Sunday, August 28, 2016

40M Junk Box SSB Transceiver

You Too Can Build A Junk Box SSB XCVR!

8/31/2016 Results Update
8/31/2016 -- Photos of the Boards Added

This is one great radio and following my success at 500 Milli-watts I added the output stage using a 2SC2075 where I am getting about 3 Watts PEP with the device  biased for 100 Ma. This in turn can drive my SB-200 to about 45 watts output. With my backyard droopy dipole on 8/30 I worked stations in Arkansas and Texas running 45 watts. So I am pleased.
At this point I am torn between putting it in a box or just leaving it "junk style" on the bench. I fear taking it apart may ruin the magic of what has been done so far. Of course if it goes into a box it will get the Juliano Blue treatment.
I am really impressed at how well the receiver is working and perhaps some of the additional refinements have helped. The transmitted signal has received excellent reports. In the bi-directional amp stage that started with 2N3904's and then went to a 2N1711 and 2N3868 has now moved on to a 2N2102 in each leg. This seems to be a very good final choice. What I like most is that in  some of the circuit blocks I have included sockets where I can test various solid state devices -- the radio after all is an experimenter's platform.
For a bit of a trip if you can find a 1966 publication entitled "The Transistor Radio Handbook" authored by Les Earnshaw ZL1AAX and Don Stoner W6TNS, there is featured a complete 40M SSB transceiver designed and built by Jo Emmett Jennings, W6EI. [Jennings invented the Jennings Vacuum Variable Capacitors].
This rig used circuit boards made by International Crystal Manufacturing that were modified by W6EI for this application. It puts out about 1/2 watt, similar to the first debut with my radio, and in the mid 1960's used a lot more parts. This book can still be found on Amazon for $12USD. [For those who are scratching their heads that is $12 dollars in US dollars. I was recently admonished that my portrayal using $12USD was incorrect as being redundant! Twelve bucks guys is what it is.]
[Follow on note: Earnshaw started several electronics companies here in the USA including Kachina which was one of the 1st computer controlled ham transceivers. Stoner teamed with Pierre Goral to found SGC which produced a line of ham equipment and of course linear amps as well as auto tuners.]
To date I have had about 20 QSO's with this radio and states worked include; California, Arizona, Nevada, Arkansas and Texas. The frequency stability using the Si5351 adds immeasurably to making contacts and getting good reports. However an Analog type VFO and separate BFO could produce good results and that is left to the builder as an option.

Other filter options include the readily available Heathkit filters at 3.395 MHz (Zin/out = 2000 Ohms) or the GQRP Club 9.0 MHz Filters (Zin/out = 500 Ohms) or Inrad sell a 4 pole 9.0 MHz Experimenter's filter (Zin/out = 200 Ohms). In fact the photo above the masthead featuring the LBS II, that radio has the 4 pole Inrad filter.

This is the Main Board with the MCL Double Balanced Mixers

 This Board is the Solid State Switch designed by me.

This Board is the Audio Amplifier Board

This is the 2N2222 & 2N3866 Driver Board

This is the 2N996 PNP Transistor Microphone Amplifier Board

This is the 2N2102 Bi-Directional Amplifier Board

This is the 40M Band Pass Filter Designed by N6QW

This board is the 2SC2075 Final Amplifier

This board is the Arduino Pro-Mini with the Si5351 PLL Clock Oscillator

This board is the TR Switch, Linear Amp Switch and the Low Pass Filter from W3NQN

The fact that there are individual photos of the boards (as above) should make it obvious the rig is headed for a Juliano Blue case.

A side note here, initially these photos had no captions wherein I was asked to include captions of just exactly what were the boards. In the spirit of cooperation I have done so. But if you could not determine what were the boards based on the schematics presented in this thread then I caution you "do not build this radio"!

Pete N6QW
This current detour in the road was the result of awaiting parts for the LDMOS Amp which are now here and the second piece was I smoked something on the Germanium PNP Transceiver. So back to the junk box to find more boards and parts to work on whilst I wait.
Having lots of old boards lying around makes it a pretty easy task of simply soldering up a bunch of wires and you are there --well almost!
First let me say that on Saturday August 27, 2016 running 500 milli-watts (1/2 watt) I made the first contact with the rig below by having a QSO with Jim W1WY, located in San Diego about 173 miles from my QTH in Newbury Park. So once again QRP does work.  But Hey I just wanted to see if it could work --more than once! You can clearly see  this is just a bunch of boards hay-wired and screwed down to the work bench. This is not how to build radios!
But lets start in a more organized fashion and we do this with a block diagram and some description of the circuits.

