Amplifier two-tone test

Helmut
Posts: 29
Joined: Sat Jul 15, 2017 3:08 pm

Amplifier two-tone test

Postby Helmut » Sun Oct 20, 2019 11:21 am

In the Groups-IO: FlexRadioSmartSDR group:

I just uploaded a PDF with the current test data. Over time there will be data on more linear amplifiers.
Likewise, we are compiling data on modern individual transceivers.
Use the link below to download the PDF file.

nc0b.com/imd

Note: all lower case, particularly the “imd” folder.

All data is dBc, NOT PEP. If you want the exaggerated PEP numbers everyone uses today, add 6 dB.
The dBc numbers are taken right off the spectrum analyzer using cursors.

There are some formatting issues with the PDF where some of the minus signs are not in the right place.

All data is –X dB below one of the two equal tones.

Please ignore this minor issue, as I don’t have time to fix that today.

Note: The test source needs to be significantly cleaner than the amp being tested. As you can see the Apache PureSignal (predistortion) provides that level of performance needed to make these tests 160m through 6 meters. The League only publishes IMD data on 20 meters using their test setup.

My memory wasn’t exactly correct as to the differences on 40 meters. You can compare the numbers in the PDF.

73, Rob, NC0B
------------------------------------------------------------------------

I made a test myself here with the 7000dleMK2 and a Gemini HF-1K LDMOSAmplifier on 40m and at

500 Watts I had around 50dB IMd3 down with Predistortion feedback from the Amplifier
1000 Watts at least 40dB IMd3 down with Predistortion feedback from the Amplifier

measured on my Flex 6500 via 90dB Attenuation from the Wavenode-Coupler (70dB) + 30dB Attenuator inline to my Dummy-Load.
Software Thetis 2.6.8 D2 and Orion_MkII_Protocol_2_v2.0_pre2.rbf / used tones 800+1800 because 700+1900 gave not the same peak.

The Anan shows the same amount of dB down on the DUP screen - so there is no difference to the external Analyzer (Flex in my case) and
the Anan measurement. I saw some complaining about differences here in the group but did not find the thread yet.

Would be nice - if you could try yourself to find the Predistortion Effectiveness with an Amplifier online and report here.

73
Helmut dl9eri/oe9eri
Last edited by Helmut on Sun Oct 20, 2019 1:52 pm, edited 2 times in total.
K9RX
Posts: 281
Joined: Fri Apr 14, 2017 3:47 pm

Re: Amplifier two-tone test

Postby K9RX » Sun Oct 20, 2019 1:04 pm

Anan 8000DLE

Using 2 tone test with Expert 2K-FA I see -65db or better. I see typically -60 on SSB.

Using 2 tone test with Elecraft KPA1500 (2XLDMOS) I see -50db. I see typically -45 to -50 on SSB.

Both at ~1200W out.

Disappointing as for other reasons I'm switching to the KPA1500, but from what I understand its the nature of the beast with LDMOS devices. Nothing the amp maker can do about it.

Gary
K9RX
User avatar
w-u-2-o
Posts: 1825
Joined: Fri Mar 10, 2017 1:47 pm

Re: Amplifier two-tone test

Postby w-u-2-o » Sun Oct 20, 2019 1:35 pm

Gary,

Those are very good numbers, which suggests you did might not have performed same test as Rob Sherwood, NC0B.

Unless I am interpreting Rob's results improperly, what they show is that Rob used PureSignal to linearize the 7000DLE only, then used the 7000 as a linearized exciter to drive the amp with the cleanest signal possible. This makes sense, as he would be hard pressed to find a lab-grade signal generator capable of 30 to 50W of output.

The amp, on the other hand, is left to operate in its natural state, i.e. without PureSignal linearization. The intent being to show how poorly, or well, the amp performed from an IMD standpoint when provided with the best input signal possible, certainly better than it would normally have from a more typical amateur radio transceiver.

The results are completely predictable. No amateur radio, or amateur radio amplifier, has good IMD performance at its specified maximum RF output power level. Obviously this makes a strong case for additional linearization methods, hardware and/or software-based, for ham radio equipment.

Indeed, Rob's results are a best case scenario, as the two tone signal will allow the power meter to read very accurately. Imagine how bad this gets when people raise their drive in order to see consistent maximum power readings when running phone on their average reading power meters.

