Full Duplex Update

[Steve Haynal]

Hi List,

I began tests with full-duplex on the new 1.22 boards today. I have RX running in full-duplex (TX is all zeroes) and things look good. No signs of signal integrity problems. This is with a SDK and the layout of the CVs is different so need to test with the CVs eventually. Hopefully I'll find the time to get the TX side running this week.

73,

Steve

KF7O

 

[Steve Haynal]

Hi List,

I just had my first QSO running full duplex - JT9 with W4UEF on 20M. Everything looks good. I'm keeping an eye on the power dissipation given the extra current used in full duplex. Hopefully this weekend I will run VNA transmission measurements on my qrp labs filters. I need a reflection bridge to see what my antenna looks like. Any recommendations or units for loan? Also, I still need to test with the CV as its different layout may still result in signal integrity problems. So far all tests have been with the SDK.

It is late and I need to go to bed. I will try to respond to other interesting posts this weekend.

73,

Steve

KF7O

[Sid Boyce]

Commercial RLB's seem pricey , I have seen $275 for one of similar design and performance to the ones I refer to here.

Sam Wetterling's RLB seems most popular one used with HPSDR users.
http://www.wetterlin.org/sam/Reflection/3BeadBalunBridge.pdf

http://www.ve2azx.net/technical/RLBridges.pdf is another design

I built my own to use with the HiQSDR - http://hiqsdr.com/index.php?title=Bridge-Circuit
73 ... Sid.

On 19/06/15 07:53, Steve Haynal wrote:

Hi List,

I just had my first QSO running full duplex - JT9 with W4UEF on 20M. Everything looks good. I'm keeping an eye on the power dissipation given the extra current used in full duplex. Hopefully this weekend I will run VNA transmission measurements on my qrp labs filters. I need a reflection bridge to see what my antenna looks like. Any recommendations or units for loan? Also, I still need to test with the CV as its different layout may still result in signal integrity problems. So far all tests have been with the SDK.

It is late and I need to go to bed. I will try to respond to other interesting posts this weekend.

73,

Steve

KF7O

 

On Sunday, June 14, 2015 at 10:52:55 PM UTC-7, Steve Haynal wrote:

Hi List,

I began tests with full-duplex on the new 1.22 boards today. I have RX running in full-duplex (TX is all zeroes) and things look good. No signs of signal integrity problems. This is with a SDK and the layout of the CVs is different so need to test with the CVs eventually. Hopefully I'll find the time to get the TX side running this week.

73,

Steve

KF7O

 

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

-- 
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
Senior Staff Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks

[John Williams]

Here is a link to many of Sam's designs, that have been formated for Oshpark fab instead of ExpressPCB.

http://www.k9ivb.net/VNA/index.htm

and lots of other HPSDR stuff...

[John Williams]

Steve,

I ordered a series fixture and the 3 bead balun boards on Oshpark from k9ivb's page. I will send you a set of boards when they come in. This will take a while so hopefully someone will loan you some bridges to allow your testing to complete.

BTW, I have your ADC done, but am testing another board with the higher power F/R sensing circuit. Am short a few parts. When I get that done will send them to you also.

Anyone else - the ADC is shared on Oshpark if you want some. The F/R sense board is also shared but untested.

John - W9JSW

[John Williams]

Steve,

Oshpark reports that boards will be back by July 1st.

John

[rodwall1]

Hi Steve and all,

You will also need a set of SOL calibration standards for proper VNA use. I built my own standards similar to the ones described in the following link.

http://www.qsl.net/in3otd/electronics/VNA_calkit/calkit.html

When making the 50ohm load standard. I used two 100ohm in parallel on a BNC connector. This gives it a bit more strength as the pins on BNC connectors tend to move. I have had a 50ohm SMT resistor crack open when the pin moved. It's a good idea to also purchase good connectors. In fact using BNC connectors is not a good idea.

When testing DUT on PC boards. I use thin 50ohm coax with short pigtails. With a suitable connector at the bridge end. And short (solder) or open the pigtails for the short and open standards. And use a 50ohm 1% resistor (solder) with zero length leads for the load standard. Keeping the pigtails as short as possible. Keeping the pigtail length only long enough that is required to connect to the DUT on the PC board. This then puts the Reference plane at the DUT and calibrates out the effect of the coax cable. I find this useful at HF frequencies.

