Dev Update 11-Sep-2020
The optional enclosure is the same aluminium extrusion as the one used in the 50W PA kit but cut shorter (92mm) and with of course, different drilling and laser etching. The QCX-mini PCB assembly slides inside along the PCB rails in the enclosure. There are NO mounting screws at all, other than the 8 end-screws that fit in each corner of the left and right panels to fasten them to the extrusion.
As of 11-Sep-2020 I am waiting for manufacture and delivery of a prototype enclosure.
Prototype boards arrived!
Five prototype PCBs arrived and now prototype assembly is underway. NOTE that I am still waiting for many of the SMD parts, as well as the enclosure prototype. The board currently boots up, but is not otherwise operational yet.
There are two PCB panels. The "bottom PCB" of the QCX-mini sandwich holds the majority of the components. Some of the components remain through-hole. In particular, note that I have kept the classic 28-pin DIP packaged ATmega328 not moved to a SMD version; therefore it is still possible to upgrade firmware in future by changing the chip. Note also that the PA transistors and key-shaping transistor are sandwiched flat against the bottom PCB as heatsink, the same as in the QCX+ kit.
The "top PCB" of the QCX-mini sandwich holds only the LCD module and associated components (contrast adjustment, and a circuit to enable switching the backlight LED on or off). There is a 2x5-pin header connector to route signals from the microcontroller to the LCD module. There is a large cutout for the LCD metal body, the LCD PCB itself bolts behind; this lines up the LCD at exactly the right "height" to slide it into the PCB rails in the enclosure extrusion. There is also a cutout in the lower half of the "top PCB" through which a small PCB holding the four controls fits. The bottom right corner of the "top PCB" is milled away, this is to accommodate the height of the metal BNC connector below it; but in the actual event, the metal BNC connector turned out to be less high than expected so this corner cut-out is not needed and I will put it back in the final boards.
The "top PCB" also has a number of smaller PCBs which break off it along lines of drilled holes. The largest of these is the specially shaped PCB that holds the controls (5K gain pot, rotary encoder, and two buttons). This controls PCB needs to be 1.6mm higher than the "top PCB" so there are small cut-out PCBs that act as a 1.6mm spacer on the 2x4-pin header connector, and the hex spacer mounting pillar.
The top and bottom boards are sandwiched together, fixed with four 11mm hex spacers. The top PCB slides into guide-rails on the enclosure. The bottom PCB is slightly smaller and is fixed to the top PCB via the spacers, rather than sliding into the enclosure PCB rails which are not at an appropriate location for the bottom PCB.
I'm particularly proud of a hole and a cutout area milled into the "Controls PCB", which line up perfectly with the adjustment controls for BPF peaking (C1 trimmer) and the three 24-turn trimmer potentiometers which adjust IQ Balance, Low and High phase shifts. These are labelled on the silkscreen of the "Controls PCB" and you can fit a screwdriver down through the gap to make the adjustments while watching the display, when the whole sandwich is properly assembled.
Views of the QCX-mini
Here are some photographs of the QCX-mini from various angles.
There was some discussion on the QRP Labs discussion group on groups.io about whether to use a BNC or SMA connector for the RF connector. A vote was held and 75% of the voting electorate chose the BNC connector. So BNC it is. But in response to some comments about the large size and possible lack of robustness of the plastic-bodied BNC connector used in several QRP Labs kits including QCX+, I have located a good source for a high-quality, all metal, PCB-mounting and chassis-mounting connector which is perfect in this application and is also smaller (protrudes much less from the enclosure).
The left photo below shows them side by side.
Notice that I left a cut-out of the bottom right corner of the top PCB, to accommodate the height of the BNC connector; but in the end I see that it is lower than I had calculated and therefore in the final PCBs I will put back that corner cut-out (which will also make the assembly slightly more mechanically rigid, not that there is any problem in this area).
Outdoor testing: bright sunshine
Since I do not yet have the SMD components to install the LED backlight control circuit, I decided to do some outdoor tests to determine the viewability of the LCD, with NO backlight (perhaps a typical portable operating condition, to reduce power consumption by not powering the backlight LEDs).
The tests were done in early afternoon on 11-Sep-2020. In Turkey where I live, the sun is very strong. Today the temperature was 38C, in the shade. Yes I boiled, taking these various photos in the sun. I'm still drying off, as I write this, and rehydrating my body.
The photographs were also very difficult to take on my Samsumg S9 phone, because it was impossible to see the OLED display in sunlight, and even difficult to see it at all when I raised my hand to shade it. The Samsung S9 being a top end phone, and with the screen at maximum brightness, this underlines the unsuitability of OLED technology for bright outdoor sunshine conditions.
The left photo is taken as close as possible to directly full-on sunlight hitting the screen perpendicularly and with the camera as close to the sunshine direction as I could get without casting a shadow. PERFECT viewability. The screen was also well readable from any other reasonable angle.
Outdoor testing: shade
In bright but shaded conditions, the screen was still perfectly readable.
Prop: The junior lab tech's sports car
The junior lab tech, being temporarily absent, had his sports car borrowed for the purpose of arranging the QCX-mini at a suitable angle to get the sun as perpenticular to the LCD as possible...
