Use of Comparative Analysis to estimate the DX Prowess of HF Receivers

By Tadeusz Raczek, SP7HT

Skrytka pocztowa 728

25-324 Kielce 25

POLAND

sp7ht@wp.pl

 

Foreword

My article, „The DX Prowess of HF Receivers”, was published in the Sept/Oct 2002 issue of QEX. The inspiration to return to this subject has arisen during my recent translation work. Not related to my ham radio hobby, note that Poland became a regular member of the European Union on May 1^st, 2004. I have been involved in the translation of papers related to Local Government Finances. These papers commonly use a Comparative Analysis to study the relationship between factors in different countries. I have quickly recognized it would be a good idea to use that analytical method to compare the usefulness of receivers for DXing.

The ARRL Technical Laboratory has published Orion test results (ARRL Website: December, 2003 and QST: January, 2004). A month later, test results of upgraded K2/100 appeared in February 2004 issue of QST followed by their review of the TS-480 and IC-7800. This triggered my second impulse to return to that subject. Being an ARRL member and having access to the ARRL Website I became a regular reader of Product Reviews Reports of HF Transceivers.

In this article you can see my findings and conclusions in this analysis and in my Table. I have also included some graphs of swept BDR, IMD DR3 and Phase Noise measurements (taken from ARRL Technical Laboratory Extended Test Reports) to point out differences between “good” and “not good” receiver front end design.

 

The Table

The far left and the far right columns in my Table perform ordinal functions: the left one designates the HF Transceiver model and manufacturer and the right one designates in which QST issue the test results have been published. The position of other columns (from left to the right) reflects “the weight” I apply in estimating the influence of a particular RX parameter on the DX prowess of HF receivers. So, in my opinion, BDR at 5 kHz (the second column) is the most important parameter for me and MDS (the tenth column) is the least significant RX parameter. To continue with ordinal matters: I have put in my Table HF transceivers of the following manufacturers (alphabetically): Elecraft, ICOM, Kenwood, Ten-Tec and Yaesu. Each of their HF transceivers are (from top to the bottom) in a historical order (as test results have been published in QST). That makes it possible to evaluate changes of some receiver parameters during the last 13 years.

My intent was to gather measured receiver parameters in the form of a Table which reflects “the weight” I apply to each parameter to estimate the DX prowess of HF receivers. After putting all results in the Table, I have been most interested in technical parameters. But, after a preliminary analysis of my Table, I have also found some historical aspects which I had not expected when I started this project.

Using the same sources of data, somebody can make a different Table, reflecting one’s priorities in our ham radio hobby. To be correct, there are many specializations in our hobby, and respectively, many different priorities concerning properties offered by different kinds of equipment. Therefore, I want to point out that my Table is constructed to reflect the DX prowess of HF receivers. I hope that DX oriented ham radio operators will be interested in its contents.

 

This place please put my Table {if possible, the whole Table on the same page (or neighboring sheets) to make comparisons easy}

ARRL Tech Lab Receiver Test Data Comparisons (Selected by SP7HT as on July, 2004)

All receivers have been measured in the following conditions: (Preamplifier OFF) / (Preamplifier ON); 14MHz; 500 Hz CW filter (or any bandwidth closest to 500 Hz); AGC OFF; High IP mode. All test results are respectively in dB, dBm and dBc/Hz.

I have used color and shading in my Table. The highest / best results of a particular parameter are shaded in gold. The next group of results are high grade results (without shading). In the third group are good results. Good results are shaded in gray. The fourth groups are poor results. Digits are in red color. In my opinion, three best receivers for DX hunting are: Orion, K2/100 and IC-7800.

 

