Warning: Creating default object from empty value in /homepages/11/d132647312/htdocs/Amfone/mkportal/include/SMF/smf_out.php on line 47
K1LI/K1KP Amplifier Design Articles




 
The AM Forum
March 30, 2017, 06:48:59 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: K1LI/K1KP Amplifier Design Articles  (Read 420 times)
0 Members and 1 Guest are viewing this topic.
WA1ICN
Member

Offline Offline

Posts: 5


« on: February 25, 2017, 10:43:19 AM »

Monitoring one of my local 2 meter repeaters, I came across a group of contesters who were talking about the last Yankee Clipper Contest Club(YCCC) meeting in Enfield, CT. They  talked about a presentation on a class D, bandswitching(2-30Mhz) amplifier using GaN devices. Being contesters, they really did not understand the technology but they said it used "polar modulation" and there would be an article in the next QEX.

That article is now out and it is worth a careful read:

http://www.arrl.org/files/file/QEX_Next_Issue/Mar-Apr2017/MBF.pdf

The article lacks a lot of detail about how the amplifier works. Specifically, it did not talk about bandswitching filters and GaN devices. I still have not secured a copy of the presentation given at the 2017 YCCC meeting but I did find the one from last years meeting. I gives a lot more detail of their project circa 2016. It is buried in this large document:

http://www.yccc.org/Meetings/presentations/2016/february/agenda_02062016.pdf

My understanding of signal re-constructive technology is that most approaches have their roots in the Doherty modulation scheme from 50 years ago. Recent implementations of this type of approach include the SGC linearizing linear that used plug in coils and the ANON SDR transceiver that processes a sampled signal from any linear amplifier, then linearizes the transceiver output signal. The SGC product was never marketed because it allegedly did not work. GaN technology, while promising, is so new that experimenters have little experience with it.

Here is also a QEX article worth noting by Hershberger, W9GR, on a related technique.

http://www.arrl.org/files/file/QEX_Next_Issue/2014/Nov-Dec_2014/Hershberger_QEX_11_14.pdf

Have a look at the articles. If I can uncover further information I will post it. Comments and any other form of peer review would be useful.


Ashley
Logged
WBear2GCR
Member

Offline Offline

Posts: 3313


Brrrr- it's cold in the shack! Fire up the BIG RIG


WWW
« Reply #1 on: February 26, 2017, 11:54:16 PM »

It's fascinating.
Kahn clearly was a genius, as his ideas from 1952 show up today in such a stunning way!

The problem, speaking for myself, is that this is not a DIY item since it involves complex math,
DSP programming, and extensive LSI/surface mount hardware. Great stuff for manufacturers.

Looks like the folks who are presenting these ideas are professionals in related fields during the
daytime.

As an aside, I had read the Sokal Class E papers back in the 70's but saw no simple way to implement them myself back
in the day (and of course the focus was strongly on SSB at that time).

This is very interesting stuff... hard to follow, will have to sit and read these for a while and see if the ideas
can possibly sink in...

Those I & Q signals! I need more IQ to figure them out!! Cheesy
Logged

_-_- bear WB2GCR                   http://www.bearlabs.com
VE3ELQ
Member

Offline Offline

Posts: 152


« Reply #2 on: February 27, 2017, 10:21:04 AM »


This is very interesting stuff... hard to follow, will have to sit and read these for a while and see if the ideas
can possibly sink in...

Those I & Q signals! I need more IQ to figure them out!! Cheesy

I agree, a good read but mostly over my pay grade.

What might be fun to try would be a Class D linear based on our PW modulator architecture. Sample a low level SSB signal at about 15 to 20 mhz rate instead of 150 khz, then low pass filter back to baseband on 160M or possibly even 80M. The new GaN FETs should handle this.  Should be very efficient.  Just food for thought.

73s  Nigel
Logged

A "Tube Free" zone.
WA1ICN
Member

Offline Offline

Posts: 5


« Reply #3 on: February 27, 2017, 05:18:06 PM »

I spoke with Tony Brock Fisher K1KP about his amplifier. He said the QEX article was
about a year old compared to the technology he is now using but he does not have a new
presentation put together. I am going off telephone notes so I will give it to you raw.

He said he has a 1KW "linearized" amplifier in operation using class D current
mode RF amplification and class D PWM modulation. Key to the performance with class D are
GaN devices. He is using 2 X GS66504(GaN systems - $13)150W devices for the RF and 
C3M0120090 (CREE -$7 devices) for PDM modulation. He is getting 10 meter performance out of these things. Mouser stocks both.

A LM5114(http://www.ti.com/lit/ds/symlink/lm5114.pdf) is used as a driver for the finals.
He says they work at 10 meters but they do get hot. GaN Systems recommended these
devices. Previous use of IXD614 drivers failed above 40 Meters but this is a longtime well known
problem.

To drive the LM5114 he uses 2 Analog devices dividers(no number yet) to develop the IQ
signals. He had an interference issue with the AD part that caused the amplifier to run
at 1/2 frequency when using class E. That is now solved by using class D.

The GaN stuff works very well in RF service as they generally have much lower gate
capacitance than silicon types. There were some reports of parasitics on the GaN
reflector but he has not had any problems with parasitics.

DSP/SDR processing is done on board with a dedicated board. One IQ signal is used for phase, the other for amplitude. Most of the signal distortion that was corrected was phase distortion. This method seems to be similar to the ANON and HPSDR pure signal type of processor. 

I have sent Tony another email to get some further detail/clarification on a couple of points. More later.

Ashley
Logged
WA1ICN
Member

Offline Offline

Posts: 5


« Reply #4 on: February 27, 2017, 07:38:22 PM »

Responses from Tony K1KP
> Tony
> In going over notes I am missing the P/N for the two analog device dividers that
drive the LM5114s.
The comparators are ADCMP600's.
> Also what is the voltage for the GS665054 X2 RF stage.
Gate drive is 0-6V. Drain voltage is around 75V DC.
> Are you still using a low level linear amp on input?
Yes. We buffer the AD9957 output with a MiniCircuits Gali-84 MMIC.

Ashley
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.091 seconds with 18 queries.