472kHz (or 500kHz) 10W Transverter

This page details the build of a transverter for the new 630m amateur band,  released in the UK Jan 1st 2013.  It will be a work in progress with latest version nearest the top of the page.

Results on 472kHz have been very encouraging with WSPR reports from many stations in Western Europe so far out to 990km with an ERP of around 10-20mW.


The lastest (Rev K) schematic with a couple of errors corrected and an additional LPF added

DX with the new transverter- Finland  22.9.12

This evening OH1LSQ (1736km) copied my 500kHz WSPR signal with the new transverter described below. In the last 3 DX seasons on 500kHz OH1LSQ has managed to copy me when I have been using different rigs and antennas. On both the previous occasions I was using an 80m sq loop antenna whereas this evening it was just the 6m vertical coax to my 10m halo tuned against the copper waterpipe ground!

Stations receiving my 500kHz WSPR this evening (22.9.12) with the new transverter described below.
OH1LSQ's WSPR screen showing my (weak) signal as received in Finland

Design 21.10.12 (less external low pass filter)

The transverter boxed and finished

The latest version of the transverter saw me go back to a design using a double balanced mixer. I had issues with the transistor mixer (stability problems that were hard to fix) and with the double balanced mixer the circuit is working very well. I have built a couple of samples with almost identical performance. The whole unit is currently housed in a small plastic box but I think a larger metal box is advisable so that the PA heatsink heat does not cause the LO crystal to drift. I am seeing about 1Hz movement during a WSPR 2 minute transmission cycle because of this. This is OK but could be eliminated with more thermal isolation between the PA and LO. There would be no problem with CW.

The transverter during final bench tests

The transverter, ferrite rod ATU and antenna current meter on the desk with the FT817

The latest version is built on a small piece of copper laminate "dead bug" style so that it will fit in a small diecast or plastic box. Note the inductors are now 5.9uH both for the internal low pass filter in the PA output and for the PA drain choke. These are wound on small cut pieces of 22mm PVC pipe. Best way is to wind turns and measure the inductance with an LC meter. This is what I did. The number of turns will vary depending on how tight the coil is wound and on the exact wire gauge. I think my wire is 0.5mm diameter enamel copper. An additional LPF has been added "outboard" to clean up the harmonics in the MW band. This may not be needed if using a sharply tuned antenna like a loop or short Marconi.

Notice I have used WIMA 600V or 1000V capacitors around the output circuit. This is probably overkill, but I had them available from my loop matching circuit so used them.

There is a video describing the earlier version of the transverter on my You Tube channel.

Converted to full transceive (21.7.12)

Added the RX part (initially without preamp). Just 4 extra parts needed to make it fully transceive. Sensitivity of the FT817 via the transverter is around -120dBm MDS on 472kHz and -124dBm on 500kHz, so arguably just about OK "as is" without a preamp, although I will try this shortly.

The next stages will be to sort out the PA stage which has a very slight tendency to self oscillation because of either incorrect gate biasing or insufficient supply decoupling, easily fixed.

This circuit is likely to change further before I arrive at a point when it is considered finished!

First breadboarded TX transverter (20.7.12)

Having had a consistent 10Hz drift every WSPR 2 minute burst I decided to go for a lower IF instead of the 28MHz IF used in my original 2010 schematic.  A 3.2MHz crystal was obtained cheaply from the GQRP club sales and this was used to breadboard the TX transverter part at the weekend. Instead of the SBL1 TX mixer I have used an inexpensive 2N3904 mixer followed with a simple low pass filter feeding essentially the GW3UEP PA design with a simple common emitter, feedback biased squarer stage. This produces a clean10W RF from the PA running from a 13.8V supply. Drift on WSPR is now zero.

This is the schematic so far.
Click the image to get a larger, clearer view.

It has been air tested with my short 6m long coax feed to my 10/6m halo as a top loaded vertical tuned with a ferrite rod ATU on the bench in the shack and central heating system ground. best DX reports on WSPR 560km. It also works fine with OPERA Op4 mode on 500kHz.

Next stage of the development is to try a higher power PA design and to add the RX preamp and switching. I will continue to use the FT817 in split mode receiving directly on 472 or 500kHz. As the FT817 sensitivity drops off in this range a well filtered, high dynamic range preamp is needed.

When I am satisfied with the whole transverter it will be rebuilt into a small box.