Measuring transmitter IMD How To's?

[Chris Wilson]

I run a couple of home built Softrock SDR transceivers and have now connected one to my old Collins valved 30L-1 linear I restored a while back. It now gives about 5 to 8 watts. (WOW! ) I would like to see what the IMD looks like though, less than one watt with a bit of IMD isn't going to worry many people, but if I up the power I feel obliged to do my best to check the signal is as clean as I can make it. I have a good spectrum analyser, (HP8568B), but I am not very good with it as I worry about blowing it up. I am becoming a bit more confident since getting a pair of DC blocks and some step and fixed attenuators. I also have a biggish 100 watt 30dB attenuator and an adjustable RF coupler. There are software programmes to inject a two tone audio frequency into the SDR transceiver, I think. I believe this is all I might need? So what I am short of is clear instruction as to what to do and how to interpret the results. I believe just looking at the signal on my `scope can be misleading as it may not show IMD well, or at all.

I looked on You Tube and Googled, but found nothing of great use. Can anyone assist please? Thanks.

[JackOfVA]

Concept is as you say, inject the two test tones in the transmitter and look at the output with a spectrum analyzer.

Obvious things to watch out for ...

1) Test tones must be clean
2) You don't want to blow up the spectrum analyzer input stage, so use the high power 30 dB attenuator. If you are really paranoid (I am when doing these tests at 100 watts output) you can also use an external 10 or 20 dB attenuator at the high power attenuator output ahead of the SA.
3) You want to keep the spectrum analyzer operating at an input level so that it does not create internal distortion. Common way to check levels to avoid this is observe the intermodulation distortion level on the SA and then increase the SA input attenuation by 10 dB. The carrier and distortion products should all drop the same amount, 10 dB. If the distortion products drop in level more than the increase in attenuation, then the SA is overloaded and generating its own internal distortion.
4) You will need a relative narrow resolution bandwidth setting, 100 or 200 Hz and preferably lower to resolve the IMD components.

Attached images are typical results I measured of my Elecraft K3 transceiver and a block diagram of the test setup.

Jack - K8ZOA

[Chris Wilson]

GREAT info Jack, thank you! My only doubts here are how to generate clean tones. I may already have downloaded a two tone test generator software app, that uses the PC's sound card. The card is a M-Audio Delta 44, so fairly decent. Do you think this is a viable way of putting the tones into the SDR transceiver, or do I need something better / different? The other doubt is how do I choose the tones frequencies and amplitude? I take it amplitude is low enough not to over modulate the transceiver, but the frequencies to use? Thanks very much, very helpful info and images!

[vk3yedotcom]

Chris, 

The tones need to be clean sine waves.  They must not be harmonically related.   Tones around 1 kHz should be OK - frequency not that critical. 

Phase shift oscillators are fairly simple - just a transistor and few resistors and capacitors.  Or an op-amp circuit (maybe use a twin or quad op amp for 2 tones generated from the one package?).

The level doesn't need to be high as the transmitter's mic amplifier will be in line and the mic gain control (presumably there is one) will help.  Tests should be done at or just under the rated power output into a dummy load.  Too high will introduce distortion while too low a power may not be as clean either.   Don't run the tone too long at full power as transmitters are not necessarily rated at 100% duty cycle.

[JackOfVA]

The test tones should not be harmonically related.  Most commonly used are 700 & 1900 Hz (ARRL) and 800 & 1800 (some commercial/military).

A sound card should be capable of putting out two clean tones, but if you prefer a stand-alone device, Elecraft manufactures an moderately priced two-tone test generator kit for 700 & 1900 Hz, model 2T-gen for US$ 69.95.

The old Motorola application note EB38 may also be worth reading: http://www.freescale.com/files/rf_if/doc/eng_bulletin/EB38.pdf

The ARRL's test procedure manual is also good reading - http://www.arrl.org/files/file/Technology/tis/info/pdf/Procedure%20Manual%202010%20with%20page%20breaks.pdf

You should also be aware that there are two ways of defining the intermodulation level. The commercial / military standard references the output intermodulation product to the power of a single tone. This is easily read on the spectrum analyzer as the difference in dB  between one test tone and the IMD product.  The ARRL method references the intermodulation product to PEP, which is 6 dB above the single tone. If you want ARRL protocol, make the same measurement, but add 6 dB to the difference.

As an example, if the 3rd order IMD product is 30 dB below one of the reference tones, the commercial/military specification will be "IMD product 30 dB down" and the ARRL's version will be "IMD product 36 dB down."

I prefer the commercial/standard methodology, but either is OK, provided you make it clear which reference is being used. If you wish to compare your measurements with ARRL data, then keep this 6 dB difference in mind.

[Chris Wilson]

Thanks for the replies. I have had a bit of a fiddle about. I made an Audacity dual tone file one at 1330 Hz and one at 550 Hz, and combined them, and ran it into the Softrock audio input. I measured the TX output at the aerial socket via a 30dB attenuator into the SA. I played with RBW and reference level settings as I wasn't sure of the best way to show the results. The image *55.jpg is with the Softrock at full 1 Watt power with sensible modulation levels. The earlier SA scans are with about 750 mW output. The `scope scan is with the Softrock at full power. I can't see any clipping. I sniffed the output into a dummy load. I am sure I am doing some things wrong, criticism welcome. Thanks! Once I am more confident about the measurement technique required I will try with the PA in circuit. Right now one Watt is enough to play with the settings.

http://www.gatesgarth.com/IMD-SCANS/imd.zip