PSK-series- Summary of known issues
1) Inconsistent T/R switching.
The first variant of the PSK-series used serial port switching, and was trouble- free. - until new computers and portable devices no longer had serial ports. The existing circuitry was reworked to sense audio from the sound card.
This had several potential issues. Given the gain distributions for the driver and PA stages, the sound card adjustment range between ‘switches to transmit reliably’ and gain compression (IMD) could be quite narrow. T/R switching transients occasionally resulted in motorboating.
Recommended fix: Establish a high audio level from the sound card and add a level adjust trimpot to the Tx audio chain to adjust output power independently of the sound card setting.
2) Transmit chain:
The driver and PA devices are obsolescent and expensive ($4 and $12 respectively. Builders seeking replacement parts may encounter counterfeit devices from questionable parts suppliers
Recommended fix: Redesign transmit chain to use currently-available parts. The RD16HHF1 ($4) MOSFET may be a viable and readily-available candidate.
3) Transmit bandpass filter:
This circuitry was trouble-free for the first two versions of the PSK-series. It used two Toko tunable inductors, nominally 1 uH value for the 20M version. Those inductors are no longer available, nor is there a drop-in replacement.
The third (and most recent) version of the PSK-series used Mouser # 42IF123 IF transformers, modified by removing the internal parallel capacitor. These have a nominal inductance of 6 uH- an impractical value at 14 MHz. I added a 33 pF capacitor in series with the two inductors to cancel much of the high reactance.
In so doing, I had to remove ‘TP1’ from the board- very crowded in that area.. Further, the series L-C combination proved to be very ‘sharp’-tuned. I got too many reports of ‘no power out’ as builders couldn’t find the sharp peaks on the tuned circuits.
Redesign tuned circuits to use toroidal inductors and a combination of fixed capacitors and trim-caps. This requires additional board real estate. Install test point diodes at both tuned circuits.
1) Solenoidal inductors (RF chokes) not recommended due to layout/proximity coupling effects and resulting instability.
2) Trim cap rotors need to be grounded; otherwise, a non-metallic tuning tool is required
4) Sideband convention:
The PSK-20 uses Lower sideband (LSB) instead of the conventional USB on 14 MHz. The two local oscillator (LO) frequencies were set at 5068.5 and 9004.5 kHz. This arrangement is fine for (B)PSK31, which is insensitive to the ‘sideband sense’. It will not work on other popular modes (e.g., JT65).
Move 9 MHz LO to 9000 kHz- this is readily achieved with an inductance in series with the crystal.
Move 5.068 MHz LO to 5.073 MHz. I doubt this will work without a custom crystal. At this point, the designer may wish to consider a DDS local oscillator. These are readily available and inexpensive (<$10) search on ‘AD9850 DDS’.They will require a serial interface to program, typically in the form of a PIC, Arduino, Atmel or other microcontroller. The ‘up-side’ is that the rig is no longer confined to a 3 kHz slice of the spectrum. It would then be frequency-agile and would work with any mode consistent with a 3 kHz passband-width. Like voice (SSB), for instance.
Note: don’t even consider an L-C local oscillator! Drift rates exceed the allowable tolerances for the narrowband modes..
There- that’s all there is to it! Feel free to use the PSK-series schematic from my website as a starting point. I now consider it to be in the public domain. I am not interested in making the necessary design changes. A new board layout will be required, and I have no interest in doing that work. Nor do I want to be ‘Mister Answer-Person’ for other people’s commercial products.
If you’d like an engineering- quality (no masks or silkscreen) set of Gerber files for the most recent layout, simply ask. You’re on your own, though!
I’m retired. It’s time to move on. 73- Dave Benson, K1SWL 29 Jan 2014