I have attempted to build 2 DSB transceivers now with limited success – a Manhattan version of G3WPO and G4JST’s DSB80 – here and here (the original kit version which Richard F5VJDD sent me for reclamation, worked fine) and the ZL2BMI rig, here and here. Both of them worked FB up to and including the TX driver stages but as soon as I added the PA, I had constant feedback/oscillation, even when not modulating the TX. In retrospect, I think a simple partition to separate the driver and final from the earlier stages of the TX would have done the trick in both cases (or even building the driver and PA on the other side of a double-sided board.)
The kit version of the DSB80 that Richard F5VJD very generously sent me was a fantastic piece of nostalgia (I owned one as a young man) and a very satisfying project, but I still wanted to be able to build at least one DSB transceiver from scratch and have it be fully operational.
Enter Joel KB6QVI from stage left. Joel is an avid homebrewer of QRP rigs – both from kits (he’s currently working on a BitX using the original board from India, which he is putting on 40M) and from scratch, Manhattan style. Joel is a fan and big user of the MePADS and MeSQUARES from QRPMe (as am I) and has constructed several QRP rigs using them. Joel and I communicate on Twitter, on which he was singing the praises of the VK3YE Micro 40 that he built. I think he was trying to get me interested in building something again, and his enthusiasm couldn’t help but pique my interest. I’ve made a number of jokes in the past aimed squarely at that trusty favorite of many a QRP homebrewer – the LM386. I usually end up using it with a 10uF cap between pins 1 and 8, which gives lots of gain but also quite a lot of noise. Joel told me several times about the configuration of the LM386 AF amp that Peter uses in this little rig which still gives enough gain to easily drive a speaker, but has much lower noise than the typical high gain configurations of this chip. Then one of my other non-ham projects came to a temporary pause and I got to looking at this enclosure which I originally made for the second beta run of NT7S’ CC-series transceivers. That beta run ended up using a much smaller board and a smaller custom case, so this blue enclosure has been sitting on the shelf for the last 2 years, just waiting for something to be built in it –
Joel got me to thinking that a little DSB rig in this case sure would be neat, so I rummaged in the parts drawers and fitted the controls and connectors I’d be using if I were to build the Micro 40. I kept telling myself that, as I was trying hard not to commit myself at this point :-) Note the little electret condenser mic insert in the middle. I thought an internal mic would make it easier to use, especially if out in the field. Also note that even though the enclosure is 2 years old, the coat of lacquer I applied has kept the copper looking pretty good –
The trouble is, on seeing a neat little case like this with a few controls and connectors installed, it’s hard not get enthusiastic about actually building it. Notice the small hole drilled above the right-hand pot for the locator lug. I used to break these little spigots off until an incident with my Fort Tuthill 80, in which the volume pot came loose and twisted round. I don’t know exactly what happened, but one of the potentiometer terminals contacted something else, causing a blue LED that was being used as a voltage regulator to blow. From that point on, I started using the locator lugs to help keep the pots in the same position –
At this point of course, I was committed, and set about building what I hoped would be the first DSB rig I’d build from scratch that would actually work. I have made a few changes to VK3YE’s schematic, and will describe them here. I hope you don’t mind that instead of using the conventional symbols for the 2 chips, I have represented them as rectangular blocks. It makes it a bit harder to figure out what’s going on with the circuit, but easier to visualize the physical layout when building –
I do have one problem with this rig. In fact, it is the only issue I have with my version, and that is a loud feedback howl from the speaker on going from TX to RX. I am thinking that Peter’s method of directly keying the mic amp with the PTT button would switch the mic amp off a fraction of a second before the relay kicked in and switched the LM386 RX AF amp on, thereby avoiding the feedback perhaps? This loud howl, which you can hear in one of the recordings linked to at the end of this post, was the only thing I wanted to cure. Everything else about the rig is great. Note – see point 10) at the end of this list.
2) I changed the value of the cap that couples the output of the mic amp to pin 1 of the NE602 from 1uF to 0.1uF (100n). On-air reports indicated that my audio was a bit bassy. Admittedly, I was using a microphone that was designed for recording and broadcasting applications, and was way overkill for this use, but I figured it wouldn’t hurt to gently roll off some of the lower frequencies in the TX, regardless of what mic was used. It did help, but I’m now thinking that the value of that 1uF cap in the base lead of the mic amp could stand to be reduced also. Feel free to experiment :-)
3) In Peter’s version, the cap that couples the collector of the BD139 final to the output network is a 47nF. I didn’t have any of those. I could have put two 100nF caps in series but figured that a single 100nF would work just as well.