TRX Producer           (1)	BDR (dB) Blocking Dynamic Range at  5kHz   (2)   Preamplifier OFF	IMD DR3 (dB) Two-Tone Third Order Dynamic Range at 5kHz   (3)   Preamplifier OFF	Phase Noise (dBc/Hz at 4kHz)       (4)	IP3 (dBm) Two-Tone Third Order Intercept Point at 5kHz   (5)   Preamplifier OFF / ON	BDR (dB) Blocking Dynamic Range at 20kHz   (6)   Preamplifier OFF / ON	IMD DR3 (dB) Two-Tone Third Order Dynamic Range at 20kHz   (7)   Preamplifier OFF / ON	IP3 (dBm) Two-Tone Third Order Intercept Point at 20kHz   (8)   Preamplifier OFF / ON	IP2 (dBm) Two-Tone Second Order Intercept Point     (9)   Preamplifier OFF / ON	MDS (dBm) Minimum Discernible Signal     (10)   Preamplifier OFF / ON	Published in QST *           (11)
K2 Elecraft	126	88	-124	No test result	136 / 128	97 / 98	+21.6 / +6,9	+75 / +76	-131 / -138	March, 2000
K2/100 Elecraft	135	91	-124	+21 / +8	134 / 126	97 / 95	+21 / +8	+80 / +79	-130 / -136	February, 2004
IC-781 ICOM	No test result	No test result	No test result	No test result	134 / 132.5	102 / 97	No test result	No test result	-134 / -140	January, 1990
IC-736 ICOM	No test result	No test result	-110	No test result	121 / 130	95 / 92	No test result	+59	-133 / -139	April, 1995
IC-738 ICOM	No test result	No test result	-118	No test result	119 / 119	94 / 94	No test result	+61	-133 / -139	April, 1995
IC-775 DSP ICOM	No test result	No test result	-117, but peak: -110 at 17.5 kHz	No test result	136.7 / 132.2	105.7 / 103.2	+20.85 / +11.6	+55.7 / +55.2	-137.7 / -143.2	January, 1996
IC-706MkIIG ICOM	86	No test result	-118, but peak: -108 at 15 kHz	No test result	122.2nl / 119.5nl	89.2 / 85.5	-1.3 / -11	+36.4 / +38.5	-136.3 / -141.5	July, 1999
IC-756PRO ICOM note 2	104	76	-130	No test result	126.6 / 125.2 / 120.3	94.6 / 92.2 / 88.2	+15.4 / +4.3 / -4.2	+63.7 / +62.6 / +42.8	-127.6 / -136.7 / -140.3	June, 2000
IC-756 PRO II ICOM notes 2 and 3	100 / 97 / 94 (after repair)	76 / 75 / 72 (after repair)	-130	-18.8 / -28.8 / -35.5 (after repair)	118 / 116 / 111 (after repair)	97 / 95 / 91 (after repair)	+20.2 / +10.2 / -4.1 (after repair)	+75.6 / +70.7 / +58.9 (after repair)	-130.5 / -139 / -141 (after repair)	February, 2002
IC-746PRO ICOM note 2	100 / 97 / 94	74.9 / 73.5 / 71	-123, but peaks: -106 at 1 kHz -120 at 10 kHz	-18.25 / -28.2 / -35.55	125 / 122.6 / 117.9	96.9 / 95.5 / 92	+20.0 / +9.3 / -1.8	+72 / +71 / +53.9	-131.9 / -139.2 / -142	May, 2002
IC-703 ICOM	95.1 / 95.4	76 / 76.3	-118	-14.4 / -20.6	120.6nl / 121.9nl	89 / 90.8	+11.1 / +1.9	+56,2 / +46,8	-131 / -140.8	July, 2003
IC-7800 ICOM	115 / 112 / 110	89 / 84 / 83	-120	+22 / +7.7 / +0.5	137nl / 138nl / 135nl	104 / 103 / 102	+37 / +21 / +11	+98 / +87 / +84	-127 / -138 / -142	August, 2004
TS-850S Kenwood	No test result	No test result	No test result	No test result	148 / 138	99 / 99	+7.5 / +17.5 (IPO ON)	No test result	- / -141	July, 1991 (8,83MHz; 455kHz)
TS-950 SDX Kenwood	No test result	No test result	No test result	No test result	131.8 / 133.9 (AIP ON)	94 / 95	No test result	No test result	-138 / -127 (AIP ON)	December, 1992
TS-50 Kenwood	No test result	No test result	-115	No test result	113 / 109	90 / 88	+3 / -7	No test result	-132 / -139	September, 1993
TS-870S Kenwood	No test result	No test result	-116	No test result	127.2 / 123	96.7 / 95	No test result	+62.7 / +63	-128.7 / -139	February, 1996
TS-570S (G) Kenwood	No test result	No test result	No test result	No test result	115nl / 115nl	98nl / 97nl	+21.7 / +9.6	+60 / 58.4	-130 / -139	May, 1999
TS-2000 Kenwood	103.4	68.9	-118 but peaks: -108	-14.5 / -28.8	125.6nl / 120.8nl	93.9 / 92.4	+18.5 / +4.2	+59 / 58.,4	-128.9 / -137.4	July, 2001
TS-480 Kenwood	98 / 91	75 / 71	-120	-18 / -32	123 / 115	98nl / 99nl	+26 / +12	+64 / +63	-133 / -141	June, 2004
OMNI 6+ Ten-Tec note 1	119	No test result	-117	No test result	- / 123nl	- / 97	- / +12	- / +58	- / -133	June, 2000
Orion Ten-Tec	130.2	93	-138	+22.1 / +11.4	129.3 / 128.3	95.3 / 94.0	+22.8 / +12.9	+63 / +63	-128 / -136	January, 2004
FT-1000D Yaesu	No test result	No test result	No test result	No test result	>143 / >154	98 / 98	+21 / +10	No test result	-126 / -137	March, 1991
FT-1000MP Yaesu	119	83	-117, but many peaks	No test result	142 (Off) 137 (Flat)	96.7 / 93.5	+15 / +5	+85.9 (Off) +87.5 (Flat) +87.5 (Tuned)	-127.9 / -135.5 / -135.5	April, 1996
Mark V FT-1000MP Yaesu	No test result	No test result	-123, but peak: -116 at 13.5 kHz	No test result	128.7 / 133	100.7 / 97.6	No test result	+68.3 /  +68.5	-126.7 / -134.6	November, 2000
FT-817 Yaesu	No test result	No test result	-117 but peaks: -113 at 6.2 kHz, -115 at 8.6 kHz	No test result	106 / 103.9	86.7 / 84.3	+5 / -5.6	+84 / +84.3	-125.9 / -134.3	April, 2001
FT-1000MP Mark V Yaesu	107	73	-117	-5.2 / -15.6	122 / 122	98 / 97	+20.3 / +11.5	+68 / +64	-125 / -133	August, 2002
FT-897 Yaesu	96nl	66.6	-108	-24.25 / -32.2	108.5 / 106.3	88.6 / 86.2	+1.25 / -6.7	+67.5 / +61.6	-132.6 / -137.2	May, 2003

