2008-04-12
PSK-31 Interface for the FT-817
Peter VK2TPM has been encouraging me into PSK-31. He loaned me his old interface to get me going, but I finally got around to building my own.
It is a simple affair, with two isolation transformers, a fixed attenuator for the RX path and a variable one for the TX path. The keying is based on a charge pump driven from the TX audio from the PC.
I based the attenuator values on the sensitivity and input impedance of my Laptop's microphone input with the software "Mic Boost" option enabled. You may need to change the attenuator values if you have a less sensitive input, or plan to use Line-In instead. Making it variable with a pot couldn't hurt, but there is of course some room for adjustment in software.
Similarly for TX you may need to change the attenuator values a bit, especially for other radios with different input sensitivities. For the VOX keying I was fortunate enough that the FT-817ND needs only a 2 mA of current sunk from its PTT line to switch, so a completely passive circuit was practical. With the values indicated the phones/line-out from my laptop drives the PTT just fine from about 1/4 of the way up the slider in windows. Normal level for TX just before envelope compression is about 80% of max, so there is adequate range for adjustment.
Notes
The transformers are a bit of a mystery. Their exact properties are unknown. I got a bag of 20 of them from Rockby in a clearance sale, their stock number was 26222, but they are no longer sold. Despite having a 3rd leg on one side, the "primary" isn't centre tapped, the extra leg is open-circuit to all other pins and the core. The "primary" has a DC resistance of 42 Ohms, the other side 9 Ohms. The turns ratio is 2.5 primary:secondary (determined using 1kHz AC injection). Secondary reactance was measured (rather crudely) at about 9+j770 Ohms at 890 Hz, which is about 137 mH. For the TX path I am using them step-up, for the RX path step-down. The voltage step-up helps with the keying circuit, without it insufficient voltage may be developed for reliable keying.
RF on the keyline was more of a problem than RF on the Data-In line, but the 10 nF capacitor specified seemed to cure it.
I deliberately detuned my antenna to get more "RF in the shack" for testing purposes. If you run higher power you may need to be more careful with the shielding. I am using a 40 metre end-fed half wave (harmonic fed at 20 metres using a new matcher) and the bicycle loop, so I likely have more RF around the shack-end than people using dipoles on long feeders with a good balun.
The matcher is just a polyvaricon (antenna gang about 120 pF max) across a T80-6 with 22 turns on it (~1.5 uH). The link coupling to the 50 Ohm feed was adjusted using my hacked-together return loss bridge and a signal generator.
I needed 5 turns to match my 40 metre EFHW which suggests it has a lower input impedance (around 1000 Ohms) than I predicted, but this is perhaps not surprising considering its proximity to the building. The Q with this load is about 10, which is still fairly sharp, but I was hoping for a somewhat higher Z load.
3 comments.