4) Peter bypasses the wiper of his tuning pot to ground with a 47nF cap. I used 100nF. No biggie. Perhaps I should have used a 10nF instead……..
5) Peter bypasses pin 5 of the NE602 to ground with a 47nF cap. I used 100nF. He couples pin 5 of his NE602 to the top of the AF gain pot track with a 220nF cap, while I used 100nF. My substitutions are based on what I have in my parts box, rather than any meaningful analysis of the circuit :-)
6) For tuning, Peter uses 2 banks of diodes, each consisting of four 1N4002’s with a switch to achieve the frequency coverage in his rig. With the switch in circuit, both banks of diodes are used, and with the switch out of circuit, just the one bank of 4 diodes are connected between the resonator and ground. He also has a 10uH inductor in series with the ceramic resonator. My 7.2MHz resonator was obtained from hamshop.cz and seems to have very desirable properties. With no series inductor, and just one 1N4004 diode (I didn’t have any 1N4002’s so I used what I had), I achieved coverage of 7207 – 7335KHz. Placing a 3.3pF cap across the diode (shown as Cx in the schematic) changed the coverage to 7183 – 7295 – almost all of the phone portion of the US 40M band. What luck! Both Jason NT7S and Joel KB6QVI did tests with 7.2MHz resonators from hamshop.cz and achieved very similar coverage. They don’t always have these resonators so my advice would be buy a small stash of them when you see them in stock. These things are like gold! Not all ceramic resonators are created equal – others have different amounts of coverage.
The key advice with ceramic resonators in rigs like this is to experiment in order to get the coverage you want. However, if you are in the US, with the band going up to 7.3MHz, and you have one of those resonators from hamshop.cz, this circuit should give you excellent coverage. Other resonators will most likely give very different results, and you may need to experiment with different diodes, different numbers of diodes in parallel, and perhaps a series inductor (which I believe has the effect of extending the bottom end of the frequency swing.)
7) Pin 1 of Peter’s LM386 is connected to ground via a 47uF cap and a 33 ohm resistor. I didn’t have a 47uF, but I did have a 33uF. Given the wide tolerances of electrolytics, it probably doesn’t matter much but I substituted a 33uF cap and a 47 ohm resistor. There is an interesting article in SPRAT 116 on page 4 that talks about the use of the feedback resistor and capacitor between pins 1 and 5, as well as the use of an RLC network between pin 1 and ground to create a high gain amp that has a peak at 500Hz for CW reception. With a resistor as low as 3.3 ohms, gains of 74dB and even higher were achieved. This configuration doesn’t use an inductor, or such a low value resistor, but still has plenty of gain without resort to the the more common method of connecting a 10uF cap between pins 1 and 8 – a method that has (in my opinion) done a great deal to give the 386 it’s reputation for high hiss. It does have a lot of hiss when used this way, so don’t do that – use this circuit instead. It is far more pleasing to listen to!
8) I added a 1N4148 diode from pin 8 of the LM386 to ground as detailed in SPRAT 155 page 26. This is designed to help with squeal on going from RX to TX. It did seem to help a bit, but my bigger issue was the squeal in going from TX to RX. Feel free to leave it out, or put it in. Whatever you’d like to do!
9) I really liked the receiver and was surprised at how good it sounded, considering the simplicity. However, at certain times of day, I did experience a small amount of low level breakthrough from AM broadcast stations in the 550-1700KHz band. Joel KB6QVI didn’t have this with his Micro 40 but then, he lives in a less built-up area, about 12 miles outside Medford, Oregon. I am in the city of Oakland, in the San Francisco Bay Area, and close to many AM broadcasters. This breakthrough didn’t actually stop me from copying any ham stations but it was there and as such, was mighty annoying. Then I noticed that while the problem occurred when I connected the Micro 40 directly to my outside antenna, it disappeared when my ATU was inline. A quick look at the schematic of the ATU revealed that it was a high pass filter (as many ATU’s are). Aha – problem solved! I installed a simple high pass filter permanently in the receive antenna lead and the breakthrough completely disappeared. The receiver now sounds great.
If you don’t live close to many powerful AM broadcasters, or you are planning to use this rig only out in the field, in the boonies, then you could most likely leave the AM BC band filter out. However, if there is any uncertainty about the circumstances under which you’ll be using it, why not install it? It’s just a couple of toroids and 5 caps (unless you have 2,000pF caps in which case it’s only 3 caps, as you won’t have to double up on the 1,000pF caps).
