[AMPS] IM, Harmonics & Spurs ( was: Quiet)

[Arlen Mendelssohn]

from Jon:

>>  If I stuff 30 MHz and 
>>30.1 MHz into the amp and I see 30.2 and 29.9, it's intermod.  If I 
>>stuff 30 and 30.1 in and I get 60 and 60.2, it's harmonics.
>
>What you are specifiying here are 3rd order intermod products.  
However, 
>intermod products are produced all across the band.  Second order 
>products 

The formula is: 

2nd order freq = First order freq1 +/- First order freq2

Yes - 2nd order products are nothing more than what you'd get from a 
mixer being fed the two signals through separate ports.

> even order products occur out of band, odd product in band.  3rd order 
IM products go by the 
>formula of 2*F1-F2 and 2*F2-F1.  There are products (not sure if they 
are 
>second order or which order) that go by the formula of 2*F1+F2 and 
>2*F2+F1.  In this case those are 90.1 and 90.2 MHz.  Definitely out of 

Oh, dear, confustion.  A third order product is the result of the second 
harmonic of one signal plus or minus the fundamental of the other, 
therefore there are four, not two, third order products.  It's just that 
two of the third order products are bothersome, and two are not.  Your 
example of:

2*F1+F2 

is not a second order, it is a third order.  However, it is out of band 
and can be filtered.  You can see what order the product is by summing 
the coefficients in the formula; thus, in this case, the order is 2 + 1 
= 3.

A second harmonic might be 2*F1 + 0*F2, so it's 2 + 0 = second order.

Another second order product is F1 - F2, so it's 1 + 1 = second order.

Not so hard.  Plenty of public-domain spreadsheets are available to 
predict where these might land.

>band.  Definitely IM.  With each succeeding order, the intermod 
products 
>get weaker and weaker.

Depends on the design.  I have measured microwave amps whose 5th order 
products are stronger than the 3rd order.


>However, if the low order products are really large and significant, 
>higher order products can be significant as well.
>
>Also, since every signal has harmonics, the harmonics of the intermod 
>products could also show up on the output as well.
>
>So....you could get lots and lots of crap on the output signal of the 
>tube that are either harmonically related or IM related.  The original 
>poster did say that the signals were all across the band.
>
>However, I would agree that at 350 MHz, any IM products that far away 
>from the original carriers would be quite small.  But for that matter, 
so 
>should be the harmonics!
>>
>>He did not say where he put the probe.  If he put it near the anode, 
>>then he's not getting the benefit of the lowpass action of the output 
>>network, and I would definitely expect high order harmonics.  The only 
>>way to use a "sniffer" and get decent results is to somehow closely 
>>couple it to the final output of the amp, and nowhere near the anode.  
>>The anode will always be worse than what you put on the air, and by a 
>>wide margin.
>
>Oh, I agree here as well.  No question about it.  Unfortunately, if you 
>are looking for oscillations, you won't find them after the output 
>circuit as they will be filtered out as well.

This is a very good point.  However, it is critical to understand 
precisely what one might find based on the input signals so that one 
does not mistake an intermodulation or harmonic component with an 
oscillation.  The first is created by expected and usually desireable 
nonlinearities (proper nonlinearity is the source of efficiency 
exceeding 50%), and the other by a circuit condition which could cause 
erratic operation and/or damage, and should be remedied.


>Quoting from "Solid State Radio Engineering" by Krause, Bostinian and 
>Raab: "Other unwanted signals include parasitic and subharmonic 
>oscillations and mixer products;  they are called spurious products or 
>simply 'spurs.'  In an RF PA, the harmonics and some of the spurious 
>products may be removed by filters..." page 349
>
>So...in order to catch higher order spurs, one would need to look at 
the 
>anode.......

Unless one's intent is to dispatch them with a filter, and the filter's 
design is alrady adequate.

Measuring at the anode characterizes performance and helps understand 
stability of the design, etc, but proof of performance and meeting 
spurious specs can only be performed at the output.


>73,
>
>Jon
>KE9NA
>
>

Arlen


______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com

--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm