PA with predistortion

[Sverre Holm (LA3ZA)]

"Apache Labs become the first Amateur Transceiver manufacturer to include Pre-distortion technology within their PA’s for ultra-clean TX".

This is something that Elecraft should have all the possibilities to follow up on: using the DSP to clean up the PA.

See http://blog.wsplc.com/?p=1701

[John Marvin]

I'm not sure how easy this would be to adapt to an existing product like
the K3 or KX3. Note that the primary change is in the transceiver and
not the PA, i.e. this is not something you would add to the KXPA100 or
KPA-500, other than perhaps the head end of the PA output sampler. The
attenuator part of the sampler could also go in the PA or that could go
in the feedback input that would have to be added to the transceiver. In
a self contained transceiver (i.e. with a built in PA) this could all be
contained within the transceiver, although it would still be nice to
provide an external feedback path for those with additional
amplification (e.g. a KPA-500). This is what Apache-Labs is doing with
the not yet released ANAN-200D. Earlier HPSDR based products based on
the Hermes board or Angelia board are also capable of supporting
predistortion, but it requires some changes, along with an external RF
sampler of some type (although some people are getting decent results
from just the crosstalk in the ANAN-100D case, although some of the
settings need to be calibrated on a per band basis since the crosstalk
level tends to change with frequency).

In order to do this you need to add a feedback path in the transceiver
and the transceiver needs to support full duplex operation, i.e. either
the feedback path needs to connect to an additional receiver or the T/R
circuit needs to connect the feedback path to the receiver while the
transmitter is connected to the PA/antenna. The DSP then needs to be
able to monitor both the outgoing transmit signal and the incoming
properly attenuated signal from the PA output. This can all be built
into the transceiver for the "barefoot" case (I'm not sure barefoot
applies in the case of a K3/100!). Of course the DSP needs to be capable
of computing the required predistortion based on the feedback input,
accounting for delay in the transmit path, and applying it to the
transmitted signal.

I will also note that the results are best with higher voltage PA
designs, e.g. LDMOS 50V amps. But the results can still be quite good
for well designed 13.8v amps. I'm working on getting this to work with
my HPSDR Hermes and my KXPA100, although it is a low priority task right
now. I need to verify that the feedback signal from the KXPA100 is
reasonably stronger than crosstalk within the transceiver, which
contains a Hermes board and the 5W driver amp from the Hardrock 50 kit,
and yet it can't be so strong that it overloads the ADC.

When it comes to IM3 distortion, applying correction from earlier in the
transmit path can still result in some improvement, i.e. the distortion
from earlier stages just gets amplified and added to new distortion
products at each stage. When I looked at the KXPA100 test results that
Elecraft submitted to the FCC, I noticed that the input test signal
wasn't all that great. I wondered how much of the IM3 product at the
KXPA100 output was contributed by the KXPA100 and how much was just the
IM3 distortion from the input amplified by the KXPA100.

Regards,

John
AC0ZG

On 2/25/2014 10:31 AM, Sverre Holm (LA3ZA) wrote:

> "Apache Labs become the first Amateur Transceiver manufacturer to include
> Pre-distortion technology within their PA’s for ultra-clean TX".
>
> This is something that Elecraft should have all the possibilities to follow
> up on: using the DSP to clean up the PA.
>
> See  http://blog.wsplc.com/?p=1701 <http://blog.wsplc.com/?p=1701>
>
>
>
>
> -----
> Sverre, LA3ZA
>
> K2 #2198, K3 #3391,
> LA3ZA Blog: http://la3za.blogspot.com,
> LA3ZA Unofficial Guide to K2 modifications: http://la3za.blogspot.com/p/la3za-unofficial-guide-to-elecraft-k2.html
> --
> View this message in context: http://elecraft.365791.n2.nabble.com/PA-with-predistortion-tp7584706.html
> Sent from the Elecraft mailing list archive at Nabble.com.
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Ignacy]

Predistorsion properties are impressive. The following presentation

http://users.burlingtontelecom.net/~[e-mail]/images/Eastern_VHF_UHF_Conf_2014.pdf

gives some information what is possible. In LDMOS amp with bias almost reduced to 0, IMD3/5 change from -22/-28db to -55/-63db. Also a crappy 13.8V CB amp with IMD3/5 of -18/-28 becomes -47/-53 after predistortion.

Class AB amps have an efficiency of around 50%. B class amps (also in drivers) perhaps can raise it to 70%. 40% less input current and 55% less dissipation, all at excellent IMD. All for nothing, oops, for lots of programming if DSP is powerful enough.

Perhaps in 10 years nearly all new rigs will have predistortion implemented.

Probably K3/KX3 do not have powerful enough DSP to do it. For example, the biggest limit in KX3 on SSB is speech processor that is currently just a treble boost.  6db lost. The fact that it is not done means that it cannot be done easily.

Ignacy, NO9E

[Jim Brown-10]

On 7/11/2014 11:01 AM, Ignacy wrote:
> Predistorsion properties are impressive. The following presentation
>
> http://users.burlingtontelecom.net/~n1jez@.../images/Eastern_VHF_UHF_Conf_2014.pdf
>
> gives some information what is possible. In LDMOS amp with bias almost
> reduced to 0, IMD3/5 change from -22/-28db to -55/-63db. Also a crappy 13.8V
> CB amp with IMD3/5 of -18/-28 becomes -47/-53 after predistortion.

I skimmed the presentation. There are no surprises. Several thoughts.

First, this is pretty complex -- that feedback (sense) loop from the
amplifier output is critical. Why not simply buy a cleaner rig and amp,
and run them under conditions that make it cleaner?  Study the
presentation that K6XX and I did for NCCC last fall.  The measures that
Bob talks about are very easy to implement.

http://audiosystemsgroup.com/K6XXAmpTalk.pdf

A decent tube amp, carefully tuned, is about as clean as the results of
predistortion, and is a lot simpler.

Also, it seems to ignore phase noise, which is produced by the rig. It
also ignores sidebands produced by distortion that results from
overdriving audio stages.

FWIW, I see no reason for Elecraft to support it other than possibly for
their own power amps. I suspect Wayne has looked at it.

73, Jim K9YC



______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]