Thursday, March 27, 2025

SolderSmoke Challenge Alternate Audio Stage

 


The audio amplifier in the SolderSmoke challenge DCR is relatively easy to build - just three 2N3904 common emitter amplifier stages, and an audio transformer.  Builders in certaing parts of the world have had challenges acquiring the transformer and have asked for a design that does not use one.  Also, many builder struggle with oscillations and feedback.  To address this I replaced the final 2N3904 stage and transformer wiht this signficantly more complex final amplifier stage from Fig. 1.17 of Experimental Methods in RF Design.  I built it mostly to the orignial schematic with a few changes that I found in an archived article from Vasily (Todd) of Popcorn QRP.  

Q1 is the gain stage for the following NPN/PNP  complimentary pair emitter follower transistors (Q3 and Q4),  Q2 provides active DC decoupling and provides the floating voltage source that establishes bias for the complimentary pair.  It uses negative feedback via R3 to stabilize the circuit and to set the overall bias for the stage.  

Todd did some analysis that showed that as-designed there was significant crossover distortion in the output wave form and I verified this to be the case. 

EMRFD Fig 1.17 Amp showing crossover distortion

 The solution was to change R6 to 4.7K which cleaned it right up.

Changed R6 to 4.7K - Virtually eliminated distortion


Clean output aftger mod

The only other change was to change R1 to 1K which increased the overall voltage gain of the stage significantly.  Here is what the final build looks like in the DCR. I've kept the first two 2N3904 driver stages from the SolderSmoke DCR Challenge and substituted the new amplifier for the final stage.


KK4DAS DCR Final Audio Amp Mod

The amp sounds great and provide room filling audio, but ut it is not a beginners amplifier.  


So it is possible to build a great sounding, very stable, all discrete component transformerless audio amplifier.  It just takes a little more work than we wanted to subject new builders to and and with the addtion of a complimentary pair of transitiors, active decoupling and feedback it adds signficant complexity to the circuit.

73 from Great Falls,
Dean
KK4DAS

Monday, March 24, 2025

KK4DAS MB 20 Transceiver - now Receiving on 20 meters

 



This week's homebrew adventure was to complete the receiver side of what I am now calling the KK4DAS MB 20 transceiver - MB is a call-out to Bill, N2CQR's Mythbuster rigs which this build is both inspried by and patterned after.  

First order of business was choosig the IF freqency which was easy in this case.  The VFO is at 9MHz so if I put the IF at 5MHz, well my 2nd grade math teacher taught me that 9+5=14 - which puts me right in the 20 meter band.   

I ordered a batch of 5MHz crystals from Mouser for the IF but they had not come in yet - and I was impatient to get started so I built the BFO and balanced modulator using a crystal from the junkbox figuring I would swap in a 5MHz crystal when it arrived.  The BFO and two diode balanced modulator is copied directly from Farhan, VU2ESE's original BITX20 schematic. It went together easily.  Despite conventional wisdom I did not attempt to mach the diodes - I just took the first two off the tape.

BFO and balanced modulator

Nulling out the carrier

To null out the carrier I injected a 2MHz signal into the AF port of the modulator and put the scope on the IF port.  Using the FFT mode of the scope I was able to clearly see the carrier and mixing products.  I was able to effectively null out the carrier using the trimmer cap and trimmer pot that are part of the design. The two peaks either side of the middle are the first sum and difference mixing products.  The display is centered on the carrier frequency and you can see that it has been nulled out.

The crystals arrived and the next job was to build the crystal filter.  I used the Dishal crystal ladder filter design software to design a 6 pole Cohn/QER filter.  The advantage of the Cohn filter is that it uses a single value of capacitor between each crystal.  The QER (Quasi Quasi Equi Ripple) modification to the Cohn adds an addtional crystal in parallel at each end of the filter and reduces the amount of ripple.  When building a filter the recommendation is to pick crystals fom your batch that are the closest in frequency to each other and the Dishal software requires you also to know the motional parameters for the crystal. I used what is called the G3UUR method to evaluate the crystals. I had previously built the simple crystal evaluation circuit and quickly sorted the crystals. I then used the Dishal software to determine first the motional parameters and then to calculate the capacitor values and input/output impedance of the filter.  The calculated input and output impedance were spot on0and I calculate the LC values needed for matching to 50 ohms.
   
