Page 1 of 1


Posted: Wed Dec 05, 2018 5:21 pm
by glovisol
This post is concerned with IIP3 (Third order Input Intermodulation) Technical Specifications and in this respect the quality and detail of those provided by SDRplay in their site under the heading “DETAILED TECHNICAL INFORMATION” are, in my opinion, among the very best I have ever seen in my professional life. ... n-R1P1.pdf


Posted: Wed Dec 05, 2018 5:23 pm
by glovisol

This is especially true regarding the Input 3rd Order Intercept Point or IIP3, which is given for every product, for each frequency and for all sensitivity levels. The meaning of IIP3 is generally known and in any case descriptions and explanations are easily found on Internet. For our analysis it is enough to understand that IIP3 is a Figure of Merit which defines the input overload power level (in dBm) at which our receivers will start producing spurious signals within their passband, or in other words, their own interference.

By using the wealth of data provided, we can easily have a quantitative idea of our receiver’s intermodulation performance across its entire frequency range. Try to obtain this info for other apparently similar products!

As an example let us consider a practical case in the 40 m band. The 8 MHz IIP3 specification for the RSP-1A is: IIP3 = -4.24 dBm at the maximum sensitivity of 0.3 uV, which corresponds to a received power of -117 dBm / 50 Ohm.
The 40 m band extends from 7.0 to 7.3 MHz. Through the day, but at night especially, at F1 =7.3 MHz operates a very strong broadcast station, sometimes coming in at – 40 dBm. At F2 = 7.25 MHz another broadcaster comes in at an average power of -70 dBm. These two stations can potentially produce two in-band intermodulation products as follows:

IIP3(1) = (2*F1) – F2 = 14.6 – 7.25 = 7.35 MHz
IIP3(2) = (2*F2) – F1 = 14.5 – 7.30 = 7.20 MHz

This second product then could appear smack into the 40 m. band as a self generated spurious signal and interfere with an on air signal at that frequency. But what would the level of this spur be? A practical rule of thumb states that the power level Ps of the spur will be three times the difference Pd between the strength of the strongest of the two incoming and the rated IIP3 value of the receiver. In our case (see SDRplay specs) IIP3 = - 4.24 dBm, thus:

Pd = 40 – 4.24 = 35.76 dB
Ps = 35.76 * 3 = -107.26 dBm

But the rated sensitivity of the RSP-1A is -117 dBm, so that this spur would be received as a real signal with a level 10 dB higher than a true weak signal at the same frequency!

In reality we must consider the noise floor and look at the real sensitivity of our receiving system, receiver + antenna. Let us look at Table 3 of this thread: ... f=5&t=3685

which I have again uploaded below. If we were in a QUIET RURAL area, our most likely noise floor would be -112 dBm, so our receiver’s sensitivity would exceed the noise by: 117 – 112 = 5 dB. If we now used a 5 db attenuator between antenna and receiver, our useful sensitivity would still be the best possible, but both F1 and F2 signal power would decrease by 5 dB with a very large effect, considering the times 3 factor:

Pd = 40 + 5 – 4.24 = 40.76 dB
Ps = 40.76 * 3 = -122.28 dBm

Thus the spur could not give us any problem because it had dropped below the level of the noise floor!

It can be argued that the same result could be obtained reducing the RSP-1A gain by 7 dB by placing the gain slider position to 1: our IIP3 would increase to +1.94 dBm and we would have:

Pd = 40 + 1.94 = 41.94 dB
Ps = 41.94 * 3 = -125.8 dBm

Again placing the spur below the noise floor.

But we must remember that reducing the RF gain can non-proportionally degrade the noise figure, so the most efficient solution is to use an external attenuator. For example, looking again at the specification, going from position zero to position 4 for the RF gain, NF goes from 18 to 50.5 dB (32.5 dB worse) while IIP3 goes from – 4.24 to +13.68, or 18 dB better.


Posted: Thu Dec 06, 2018 9:55 am
by glovisol

Uploaded below a worked out IIP3 example with actual signals received by the RSPduo on 11/07/2018. See comments here: ... f=5&t=3436


Posted: Fri Dec 07, 2018 9:24 am
by glovisol

Uploaded below are two tables showing IIP3 levels for the RSP-1A calculated for a strong interfering signal F1 at -40 and -50 dBm on most HF frequency bands and for the QUIET RURAL and for the RURAL situations. Based on data provided by the SDRplay detailed specifications, we can conclude that the critical IIP3 input level for the RSP-1A lies somewhere between -40 and -50 dBm. In fact at -50 dBm there is no danger of IIP3 at any HF frequency. The figures in red in the -40 dBm table show the cases where interference would be above the likely Noise Floor, requiring either an internal RF gain reduction or (better) an external attenuator with a value given by column E - EXCESS SENSITIVITY. The two figures in red in the -50 dBm table merely show only what would happen if we tried to improve the sensitivity (excess sensitivity is negative with respect to Noise Floor) by adding an external low noise amplifier. See also on the external attenuation question: ... f=5&t=3723

Since, from a practical standpoint, it is extremely unlikely to experience on air signals with levels above -50 dBm and extremely likely to have a higher Noise Floor, we can conclude that the RSP-1A and its companions of the RSP class, represent an extremely good and practical compromise among available maximum bandwith (10 MHz) sensitivity and resistance to in-band overload.

Out of band overload is wisely dealt with by means of the array of built-in low pass and notch filters (also extremely well specified by SDRplay) so, both in theory and in practice, it is very difficult to find fault with the RSP class receivers and probably impossible to find similar units so well specified.


Posted: Sat Dec 08, 2018 11:54 am
by glovisol

The specifications for the RSP2 are those published by QST magazine and are based on measurements done in the ARRL labs. For the IIP3 specification, ARRL gives the following measured data taken at 14 MHz:

Ps = -97 dBm with F1 at a power of -44 dBm

Note that -44 dBm is half way between -40 and -50 dBm, as deterrmined in the previous posts for the RSP-1A & RSPduo. Using the previously described rule of thumb, we can find the value of IIP3 as follows:

-97 = -3*(44 - IIP3)
-97 = -132 + 3*IIP3
97 = 132 - 3*IIP3
IIP3 = -(132-97)/3 or IIP3 = -11.66 dBm
PROOF: Ps = -3*(44-11.66) = -97 dBm

This result is also confirmed by the ARRL measurement of the second order intercept point IIP2. The rule being that the second order intercept point is located at 9.6 dB below IIP3. In our case:

IIP2 = -11.66 - 9.6 = -21.66 dBm

The IIP2 measured by ARRL is: IIP2 = +21 dBm measured at zero gain, out of a total gain of 40 dB so this measurement must be corrected by 40 dB:

IIP2 (ARRL) = +21 - 40 = -19 dBm which is very near our value of -21.66 dBm.

We can conclude that, considering IIP3, the RSP-1A and the RSPduo have a better IIP3 performance than the RSP2/RSP2PRO.