Some notes about commercial HF amateur radio transceivers for CW/SSB  Contest operations focusing on a relevant parameter of third order CLOSE-IN IMD DR (a two RF tones test)

Take note, from ARRL, RSGB, or personal laboratory test report data of HF Transceivers,   about close in IMD figures DR (two tones spaced apart between 2 to 20 Khz,  that fall just inside the first roofing filter in most commercial upconversion transceivers that, until now, make use of wide first IF "roofing" filters). From 20 khz to 50 Khz usually they are better.

Some relevant links for laboratory reports on Amateur Radio Transceivers:

ARRL (members only)  (ARRL members login only, there is a big document available about Test Procedures)
available some expanded test report and see QST Magazine Product Review, on paper or CD ROM from ARRL.

RSGB (members only) (RSGB members login only) and on paper or CD in RADCOM (RSGB) Magazine

W8JI, Tom Rauch   and Receiver Tests sorted by IM3

Sherwood Engineering Rob Sherwood NC0B  and  Receiver Performance White Paper (I suggest this reading)

Elecraft comparison charts (test data from ARRL)

Some data taken from Peter Hart G3SJX test reports presented in RADCOM (RSGB) Magazine Oct-2000 and June-2002.  Close-in IMD DR  or  IIP3 (dBm) e IMD3DR (dB) with two RF tone test, spaced closely from 3 Khz to 50 Khz  of 2 well known transceivers  (7Mhz, preampl. OFF):

Analysing RSGB, Peter Hart G3SJX test report of a modern rig as IC756PROIII  from a critical “contester”  crowded in-band  signals  point of view, it appears a modest or not at all  improvement over IC756PROII, in this CLOSE-IN IMD behaviour area, until  there will be a mode switchable sharper first roofing filter. Receivers  with sharper and mode related first IF filters are potentially better in this Close-In IMD area.  Even modified up conversion rigs with original broad roofing filters end-up to be improved with a sharper roofing filter placed inside. This is a nice  experimental  amateur radio exercise seeking for any  improvements.

Here there are some recent examples officially documented by ARRL and RSGB test report:

Some data resumed by QST, RADCOM for comparison  about CLOSE-IN IMD:

ARRL laboratory, February 2005 QST,  tested Inrad roofing filter mod inside an FT1000MP and Mark V two tone IMD DYNAMIC RANGE (14 Mhz, pre off) for:

(*) measurement is reported to be noise limited at that value)

Also Peter Hart, G3SJX in January 2005 Radcom (RSGB magazine) (FT1000MP, 14 Mhz, CW 500 Hz, preamp OFF)

(Inrad roofing filter is  supposed to be around 4880 Hz wide since AM position bandwidth has been measured by Peter Hart to be narrowed from 7 Khz to 4.880 Hz - See INRAD web site and the pdf document "a few words about Roofing Filters) which contains the following grafics about INRAD 70.455 MHz first IF roofing filter mods for FT1000MP and FT1000MP-MARK V and FIELD version Yaesu TRX:
 

                                        Inrad 70.455MHz Roofing Filter:

If you look carefully to Ten Tec ORION close in IMD data (3 Khz to  20 Khz   2 tone test) from ARRL QST report

it appears that, in this area, it really tops every other rig now available (FTDX9000 not tested yet) and it has also a much better LO Synthesiser Phase Noise which really tops almost all other rigs and  a sharper ultimate bandwidth is achievable, see bandwidth figures at attenuations below 60 dB.

Anyway a nice rig as IC756PROIII has so many features that it looks really a desirable rig especially now with the improvements released in MK III (really missing only the IC7800 switchable sharper first Roofing filter), so I suggest to get and read the complete test report in RSGB or ARRL magazines and final choices are personal and you should always listen to a rig personally to satisfy your needs and tastes.

Third order IMD DR with two RF tone at antenna input  is mainly a FRONT-END (early input RX stages) test and

stresses all stages before the main sharp IF selectivity, with a close in imd test also the roofing filter, and it depends a lot to how  sharp is selectivy in the early stages and to the linearity of the components (also passive one like inductors in front end filters can contribute to deteriorate or improve IMD immunity, even the roofing filter should be well implemented, see the larger coils used in IC756PROIII or coil selection in CDG2000 home-brew transceiver and IMD test report by G3SBI in Radcom June-Dec 2002).  The CDG2000 9Mhz Roofing Filter is very much interesting. It basically a four pole, 4 quartz, for the in series 9 MHz IF signal but in parallel with others 4 quartz with an input  and output splitter

so improving 3rd order in band IMD, and it provides also a good 50 Ohm match to the mixer - diplexer. Here is our duplicated unit and how it performs, tests are performed with a Vector Network Analyser VNA by N2PK with the professional  help of its builder  Eraldo, I4SBX.

CDG2000 9MHz Roofing Filter by I4FAF and IK4AUY  Total Span is 10KHz  or 1KHz per division Insertion loss is - 2.5dB to max - 4 dB in the upper portion. Bandwidth at -6dB is 3.4KHz,  at -60dB  BW is 17.8KHz

In the above Return Loss picture you must add 12 dB to take in account the measurement head system loss so max Return Loss  is -12dB and minimum -32, that means VSWR is  better than 1.5:1  a very good match to 50 Ohm for a quartz filter input and output splitters work well.

CDG2000 9MHz Roofing Filter by I4FAF and IK4AUY    50KHz total Span or 5KHz per division

IN-BAND AUDIO IMD

To get an idea about the receiver quality delivered to our ears, in good phones,  it is interesting to see what happens in the  following inside RX stages like  IF Amplifiers, AGC, Detector and Audio. In the ARRL Expanded tests Report you can often find a two tone IN-BAND IMD test  in which the 2 RF signals are applied at the RX antenna input at very close spacing for examples  200 Hz apart and the unwanted IMD products  at base band audio frequency should be attenuated as much as possible  (to say an higher absolute value in dB is better).

If the IF AMPLIFIER chain is contributing to the inband IMD more than the Detector or the Audio chain the IMD has

seen at audio is the poorest of the poorer stage that contributes to it.
Audio quality is also affected by  frequency bandwidth response so it is worthwhile to measure it.

 last update 02.05.2005