 

Notes:

*              QST Product Review columns and ARRL Website: www.arrl.org/members-only/prodrev

No test result: some parameters have been measured from mid of 2001, some 1996 and later.

Note 1: only one set of numbers is listed because that radio uses a fixed gain front end which is not switch-able. Compare their BDR, IMD DR3, IP2 and IP3 results to other receivers by using similar MDS to determine which set of results to use (Preamplifier OFF / Preamplifier ON) for comparison.

Note 2: the IC-756PRO & PROII, IC-746PRO and IC-7800 have two different gain Preamplifiers. Results are for Preamplifier OFF / Preamplifier #1 / Preamplifier #2. Preamplifier #2 adds more RX gain than Preamplifier #1 at the expense of dynamic range. Higher Preamplifier gain reduces these receivers dynamic range.

 

Why Are The Parameters in The Order They Are?

Since my early beginning as a ham (I got my first license in 1957 as a young boy) till now (actually a retired old man), I am entirely devoted to DX hunting. I estimate the value of a rig in respect of its usefulness for DXing. That is natural. I am not only a ham radio (HF) educated person, but I have also spent the last 35 years in professional ground and satellite microwave telecommunication. Assuming my life experience in radio communication I feel free to have my own personal opinion in estimating the usefulness of HF receivers for DXing.

 

The Four Most Important Receiver Parameters:

·       BDR at 5 kHz,

·       IMD DR3 at 5 kHz,

·       Phase Noise at 4 kHz,

·       IP3 at 5 kHz (if measured, because usually is known only IP3 at 20 kHz).

There are also other receiver parameters influencing the ability to copy weak DX signals in the presence of strong signals near the DX frequency, which I recognize being less important compared to these four. I will discuss them later.

I do not explain the meaning of these parameters. I hope anyone interested in DX hunting knows them. If anyone would like to refresh his memory, he can look at the ARRL Website (shortened version: QST October 1994, pages: 35 – 38; detailed version: Test Procedures Manual – 159 pages / 1,47MB).

I have used in my Table point test results closest to the listening frequency. In case of the ARRL Technical Laboratory tests it is 5 kHz spacing for BDR and IMD DR3 manually point measured parameters. DXers wants to know these parameters also for closer spacing approximately 2 or 1 kHz from the listening frequency. Such results are presented on several Websites.

But it is difficult to compare test results measured with different test procedures, measurement test set-ups and so on. Trying to be consistent, I have used only one source of data. There is also a danger “deriving directly” 2 or 1 kHz “point” test results from Swept BDR or IMD DR3 graphs. The ARRL Technical Laboratory specialist states: “The computer is not skilled (yet) at interpreting noisy readings as a good test engineer, so in some cases there are a few dB difference between the computer generated data and those in the "Product Review" tables. Our test engineer takes these numbers manually, carefully measuring levels and interpreting noise and other phenomena that can affect the test data." And there is also second statement from the ARRL Technical Laboratory: “We are still taking the two-tone IMD manually."

These statements mean for me that somebody unfamiliar with all measurement set-up and procedures nuance cannot use data derived “directly” from Swept BDR or IMD DR3 oscilloscope graphs close to the listening frequency instead of point results measured manually. Therefore I have consequently used in my Table only manually measured test point data at 5 kHz and 20 kHz spacing.