A certain pair of Ray Ban sunglasses having polarized lenses were used to determine if the screen is still easy to read even wearing sunglasses, and I attempted to take photographs through the sunglasses too, with a reasonable degree of success.
Yes the sunglasses are horribly scratched up. In my defence they are many years old. Apparently I am not entitled to new sunglasses because I am not properly able to look after anything well enough, so I am told anyway.
The screen remained well readable through the sunglasses as you can see. When rotated, at 45-degrees relative to the line of the text, the display looks black completely (see the last of the five photographs). At 0 and 90-degrees it is fine. The blackout is a few degrees either side of 45-degrees. I don't think in practical use this would be an impediment since normally the operator would probably not rotate his operating position at 45-degrees to the line of the display text.
More sunglasses testing
On pain of death - to take the utmost care without inflicting scratches or even worse crimes such as dropping them - I also tested a pair of Gucci sunglasses and a pair by Chopard. Since there were two visiting YL's at the time of my experiments I also tested on their sunglasses too (though somewhat nervous and unwilling and watching with a degree of trepidation). Through all of these sunglasses, the display remained clearly readable. They did not exhibit any change or any black-out at 45-degrees or any other angle so I gather they are not using polarized lenses like my Ray Bans.
The uSDX modification
uSDX (also known as QCX-SSB) is a modification to QCX developed by Guido PE1NNZ. It turns the QCX into a 5W CW/SSB transceiver utilising digital techniques: SDR and DSP functions, using the existing 8-bit ATmega328 processor. It keeps the QCX Class-E final and implements EER (Envelope Elimination and Restoration) to provide SSB funcationality. This modification is described in full here: https://github.com/threeme3/QCX-SSB .
This modification is entirely supported by its own discussion group of interested constructors. It is NOT an official QRP Labs modification and there is no technical support from QRP Labs in anyway for this project.
I think Guido's work is interesting and inspiring and so I decided to provide an easy way to modify QCX-mini into a uSDX. This is implemented by a tiny daughterboard, which is specially shaped with an arc cut-out and a corner square cut-out, so that it fits snugly between the paddle socket to its West, the 28-pin DIP ATmega328 socket to its North, a 470uF capacitor to the South, and the I-Q balance 24-bit potentiometer to its East. Although in practice if starting out to build a uSDX, one may omit the installation of the not-needed 470uF capacitor and the I-Q balance pot. This PCB is expected to be installed at a height of a few mm from the QCX-mini bottom PCB at 8 points (indicated in red in the component layout diagram below), perhaps by component lead offcuts or by sawn-off pin headers used as 2.5mm spacers. This little daughtercard is fit into the QCX-mini top PCB, in the cut-out area where the Controls PCB is finally installed; it is simply snapped off from the main QCX-mini top board and the edges filed neatly. Since the space was already available, it costs nothing to include this daughterboard in the hope that it may be of interest to some interested constructors.
The uSDX modification to QCX-mini would involve:
- removal of some 0-ohm SMD resistors (that disconnect parts of the QCX-mini circuit, including removing power to op-amps IC6-10);
- removing some other SMD components;
- changing out a few others to different values;
- populating the bias circuits components (all 0603 SMD) which are on the little daughtercard and installing the daughtercard;
- Adding three jumper wires (from points shown in pink on the diagram below) to connection points provided on the QCX-mini PCB (Dah, I and Q).
- Changing to a uSDX firmware chip, naturally ;-)
Basically the changes are similar to those documented on the uSDX GitHub page https://github.com/threeme3/QCX-SSB. The T1 transformer (with its Band Pass Filtering) can be left in place or further modifications undertaken to eliminate it if required. My own recommendation is to keep T1 because the BPF does improve performance, so does using the Quadrature Sampling Detector (QSD) in its double-balanced configuration.
Doubtless when QCX-mini becomes available, more detailed instructions may be provided by the uSDX community.
AGAIN: this modification is entirely supported by its own discussion group of interested constructors. It is NOT an official QRP Labs modification and there is no technical support from QRP Labs in anyway for this project.
Price and availability
The price is of course not finally known or decided but is expected to be the same or similar as QCX+ (which is US$55), and the optional enclosure a bit cheaper than the QCX+'s $25. "A bit" because though there is significantly less aluminium involved, the bulk of the production cost is in the cutting, CNC milling, drilling and laser etching, rather than the raw material itself.
I do not anticipate any major issues in completion of the development phase of the project, only minor PCB layout tweaks. If all goes well it should be available in November.
Conclusion, next steps...
I am pleased that so far all the boards go together exactly as I had planned. I had to choose the location of through-hole components carefully so that anything high did not interfere with the metal tabs on the rear of the LCD module, and nothing on the controls board interferes with the connectors or any other components beneath it, etc. So far it has all fit together perfectly. I have found only minor tweaks needed to the PCB.
The yellow/green LCD without LED backlight, is daylight readable both in bright shade conditions and in full strong sunlight, as well as through sunglasses.
I expect to receive the SMD components from Farnell UK in about 1 week; and the prototype enclosure maybe two weeks from now. Then I can complete the assembly and testing, make any modifications, and if they are minor and I feel confident, proceed with manufacture.