Tuesday, February 22, 2011


I've been playing with the new 2.0.19RC1 PSDR release candidate from Flex. On the CW front it is fantastic, and if you have any interest in CW I hope you download it and give it a whirl. On its surface it looks like any other version of PSDR but underneath the hood, the CW code was completely ripped out of the radio, and at each stage from the keyjack to RF out it was optimized. Some aspects were moved into firmware. I don't know the exact details, but whatever they did I like!! I normally use and external keyer which is inputed into the radio over a serial port in the computer, but in the original version I tested (2.0.18) this mode of input was broken so I was forced to use the internal keyer. The keying of the internal keyer was fabulous. I was able to take the TR turnaround to less than 5 ms

There have been several knocks to PSDR over the years as a CW radio. If you used the internal keyer there was variability of timing on some of the elements, there were pops and clicks, and there were first dits being cut off, and above all there were complaints of lag. When I first bought a SDR-1000 it was almost unusable as a CW radio for DXing because of the lag. All of that has been eliminated in this version. In addition the radio can do QSK up to about 39 wpm. At 40 wpm it automatically switches to semi-breakin. I ran the radio down to 5ms TR turnaround. The relays are rated to 4ms.

Here is an audio recording of some of what 2.0.19 sounds like. When I set the TR turn around to 5 ms if you listen closely to the background you can hear the relay switching with each character transmitted. This relay sound is NOT part of the audio or RF, but is due to my recording technique which uses a microphone. The set up is a pair of headphones placed on top of an Elecraft K-1 which I use as an external off the air receiver. The K-1 has no antenna. I set the K-1 so the agc was not hard limiting by running 4 watts. You can hear receiver noise in the background. I set PSDR to worse case lag, 48khz audio bandwidth with 512 buffers. CW in DSP was set to 4096 on CW RX and 512 on CW TX. With these settings the filters are at their sharpest but if there is lag to be heard it is also at its worst.

I run first the external keyer (K1EL winkey USB) into PSDR via a usb to serial port converter. I use RTS and as a buffer I use a 2n3904 PNP transistor. The transistor is driven through a several hundred ohm resistor on the base. I have found this to be the cheapest and most effective way to get a CW signal into the radio over the years. I normally work semi-breakin with the TR turnaround set to 60ms. This is how the radio is set up during the first part of the test. Next I switch to the internal keyer, and its behavior is virtually identical to my K1EL keyer. I play the sidetone from the radio then add the received signal from the K1. The K1 is the higher pitched tone. Next is the internal keyer at 60ms TR turnaound. Next I set the radio to 5 ms turnaround and go back and forth between external and internal keyers increasing the speed on each keyer My K1EL is set for a top speed of 45 wpm the internal keyer can go to 60 wpm

2.0.19 audio

I also created an example of how the radio sounds when transmitting on top of a station to simulate QSK. For this demonstration no animals were killed and QRM was caused. I merely set the radio to zero watts output and transmitted over another station on 40M I sent random characters over the other station and my sidetone is slightly louder and slightly lower in pitch than the CW station. If you listen closely to the background you can hear the relay but this is not part of the audio just my recording technique.

2.0.19qsk demo

The release candidate is downloadable on the Flex website down the page to the left.

So there ya go!! Hope you have some fun with the new CW aspect of the radio. I certainly am having a blast..