Let us begin with the four most important parameters. Each DXer does remember: “5 up please”. 5 kHz from DX frequency is the place on the band, which is usually heavily occupied during a DX pile-up on HF bands. A typical DX pile-up is 5 to 10 kHz wide. In the case of very heavy DX pile-ups, such as in case of a new DXCC entity, the DX pile-up could be wider than 10 kHz. For any DXer the most important information is: “how a particular receiver responds to many very strong signals 5kHz (or so) apart from the frequency of the weak DX signal?”

 

I think, BDR at 5 kHz is the most important RX parameter. Measured values are in the second column in my Table.

In a typical DX pile-up there is a quite big probability that any signal, strong enough, situated close to the DX station’s frequency, can cause receiver gain compression or receiver desensitization. Receivers have different levels of immunity against these negative effects. A particular receiver’s immunity depends on its front end design. BDR - Blocking Dynamic Range - describes a particular receiver’s ability to maintain its sensitivity (or not to become desensitized) in the presence of a strong undesired signal on an adjacent frequency. HF transceiver producers prefer this parameter to be measured far from the listening frequency (wide spacing measurement produce “optimistic” results, giving higher BDR values, compare BDR measured close or very close to the listening frequency). On the contrary, DXers want that parameter to be measured as close to the listening frequency as possible. Why? Because measurements made close to the listening frequency reflect real receiver abilities to copy extremely weak signals from a DX station in the presence of strong signal (signals) on adjacent channels. Since mid-2001 the ARRL Technical Laboratory has measured BDR not only at 20 kHz, but also at 5 kHz spacing from listening frequency. For HF bands these measurement are made on 3.520MHz and 14.020MHz. I use in my article results measured on 14.020MHz.

It is not true what some Ham Radio operators claim: these effects are “not present” in their receivers. Such statement is false! Blocking / desensitization effects could appear in any receiver, even in such superior receivers as Orion or upgraded K2/100. In a real DX pile-up blocking / desensitization effects appear as less comfortable receiving conditions of weak DX signals in the presence of strong signals on nearby frequencies. This is only a matter of how strong the signal on a nearby frequency is. The blocking / desensitization level is very high for some “good” receivers (for example: K2/100 and Orion) but in “other” receivers, (for example IC-706MkIIG, FT-897 and others – see Table) blocking / desensitization level is significantly lower.

Why do I recognize BDR at 5 kHz as the most significant parameter for estimation of the DX prowess of HF receivers? Because the probability of blocking / desensitization effects, even by a single but strong signal, is higher than any other negative effects caused by strong signals on nearby channels. Any strong signal anywhere outside of the listening channel receiver selectivity skirt, can cause this effect if it is strong enough to surpass a particular receiver’s dynamic range. In such a case, it is a fact that with the appearance of that strong signal, receiving conditions of a weak DX signal will be deteriorated.

Blocking Dynamic Range = Blocking Level – Noise Floor. BDR is expressed in dB relative to receiver Noise Floor. BDR at 5 kHz spacing is usually measured with Preamplifier OFF. Receivers offering BDR more than 120dB at 5 kHz are recognized as high grade receivers for that parameter. BDR values above 130dB are actually top grade ham radio HF receivers. Only receivers in the Orion and K2/100 HF transceivers have such high values of BDR at 5 kHz spacing. The two next receivers in turn are: FT-1000MP and OMNI 6+ HF transceivers with the BDR at 5 kHz value of 119dB. The OMNI 6+ result is respectable; because it is measured with Preamplifier ON (there is no possibility to turn Preamplifier OFF in OMNI6+). The next one is the newest IC-7800 with BDR at 5 kHz value of 115dB. Receivers with BDR at 5 kHz around 110dB and below that value (see Table) are less useful for DXing in today’s DX pile-ups.

Receivers equipped with narrow band crystal filters in the first IF (among others: K2, OMNI6+, Orion) have a much higher value of BDR at 5 kHz than general coverage receivers equipped with a 15 kHz wide filter in their first IF. Except FT-1000MP (worse by 16dB) and IC-7800 (worse by 20dB), all receivers with wide filter (15 kHz) in their first IF have their BDR at 5 kHz significantly worse than the upgraded K2/100. In the case of IC-756PRO it is worse by 31dB (more than 5S units on the S-meter scale) and in the case of IC-706MkIIG worse by 49dB (more than 8S units on the S-meter scale). Other receivers in my Table are contained between 31dB and 49dB marks. In case of a wide first IF filter, BDR is quickly decreasing as a strong signal is passed inside the pass-band of that filter. That is clearly seen from the Swept Blocking Dynamic Range graph 2 (for TS-2000) and from data in the second column in my Table.