Sorting the crystals using the G3UUR circuit

The completed 5MHz, 2.5KHz filter

The last board I needed for the receiver was an audio amplifier.  I opened up by box of "boards that could be rigs someday" and found an audio amplifier stage I had built for Pete, N6QW's SimpleSSB a few years ago.  Its a single 2N3904 pre-amp followed by an LM-386.   With all the boards needed for the receiver at hand it was time to connect everything up.  

The completed KK4DAS MB 20 Receiver

I placed the crystal filter at the rear center with the MostlyDIYRF TIA IF amplifier boards on either side. I left space adjacent to the TIAs for the transmit IF amplifiers. The audio amplifier is on the lower right and just above that is the BFO / balanced modulator board.  The receive band pass filter is on the upper left below the first IF amp.  At the far left is the future Tx RF amplifier stage - not yet wired into the circuit.  I mounted the antenna connector on a piece of single sided PCB.  To keep the DC wiring straight I decided to use blue wire for DC on Rx and yellow wire for the DC on Tx.  Red is used for always on DC.   

A big advantage of modular construction is that since all of the modules have been tested individually it increases the liklihood that the build will work well when everything is connected.  To my great pleasure that turned out to be the case and the receiver woked perfectly on first power up.  There was one glitch - the tuner maxed out at 14.260 MHz about 100KHz short of covering the entire 20 meter band.  I thought I was going to have to rebuild the IF with 5.2MHz crystals but Bill pointed out that Yaesu and built in a variable cap to move the VFO passband up or down. I zeroed out the capacitor and that gave me full coverage across 20 meters.

Adjusting the Yaesu VFO for full coverage on 20 meters

Now that the receiver is built I will spend a few nights following Farhan's advice and take pleasure in listening to the receive I have just built.  

All I need to build for the transmitter is a microphone amplifier and the final low  pass filter.  Then add some relays for Tx/Rx switching and then I can work on fit and finish of the cabinet.  

73 from Great Falls,
Dean
KK4DAS

Sunday, March 16, 2025

New Build - 20 Meter SSB Rig


After spending months on Vienna Wireless Society Makers projects and the SolderSmoke Direct Conversion receiver project I wanted to get back to building a rig just for me!  Several months ago Pete, N6QW sent me an ebay listing for a very nice Yaesu VFO module, for only $35.  I jumped on it, not knowing what I was going to do with it.  The module is complete with reduction drives and dials.  It even came with the dial window that proclaims it to be from a Yaesu FT-401B.


Earlier versions of the VFO run on 6V but this one needs 9 to fire up and oscillate.  After getting the voltage right it fires right up and tunes from about 9MHz to about 9.4MHz.  So with a 5 MHz IF this will make a 20 meter SSB rig.  Bill, N2CQR has built two of his Mythbuster rigs using this VFO module and I am following in his footsteps.  Like Bill I found the output was too low to drive a 7dBm diode ring mixer so I build a 6dB gain Termination Insensitive Amplifier (TIA) to boost the drive to the required level.

I'm doing things deliberately this time around - I'm starting with the end-layout in mind including placement of all the modules.  Its going to be in a homemade plywood case.  I'm using copper plumbers tape to provide the ground plane and case shielding.  After 5 years of continuous homebrewing I have built up a collections module boards that are just waiting for the right project come along.  Bill calls these "boards that could be rigs."  I already have the transmit chain built - it is two stages of 2N3866 drivers followed by a Mitsubishi RD16 MOSFET final - which should give me 10-15 wattts or more on 20 meters.  This was the transmitt module I had built for my SBITX build before I went to a push-pull final for 25-30 watts.  I have the diode ring mixer built (one of several build during the DCR projectg testing.)  For the IF I have ordered 5MHz crystals from Mouser and am using 4 TIA boards that Todd, K7TFC of mostlyDIYRF kindly sent me some time ago.  I assembed the TIA boards the other night. In addition to being termination insensitive which means that what you hang off the output doesn't load down the amplifier, they are also variable gain - anywhere from 7dB to 24dB by changing a three resistors.  I had a 20 meter BPF built that was looking kind of ratty so I build a new one last night - I just used the component values from the QRPLabs BPF filter kit instructions.  Tested on the NanoVNA less than 1dB of insertion loss and almost no ripple across the entire band.

