# Rx Intermodulation Implications

I am looking at the product spec of NRF52832, in particular the Radio Electrical Specifications (here, the information is on page 232 of v1.4). Within it is a section called Rx Intermodulation that shows this:

IMD performance, 1 Msps (3MHz, 4MHz, and 5MHz offset)        -33 dBm


I have talked to a few people, but the opinions as to what this means practically vary.

My question is: If I have this receiver set to listen on frequency X, and I start a transmitter on frequency Y = X + 3 (and all settings allow the receiver to decode the transmitter's messages), does that mean that intermodulation will make it possible for me to receive the transmitter's messages with the received signal strength reduced by 33dB?

• @BrianDrummond, your comment, while valid, could be more tactful. – Catsunami Jan 30 '18 at 21:22
• The dB is a gain ratio. Henceforth, -33dBm means a very very tiny gain. – KingDuken Jan 30 '18 at 21:22
• @KingDuken, in wireless systems, a received signal at -33dBm is pretty strong. What does it mean in the case of intermodulation though? – Catsunami Jan 30 '18 at 21:25
• @KingDuken -33dBm means a very very tiny gain A gain can be in dB but dBm is always a power level, not a gain. – Bimpelrekkie Jan 30 '18 at 21:28

In the datasheet there's a note regarding that section

RX intermodulation

The note says:

Wanted signal level at PIN = -64 dBm. Two interferers with equal input power are used. The interferer closest in frequency is not modulated, the other interferer is modulated equal with the wanted signal. The input power of the interferers where the sensitivity equals BER = 0.1% is presented.

So then

PIMD,1M IMD performance, 1 Msps, 3rd, 4th, and 5th offset channel: -33 dBm

means that the power level of both interferers can have a power of up to -33 dBm.

Since the wanted signal is at -64 dBm, those interferers can be 64 - 33 = 31 dB higher in power than the wanted signal.

Under those conditions the BER (Bit to Error rate) is 0.1% (1 bit per 1000 bits can be wrong). More accurate: the chip's manufacturer guarantees that the BER will be 0.1% or better under these conditions. It is impossible to design for an exact BER, I know, I worked on BER and Bluetooth many years ago.

You cannot draw conclusions from this for your transmitting at Y = X + 3 case because in the specified case the carrier at + 3 MHz is unmodulated. You ask about a modulated carrier which makes it a different scenario.

Also there is no 33 dB in this. What is mentioned is -33 dBm so 33 dB less power than 1 milliWatt. -33 dBm is an absolute power level, the level which the interferers can be for the stated conditions for this specification point.

Such a specification point might look weird, like it is some "random" condition. Well it isn't. It is usually a result of the Bluetooth specification and then translated to include antenna gain and antenna switch losses and such.

• Hmm, thanks! In the little description it mentions one of the interferers is modulated the same as the desired signal. So does that mean they were doing tests with +3 and +4 (for example) and not modulating +3, but modulating +4? And same with +3 and +5? I am sort of curious what the point of doing unmodulated measurements is since noone would use their radio unmodulated unless they are doing compliance testing? I know what the difference between dB and dBm is, my brain just blanked that out while I was trying to explain my X + 3 problem. Thanks again. – Catsunami Jan 30 '18 at 21:39
• You're right that using an unmodulated interferer might not sound realistic. Indeed it is probably specified like that in the Bluetooth specification. Also it is easier to generate an unmodulated signal during measurement. If the closest carrier was modulated it would need to use different data otherwise a bad receiver could just use the interferer's data. Also the unmodulated interferer does influence the receiver circuits in a way that can affect reception. So there are many reasons why an unmodulated interferer is used and is a fair test of the system. – Bimpelrekkie Jan 30 '18 at 21:45
• Ah okay. So given that my receiver is picking up transmissions on a different frequency, does that generally indicate a hardware problem? I have checked that the software is not changing the frequency. I have also monitored the Rx spurs on the SA and I see some spurs nearby, I just don't know the source for them. Is there a name for this sort of problem so that I can look into it better? – Catsunami Jan 30 '18 at 22:28
• As far as I know, BT doesn't work on a fixed frequency, it does frequency hopping. Not sure if BLE is also like that though, I did not read the spec. of BLE yet. If you have a proper BT link then other frequencies should not be picked up. It is unclear to me what is happening in your setup. – Bimpelrekkie Jan 31 '18 at 6:38
• It's not actually BT or BLE. The protocol I am using does not hop. – Catsunami Jan 31 '18 at 14:22