Trying Radioberry preamp measurements

[Yado-san]

Hi Group,

I’v made trying the Radiobelly preamp measurements. The report is attached.

It took me a long time because I did it by trial and error as shown below. Therefore, there may be some mistakes.

But I'm very happy if this report can help anyone who consider the Radioberry preamp.

Note:

p.1 [fig.2]: Radioberry2 TRX overall configuration

　 Output 5W. used TS680V junk LPF board. 

p.1 Note, p.3[Fig5]:  RF-OUT circuit modification

   RF-OUT AD9866-IAMP has Spurs.(2Fs-5Ftx), so AD9866-TxDAC used.

   RF-Amp was inserted. Because TxDAC output was less than 1/10 of IAMP.

p.2 Noe*2:  N2ADR LPF relay control

   In piHDSDR, I set OC, but LPF relay control did not work.  Quisk is ok standalone.

   Checked with other SDR software. Win10PC.

 

p.4 [Fig.6,7]:  bias setting.

   It was possible to set the bias in SpakSDR and Quisk.

p.6 [Fig.14-16] Two-tone IMD

   I used SDRConsole because PowerSDR's Two-tone was unstable.

   PowerSDR sometimes increased the RX noise level suddenly. I had to re-power.

p.7,8: Output Spectrum

   At the TxDAC output, a spurious of AD9866 source osc of 38.4MHz-ftx was observed.

   The value is below -60dB, so it meets the regulation, but may need to adjust the load resistance of the TxDAC.

73

Yado-san

jg1twp

[pa3gsb]

Hi Yado-san

Thanks for sharing your experiments.... 

Some remarks:

-   Another way to solve this problem is to insert a BPF instead of an RF-Amp. and don't need to modify RB2 to pull out TxDAC pins.  What are the specs of this filter? Are the measuring results about the same ?

- If you are willing i can add  the    MCP23018 control  in the filters.h  ;   we need a seperate i2c address.

- Using puresignal in pihpsdr using 48K or 96K should be possible?; control of N2ADR ; i did recently did a bug fix.... hope this solves your problem.

I like your exeperiments and enjoyed reading your document.

73 Johan

PA3GSB

Op zondag 4 juli 2021 om 07:49:20 UTC+2 schreef Yado-san:

[Yado-san]

Hi  Johan

Thanks for your message to my experiments report.

The spurious results before the circuit modification are attached.

I have also included information on the BPF used. 

MCP23018 was used because I had one that I had got before.

But I would be happy if MCP23018 is also included 'filters.h'.

And I hadn't tried puresignal yet. I will try it.

Tnx

Yado-san

jg1twp

2021年7月4日日曜日 20:44:43 UTC+9 pa3gsb:

[Yado-san]

Hi Johan, All

Using the latest firmware, I confirmed that there is no problem in controlling of the N2ADR with the piHPSDR.  :)  

Also tested the PureSignal function that seems to be working fine. (sample rate 48kHz)

-> Attached a results.

ps. 

About the PureSignal, I referred to 

  https://groups.google.com/g/radioberry/c/oxOmofqrsbI  : "Rpi4 and PS/piHPSDR"  ja1cca Tokio-OM

  https://github.com/softerhardware/Hermes-Lite2/tree/master/hardware/companions/psfeedback

  https://github.com/TAPR/OpenHPSDR-PowerSDR/blob/master/Documentation/Pure%20Signal/PureSignal.pdf

Tnx.

Yado-san

jg1twp

2021年7月4日日曜日 23:35:18 UTC+9 Yado-san:

[pa3gsb]

Yado-san

Thanks for sharing your results.

It does not seem to work fine... i actually see it working, also based on your external SA  measurements.

Very fine to see it working all together.

Have fun

73 Johan

PA3GSB

Op woensdag 7 juli 2021 om 16:40:42 UTC+2 schreef Yado-san:

[Yado-san]

Hi Johan,

Thanks for your comment.

I'm sorry, I thought it worked and posted it.

What is the result if it works correctly?

Tnx

Yado-san

jg1twp

2021年7月8日木曜日 0:57:39 UTC+9 pa3gsb:

[pa3gsb]

Yado-san

Iam sorry ; i did make a joke....

... it works very fine in mine opinon.  You wrote in your post it seems.... and i tried to say:  No It is actualy  working.

English is for both of us not ournative language; which can lead to such misunderstanding.... i will try to keep it serious, hi.

Have fun

73 Johan

PA3GSB

Op donderdag 8 juli 2021 om 05:20:27 UTC+2 schreef Yado-san:

[Yado-san]

Hi  Johan,

Okay, I understand!

VY SRI, "DeepL translator" that I normaly using was not tell me that are Jokes. Hi.

The actual reason is that I didn't really understand the PureSignal function.

I have to learn and fun more...

Tnx!

Yado-san

jg1twp

2021年7月8日木曜日 14:57:18 UTC+9 pa3gsb:

[Yado-san]

Hi Johan, All,

I experimented with measuring power with a Radioberry preamp board.  Attached file.