For through calibration I use another thin 50ohm with short pigtails.  With the other end connected to the detector (receiver). And solder the pigtails ends together during calibration.

Regards,

Roderick, vk3yc.

[Steve Haynal]

Hi John,

Does this mean you are sending me a bridge and a ADC board? Also, with the ADC board, is it possible to run the other four IO signals to some sort of connector for a CW key? I think we will need to get Hermes CW working to help round out the QEX article.

I hope to make a transmission VNA measurement this weekend and will release the code if that is okay.

73,

Steve

KF7O

[Steve Haynal]

Hi Rod,

With my pysdrvna, I made my own OSL standards seen here. I have my software setup so that the load can be other than 50 Ohm, although near 50 Ohm. I measure my load precisely with an accurate ohmmeter and then just enter that value in a config file. Is that supported in myVNA?

What are your concerns with BNC? I can see issues with very high frequencies, but for the frequencies the Hermes-Lite will support I would hope that BNC is fine.

Although pysdrvna only supports reflection measurements, one thing I like about it is that it works without a bridge for basic antenna measurements. The accuracy is still pretty good. It would be nice if Hermes-Lite version 2.0 can measure an antenna without a bridge even if accuracy is diminished as a convenience for users. This would not exclude more conventional VNA use.

Also, to continue the discussion on software, I really like the pysdrvna software as it is basically a Python programming shell. Python is my language of choice, and one has total flexibility in an environment like that where you can use packages like numpy and scipy. I realize I'm in the minority here and am glad to see potential myVNA support.

I will try to make some measurements this weekend and report on this list.

Thanks for all your help.

73,

Steve

KF7O

[rodwall1]

Hi Steve,

OK on pysdrvna. That is good work that you have done there. Yes you can set the load along with other parameters.

BNC connectors.
I guess I am looking at it from a VNA instrument view where high quality connectors are used. The problem with BNC connectors is that the centre pin can move and the load surface mounted resistors cracks open. I used BNC connectors on my N2PK VNA and wished I had used N connectors. I have had the load calibration standard resistor crack a number of times. What I ended up doing is to use two 100ohm or four 200ohm in parallel. But using BNC connectors would be OK if the Load standard is not used often.

Pysdrvna is a good idea and makes a simple reflection mode instrument. Even though a bridge is not required. It requites extra circuitry that can easily be added. Good idea.

myVNA status. Dave has indicated that he will build a interface to the old standard and change it when the new protocol standard has been tested and released.

Can you tell me where can we get the Hermes specifications for the old protocol standard. Dave needs this to build the interface for myVNA.

Regards,

Roderick, vk3yc.

[Steve Haynal]

Hi John,

Pins 9,11 and 12 have pads on the Hermes-Lite board to connect to power/pull-ups for a possible Alex interface. If you run those also to the edge or through-hole connectors on your board that would be great.

I've been using the FB for noise suppression. It would be good to have noise suppression on these signals. Experimenting with what Hermes uses would be interesting too.

I am debating whether CW connections should really be on the unit. I like the idea of a black box that can be put far away from the operator and used remotely. CW connections breaks this paradigm. I will try with fldigi sometime soon, although I am sure that will not satisfy a CW purist.

73,

Steve

KF7O

On Fri, Jun 26, 2015 at 5:07 AM, John Williams <jswi...@gmail.com> wrote:

Steve,

I see a ground (p1), and 4 signal pins from CN3 - pins 3, 5, 7 and 10. They attach to the MEC6 at P39, p41, P45 and P3 respectively. I see no power on CN3. Where do I pick up power? Looks like pins 9, 11 and 12 do not connect to anything? I can pick up 3.3V and 5V elsewhere on my board. Which voltage do you want at the connector?

