IMD and S/N

This is about the numbers reported by programs such as HRD and others.

Let me see if I have this correct.

1. The S/N measures the difference between the signal power and the (average?) noise power with the noise power as the reference?

2. The IMD measures the difference between the signal power and harmonics power with the signal power as reference?

3. IF the harmonics power is, in fact, below the noise power, can the programs that report IMD mistakenly assume that the noise power is the harmonics power because the noise masks the harmonics?

If so, when the IMD is really -40 dB but the S/N is 20 db, the IMD number reported by these programs should be about the same as the S/N give or take a few dB and the IMD number reported is, in fact, meaningless.

Unless, of course, the S/N is 60 or 80 dB or more.

 

Your understanding is 100% correct!

73,

fp

 

But it looks impressive when someone tells you your PSK signal has an IMD of -22. The big problems are the digital DX powerhouse operators who "black out" the screen every time they put out a call. Conditions improve enormously when they get tired and bored with no response, and amazingly the screen turns blue again, sanity returns.

G0GQK

 

IMD and harmonics are usually specified separately, Imd is (usually) frequency's close in to your transmitter.

For instance, you are on 14.1 mhz transmitting a 2 tone USB signal of 1 and 2 KHZ

You will have signals at 14.101 (F1) and 14.102 mhz.(F2)

Due to non linearity in a "linear" amplifier, you will generate "third order IMD" at 2F2-F1 (14.103mhz) and 2F1-F2 (14.1mhz).

Similar frequency’s are generated at the 5th, 7th 9th etc order.

Harmonics are multiples of the transmitter frequency, the 2F2 and 2F1 above.

Yes, any IMD or Harmonic power that is below the noise power will not be detected.

HOWEVER, Keep in mind that the reading will be dependant on the "detected bandwidth", and what you see on a wide band (3khz) display may be VERY different than what the computer sees in it's detected bandwidth.

For example, suppose the detector uses DSP techniques to separate the receiver into 10 hz "channels"

The noise power is (usually) directly correlated with the receiver bandwidth (Google "KTB Noise"), in this example we have 2 B/W's, 3000HZ, and 10 HZ

There will be 300 times the noise power in the 3000hz b/w than the 10hz b/w, or about 24.77db. This is the improvement you would see between the 2 b/w's.

You can see why it is so important that your receiver b/w exactly matches your transmitted b/w, you are getting rid of the "extra" KTB noise.

(KTB noise also explains why we cannot keep adding "preamps" forever...)



Rege

 

This effect is caused by the bandwidth in your receiver being much wider than necessary, and the action of your rigs AGC.

Your receiver passes a slice of the spectrum about 3 khz wide (for a SSB voice filter) to the computer.

The strongest signal in this 3 khz will "capture" the AGC, and reduce the gain of the receiver to keep the audio level constant. This of course makes all the other signals weaker.

Switching in a receiver filter with a bandwidth just wide enough to pass a single PSK signal will prevent this from happening, The downside is you will not have the pretty spectrum display on your computer screen.

Rege

 

Rege is correct, but you can also just use the notch filter if you have a single "offender."

73,
Bill

 

Thanks to all of you for your comments.

I'm married and it's nice to know I can be right once in a while, at least.

 

I was just on 20 with DM780 and this guy looks like he was splattering pretty good.

Can someone possibly diagnose what the problem with this signal is, and I would love to get a screen capture of my signal, I have been told my IMD is usually -28 to -30db and is clean.

View attachment 43423

 

My two cents worth...

Looks like a combination of too much power and the op's sound card settings are too high. A lot of hams just don't understand (or don't want to change their ways) on digital modes more power is not better.

 

I made the mistake of mentioning to one guy that he was painting the entire waterfall, that was a mistake. Then I got a lecture on how great his signal was because he has an IMD meter and its always -30db blah blah blah then he sends me E-Mails to educate me. Maybe he should read the references he sent me. Funny thing was two other guys told him he had a crappy signal and he actually got into an argument with the guy on the waterfall hehehehe. It was pretty funny.

I dont know if this is right or wrong but I always set my RF Power to 100 watts and adjust with my sound card volume to about 40 watts of output and no more than 3 bars of ALC on my Icom IC-7000. I always get good reports so I am hoping I am doing it right.

Thanks

Craig