I couldn't understand the order in which they are sent to the packets.

Is there a specification for this somewhere? (protcol?)

 for example:

  hpsdrdata[4,5,6,7]    : last_sequence_number 

  hpsdrdata[11,12,13,14,15]    : measured values

 

The Radioberry preamp board is a smart and nice QRP amplifire for a SDR. I like it very much.

Thanks you Johan PA3GSB, Radioberry Group and Hermes-Lite Group, all the amazing work !

73

Yado-san

jg1twp

2021年7月8日木曜日 20:48:36 UTC+9 Yado-san:

[pa3gsb]

Hi Yado-san,

Nice work.

Looking in the attached document, you have done the right steps following the protocol: 

https://github.com/softerhardware/Hermes-Lite2/wiki/Protocol

https://github.com/g0orx/pihpsdr/blob/3bc649869b24967b4785dcb489d402449c8a44c5/transmitter.c

1.61 * 1.61 / 1.4 = 1.8 W

The 1.4 is the constant value for the N2ADR bridge for the HL-2.

I did not understand your problem; the steps you have done are correct.

The bridge you are using is having maybe a different bridge constant?

Some time ago i have also created a vswr:  https://github.com/pa3gsb/vswr/tree/master/software/vswr-monitor

This bridge had constant value of 0.09

Hope this helps in figure out your problem (which i not directly understand)

73 Johan

PA3GSB

Op zondag 18 juli 2021 om 16:34:34 UTC+2 schreef Yado-san:

[Yado-san]

Hi Johan,

Thanks infomation.

There is a correction value for the each bridge, I see.

" case DEVICE_HERMES_LITE2:

            constant1=3.3;

            constant2=1.4;

            fwd_cal_offset=6; "

I haven't used the SWR bridge yet, but the FT680-LPF board has the same bridge as the N2ADR board. (CM coupler using a toroidal core)

And I did not understand how the SWR value was calculated when the voltage was simply applied Vfwd and Vref.

so I will look into piHPSDR source code.

Tnx,

Yado-san

jg1twp

2021年7月19日月曜日 1:42:31 UTC+9 pa3gsb:

[Yado-san]

Hi Johan, All,

I understood how to calculate the SWR of piHPSDR at last.

Problem is 

 The piHPSDR's SWR value is smaller than the calculated(actual) value. when the voltage was applied AIN2(Vfwd) and AIN1(Vref).

  -> The parameter of the meter_update() function is RF power, but the SWR was calculated in the voltage. 

Therefore, need to modify 'meter.c' as follows.

$ nano /tmp/pihpsdr/meter.c

    446       double swr;

    447       if (max_level > reverse) {

    448       //swr=(max_level+reverse)/(max_level-reverse);                                  // SWR calculated by fwd,rev-Voltage.

    449       //fprintf(stderr,"SWR=: %f fwd= %f rev= %f \n", swr, max_level, reverse);  // for checking values.

    450         swr=(sqrt(max_level)+sqrt(reverse))/(sqrt(max_level)-sqrt(reverse));     // SWR calculated by fwd,rev-Power. (OK)

    451       } else

    452         swr=999.9;

    453       }

73

Yado-san

jg1twp

2021年7月19日月曜日 18:02:13 UTC+9 Yado-san:

[pa3gsb]

Yado-san

Seems you discover a bug in pihpsdr.

2 formula's for vswr :

vswr = Vfwd + Vref / Vfwd - Vref 

vswr = 1 + sqrt (Pref / Pfwd) / 1 - sqrt (Pref/Pfwd)

Value of 1.36 for 5W Pfwd and 0.77 Pref is indeed wrong.

Maybe a good idea to report this as an issue at https://github.com/g0orx/pihpsdr 

73 Johan

PA3GSB

Op donderdag 22 juli 2021 om 15:10:37 UTC+2 schreef Yado-san:

[Yado-san]

Hi  Johan,

I understand!

I have registered on github and posted the Issue on g0orx/pihpsdr.

Tnx

Yado-san

jg1twp

2021年7月23日金曜日 2:06:23 UTC+9 pa3gsb:

[softerh...@gmail.com]

Hi Yado-san,

When you switched to the TxDAC output, did you also update the gateware to use the TxDAC? I think there will be signal on the TxDAC even when the IAMP is used, but it may not be clean. There are AD9866 settings which turn of the IAMP (saves power) and exclusively uses the TxDAC output for the cleanest signal. The HL2 makes these settings but Johan's variant changes them to enable the IAMP.

To address the spur you also see on the IAMP output, the HL2 uses the TxDAC low power output plus OPA2677 preamp.

One direction I wanted to explore more with the HL2 was to still use the IAMP output but decrease the power and add a capacitor trap across the TxDAC output to reduce/eliminate the spur. I am not sure that you still get enough power out of the IAMP to make this worth it. You can find some details here:

https://groups.google.com/g/hermes-lite/c/dmHNXZ9JqRY/m/EHD760SoBAAJ

and search https://groups.google.com/u/1/g/hermes-lite

for IAMP and/or TxDAC for much more discussion.

