开云体育

I'd like to consolidate this discussion in a single place as multiple lists are getting confusing.

I'm quite impressed by the uSDX design concept, though less so about any of the implementations. This will proceed to modularize the radio in a series of steps.

First up is a 50 ohm stripline relay switched suite of filter boards with 2-14 filters per board in even numbers. The filter footprint will accommodate 2 fixed SDM caps and a trimmer per leg with shunt inductor pads for each element of a 5th order LC filter. Connections will be SMA and IDC ribbon cables. I plan to design both LC Rx roofing filters and Tx LP filters for all the bands from 160 to to 70 cm. I am pressing a friend to do the PCB work for this board. A full fat version will have both Rx & Tx filters.

The 2nd board will be the main RF board with provision for choice of crystal, internal TCXO and external clock or sine wave references via U.FL for the Si5351. There will be also be a precision PA drive attenuator and a diode noise source in addition to the current variable power out.

All control functions will be performed by a front panel with a Nano compatible pinout via a mezzanine board. The panel will have an as yet unknown number of switches, a 240 x 135 or 320 x 240 OLED display, one or more encoders, etc. Provision will be made for using an OXCO to autocal the TCXO and then turn off the OXCO. . By using the Nano as the base, the existing FW can be used followed by a port to the Mega256. Goal is to internally implement all the digital modes on the MCU with Bluetooth and WiFi control interfaces by using something like a Heltec ESP32 module with a radio chipset

Internal power will be supplied from a 3-5 Ah LiPo with built in charger operating off the single 13.8 V power input.

The aim is to produce a set of OSHW modules which can be connected to form an enhanced uSDX+ with proper RF shielding on all boards. Or alternatively some other architecture such as a superhet by adding crystal IF filters ahead of the quadrature. I have enough experience comparing the sensitivity of the uSDX+ to my Icom 705 to I think that a sub $300 retail radio with lower current consumption and comparable performance to the 705 is viable with a bit of community engineering effort.

As it stands, the uSDX+ has better FW than the $600 Xiegu X6100. So I'm returning one of those and will exclusively focus on the uSDX concept until such time as the X6100 becomes a viable radio. the current FW is pretty useless.

I have the design and programming skills, but no PCB layout skills and not enough time to complete other projects and learn KiCAD. I also have a very well equipped RF lab and would prefer to spend time on extensive testing.

I'm hoping that someone else will take an interest in this and help with the design and layout. JLCPCB has cheap multilayer boards so real microstrip is inexpensive.

Have Fun!
Reg

I have made many PCBs, and I have 40+ open source hardware projects, and I am currently working on my own QRP rig



The rig uses the USDX finals, but currently I use pluggable band modules rather than relay switched modules because I don't really find changing modules a big deal.? I could be persuaded though otherwise.

My interests in making a QRP rig are cheap and easy to make and longevity, that the design should be buildable into the future.? That has informed my choice of parts.? I use the STM32F103CBT6 processor, which while not the most modern processor, is still an ARM processor, and has many clones.? I am using the SI4735 RF receiver integrated circuit, because it will be much easier (in my opinion) for someone building a radio to get good results out of this circuit.? It has been around for awhile and probably continue to be so until AM/FM radios are no longer a thing.? I am using the SI5351A because, well, it's the best choice of frequency synthesizer still.? I am using a HD44780 alphanumeric LCD because they are so easy to get and cheap.?? I also decided to add a SD card to it so that the radio can receive and store data that it decodes, or it can store configuration information on the card.

I will say that the SSB phone output of uSDX would probably not be acceptable in any commercial product, especially one that has to comply with FCC regs.? The audio output quality is not great and is much broader in frequency that the audio signal.? One of the major reasons uSDX is so cheap is because it doesn't have quadrature encoders for the SSB as well as the needed DACs for phased SSB output.? I am not sure how much that would cost.

Spending $300 for a somewhat better uSDX quality radio may not be what many operators are willing to do if the alternatives are not significantly more expensive.

73,
Dan
KW4TI

Pluggable band modules make a lot of sense, especially in an experimental radio.

I need to look at the 4375 as I'm not familiar with it.

The Mega328P has a 15 kSa/s 10 bit ADC. That provides a 3.75 kHz Nyquist for each I &Q stream. A commutating mixer aka Tayloe detector, provides a precise quadrature output at very low cost. The effective Nyquist is 7.5 kHz. But if you don't have the compute to process them all is for nought. So adding discrete calculation of the Q stream by Hilbert transform is a complete waste of already very limited resources. There is no increase in complexity and it frees up flash.

The output of a Tayloe detector is the basis of all the sound card based SDRs and the failure to take IQ outputs from a quadrature detector a significant design gaffe in my view.

For $300 I'd expect a 2-4" 320x240 OLED, several push switch encoders and a 16 switch keypad in an extrude enclosure with metal panles. I will not use toroids in the filters because they are expensive to wind and simply not needed The entire filter block should be shielded and there is only a single filter active at any time, so the coupling problem associated with solenoid form coils doesn't exist. That's a huge cost savings and increase in manufacturability simply by using SMD standard inductors.

The key goal is a 320x240, pixel per RBW waterfall with a 1, 5, 10, 50, 100... sequence from 1 Hz to 1 kHz


Have Fun!
Reg

--
Dale Hardin KS4NS
Elberta, AL

--
Dale Hardin KS4NS
Elberta, AL

Hi Reg,
New member here, g’day to all.
Have been playing with the QCX-SSB since it’s debut, I realised that I can “gate” modulate a zero voltage “drain-source” switching mode class E amplifier with distorted SSB, -5dBc IMD3. Have you experienced with uSDX+ SSB distortion ( I don’t own ?an uSDX+) and any thought how to integrate the SSB envelope to the SSB phase modulated carrier. Hans G0ULP in his QMX (QRP Labs) uses drain modulation for his switching mode class D amplifier, but the SSB firmware has not been related yet.
Cheers and thanks for let me participate in this great group.
VK2IHL, Pascal

This thread is really aimed at the component & PCB level issues. So I'm moving Pascal's SSB issue to its own thread.

Have Fun!
Reg