Saturday, November 13, 2021

QRP digital transceiver kits compared

The technology behind WSJT modes & JS8, and the evolution of various kits that take advantage of them, is something I find very interesting. It's amazing to see the innovation of design that has occurred, in such a very short time, in the four kits described here.

It all started with the Phaser, and it was stunning. Designed not by Mr. Spock, but by Dave Benson K1SWL. I built one for 20m, converted it to 60m and had a lot of fun with it.

Phasers were (they are no longer available) monobanders, available initially for your band of choice, 17 to 160 meters. Later versions made 10 and 15 meters available. Each kit would operate on either of two discreet frequencies to allow FT8 and (usually) JS8.

Power output was ~4 watts and audio cables were required to/from the PC. VOX keying took the place of CAT and a single-step alignment procedure was required for frequency calibration. Modulation was SSB.

Price with a case made from circuit-board material was $80.


Arriving about the same time as the Phaser, and still available, unfortunately, is the D4D from CR Kits.

This kit uses double-sideband modulation and therefore produces as much QRM as usable signal. Just what we need on crowded FT8 bands. The D4D is a monobander for 20, 40 or 80 meters and output power is 1 watt. 

Audio cables to/from PC and VOX keying. The price, with aluminum housing - $35.

CR Kits also makes a new 13TR kit (pictured, and same website as above) for 40m FT8 - this time using SSB.

Output is 1 watt, price with housing is $59. I am unable to find a manual for either of these kits without being a member of the relevant


And with the next two kits we step into a different league:

The new AFP-FSK from QRPGuys and the new QDX from QRP Labs both use a novel method of frequency generation. The QRP Guys explain it in text here and QRP Labs does so in a video here (beginning at 2:10).

Although the two kits share that unique feature, there are some significant differences between them.

The AFP-FSK can operate on any band, 10-160 meters. Plug-in filters define the band in use, with three included in the kit - for 20, 30 and 40 meters.

Four additional filter boards, minus the components, can be bought for $10. Output power is 4-5 watts, depending on band, and audio is routed to/from the PC with two standard audio cables.

The price is $80 with 3 filter boards, no case included (or available).


Finally, the hard-to-obtain QDX. Hans is hoping to be able to offer 375 more of them next month and is already making changes to the board to do away with a mod necessary on the Rev1 kits.

What makes the QDX unique, even compared to the AFP-FSK, is its onboard soundcard for USB connectivity of both audio and CAT, 5-band operation (20, 30, 40, 60, 80 meters), tiny size and no alignment. And nothing compares to a G0UPL-written manual.

Power output is variable (in my case, from 750mW to 5 watts), based on input DC voltage. The QDX can be built to operate at full output with either a 9V or 12V input.

Discussion of the QDX takes place here.

Price with aluminum/aluminium case is $80.




Anonymous said...

When operated on USB, the DSB radios put the LSB signal right in the middle of the PSK31/Olivia/THOR sub-band. My personal opinion is it is very reckless to market something like that to a new non technical ham population. OK for an experimentor to build and use, but not to market.

Dave kx3dx

John AE5X said...

"reckless to market something like that to a new non technical ham population. OK for an experimenter"

I agree 100%.

John AE5X said...

From Hans:

"What they (QRP Guys) call AFP-FSK is the same concept though implemented differently. They use the ATmega328's analog comparator on the incoming audio and "time" a cycle using the timer input capture register, clocking the time with the system clock which is 16MHz. Then add that to the FT8 "USB dial" frequency to get the RF frequency, and command the Si5351A to make it so. It's a neat solution that fixed the problems in the transmit side, at least, of their transceiver.

QDX implements the concept differently because it has the built-in USB sound-card emulation which is known to be running at 48kHz sample rate. Rather than a comparator for zero crossing detection, I have the precise numeric values straight from WSJT-X via the noise-free, lossless digital transfer of the USB Audio interface. The zero crossing is calculated by finding the samples above zero followed by below zero, then doing a straight line linear interpolation between them to find the zero crossing point. No timer clocked by the system clock is used, because I already know the sample rate is 48kHz so can count the cycle duration (time between zero crossings) in terms of that constant 1/48000th second interval then simply convert this to frequency."

Anonymous said...

Today is another banner day for interference on 20m in the PSK and JS8C sub-bands near the FT-8 sub-band. It's definitely someone trying to run FT8 because the timing makes it obvious. I do not know if it's an overdriven 100W box radio or a kit radio or what. A couple days ago I saw a slow locking PLL on transmit with a wide frequency excursion - today it's just plain hash. Hard to track down and notify the op. And if it's someone experimenting - great - but I suspect that its just someone who is not monitoring their signal and they have a problem.

73 Dave kx3dx

BD6CR said...

Thanks for your introduction to my products. I agree with your point that DSB transceiver has the bandwidth issue although I have designed to be QRPp and the decoding threshold of FT8 is much lower than other modes like PSK31 in its LSB portion. I have added some information telling potential customers about the concern. And, I have added more info about specifications of 13TR kit.

John AE5X said...

Thanks for your comments. Would you consider putting a link to the 13TR's manual on your website?


BD6CR said...

I think a schematic will serve you guys better. I will add it in a few minutes.