73,

Steve

kf7o

[Yado-san]

Hi Steve,

Thank you for the information. I will read the discussion.

I did not make any changes to turn off the AD9866_IAMP.

 It's better to set IAMP to Disable, I see.

I think the following changes are necessary, compared to the HL2 Verilog. 

 I will try to compile the FPGA.

- ad9866ctrl.v

  initarray[14] = {1'b1,8'h81};  //  <-  {1'b0,8'h01};   Address 0x0e,  Disable/Enable IAMP

  initarray[16] = {1'b1,8'h80};  //  <-  {1'b1,8'h84};   Address 0x10,  Select TX gainAddress 0x10, Select TX gain

  IAMP-Disable  https://github.com/softerhardware/Hermes-Lite2/tree/master/gateware/rtl/ad9866ctrl.v

  IAMP-Enable   https://github.com/softerhardware/Hermes-Lite2/tree/master/gateware/rtl/radioberry/ad9866ctrl.v

Tnx!

Yado-san

jg1twp

2021年7月26日月曜日 14:56:58 UTC+9 softerh...@gmail.com:

[pa3gsb]

Hi Yado-san, 

Good experiment.

Yado-san if you need help do not hesitate to ask.

Have fun,

73 Johan

PA3GSB

Op maandag 26 juli 2021 om 16:22:23 UTC+2 schreef Yado-san:

[pa3gsb]

Hi Steve,

Steve G6ALU did the following :

Loading the txDAC output with 100R across the TxDac output made a significant difference to the spur. 

I initially tried a 33R between each output and ground but this reduced output to ~ -9dBm.

Original design 15dBm output - 24.990MHz TX, spur at ~ 28.66MHz 55dB below carrier

100R between TxDac pins gave 13dBm output - 24.990MHz TX, spur at ~ 28.66MHz 64dB below carrier

https://groups.google.com/g/radioberry/c/rUA8bocApNg

73 Johan

PA3GSB

Op maandag 26 juli 2021 om 07:56:58 UTC+2 schreef softerh...@gmail.com:

[softerh...@gmail.com]

Hi Johan,

I am experiencing a bit of dejavu here. We looked in depth at these spurs with the Hermes-Lite. The best results were with a trap across TxDAC. In the post I linked to, there is a link to Claudio's blog. This is the most thorough analysis of the spur and ways to reduce it that I've seen:

http://www.qsl.net/in3otd/ham_radio/Hermes-Lite/TX_out.html#specouttrap

Claudio is the designer of the PA used in the Hermes-Lite 2 and now on the radioberry preamp board.

My decision with the Hermes-Lite 2 was to add the additional OPA2677 stage (designed by Jim N2ADR) and use the cleaner TxDAC output. I wanted 5W. The gain of the AFT05MS003N (and most similar devices) is not enough to realize a clean 5W with the reduced output of the IAMP (15dBm or less) when any spur reduction technique is used.

If radioberry users do go with the IAMP output and spur reduction techniques, I'd suggest lowering Vpa. You won't realize 5W output and the extra voltage is just burned by the LDMOS devices and converted to heat. We originally used a Vpa of 10V for the HL2, but lowered that to 8V and saw a reduction in heat. Too low of a voltage may lead to distortion.

 73,

Steve

kf7o

[Yado-san]

 Hi Johan, Steve, Group,

 Sorry for the delay in reply.

1. I have examined the RF-OUTPUT(TxDAC) change with IAMP Enable/Disable which was pointed out.   Attached (1,2,3)

   Results:  There was no change in spurious. (What a relief..) 

2. I also took a simple measurement of power consumption when paired with Rpi4B.  Attached (4)

   Results:  About 5W. TX/RX with PiHPSDR(SPI) increased by about 1W. ( It can run on 18650 batteries.)  

3. As for the FPGA compilation environment, I found Makefile in the release directory .  Attached (5)

   Are you compiling on the command line in Window10 (WSL)?

   

   I compiled it with Qurtus prime Lite18.1 for Windows10 (GUI), just like Radiobery2_beta3.

Note:

 I think RB2 is an SDR that can be create a small Raidio at a low cost to suit everyone's preference.

 I posted this topic as an experimental example of using RB2 as a TRX for QRP. 

 I was hoping to get some examples or specific measurement data.

 So I am NOT trying to tell you that you should use TxDAC for RF-OUT. Becase It's so difficult to solder at the AD9866-QFN pins. hi!

ps.

 Now trying a beta version of the Radioberry-Juice board.

 It can be realized Multi-band RX with 192kHz or 384kHz sampling. Just connect to the USB-IF that is available on your PC.

 Please try it when it is released.

73

Yado-san

jg1twp

2021年7月27日火曜日 14:47:36 UTC+9 softerh...@gmail.com: