AM Operation Using Power SDR with Flex and Other SDR Platforms

[n1eu]

I wanted to add a few additional tips using PowerSDR OpenHPSDR with ANAN/Hermes/Angelia.

Leave the front panel Compressor OFF.  It causes serious audio degradation.  Don't even use it in ssb.

DO use the Leveler.  Enable it under setup-dsp-AGC/ALC.  I set it for 10dB and didn't touch the default timing settings (attack 2msec, decay 500msec, hang 500msec).

I have found very mixed results using the ANAN mic and line inputs.  If it works for you, great.  What I did find to work VERY well and delivered a noticeably cleaner sound was to use a Focusrite 2i2 external audio interface.  It is a high quality mic preamp coupled with a high quality soundcard in a small box with 2 XLR inputs and a USB cable to connect to the computer.  It also has balanced line level inputs/outputs and phantom power that I don't use.  I connect it via VAC to PowerSDR under setup-audio-VAC1.  Enable it, shoose ASIO driver and Focusrite Input.  My other settings are Buffer 2048, 96000 Sample, Stereo, 8dB TX Gain, Combine VAC Input Channels, Stereo and 120msec Buffer Latency.  I'm admittedly no expert on these settings but this is what works well for me.

Here's a video of Carlos CT4RK's very cool vintage shack while he tuned my ANAN's s9+ 10M signal in on his Geloso G4/214 :  http://www.youtube.com/watch?v=-rNzDCw8oZU

73, Barry N1EU

[N2DTS]

Nice shack, nice receivers!

[W1AEX]

My ANAN-100 Hermes based Apache Labs rig has been on my operating bench for 10 days now and I've had the chance to run quite a bit of AM with it. Unlike the VAC approach that Barry - N1EU is using, I ran my hardware audio chain through an Edcor 1:1 600 ohm transformer into the unbalanced rear panel audio input connection. That approach has worked cleanly for me and I have experienced no issues with RF feedback on any band at any power level. The performance has been every bit as good as I had hoped it would be and the changes that Warren NR0V and others have made to OpenHPSDR mRX have been really good ones for AM. The audio levels could not be simpler to optimize and as far as I can tell it's close to impossible to drive the transmitter beyond 0 dB on the ALC meter with any voice mode. As long as you set the mic/line input level, eq level, and the leveler to undistorted settings it will cleanly and faithfully transmit whatever your TX audio source is.

The links below will take you to a pair of videos with the first one showing my ANAN 100 injected with a 1000 cps tone to produce a sine wave and the second showing obvious asymmetry while transmitting with voice. I chose an AM Carrier Level setting of 70 for these videos but have found anything in the range between 70 - 80 to be very clean and loud sounding.

https://www.youtube.com/watch?v=dCJpenhEyx4

https://www.youtube.com/watch?v=aNrKCXeiveY

The pictures below are snapshots from the two videos. The first one shows a 1000 cps sine wave reaching beyond 100% positive as it approaches the baseline in the negative direction. Note that when you select an AM Carrier Level that is less than 100 you enable the transmitter to produce this kind of asymmetry. The second picture shows the asymmetry that can be developed effortlessly with voice modulation. I haven't yet received an audio "air check" recording of the rig during a QSO on AM but the reports I've received have been very positive. The third attachment below is a short mp3 made with the audio recording function of the SDR-Console software that I use to run my Fichten Field Day SDR. It gives a pretty good idea of what the transmitted signal sounds like in spite of the heavy AGC action due to proximity effects.

Rob W1AEX

[W1AEX]

A remarkable feature that has been implemented in the OpenHPSDR mRX software is the ability to run pre-distortion linearization with AM and SSB. This protocol is named "Pure Signal" and its ability to improve transmitter IMD is impressive. Once the user has calibrated the power table in OpenHPSDR mRX so that the PA produces 100 watts output from 160 - 6 meters all that is required is to optimally set the sampling level for each band to allow the algorithm to remove transmitted distortion products. The effect is very dramatic with 3rd order IMD improving from a fairly typical level of -30 dB to around -50 dB.

The video below shows my ANAN 100 transmitting a 10 watt AM carrier while being received by an SDR across the room. If you watch above and below the 10 KHz wide transmit bandwidth, you can see the effects of pre-distortion linearization on the signal. Note that the received audio has some grit due to a recording level issue in Windows 7 that I have not yet resolved. The ANAN actually produces very clean audio when transmitting. At any rate, it would appear that the pre-distortion linearization protocol really helps to produce a very neighborly AM signal.

https://www.youtube.com/watch?v=o8h6ParzsAs

The two pictures below are screenshots from the video that demonstrate the difference when "Pure Signal" is toggled on and off.

In my setup, at this time I am only optimizing the IMD of the ANAN itself. By using an RF sampler at the output of a linear amplifier and bringing that sample back to an RX port on the ANAN rig the "Pure Signal" algorithm will also correct distortion produced by the amplifier as well. Those who are using receivers with spectrum displays can easily spot the amazingly clean signals where this has been implemented.

Rob W1AEX

[W2VW]

Great report as always Rob.

Do they have a way to include an outboard final AMplifier within the pre-distortion loop?

[n1eu]

Yes, but you have to fashion some sort of RF Sampler at the output of your amplifier.

[W2VW]

Thanks Barry.

[W1AEX]

After doing some troubleshooting I was finally able to resolve the distortion issue that plagued my video capture software after my migration to Windows 7. The direct stream audio captures heard in the videos are a good representation of what the ANAN sounds like and the video capture of the GSDR panadapter gives a good representation of the signal's occupied spectrum characteristics. The two links below are videos that were made with the ANAN-100 transmitting into a dummy load as the FiFi SDR received the signal while the Video/Audio capture software was used to make direct stream recordings.

ANAN-100 transmitting AM with a TX bandwidth of 10KHz  https://www.youtube.com/watch?v=tJk36tWxmpw&feature=youtu.be
ANAN-100 transmitting SSB with a TX bandwidth of 3KHz  https://www.youtube.com/watch?v=MCm3mKHyuAo&feature=youtu.be

Keep an eye on the filter edges and on the spectrum around the transmitted signal when pre-distortion is active. The pre-distortion linearization correction can be wrapped around the output of an amplifier by inserting a directional coupler at the output of the amplifier and bringing an attenutated RF sample back into one of the ANAN's RX ports. Last August the ARRL acknowledged Warren Pratt - NR0V for his development and implementation of the pre-distortion linearization module in the OpenHPSDR mRX software as they presented him with the 2014 ARRL Technical Innovation Award. I definitely love what he has done and also love transmitting splatter-free!

Rob W1AEX

[W2VW]

love transmitting splatter-free!

Rob W1AEX

Great post as always and maybe the concept will spread to the rest of your state 

[n1eu]

Wow, nice job with the video and the transmission Rob!

I'd love to see/hear how a Flex 6000 would do on AM in comparison to the ANAN-100.

73, Barry N1EU

[W1AEX]

Thanks for the comments Dave and Barry. From what I have heard on the air the Flex 6.xxx rigs have the potential to sound every bit as good as anything out there. The only Flex 6.xxx owner that I have heard on AM is Alan K2WS and he sounds very nice on his. There is a recording of his signal as it was heard last winter on 10 meters at this link:

https://www.youtube.com/watch?v=3zNhBRa0imU

That was recorded almost a year ago (last October) so SmartSDR was not as evolved as it is today. If I can catch Alan on the air I'll definitely ask him if I can make a recording so that I can add the Flex 6.xxx to my collection of videos. Depending on how SmartSDR comes along I will probably end up playing around with one of the Flex 6.xxx series but I have to admit that the ANAN is so feature rich that I have no interest in parting with it yet!

73,

Rob W1AEX

[W1AEX]

The video at the link below shows CuSDR 64-bit running my ANAN 100. This is version 0.3.2.14 which is the latest publicly available beta copy of this software that at this time is capable of only receive functions. This beta version is a little rough and was simply released to give people an idea of what is to come when more mature versions are forthcoming. The developer is currently implementing TX functions into the software, complete with the pre-distortion linearization protocol that is currently available in OpenHPSDR mRX. In the video, the software is running 3 slices that include an AM broadcast station, an AM QSO on 40 meters, and some long haul utility aviation traffic being handled by one of the NY Advisory stations in the 8 MHz band. In the first screenshot the slice interface is opened up to change the bandwidth of the slice tuned to the AM broadcast band. In the second screenshot you can see some pretty bad IMD present in the waterfall that is displaying an upper sideband signal from an aircraft communicating with NY Advisory. Under the link below are some notes about the contents of the video.