This place please put Swept Blocking Dynamic Range graphs from Extended Test Reports:

1.                   K2 (page 19)

2.                   TS-2000 (page 28)

Figure 1 is a representation for receivers equipped with a narrow bandwidth crystal filter in first IF (K2 is an example). There is a very narrow, deep null only around the proximity of the listening frequency. Figure 1 shows a deep null on 14.020MHz of approximately 12dB down compared to the flat part of the graph.

Figure 2 is an example of a receiver equipped with a wide filter (15 kHz) in its first IF. For that case, the deep null in the proximity of the listening frequency is not only significantly wider (about 45 kHz wide!) but is also remarkably deeper: almost 53dB down from 130dB level. Additionally, the flat part of the graph does not exist at all: we can see BDR fluctuations even 150 kHz apart from 14.020MHz.

To illustrate that relationship I have intentionally chosen the worst Swept Blocking Dynamic Range graph measured since mid of 2001 for TS-2000. Other receivers equipped with a wide filter (15 kHz) in the first IF have a little bit “better” graph (a little bit narrower deep null around the listening frequency and there are flat parts far away from the listening frequency). By comparing these two different approaches of receiver front end design, anyone can judge which one is better for DX hunting. I can precisely say that K2, OMNI6+, Orion and other receivers equipped with narrow bandwidth crystal filters in their first IF are narrow as “needle eye” and open only for desired signals somebody wants to receive. In contrast, receivers equipped with wide filters (15 kHz) in their first IF “are as wide open as barn doors”, not only for the desired DX signal, but also any other disturbing signals (including DX pile-up signals) in the neighborhood (up and down) of the listening frequency. All signals passed via wide first IF filters are amplified by the receiver’s first IF circuitry and can cause blocking, desensitization and inter-modulation (see below) at the receiver’s second IF mixer.

 

IMD DR3 at 5 kHz is second in importance (third column in my Table).

ARRL Test Procedures Manual enumerates this parameter as Two-Tone Third-Order Dynamic Range. IMD DR3 is an indication of a receiver’s ability not to generate false signals because of the two strong signals on different frequencies outside the receiver pass-band. IMD DR3 is expressed in dB relative to receiver noise floor.

Two strong parent signals f₁ and f₂ of the same amplitude can produce third-order intermodulation products in the receiver. If f₁ and f₂ signals are strong enough, any receiver will produce intermodulation products (2f₁ - f₂) and (2f₂ – f₁) that are generated inside the receiver due to its limited immunity against Two-Tone Third-Order Intermodulation Distortion.

Special conditions can be met to produce an intermodulation product exactly on a specific listening frequency, f_DX. Two strong parent signals f₁ and f₂ have to have a special frequency relation to each other and with that specific listening frequency, f_DX. Therefore, for a typical DX pile-up, only some specific pairs of f₁ and f₂ frequencies can produce third order intermodulation products exactly on the frequency of the weak DX station, f_DX. That is why the probability of appearing Two-Tone Third-Order Intermodulation products on the listening frequency, f_DX is smaller than blocking or desensitization of the receiver by any single signal. That is why I have ranked that parameter in second place, after BDR at 5 kHz. In case of receiver blocking / desensitization, any single signal, anywhere, but close to the listening frequency, f_DX, can inhibit the reception of a weak DX station, if that signal is strong enough to produce the receiver blocking / desensitization effect.

The higher the IMD DR3 parameter, the higher is the particular receiver’s immunity against intermodulation. For parent signals f₁ and f₂ spaced by 5 kHz, a value of 88dB means a high grade receiver for that parameter. Three best receivers for that parameter are the Orion (93dB), upgraded K2/10 (91dB) and newest IC-7800 (89dB). General coverage receivers with wide (15 kHz) first IF filters have IMD DR3 parameter worse than Orion ranging from 10dB (FT-1000MP) to 26dB (FT-897).

We can observe the same rule as for BDR at 5 kHz: receivers equipped with a narrow crystal filter in their first IF have higher values of IMD DR3 at 5 kHz than general coverage receivers with wide (15 kHz) first IF filters. That is clearly seen from Swept Two-Tone, Third-Order IMD Dynamic Range graphs (Figure 3 for K2 and Figure 4 for TS-2000) and from data in third column in my Table.

This place please put Swept Two-Tone, Third Order Dynamic Range graphs from Extended Test Reports:

3.                   K2 (page 20)

4.                   TS-2000 (page 29)

And the same conclusion (as in case of BDR at 5 kHz): comparing these two graphs anyone can judge which receiver front end design is better for DX hunting.

 

I have placed the influence of Phase Noise third (fourth column in my Table).

