Amps
[Top] [All Lists]

Re: [Amps] IMD test question

To: Jim W7RY <jimw7ry@gmail.com>
Subject: Re: [Amps] IMD test question
From: Larry Benko <xxw0qe@comcast.net>
Date: Wed, 01 Aug 2012 23:39:45 -0600
List-post: <amps@contesting.com">mailto:amps@contesting.com>
Jim,

There are NO radios currently made (that I know of) that are capable 
with 2 audio tones input of having output IMD <-60dBc when driving even 
a dummy load.  Even when driving the radio to 1/2 or 1/4 power output I 
didn't see any radios with IMD that was even close to -60dBc.  I have 
not measured but am told that some of the Yaesu class A output radios 
can get close to -50dBc.

Philips (now NXP) has a pdf document titled "Philips RF Transmitting 
Fundamentals" where they derive the necessary input IMD to an amplifier 
in order to be assured of an output IMD accuracy uncertainty.  The data 
is on page 35 but the bottom line is that for a 1.4dB error in measured 
output IMD the input must be better than the output by 15dB.  Given that 
the ARRL measured the IMD of the Alpha 9500 at -45dB PEP worst case 
which is -39dBc I would want the source to be -54dBc or better.  
Additionally amplifiers even with swamped resistor or tuned inputs have 
some small input impedance variation throughout the RF cycle which can 
cause the exciter to generate IMD.  The ARRL document version 2?? 
outlines a 2 class A exciter source with a 5/4 wavelength 20m coaxial 
power combiner to make their IMD measurements (ARRL IMD msmts. made on 
20m only).

My circuit with the 2 exciters and the 3dB loss power combiner and the 
tuner is capable of producing 200W peak for driving an amp like a GS-35 
or 2x 3-500Zs where 100W is not quite enough.  The tuner provides some 
additional Q between the exciter and the amplifier under test and is 
adjusted so that the isolation of the power combiner is so large that 
the 2 exciters do not produce any significant IMD due to this 
isolation.  I still can have issues with input impedance changes in the 
amplifier but with 200W of drive I can put a 3dB pad in series with the 
tuner if needed which reduces these impedance variations by 6dB.  In all 
the tests I have done to date I can drive the amp under test with IMD 
<60dBc easily (on any band) and usually <-70dBc with more careful tuner 
adjustment and a couple dB pad.  Tests on my power combiner show no 
significant IMD due to the combiner itself.  The output monitoring port 
on the power combiner serves as the proof of the IMD actually going into 
the amplifier under test.

I hope this makes sense and I'm sure there are other ways to make 
accurate IMD measurements but this way worked well for me.  I learned a 
lot doing all these experiments and tests.  I talked with some industry 
folks who told me they use multiple single frequency class A exciters 
for IMD tests for exactly the reasons I mentioned above.  Unfortunately 
I didn't know enough to ask before I did the work but if I did I 
probably wouldn't understand the issues as well.

73,
Larry, W0QE


On 8/1/2012 10:07 PM, Jim W7RY wrote:
> Larry.. Why not just use one transmitter with two audio tones 2 KHz 
> apart? Can you expound on this a bit?
>
>
> 73
> Jim W7RY
>
>
> -----Original Message----- From: Larry Benko
> Sent: Wednesday, August 01, 2012 1:36 PM
> To: amps@contesting.com
> Subject: Re: [Amps] IMD test question
>
> Bill,
>
> I have made amp IMD measurements with the circuit shown at
> http://www.w0qe.com/Technical_Topics/imd_testing_of_amplifiers.html
> which can measure any amp currently made.
>
> As I'm sure you are aware there are not any exciters made today with IMD
> in the -60dBc range.  This is why I did the test with 2 exciters. Also
> you need to check the final input to the amp to be sure that any IMD is
> due to the amp and not the driving circuitry. An amp which has an
> impedance which varies throughout the RF cycle (any non class A amp) can
> cause the exciter to produce IMD which may improve or reduce the IMD of
> the amp.  The circuit I use can drive a 2 3CX800 amp with IMD levels
> -70dBC so that the amp IMD is easy to see.  Also theory and tests I have
> done say that the IMD doesn't change for carrier separations of 1kHz
> thru 100kHz.  In your case you might want to use 10kHz based on the
> receiver you are using.
>
> Have fun.
>
> 73,
> Larry, W0QE
>
> On 8/1/2012 2:15 PM, Bill, W6WRT wrote:
>> Is there a way for a homebrewer to accurately measure IMD without 
>> expensive
>> equipment? I'm thinking of doing two-tones at full power into a dummy 
>> load while
>> listening on a second RX nearby. How does one accurately analyze what 
>> one hears?
>>
>> 73, Bill W6WRT
>>
>> _______________________________________________
>> Amps mailing list
>> Amps@contesting.com
>> http://lists.contesting.com/mailman/listinfo/amps
>>
>
> _______________________________________________
> Amps mailing list
> Amps@contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>

_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

<Prev in Thread] Current Thread [Next in Thread>