10) A word about the bypass cap on the TX +ve supply line – the one marked Cy. In Peter’s version, this cap is 220uF. His mic amp is permanently connected to the +ve supply and switched off by a 100nF cap in the emitter lead, which is shorted out by the PTT button on TX. To achieve this, his PTT button keys the -ve side of the TX/RX relay. I understand now why he did this but in my “wisdom” I decided to permanently connect the mic amp +ve supply line to the TX driver final supply line and key them together. A side effect of doing it this way is that when the PTT is released, the remaining charge in the 220uF bypass cap on the TX supply line keeps the mic amp energized for about a second, causing a loud squeal in the speaker. I found that decreasing the value of Cy to 10uF gave a much shorter squeal that I could live with. I am hoping that this lower value of capacitance will still bypass any audio on the TX DC supply line.
As is usually the case with such projects, I built the AF amp first. Touching the input of the amp chip (in this case an LM386) to hear a loud buzzing sound always provides good positive feedback (pun intended :-) ). In the following 2 pictures, the AF gain pot hasn’t been hooked up yet. The curved red lead is a temporary power connection –
The MePADS and MeSQUARES from Rex at QRPMe have become a firm favorite of mine. Every Manhattan project I build uses them. I just realized that I can buy SMT chips from now on if I like, as the sheets of MePADS contain pads for mounting SMT devices too.
The next stage was the point at which things started to get interesting. This is the VXO using a 2N3904 and a 7.2MHz ceramic resonator. Thru-hole resonators for frequencies such as 3.58 and 3.68 are easily available, but ones for 7.2Mhz are a little harder to come by. When I discovered that http://www.hamshop.cz stocked them, I ordered 3 and gave away one, leaving me with just 2. Now I’m realizing that I should have ordered more, because on firing up the VXO, I found that the coverage with just one 1N4004 diode used as a tuning diode and no series inductor, was 7220 – 7335KHz. Of course, 115KHz of swing is quite a lot but what surprised me more was the fact that this 7.2MHz resonator was happily being pulled so high above it’s nominal frequency. A 3.3pF capacitor placed in parallel with the tuning diode brought the tuning down to 7169 – 7297KHz, which I consider very satisfactory, encompassing as it does the majority of the phone portion of the US 40M band. I like that the upper limit is 7297 as this means I won’t inadvertently transmit out of band. What a cracking little resonator! The resonator is the blue thing just below and to the left of the tuning pot (the top pot) in the photo below –
Then, things started to get really good, because I built the VXO buffer (an MPF102) and installed the NE602. At this point, I could connect an antenna to determine whether I would be able to hear signals. The first thing I usually do at this point is to turn the power on my K2 right down to 01.W and give a few short bursts of carrier. Even without an antenna attached, the little DC receiver picked it up with no problem and I knew we were in business. The antenna input coil is on the lower left of this next picture. You’ll notice that I have also built the 2N3904 mic amp. The blue wire was a temporary connector to the BNC at the rear of the case, so I could plug in the antenna for listening. If you look closely at the AF gain pot, you’ll see that I soldered a short grounding wire from the body of the pot to the chassis. Without this lead, you may get hum whenever your hand comes close to the pot –
This is always the point at which building transceivers gets tricky for me, as I spend so much time listening to the receiver, I lose momentum. I was even beginning to wish that I had set out to build just a receiver.