For some receivers BDR and IMD DR3 results are “noise limited” (for instance BDR at 5 kHz is 96dBnl in the case of FT-897). A particular receiver’s phase noise contribution is masking and interfering through desensitization (during BDR measurement) or intermodulation products (during IMD DR3 measurement). The normal 1dB decrease of receiver sensitivity (during BDR measurement) cannot be measured due to a 1dB increase of phase noise. In all probability the receiver has greater BDR than the measured noise-limited value.

The third limiting factor of modern receiver performance is local-oscillator phase noise. Phase noise contributes to poorer receiver BDR in the form of desensitization by nearby strong signals resulting from reciprocal mixing. In a real DX pile-up, a receiver with poor phase noise will not only receive signals the VCO is tuned to, but also some very strong signals, which could be several kHz away from the listening frequency f_DX, will be heard (like a crosstalk) on the listening frequency due to reciprocal mixing.

Frequency synthesizing VCO oscillators are more prone to generate phase noise sidebands than ordinary LC oscillators. Strong signals from the antenna can produce in such a receiver IF signals not only when mixing directly with the VCO carrier, but also when mixing with noisy sidebands of the VCO. Oscillator power noise density is expressed in dB referred to the carrier in 1Hz bandwidth (dBc/Hz). The measurement is usually made as Transmitted Composite Noise Test from the carrier up to 22 kHz apart. Only in case of the Orion VCO oscillator can its phase noise be measured directly.

This place please put Transmit Composite Noise /Receiver Phase Noise graphs from Extended Test Reports:

5.                   Orion (on page 22)

6.                   K2 (upper right corner on page 16)

7.                   TS-2000 (lower left corner on page 19)

A new record of VCO low phase noise in ham radio HF equipment has been achieved recently in the Ten-Tec Orion HF transceiver: only -138dBc/Hz at 4 kHz from the carrier. But there is also some increase of phase noise 45 – 70 kHz from the listening frequency (see Figure 5). Very good synthesizer design is implemented in the ICOM IC-756PRO and IC-756PRO II HF transceivers (-130dBc/Hz). A typical phase noise graph for good VCO design is demonstrated in Figure 6. And finally an example of quite poor design referring to VCO phase noise is shown in graph 7. There is more than 20dB difference (in favor of Orion) in phase noise contribution between Figures 5 and 7. For DX hunting oriented operators it is clear which VCO oscillator design is better.

 

And finally IP3 at 5kHz is fourth in importance (fifth column for 5 kHz and eighth column for 20 kHz spacing).

This is Third Order Intercept Point at 5 kHz spacing (if measured, because usually only IP3 at 20kHz is known). IP3 is an extrapolated point at which the desired response and the Third Order IMD response intersect. The Third Order IMD products increase three times as fast as the pair of equal amplitude parent f₁ and f₂ signals. Increasing the power of parent f₁ and f₂ signals results in an increase of fundamental output signals in one-to-one ratio (until saturation will destroy the linear relationship).

The frequency separation of the two equal amplitude signals f₁ and f₂ can greatly influence intermodulation in a particular receiver. The worst situation is when there is not enough selectivity at receiver front end and both equal signals f₁ and f₂ can pass through the first IF filter. That causes intermodulation in first IF circuitry and in the second mixer.

IP3 parameter for two best receivers in my Table deteriorates slightly when changing f₁ and f₂ spacing from 20 kHz to 5 kHz:

·       Orion goes from +22.8dBm / +12.9dBm at 20 kHz spacing to +22.1dBm / +11.4dBm at 5 kHz spacing,

·       The K2 goes from +21.6dBm / +8dBm at 20 kHz spacing to +21dBm / +8dBm at 5 kHz spacing.

 

On the opposite side are behaviors of general coverage receivers equipped with wide (15 kHz) first IF filters. Let us analyze (for example) IC-756PRO II Extended Test Results of IP3 at 20 kHz and 5 kHz spacing: IP3 is +20.2dBm / +10.2dBm / -4.2dBm at 20 kHz decreasing dramatically to -18.8dBm / -28.8dBm / -35.5dBm at 5 kHz.

This is an enormous degradation of receiver performance in the presence strong signals close to the listening frequency. 20 kHz and 5 kHz IP3 difference in IC-756PRO II receiver is reaching an enormous value of 39dB! That is almost 38dB worse compared to Orion (0.7dB) and K2/100 (0.6dB) test results. When comparing test results from fifth column for 5 kHz and eighth column for 20 kHz spacing anyone can see similar IP3 deterioration in other general coverage receivers equipped with wide (15 kHz) first IF filter. That is why I recognize such design as a “not good” one for DX oriented ham radio operators.