The architecture of this rig is not unlike many of the radios I have built in the past--it works so why not use it! Of note is that most of the boards have been recycled. The Solid State switching board was one I built for the 30M CW transceiver and the driver stage board was actually the prototype board for the BFO/CIO used in the 2012 KWM-4 build. I had made an error on the layout and just used the available islands for this build. The Bi-Lateral amp stage was another prototype build from about three years ago. The LPF board was in the junk box. The Band Pass Filter board was built on a board that originally housed some external low pass filters used with the softrock SDR transceiver.
Some notes the bi-lateral stage mainboard as I mentioned was built several years ago and while it worked seemed a bit deaf. Now I know why -- I had failed to install two 100NF caps on 2N3904 Emitters. Boom fixed that and the receiver is really hot!!! Normally the supply voltage is 6 VDC -- I upped it to 8 Volts by replacing the on board 6 VDC regulator with an 8 VDC regulator. Big improvement. I heartily recommend that change. The bilateral stage was developed by Plessey and is a direct lift from EMRFD -- finally found something useful in that publication.
A special comment about the audio amplifier. When I built it and connected earphones to the output at first I thought it was not working as there seemed to be no sound. I used the Juliano special audio test which consists of placing your finger on the input and BOOM -- A Loud Sound! This is a great amplifier circuit as it has a lot of gain with very little noise. FORGET building those discrete part amplifiers --you are wasting your time and using a lot more parts!
 In the bidirectional stage which is something I developed I wanted to try some transistors other than the 2N3904's as they get a bit hot. On the receive leg I used a 2N1711 and on the transmit leg a 2N3868. These are TO-39 devices and a bit more amenable to a heat sink versus the plastic 2N3904.

The driver stage is another useful lift from EMRFD --that makes two now. However I substituted a TO-18 style 2N2222 for the 2N3904 as that will work with a heat sink.

For the Band Pass Filter I used what I designed with the help of LT Spice for the Simpleceiver project. It worked there so why not here?

C1,C2 in this schematic is comprised of a 150PF NPO in parallel with a 15-60 PF Trimmer Cap. C3,C5 are fixed 47PF NPO caps and C4  is a 5 to 15 PF Johnson Air Trimmer Cap.

Cautionary note here many of the schematics were used in other projects and the basic circuitry is solid --some of the references to in/out connections may not apply.

The IF amplifier is similar to that shown below. Changes include using the Yaesu Filter at 3.180 MHz and the matching transformers are 6 turns #26 on the primary (50 Ohm side) and the secondary is 19 Turns of #26 and connects to the filter (500 Ohms) [ 6^2 = 36 and 19^2 =361 -- 361/36 = 10 and thus a 10:1 transformation that step up 50 Ohms to 500 Ohms.] So just replace the home brew crystal filter with the commercial unit and rewind the transformer for 10:1. Finally Change the source voltage to 8 VDC and don't forget the 100 NF caps on the 2N3904 emitters.

Solid State Switching Circuit is shown below. Ignore the circuitry below the 4N35 as that was used in the KWM-4 for CW operation. That part is however useful if you would like to use this circuit for CW break-in operation.

I am pondering building an IRF510 final RF stage or using a real RF amplifier transistor. The Si5351 uses basic code from AD7C which was modified by me for this application. Send me an email to if you would like the sketch. I am almost tempted to switch over to a color display -- but I had the LCD in the junk box.

The front end use the TUF-1 and the back end an SBL-1-just what I had. The Low Pass Filter is a direct lift from W3NQN.

Finally the Microphone amplifier --yes Virginia there is a Santa Claus and you can build audio amplifiers with PNP transistors.

I decided to build a real RF amp stage found a 2SC2075 lying in the junk box and that is what I used as the circuit for the final amp is the same one I used in the JABOM built about 5 years ago and is reproduced below. The 2SC2075 has the same pin out as the 2SC2166.
Thee rig now puts out about 3 watts and tickling the SB-200 is good for about 40 Watts. Plenty good.
That is it -- have fun building this 40M SSB rig.
Pete N6QW