It would have been interesting to see what the performance was for each amp running at half specified maximum RF output power, just to see if there was any improvement.

Also, it's a bit confusing that there is only a single table of 7000DLE measurements. One would think the drive conditions for each amp would be a little different.

73,

Scott
Helmut
Posts: 29
Joined: Sat Jul 15, 2017 3:08 pm

Re: Amplifier two-tone test

Postby Helmut » Sun Oct 20, 2019 2:37 pm

Gary and Scott,

yes, Rob Sherwood did this test to show the IMD-behave of the Amplifiers themselves - not using a feedback-output of the amp.
The Anan with its own predistorted signal is a good source to do this.

Comparing the IMD at the output of the amplifier with an Analyzer or other SDR-Radio with sufficient attenuation to the RX in DUP mode on the Anan - should give the same results when RX Bypass on TX is checked ? The Anan gets the feedback via RXBypass input (7000dle) and the internal coupler is bypassed - so it displays the real output IMD after the amplifier on the Anan RX.

Therefore it is no need the check the output after the amplifier with an Analyzer or SDR ?
That would confirm my measurement from today with no difference on Analyzer to Anan RX.

If you by mistake have the amp online and RXBypass on TX unchecked you see the good Anan IMD and after the amp it looks not so nice - hi,,,,

73, Helmut
User avatar
w-u-2-o
Posts: 1825
Joined: Fri Mar 10, 2017 1:47 pm

Re: Amplifier two-tone test

Postby w-u-2-o » Sun Oct 20, 2019 3:39 pm

Helmut wrote:Comparing the IMD at the output of the amplifier with an Analyzer or other SDR-Radio with sufficient attenuation to the RX in DUP mode on the Anan - should give the same results when RX Bypass on TX is checked ? The Anan gets the feedback via RXBypass input (7000dle) and the internal coupler is bypassed - so it displays the real output IMD after the amplifier on the Anan RX.

Therefore it is no need the check the output after the amplifier with an Analyzer or SDR ?
That is actually incorrect. As soon as you check Bypass on TX then RX1 looks at the Bypass input for PureSignal corrections. If you are using a coupler on the output of an external amp to look at this signal via the Bypass port then PureSignal will be linearizing the external amp.

The only way you can use ANAN series hardware to do this experiment properly is to a) use a model with a second ADC (200D, 7000, 8000), attach the external coupler output to RX2 (the second ADC input) and watch on RX2. Meanwhile you must leave Bypass on TX OFF.

73,

Scott
Helmut
Posts: 29
Joined: Sat Jul 15, 2017 3:08 pm

Re: Amplifier two-tone test

Postby Helmut » Sun Oct 20, 2019 3:51 pm

VY confusing for me, Scott

what I see on the Anan RX with DUP checked and RXBypass on TX checked for Amplifier use - is not the truth of IMD on the way to the antenna or dummy load ?

My Amplifier has a coupler built in with +10dBm/1KW which is on the RXBypass input of the 7000dle to linearize. The next coupler on the way to the antenna is from the WaveNode Powermeter (-70dB) + additional 30dB to the Analyzer or SDR-Radio to check the real IMDs at full output incl. linearized Amplifier. What I don´t know: is the spectrum on my Anan RX when TX with Predistortion the same as on my Analyzer ?

Thank you for your patience
Helmut
User avatar
w-u-2-o
Posts: 1825
Joined: Fri Mar 10, 2017 1:47 pm

Re: Amplifier two-tone test

Postby w-u-2-o » Sun Oct 20, 2019 6:53 pm

Helmut,

We may each be writing about two different things. To clarify:

If you select Bypass on TX and have the Bypass port connected to an appropriate coupler, then you are indeed seeing what is coming out of the coupler when DUP is ON.

The problem is that you cannot do that and perform Rob's experiment. Bypass on TX must be OFF on the 7000 in order for the 7000 to switch to the internal coupler during TX so that only the 7000 is linearized, as the experiment requires that only the 7000 (or whatever hardware you are using) is linearized, not the amplifier.

If you wish to perform Rob's experiment, then Bypass on TX must be OFF, only the radio is linearized, but you can monitor the output of the amplifier coupler on RX2 if you wish.