 

By evaluating “good” design of receiver front end, we can assume that only receiver input circuits, including the first mixer, are “responsible for” its IP3 properties. But this is not true in case of “not good” design with wide (15 kHz) first IF filters. In that case, first IF circuits and second IF mixer add their contribution of IP3 deterioration for a narrow spaced pair of equal amplitude f₁ and f₂ signals.

“Good” receivers have IP3 at 5 kHz in the range of +15dBm. Values around +20dBm characterize a receiver as high grade. Any dB more above the +20dBm level puts a receiver in the top grade group for that parameter. At this point I have not mentioned that it is possible to “improve” the IP3 parameter artificially by making a receiver less sensitive, or by putting an attenuator before the receiver front end. For example, turning on a 20dB attenuator results in an improvement in IP3 from “poor” -3dBm to “good” of +17dBm.

 

I have recently seen discussions about front end design offering IP3s in the range of +30dBm and even +40dBm. I am not enthusiastic in such upgrading. I recognize such designs are important only for some specific locations, like wireless communication command centers or for air / sea mobile use, when different transceiver antennas “almost touch each other” and there is a great possibility of electromagnetic induction of big signals in neighboring antennas. But this is not the case for shack-style ham radio practice. Ham radio antennas are not so densely populated.

Let us estimate if a front end design offering IP3 = +40dBm is the solution for receivers equipped with wide (15 kHz) first IF filter. From the DX hunting point of view, I think not. If such a receiver front end has (for example) IP3 = +20,2dBm at 20 kHz and decreases down to -18,8dBm at 5 kHz, what would be the real improvement replacing that front end (including the first IF mixer) with a design offering IP3 = +40dBm at 20 kHz? Not too much in case of a typical DX pile-up. Still, for narrow 5 kHz spacing, deterioration of IP3 (including first IF circuits and second IF mixer) is 39dB. This means that for 20 kHz spacing IP3 would be excellent, but for 5 kHz quite miserable +1dBm (+40dBm - 39dB = +1dBm), which is not an impressive value and worse by 20dB than 5 kHz IP3 achieved in Orion and K2/100. Such a low value of IP3 at 5 kHz spacing I recognize as “bad front end design” for DX-oriented operators. This is my personal point of view: for me, most important are receiver properties at 5 kHz spacing (and closer). For 5 kHz spacing none of the general coverage receivers with wide (15 kHz) first IF filter can offer as good parameters as measured in the case of Orion and K2/100.

 

The main disadvantage of a general coverage up-conversion concept is its very wide first IF filter. For DX hunting, contesting and digital modes operation a much narrower filter shall be used. But a general coverage receiver design is forcing up-conversion in the first mixer. That results in quite high first IF (45MHz – 70MHz). It is not easy (and not cheap also) to replace the 15 kHz wide first filter in the 45MHz – 70MHz range with a narrow one, offering a bandwidth adequate for SSB. But some ambitious ham radio constructors follow this way: I know one Polish ham radio constructor (SQ9GAT) who has upgraded his old FT-1000MP with a narrower first IF filter.

 

Other Receiver Parameters (Less) Important For DX Hunting

Receiver sensitivity measures its ability to hear weak signals. Sensitivity is expressed in µV, dBm and sometimes in dBµV. MDS is the input level to the receiver that produces an output signal equal to the internal noise generated in the receiver. Therefore, MDS is sometimes referred to as the receiver noise floor. The typical MDS of a ham radio HF receiver is in the range -135dBm to -140dBm at 500Hz bandwidth.

By analyzing values from column 2 and 10 in my Table we can find that high sensitivity is sometimes an enemy of good dynamic range of the receiver. It counts for IC-706MkIIG and IC-756PRO, IC-756PRO II and IC-746PRO, when Preamplifier #2 is ON. Adversely, the three best dynamic range receivers in my Table are rather moderate in sensitivity.

Only in the case of a very wide DX pile-up and main World-Wide Contests 20 kHz BDR (column 6) and 20 kHz IMD DR3 (column 7) shall be taken into account. The same counts for IP3 at 20 kHz (column 8). Generally, these parameters are significantly better for 20 kHz spacing than for 5 kHz. That is why I have put them further down on my list of importance.

IMD DR2 and IP2. They are maladies of “not good” receiver front end design (equipped with semi-octave wide RF filters at the front end of receiver). I don’t care about them (I am not going either to use or to buy any “not good” design receiver).

 

An attempt to analyze the historical aspect

After putting all these parameters into my Table I have realized there are not only strictly technical relationships but some historical dependencies could also be found. I was really surprised about that!