Rob W1AEX

https://www.youtube.com/watch?v=jlJZEft1Ci4

Watch at 720p for best clarity. Currently, CuSDR64 can run 5 RX slices along with the full spectrum panadapter under the frequency and s-meter panel. When the 1 gigabit ethernet firmware upgrade is completed my ANAN-100 will be able to display 7 slices. Each slice can be un-docked from the main display if desired and placed anywhere on your desktop. To switch to one of the RX slice audio streams simply click on that slice display. In this video, the main window is tuned to 0.880 MHz WCBS in NYC with one slice set to receive a 40 meter AM QSO and the other slice set to receive NY Advisory as it directs air traffic through its air space. There are differences in RX gain as you move from slice to slice which I believe is due to different bandpass filters being switched into line in the ANAN-100 as it moves to different frequency ranges. Note that at 5:30 into the video you can observe an aircraft transmission that exhibits severe splatter issues due to an issue with the transmitter in the plane. Looks like something is overdriven with resultant IMD problems.

[N2DTS]

That pre distortion looks amazing!
Pick the bandwidth you want and no matter what you do its not going wider.
TX 1 Hz square waves and so on, truly amazing.

[n1eu]

Rob, as far as I can tell, tx capable cuSDR is pure vaporware and some folks have privately told me they don't think it is ever going to happen.  I guess we'll wait and see.

73, Barry

[W1AEX]

Brett - I am amazed at the way pre-distortion removes distortion inside the bandpass as well as outside the intended transmit bandwidth. I am hoping that the developers are able to refine the implementation to make it easier for the non-technical guys to run it. Not that it's difficult to set up but it's a little intimidating for the plug and play operators.

Barry - A couple of months ago I found a small bug in CuSDR and posted it to the yahoo group to see if anyone else noticed it. Within a couple hours of posting it I received an email from Hermann, who is leading the development of CuSDR, thanking me for bringing it to his attention. He mentioned that the only major obstacle to quicker development is the amount of time he can devote to the project. It is my understanding that he is collaborating with Warren and others in the OpenHPSDR community to implement the new DSP engine and the pre-distortion protocol. The actual interface for CuSDR is pretty much done. The release of the gigabit ethernet driver/firmware for the ANAN main boards is also an important factor in pushing the functionality of CuSDR (and OpenHPSDR) to its maximum. I know they have a working gigabit driver and firmware but have no idea if they have it out there for the alpha testers yet. There is no timeline but I would not be surprised if an updated pre-alpha type release comes out in the next 6 months for people to play around with. I'm very content to play with OpenHPSDR for the time being!

73,

Rob W1AEX

[n1eu]

Just a note that in recent versions of OpenHPSDR PowerSDR (3.2.19, 3.2.20), I don't seem to see the need to adjust carrier level.  With the default setting of 100, I'm seeing 100% negative and 150% positive modulation on the REA mod monitor.

73, Barry N1EU

[WB2EMS]

recent versions of PowerSDR (3.2.19, 3.2.20

Where did you come across these versions? The latest I've seen on the flex site is 2.7.2. Is there another source for early release versions?

Kevin, WB2EMS

[n1eu]

Sorry, I'm speaking of the OpenHPSDR version of PowerSDR that runs on Apache Labs and OpenHPSDR hardware.

Looking back at my posts in this thread, there's some outdated info.  I've got the latest audio setup info that is working well for me and recommended by others at http://anan-100d.wikidot.com

73, Barry N1EU

[WB4AIO]

Free Band Text Editor for PowerSDR from Ray - K9DUR

This is a very slick program that allows you to insert the established AM operating frequencies or AM windows on any band. There's nothing worse than seeing 75M SSB in the "Band Text Field" as you're having a nice AM conversation with your buddies. Now you can set ranges of frequencies or individual frequencies to be labeled however you want them. It took me less than 10 minutes to label every band with my AM operating areas. Note that the "BandText Editor" only works with versions of PowerSDR that use the .xml format file for the database. This means it is compatible with PowerSDR version 1.18.1 and all the 2.x.x versions that followed.

Over the years Ray has written some very nice utilities for managing the PowerSDR database files and this is another fine piece of work. The software is available for free from his site at the following link:

http://k9dur.rnaconsultingservices.com/Downloads.aspx

Ray generously offers his software utilities for free but you might want to consider sending him a few bucks if you enjoy what he has created.

Rob W1AEX

Using the excellent K9DUR software, not only did I change the band text for 3870-3890 and 7280-7300 kHz to "AM Window" -- I also replaced the term "SSB" wherever it appeared with the more appropriate term "Phone."

