy. You’ll notice that I ditched the nylon mounting hardware for the VFO toroid in favor of a single nylon strap. I figured it would be one less material in contact with a frequency-determining part of the circuit that might cause drift. I didn’t clean up the board for it’s photo-op, as this is a work-in-progress that may not make it out of the emergency room -

VFO drift was steadily downwards after the initial warm-up and probably something that could be brought to within useable limits with some temperature compensation work.  There are almost no microphonics to speak of (always a good thing in a direct conversion receiver), and only a small amount of broadcast band break-through which is only coming through at the kind of high volumes that will rarely be used. This breakthrough is not coming from the DBM, but from the preamp, which is set for a gain of 1,000 (60dB). If this rig makes it to a later stage of building, I may reduce the gain of that preamp just a bit – we’ll see. The receiver sounds pretty good on 80M SSB with one big problem – the VFO FM’s when receiving signals, and that IS a problem.

The documentation that came with my kit for this rig in 1983 had something to say about  FM’ing of the VFO:

Hmmm….but I was using J310’s in this re-creation and was still getting FM’ing of the VFO.  As far as I can remember, it was not happening in the original version I had.

I decided to build the original VFO circuit with the zener diode regulation and with an air-spaced variable capacitor instead of a varactor to tune the circuit (the first schematic in this post and the second image down). I wanted to do this on a separate board, before connecting it to the rest of the receiver.  This is where it started. It sure was exciting looking at a bare board (blank canvas) with just an air-spaced variable capacitor. The variable capacitor was given to me by a friend and boy, what a great gift. Thank you – you know who you are. I was looking at this thing and thinking of all the possibilities – a signal generator, crystal set, or a regen perhaps?  Air-spaced variable capacitors are very inspiring to me -

Here’s the VFO circuit built – no buffer yet, and you’ll notice that I have not yet installed C2 and C3.  I wanted to see what the frequency coverage was without those capacitors first.  It was pretty wide, so I ended up installing C2 and C3 in the values suggested, and removing a few turns from L1 to achieve coverage of 3600 – 4000 KHz with one gang of the capacitor, which was about 330pF at full capacitance -

And with the buffer added, and temporarily terminated with a 51 ohm resistor for drift testing:

I noticed that on touching the output of the buffer with a small metal screwdriver, the frequency of the VFO changed by about 600Hz when terminated with the 51 ohm resistor. I wonder if this is the reason the VFO I built on the main board FM’s when receiving signals? The only difference between this one and the one that is incorporated into the rest of the circuit as pictured 5 images above, is that the one directly above is tuned by a variable capacitor, whereas the other one is tuned by a varactor. Either way, it suggests to me that I need another stage of buffering. Before I even look at the drift and figure out how to compensate for it, I need to tackle this issue.

To be continued……..(unless another project derails this and it ends up on the shelf, with the variable capacitor used for something else…..)

The Rubric Theme. Create a free website or blog at WordPress.com.

Follow

Get every new post delivered to your Inbox.

Join 1,103 other followers