Here is a synopsis of what had been built up to this point. I had removed the PTT pushbutton to make soldering in that area easier –
You know how when you move house or apartment, you reach a point where you feel as if you’re very nearly done? That’s usually the point at which you are only halfway through (or even less.) All I had to do to turn this rig into a full transceiver, was add driver and PA stages, and I was in business. It wasn’t quite that simple, as I also had to cut and fit a partition, and wire up the transmit/receive switching. Here’s the first view of what I thought at the time was the completed rig. If you’re sharp-eyed, you’ll notice that the electret mic has been replaced by a phono socket. This was because I kept getting a motorboating sound on TX which was coming from the mic amp. Peter VK3YE said that I should either try a dynamic mic, or try lowering the gain of the mic amp if I wanted to use it with an electret mic. I decided to take the easier route, and replaced the internal electret mic with a mic socket. That way, I could experiment with different dynamic mics to find the best one. Also, the 2N3053 driver is fitted with a heatsink, wheras the BD139 final is not. KB6QVO said that his driver got warm, while his final ran cool. For this reason, he used a heatsink on his drvier, but allowed the final to go au natural. I simply copied him ( it was easier than doing my own research!) –
Well, this little rig works well. See the video at the end of this post to see and hear it in action. As mentioned before, the only issue I was having with the receiver was low level breakthrough from local AM broadcast stations. It was the only downside to what was otherwise a neat little receiver. I won’t retell the story related in point 9) near the beginning of this post, but the simple high pass filter I installed to attenuate signals in the AM BC band did the trick. Here’s a view of the completed transceiver with the high pass filter installed in the receive antenna line. The 2 toroids wound with green wire and the 5 blue caps in the upper left-hand side of the picture are the receive-only high pass filter –
I cut two small triangular pieces out of the bottom of the partition on both sides to allow wires to pass through. One cutout was a little bigger, as it had to allow more wires through. In the following picture, you can just see one of the triangular cutouts (I cut the pieces out with a flush wire cutter – perhaps not the best idea, but the cutter seemed to be undamaged) –
I suppose that at this point extra images just seem gratuitous, but perhaps one of them will contain an extra detail revealed by a slightly different camera angle that will help a hopeful builder somewhere –
This one might be useful in determining what goes where –
Here’s a view of my VK3YE Micro 40 from the back. The hole on the right is unused. I will cover it up with a piece of electrical tape on the inside –
And here is what this little DSB beauty looks like with it’s cover on, and viewed from the front –
You might wonder about the practicality of covering a tuning range of over 100KHz with a 1-turn pot. In fact, you are covering this range with just 300 degrees of rotation, which is not much. The tuning is a bit touchy but I was surprised to find that I got used to it. If you want, you could use a 10-turn pot for tuning, or a 1-turn pot fitted with a turns counter. The value is not critical – I often use 10K pots for tuning. One advantage to a 1-turn pot for tuning in simple rigs is that you can see roughly where you are in the band with a quick glance. Also, it is great for quickly scanning the band for activity. A second 1-turn pot for fine tuning would make it easy to exactly tune stations, while still keeping the cost lower than a 10-turn pot or a turns counter. I’m thinking a 100K pot for rough tuning and a 5K for the fine adjustments.
I plug a little MFJ-281 ClearTone speaker into the phone jack and it sounds great, with easily enough volume for comfortable listening. Current consumption is about 30mA on receive with no signals, peaking up to 100 – 125mA on very loud signals. I didn’t measure the current consumption on TX.
Here’s a recording of me in QSO with KE7NCO 180 miles away. At this point, the 220uF capacitor was still in use bypassing the +ve supply line to the TX, causing the very noticeable feedback when switching from TX to RX –
On changing that bypass cap from 220uF to 10uF, the feedback reduced considerably. Here I am in QSO with N7UVH. He is my greatest DX to date, being 736 miles away – not bad for 800mW of DSB (equivalent to 400mW of SSB) -
Here’s a video demonstration of the receiver (boy, I really need a new video camera. This one is 10 years old and limited in resolution!) –
This was me checking into the daily Noontime Net on 7268.5KHz with Jim W6FHZ, who is 180 miles from me –
On the scale of dollars spent for fun and satisfaction had, this little rig is high up there on the list. I built mine with components I had on hand but even if you had to purchase all the parts, I calculated it would cost you around $23 (not including shipping from the various different suppliers.) This is one fun little rig – and it wouldn’t be hard to whip up a simple matching network for an end-fed halfwave antenna, and take a small battery with you for some portable fun.
As an aside, I went to Pacificon last weekend and had the pleasure of meeting Steve the Goathiker, WG0AT. Here he is at the Buddipole booth holding the packet that contains his entire portable station – a KD1JV MTR with a small key and end-fed half-wave antenna. Fantastic!
Note on Ceramic Resonators – sourcing suitable resonators for projects like this can be tricky. The supply of hamshop.cz 7.2MHz parts seems to have dried up. Mouser have some 3 terminal 7.2MHz ones that, even with the internal capacitors out of circuit, wouldn’t resonate much higher than 7.15MHZ (myself, NT7S and KB6QVI all got the same results). More recently, Patrick W9PDS found some 7.3728 MHZ resonators from Mouser that seem to fit the bill. Joel KB6QVI just reported that with a 27pF cap across this part, and using a polyvaricon for tuning, he is getting a freq coverage of 7.175 – 7.303MHZ. It sounds like this would work in the Micro 40. You’ll need to experiment a bit with parts values to get the coverage you want, but suitable ceramic resonators for 40M are getting hard to come by, so you might want pick up a few while you can. You can find these ones (while supplies last) here.