20 Meter BPF - less than 1dB of inserstion loss acrosss 20 meters


20 meter SSB rig in progress
VFO, buffer amp and diode ring mixer

VFO and mixer test - 14Mhz RF, 9Mhz VFO, products at 5MHz and 23MHz! 

I have a couple of different ideas for the audio chain - I might use a kit (shocker) from Kitsandparts - as part of my ongoing desire to master surface mount construction.   The balanced modulator will be a 2 diode version straight out of Solid State Design for the Radio Amateur.  The mic amp will be identical to the electret version I build for the SBITX.  The transmit LPF will be a W3NQN filter again with the values from the QRP labs LPF kit instructions.  

After five years of building I am finally starting to feel like I know enough to do stuff!  Thanks Pete and Bill for all the encouragement over the years.

73 from Great Falls
Dean
KK4DAS

SolderSmoke Direct Conversion Receiver Challenge

It's been a busy several months - all of my recent internet activity (and much of my homebrew time has been spent on the SolderSmoke Challenge direct conversion receiver project that I am working on with Bill, N2CQR.  Since January more than 40 homebrewers from all around the world have completed the challenge and more are finishing up every day.   If you haven't followed it check out the SolderSmoke blog and YouTube channel and search for SolderSmoke Challenge.  Join us on Discord if you too would like to complete the SolderSmoke Challenge an prove that "When you know stuffm you can do stuff!"



Schematics, instructional videos and group support on the SolderSmoke Discord Server:

https://discord.gg/Fu6B7yGxx2


 

SolderSmoke YouTube channel:

Saturday, September 21, 2024

2000 47pF Caps ...

 


An unexpectged package arrive in the mail today.  Did you ever wonder what 2000 47pF NP0 capacitors look like?  Thanks to John, AB2XT I will never run out of 47pF caps. These caps will find their way into future projects and inject a little more soul into the machine. Thanks John.

Monday, August 12, 2024

Homebrew sBitx - Fldigi

 

Homebrew sBitx - Fldigi

Now that I have the hardware tamed I'm turning my attention to some of the more itnteresting and useful features of the sBitx.  With the sBitx software running on a Raspberry Pi 4B there is plenty it is possible to run a variety of digital mode software programs right on the Raspberry Pi itself - there is no need for a separate computer. FT8 is built directly into the sBitx software so there is no need to run WSJTX, but for other digital modes including RTT, FSTK, Olivia, etc. you need to get the Linux versio of those programs and configure them to talk to the sBitx via the Hamlib interface.  

This was my first attempt at using Fldigi and I managed to make several good connections with myvfriend Don. KM4UDX.  The one pictured above was using the Olivia 16/1K mode.  Pretty impressive, but I still have more tuning up to do. While Olivia modes decode flawlessly - I was not able to receive and decode PSK, RTTY, and others.  I need to work on getting audio at the  correct level to not over-drive the SDR. 

I can turn the IF down on the sBitx.  I can also make some adjustments to the filtering in the Fldigi softwatre itself.

More to come on this one.


73 from Great Falls,

Dean 

KK4DAS


Tuesday, August 6, 2024

Homebrew sBitx - LPF Leak Stopped!



Quick update - I have resolved the problem with the LPF module in my homebrew sBitx.  The observed problem was that the transistors that control the relays sometimes were conducting when they were not supposed to resulting in more than one filter being switched in at a time.   The solution turned out to be adding a cap to ground at the low side of the relay coil where it connects to the transistor.  The hypothesis was that RF was being induced onto the inexpensive non-RF-rated relay coils.  I also swapped the NPN transistors for 2N7000 MOSFETS - the 3V threshold voltage guarantees that they won't turn on from minor fluctuations on the Raspberry Pi GPIO lines.   Reliable informaion on the high and low logic values for the Pi GPIO have been difficult to find.  Data sheets for the Pi 1.0 suggest logic "low" could be as high a 800 mV which could turn-on an NPN transistor, so the MOSFET seemed a safer bet.

Here is the final circuit:


 

Thanks to everyone for their suggestions and ideas.  The wisdom and entrhusiasm of the crowd really helps.


73 from Great Falls,

Dean

KK4DAS