73,

Scott
Helmut
Posts: 29
Joined: Sat Jul 15, 2017 3:08 pm

Re: Amplifier two-tone test

Postby Helmut » Sun Oct 20, 2019 7:34 pm

w-u-2-o wrote:Helmut,

We may each be writing about two different things. To clarify:

If you select Bypass on TX and have the Bypass port connected to an appropriate coupler, then you are indeed seeing what is coming out of the coupler when DUP is ON.

73,

Scott


Thank you Scott ...

that´s what I wanted to know - when we are testing Predistortion with the Amplifier that has feedback (via internal or external coupler) - we see the result on the Anan RX with DUP on - that reflects the IMDs going to the antenna.

So there is no need to check after the Amplifier with another coupler on the Power-Meter to an Analyzer or second SDR-Radio - as the results are the same - as I measured today and posted in this thread before.

73
Helmut
User avatar
w-u-2-o
Posts: 1825
Joined: Fri Mar 10, 2017 1:47 pm

Re: Amplifier two-tone test

Postby w-u-2-o » Sun Oct 20, 2019 9:26 pm

Helmut wrote:So there is no need to check after the Amplifier with another coupler on the Power-Meter to an Analyzer or second SDR-Radio - as the results are the same - as I measured today and posted in this thread before.
Unless you are doing Rob's experiment. Then you have to use another path other than Bypass to monitor the output of the amplifier. I hope you understand that.
Helmut
Posts: 29
Joined: Sat Jul 15, 2017 3:08 pm

Re: Amplifier two-tone test

Postby Helmut » Sun Oct 27, 2019 6:07 pm

K9RX wrote:Anan 8000DLE

Using 2 tone test with Expert 2K-FA I see -65db or better. I see typically -60 on SSB.

Using 2 tone test with Elecraft KPA1500 (2XLDMOS) I see -50db. I see typically -45 to -50 on SSB.

Both at ~1200W out.

Disappointing as for other reasons I'm switching to the KPA1500, but from what I understand its the nature of the beast with LDMOS devices. Nothing the amp maker can do about it.

Gary
K9RX


Gary,

the KPA1500 and my Gemini HF-1K have both PIN-diode T/R switching,
could be the reason of 10-15dB lower IMDs than other Amps with Vac-Relais or contact switched ones ?

What do you think ? Today I saw that there is a Gemini DX1200 now, similar to the HF-1K but with vacuum relay and Roger posted on twitter
the two-tone-test with rf-generated tones (no audio) generated from Dual-Signal-Generator R&S:

https://pbs.twimg.com/media/EH5Eie6WoAAskOZ.jpg

Impressive for Dual BLF188XR LDMOS.

73 Helmut
K9RX
Posts: 281
Joined: Fri Apr 14, 2017 3:47 pm

Re: Amplifier two-tone test

Postby K9RX » Sun Oct 27, 2019 7:31 pm

Helmut,

I can't say. I don't know if the issue is something associated with the structure of the LDMOS devices or not. That was my thought, that it was associated with the LDMOS construction - its a hell of a lot of power in one very small package vs. with the Expert 2K at least there are 8 MRF151's.

Assuming the PIN diodes are biased off when transmitting I'd not think they'd have an impact - or if so only at higher powers ... if they are conducting at all they'd be damaged in an instant so one has to believe they're doing their job and are off when they need to be. Are there other suggestions/tests done that suggest PIN diode switching increases IMD?

Gary
K9RX
Helmut
Posts: 29
Joined: Sat Jul 15, 2017 3:08 pm

Re: Amplifier two-tone test

Postby Helmut » Sun Oct 27, 2019 8:29 pm

Gary,

I think it is not a noticable difference - if you compare the results of Rob Sherwoods PDF at

nc0b.com/imd

you can notice, that the IMDs are between 20 and 37 dBc on both the PowerGenius XL and the KPA1500.
Most visible is the huge variation from band to band. I see also the variation of the consumed current
at the same output power from band to band on my Gemini HF-1K and hands down we will see it on more
LDMOS Amps.
I asked Roger Banks from Dxshop if he could do a test of the Gemini DX1200 with the Anan (he has a 100d i think)
in predistortion mode - curious on the results.....

best 73
Helmut

Return to “Amplifiers & Tuners”