Tuesday, August 23, 2016

Homebrew Germanium PNP Transistor 20M SSB Transceiver

UHF Germanium PNP Transistor SSB Transceiver

Whilst I await some parts for the LDMOS amplifier and Arduino Control system, I began to search my large box of projects that "sorta worked" , worked only once or never worked looking for something I could fix in short spurts of time. Somehow those who possess the "knack" for not having a project work or work properly are always drawn back to the bench in hope of finding the magic pill to bring the "pile of parts" back to life.
Mention of Germanium PNP  transistors must seem odd today. For some it must seem like a cruel April Fools joke suggested by Bill, N2CQR. Yet for other old timers who really do know about such devices invariably must think N6QW has passed into senility.  Yes there really are (or were) Germanium transistors (the CK722 was one of the first) and most early transistors were of the PNP variety. One of the early RF grade Germanium Transistors was the Philco SBT-100.
Interestingly enough today audiophile homebrewers and DIY musicians love germanium transistors as applied to "fuzz boxes".
Today by and large most transistors are silicon and NPN. Some 60 years ago most transistors were limited in power and frequency range; but today cheap UHF silicon NPN  transistors can be had for pennies.
In about 2011 I had a grand idea to use Germanium PNP  transistors in a SSB transceiver. I was not the first, as Ben Vester, W3TLN in a 1963 QST article did just that, and it was a small radio, roughly 5X7X2 inches. His rig was on 20M and ran about 1/4 watt. W3TLN's article is what prompted me to build the now "Juliano Blue" shirt pocket SSB transceiver. My project was 2X2X4 inches. In preparation for that project which ultimately used no Germanium PNP  transistors, I did purchase a 100 piece box of Russian Germanium PNP UHF transistors. At the time I paid about 10 cents a piece for them. These transistors were weird looking and the in line pinouts were BCE.
In 2013 after moving to Southern California, I decided to attempt to build the Germanium PNP UHF transistor SSB transceiver using the Sideband Engineers Model SBE-33 as a template. You might be surprised to know this radio vintage 1963 used bilateral circuitry. I did build the radio and the results were less than spectacular --so into the large box it went. Fast forward to today and I found that radio and cannibalized the parts. Using Vester's approach I built a new Germanium PNP UHF Transistor SSB transceiver. There are still lots of bugs but I am on the path to success. If you look closely in the video you will see some of the strange Russian transistors.
Some changes I made to Vester's deign include the following:
  1. Vester used a tuned Germanium transistor RF Amp Stage that fed a Germanium Transistor Mixer Stage for the front end. Using a 4 Pole Slide switch whereby using one set of the poles, he switched that front end output to the Crystal Filter on Receive and on Transmit fed the Balanced Modulator output  into the filter. In my scheme on receive I have a broadband Germanium transistor RF amplifier feeding into an SBL-1 Mixer stage. Utilizing a pair of ultra miniature relays ahead and following the SBL-1 on transmit the Microphone amplifier (a Germanium PNP audio transistor 2N996) is switched in place of the RF amp. On receive the LO (an AD9850) is fed into Pin 8 of the SBL-1; but on transmit the BFO is now switched into Pin 8. Cheating a bit? I don't think so --just using what I have.
  2. Following the SBL-1 is a post mixer amplifier (Germanium UHF PNP transistor) and that feeds the 9.0 MHz commercial crystal filter. Both the broad band amp and the post mixer amp are simply my standard 2N3904 general purpose amp stage that has been modified for the Germanium PNP transistors. The output is adjustable so we don't overdrive the follow on stages. Works perfectly!
  3. In W3TLN's design he has a two stage IF amp chain following the filter and that feeds a diode ring. On receive it is used as the product detector and as the transmit mixer on transmit. Another small relay will switch the BFO/LO signals into this diode ring.
  4. I currently have some dreaded feedback/oscillation issues in the IF amp stage which I hope to cure. Shielding and grounding are key no matter what kind of transistors are used.
  5. The video was made in the early evening and 20M signals were not too strong. The radio does have AGC (one of the few NPN devices --and not Germanium) so it does have some frills.
Stay tuned.
Pete N6QW

Thursday, August 18, 2016

LDMOS Amplifier ~ Next Steps

More Revelations about the LDMOS Control System.

Rev 8/19 Added a second video of the three light sequence.

My efforts over the past week or so were to further refine and explore the control system functionality. It is has truly been an exercise in "non-linear" thinking which by my definition is not a simple straight line from step A to B to C and so on. Instead it is the need to see that it is A to B & C followed by D and then perhaps and E & F & G. What may seem straight forward and direct often is not that simple based on the complexity of what you want to happen.
Above all, the complexity of the control systems lies in the three basic functions of the Arduino control system those being: 1) Pure Control, 2) Supervisory Oversight and 3) Hardware Protection. The Supervisory and Protection aspects frequently drive and override the Control functionality. Overlaying this is the speed at which events must occur to failsafe the very expensive RF device.
I cannot overstress the value of having built a breadboard of the Arduino Control system as this has been an invaluable tool. From this effort I uncovered that the 48 VDC power supply had to be essentially shut off immediately through the use of a FET power switch.
Using the on board LED indicators helped in "physically seeing" sequenced actions. To this end there is a sequence where the output of the Low Pass Filter banks are connected to the antenna (step one), a bias is applied to the LDMOS amp (step 2) and finally the transceiver output is connected to the LDMOS amplifier input (step 3). These steps are reversed when going back from transmit to receive. The Arduino Code has the timing (time delays) and forward/reverse sequencing built into sketch. However the time delays were more or less randomly picked based on a notional gut feel. When the hardware was built and the LED's sequenced I could see one of the delays was too short and the other too long.
Exercising the keypad buttons has shown where conditions could exist to essentially cause an unexpected event to occur. This is valuable in that such events are discovered and corrected before any expensive hardware is actually connected.
To date we have the following in operation:
  1. Basic Start and Stopping of the 48 VDC Power Supply
  2. Low Pass Filter Selection
  3. TR Circuit Tied to LPF Selection ( No Filter selected, no TR)
  4. Sequencing of the TR Relays and Bias Circuit.
  5. The FET Power Switch is built (awaiting the heat sink)

See the latest progress as of 8/18/2016
Pete N6QW