Checking amplifier linearity with a set of three circuit boards?

[K1JJ]

While on 75M ssb last night working DX, I got a call from a good friend, Nick, LZ1JY in Bulgaria. He's the guy many AMers have worked split freq on 75M AM years ago.

Anyway, we got to talking about amplifiers and linearity. He had some ties with the Acom International amplifier company that makes FB linear amps. He mentioned the amplifiers used a set of built-in PC boards that measure the input of the amplifier RF, the output RF and then compares the results to give a + - indicator of linearity. (three PC boards)

I have seen detectors used for RF negative feedback that let you know when the input tuning was the same as the output tuning as a phase detector, but never heard of this used for a linearity check.

Nick said he will be sending me info via email and the actual boards to use in my linear amp... but in the meantime, anyone have a technical idea on how a system like this would work to show linearity using analog and/or digital technology?

I realize I cud use a trapezoid or spec analyzer to the same thing, but this seems like a cool idea if it gives similar linearity error resolution.

Thanks.

Tom, K1JJ

[Steve - WB3HUZ]

RF tap off the output, and one off the input. Set the levels and feed a comparator.

[K1JJ]

RF tap off the output, and one off the input. Set the levels and feed a comparator.

Yep, Nick said it used a comparator.

So how would you "set the levels?"  Do you need a known linear signal to put through the system to calibrate it or some other method unrelated to the actual amplifier system itself?

T

[WA1GFZ]

I'm thinking you need a phase comparator not a typical level type.

[K1JJ]

Well, it dawned on me....

If the input RF generates 1 volt from the first board and the output RF level is also set to 1 volt from the second board when using a small signal into the amp.... THEN as the signal is increased, the in and out voltage ratio shud remain the same. IE, 2 volts in shud give 2v out, etc. Any departure would mean negative or positive non-linearity. (As Nick called it)   If the comparator sees the signal levels remain the same, then that is perfect linearity.   Maybe no phase is needed here, just detected RF levels.

Does this make sense?

T

[WA1GFZ]

How would you tell the difference between -10 and -40 dB IMD. I supposed you could integrate the pulses to a DC level. How do you deal with the  degree phase shift through the amp which must have some dealy? Seems like way too much work when you can just look at a spectrum display on a soft rock and see it calibrated in dB. Remember when I was doing the test with two signals into the spectrum analyzer checking the erb board. The soft rock and and good RX would do the same thing.

[K1JJ]

I wud think the zero indicator wud have to be calibrated for IMD. Maybe it has a procedure already for this.

Yeah, I especially like measuring relative IMD with my other FT-102 receiver. It really shows the differences between amps once everything is normalized to S9 +40 over for all amps.

I think the advantage of these boards is that once you determine the amp is a certain level of cleanliness, you can then quickly tune it up on any band or freq, on the fly, and adjust the loading, etc, quickly against the zero indicator.  At least that's the impression I got from Nick.  Tuning it up with a spec analyzer or another RX takes some set up time and interpretation. A zero indicator, assuming it works well, seems ideal.

I'll show you the schematic info once I get it and see what you think.

T

[WA1GFZ]

no brainer with soft rock. Set the cursor on the carrier peak, write down the number that comes up then move it to the spur and do the same thing. then you have to do this very complicated math function....subtract
I found it easier to do a two tone input so used a pair of generators. Plenty of sound card two tone generators you can pump through the rig to do the same thing.
There is nothing cooler than an FFT spectrum analyzer for sorting things out.
OH and you will find the spectrum display is very accurate. It tracks the attenuator on an HP8640B very well.
  

[DMOD]

Does this board set have a DSP or a PIC with an FFT analyzer?

I would think one would have to a fast FFT algorithm to measure nth order IMD products.

Or am I thinking to complex here?

Phil - AC0OB