New models of HF transceivers of both American producers (Elecraft and Ten-Tec) have usually offered higher values of receiver dynamic range parameters. From March 2000 until December 2003, a prototype of K2 (serial number 00495) was measured with the best dynamic range receiver at 5 kHz spacing. An Orion manufactured by Ten-Tec beat this record in December 2003. I have found in the February 2004 issue of QST test results of the same K2 (serial number 00495) but upgraded to K2/100 version. I have noticed that the 5 kHz BDR record now belongs to K2/100 (almost 5dB better than Orion). I have followed modernizations and upgrades implemented by Elecraft in recent years. Therefore I can say “K2 is alive”, almost every month some new upgrading is announced. There is sensitive feedback between K2 users and Elecraft designers and proposals of improvement are checked by Elecraft staff and many of the ideas have been implemented already. To be specific, between March 2000 and February 2004 some K2 receiver parameters have been significantly improved:

·       5 kHz BDR has been improved from 126dB to 135dB, which is a new record value for ham radio receivers,

·       5 kHz IMD DR3 has been improved from 88dB to 91dB (only Orion is better by 2dB),

·       IP2 has been improved from +75dBm / +76dBm to +80dBm / +79dBm.

The same story is in case of Ten-Tec’s last two models: dynamic range parameters of Orion have exceeded previous OMNI 6+ results.

 

My historical findings in case of ICOM, Kenwood and Yaesu over the last 13 years are not so positive. I have found a similar approach concerning receiver front design. Approximately until 1990 they have manufactured HF transceivers designed only for ham radio bands (some models included also the 160 meters band). Receivers have been generally designed as a double super-heterodyne (for example, Kenwood used IF: 8,83MHz and 455 kHz). What was important in these receivers was the fact that they were equipped with narrow filters after the first mixer, with a SSB crystal filter as a standard. It was possible to install also an optional narrow crystal filter for CW (Kenwood: 500Hz or 270Hz).

Later ICOM, Kenwood and Yaesu started to implement the general coverage receiver design. When comparing the dynamic parameters of previous and new generation of receivers, I have found (you can check in my Table) that changing the receiver front end design concept has brought a deterioration of receiver’s dynamic parameters. Specifically, ICOM top values of BDR and IMD DR3 at 20 kHz were measured for IC-781 and IC-775DSP. None of the later (new generation) models beat (or even draw near) IC-781 and IC-775DSP records. Only the newest IC-7800 offers these parameters in top grade class. The Kenwood top values of BDR and IMD DR3 at 20 kHz were measured for TS-850S (model from early 1990). None of the newer (new generation) models beat TS-850S records (only the latest TS-480 in case of IMD DR3 at 20 kHz parameter comes near). Finally, Yaesu top values of BDR and IMD DR3 at 20 kHz were measured for FT-1000D (model from early 1990) and FT-1000MP (model from mid 1995). None of the newer (new generation) models beat these records (except Mark V FT-1000MP in case of IMD DR3 at 20 kHz).

 

At this point I have to underline that we can use only some parameters for historical comparisons over the last 13 years. 5 kHz dynamic range measurement results are accessible only for models measured since mid-2001 and later. A similar situation is found with wide band swept dynamic range measurements (since 1996). I have also found some difficulties to compare composite noise data measured recently and 10 years ago (different graphs). Beginning from mid-2001, the testing range is wider than previously and since that time we have access to more complex test results.

 

Conclusion

First, from a technical point of view I have tried to demonstrate that comparative analysis is a valid tool to estimate the DX prowess of HF receivers. I have concentrated only on receiver dynamic parameters having in mind the usefulness of the RX for serious DXing. I have tried to point out that – thanks to ARRL Technical Laboratory good job done – any ham radio operator can judge for himself about the usefulness of a particular piece of equipment for his particular needs and habits as a ham radio operator. There are many other points of interests in our hobby. Using ARRL Technical Laboratory test results someone could make his own analysis, reflecting one point of interest.

Second, unexpectedly the historical aspect has appeared during my evaluation. With respect to that factor, I want to point out that ham radio operators shall be cautious with marketing claims about “next top achievements” offered in new models. As you can see from the above historical comparisons, these claims are not always true. Sometimes, they are contrary to measurement results. Being pragmatic, it is better to have “limited confidence” for such claims, until we have seen a report in the Product Review column in QST or at the ARRL Website and have a chance to make our own “comparative analysis”. That is my friendly advice.

 

Acknowledgements

I want to thank Wes, SP2DX and Charlie, W0YG for their encouragement and useful comments during preparation of the final version of this manuscript in English language.

 

References: ARRL Website, Elecraft Website, Ten-Tec Website and „Product Review” columns in QST.

 

SP7HT has been involved in DX hunting for the last 47 years. SP7HT was the very first DXer from Poland to reach DXCC Honor Roll (1981) and DXCC Honor Roll #1 (1986). For the last 35 years his occupation has been associated with ground and satellite microwave telecommunication. Actually, retired since January 2003.