Do you want any of these signals to go thru noise suppression filters. I like to use these for all off-board signal connections. Hermes uses them quite a bit...

http://www.digikey.com/product-detail/en/0/490-6981-1-ND

John

On 6/23/2015 9:42 PM, Steve Haynal wrote:

If you run all the rest the CN3 pins that are not used by the SPI, including the power, to through-hole pads on your board that should be sufficient to test CW.

Thanks,

Steve

KF7O

On Sun, Jun 21, 2015 at 7:49 AM, John Williams <jswi...@gmail.com> wrote:

Let me know what signals you want to use on CN3.

On 6/20/2015 4:27 PM, John Williams wrote:

Steve,

4 boards. ADC, FWD/REV sense, 3 bead balun and series board. Not sure if the series board is useful but what the heck, it was only a few bucks to order. The ADC plugs into the Hermes lite across all of the 12 pins so you can bring out any signal you like on a cable. I can add a revision if you like later to make it official...

John

[Zl2APV]

Hi Steve,

Regarding the CW interface, it really has a lot of pros and cons. To have CW (and PTT) inputs on the hardware is a big advantage when experimenting on the bench as getting a signal out is pretty much SDR independent and of course the minimum of latency applies. It is possible to work full break in this way and gives the hottest performance. On the other hand most of us don't work full break in and can't send fast enough for latency to be a big problem. The other situation is to have the radio running from a laptop etc with a key plugged into the com or usb port and operate it over the local network anywhere on your site or more ambitiously as done by ghpsdr3-alex anywhere in the world. Latency rears its ugly head and a lot of the software options have not addressed CW very well. Big difficulties exist in going cross platform. It seems every time I suggest something would be good I find something wrong with it. Am I confusing you yet Steve - no! oh well I'll keep trying.

The ultimate SDR station would have the hardware in a hut/box at the base of the antenna system with a gigabit ethernet back to the shack's switch. No long antenna cables into the house, just a power cable to a float charger system with a battery supply all setup for emergency operation. This of course suggests that the key and PTT switch will connect to the computer but to get best latency I would suggest that the tone be generated by the fpga complete with its exponentially rising and falling leading edges which is turned on and off from a binary signal. Now if you are going to do this you might as well or this with an input pin and provide a hardware key input on the HL board but wait, you are already crystal ball gazing and looking at the potential of the fpga surrendering a lot of its functionality and running it on a server with a gigabit ethernet connection and sending processed IQ back to the radio software via localhost or medium speed lan. A lot to sort out here.

I would like to see all tone generation etc. stay on the fpga
Signal manipulation with a dsp server
The sdr radio for display and function selection.
The keying done from the sdr radio via serial/parallel/usb port and is an on/off signal to fpga.

Just a couple of points to round off. The HL is probably going to spend much of its life on the work bench as with its low cost you'll be prepared to mess around with it whereas it is too financially risky to play with a high end radio like Anan, Flex or HiQSDR. I was going to say the HL probably wont be the primary radio but now I am having doubts about that as recent performance measurements show it is not to be sniffed at.

73 Graeme ZL2APV

[John Williams]

Steve,

See attached schematic and jpg of the board. If this is to your liking, you can order some boards... I don't plan to build any since I have enough prior version boards and no interest in CW...

John

[Steve Haynal]

Thanks John! 

73,

Steve

KF7O

On Sunday, June 28, 2015 at 6:19:35 AM UTC-7, John Williams wrote:

Steve,

[John Williams]

I will work on a bom while you await the boards.

[Steve Haynal]

Thanks for the input Graeme!

[John Williams]

Steve,

Here is the BOM. Note that I made a footprint error on D3. See the datasheet on Digikey to orient properly. You can omit the 4 components around this Zener on on the schematic if you do not need the 13.8v sensing. John Laur wanted me to include that feature on the design. You may want to share a board with him for testing with PowerSDR.

John

[John Laur]

I may not have time to build one. I am trying to focus on software with the limited time I have available. But I still do think that the input voltage sensing is a very good feature to have; thank you for including it. A lot of the problems related to the commercial ANAN radios at 100W have been due to supply issues or issues with poorly assembled power cords. The ability to have a measurement at the radio input is very nice even if the PA is external it can help when using a common supply.

73, John K5IT