73,

Kevin, WB4AIO.

[n1eu]

I get a 404 error with that link.  I believe the correct link is http://k9dur.rnaconsultingservices.com/powersdr_utilities.html

[WB2EMS]

Sorry, I'm speaking of the OpenHPSDR version of PowerSDR that runs on Apache Labs and OpenHPSDR hardware.

I thought it might be something like that. Can Open HPSDR run on the flex radios?

Kevin, WB2EMS

[n1eu]

No, and you have to appreciate the irony in Flex creating something that other folks ran with and Flex cut themselves off from.

[W1AEX]

For those who are interested in learning more about the hardware, software, setup, maintenance and modifications for the Apache Labs ANAN transceivers visit the Wiki that Barry - N1EU has put together at the following link:

http://anan-100d.wikidot.com/

The volumes of information he has collected there would be helpful to prospective purchasers as well as current owners.

Rob W1AEX

[W1AEX]

The OpenHPSDR mRX PS software used with the TAPR and Apache Labs ANAN rigs offers a feature called adpative pre-distortion linearization (APD) referred to as "Pure Signal". When an RF coupler is installed at the output of a station's linear amplifier and an attenuated RF sample of the amplified signal is returned to the ANAN hardware the algorithm used in OpenHPSDR is capable of substantially reducing IMD throughout the entire transmitter chain. It is not unusual for stations employing APD to produce signals exhibiting IM3 better than -50 dB.

In the video at the first YouTube link below you can see what Pure Signal does to the 10 kHz wide AM signal of Chuck K1KW following his installation of the "wrap around" hardware and the proper adjustment of the Pure Signal software settings. Chuck is running his ANAN-100D into an 8877 and producing an amazingly clean signal on 40 meters. In the panadapter, watch the area on both sides of his transmitted signal when Pure Signal is toggled on and off. The waterfall view presents a nice view of the absence of IMD artifacts when Pure Signal is engaged. The second link shows what Chuck's signal looks like with and without Pure Signal APD while running a 3 kHz bandwidth on SSB.

The link directly below shows the components I needed to implement pre-distortion linearization. The block diagram illustrates the station setup with an external coupler and relay installation that is used to wrap Pure Signal around my entire transmitter chain. Note that current production ANAN rigs no longer require an external relay as the newly designed PA board uses an internal relay to route the sampled signal to the RX input:  

http://www.w1aex.com/anan/anan.html#relay

With Pure Signal APD a big signal can also be a very neighborly signal to stations on adjacent frequencies!

73,

Rob W1AEX

https://www.youtube.com/watch?v=TtbNBHhVqeo    (AM video)

https://www.youtube.com/watch?v=xIimiVMmhmM    (SSB video)

[W1AEX]

In the first week of August (2015) the Zeus Radio SDR manufacturers announced their intention to make the SDR software used by their Zeus Radio hardware compatible with the Apache Labs ANAN hardware. Since then several beta versions have been released with added functionality for the ANAN products. The software is very nice looking and has a lot of clever GUI enhancements such as being able to undock the panadapter/waterfall from the control panel and on-the-fly sampling rate changes (from 48k to 384k) simply by grabbing the frequency band between the panadapter and waterfall with a right mouse click and moving left or right. The current beta has TX and RX capability and offers a very complete built-in voice processing package. Transmit bandwidths are variable from 2kc to 10kc for AM and 1kc to 5kc for sideband. In addition, there is native support for VST plugins in the Zeus Software. There is also a pre-distortion linearization module implemented in the software but I am not aware of anyone with ANAN hardware that has tried that out yet. Note that the software does not appear to recognize the hardware MIC and LINE transmit audio inputs of the ANAN radios but instead relies on software derived audio. I'm certainly no expert with this software package but I did mess around with it a bit and made a few videos for those who are interested:

Setting up the TX audio levels for AM:     https://www.youtube.com/watch?v=-0IFU2bcGX0

Tuning around the AM broadcast band:      https://www.youtube.com/watch?v=t4Q1EHi21iE

This project is kind of a moving target as Zeus Radio coders release new beta versions with more features enabled. Should be fun to explore what they are able to do with this. Another interesting development is the announcement by Simon Brown that he is implementing full TX/RX support for the ANAN line of transceivers into SDR-RADIO version 3.0 with a targeted release in the first quarter of 2016.

Rob W1AEX