OpenHPSDR PI4 viable?

227 views
Skip to first unread message

pete...@gmail.com

unread,
Nov 25, 2020, 3:29:10 PM11/25/20
to Radioberry
I did not work on my Radioberry for some time. I was needing time for many more project likes a UHF repeater, a 900 mhz quantar repeater , duplexer tuning and much more. 

So I lost track of the advancement of the project.

I have seen a new release point with a new PCB (beta4) 
From what I was able to read. It is the same pcb with the jumper for the second clock on the fpga. 

But I dont have much info on the possibility to run the RB with OHPSdr on a IP link from the RPI to the computer running the software.

Does it work well? Is there some bugs to fix? Do we need "special" programming? Will I lose some fonction compared to PIhpSDR? 

THnaks for the feed back.

Pierre
VE2PF 

Radioberry

unread,
Nov 25, 2020, 4:24:38 PM11/25/20
to Radioberry

Pierre

A lot is changed after focussing on RPI-4.

  • The gateware is now a variant in the HL-2 
  • Linux version 5
  • Radioberry uses device driver (including loading the gateware)
  • Firmware, which implements openHPSDR protocol is now started as a service 
  • Installation by shell scripting
  • local CW using pihpsdr
  • soapy support
  • openwebrx support
  • https://github.com/pa3gsb/Radioberry-2.x-extensions 
  • received also a lot of direct messages of builders
  • one solution for local and remote operation using SDR app at other PC 
  • working on 8 rx 4K channels for decoding like ft8 and
  • connection to gnu radio using 0MQ in the experiments with 4K channels
  • gnu radio release 3.8 i like the soapy connector


Most information is present at the wiki ; https://github.com/pa3gsb/Radioberry-2.x/wiki

Iam also thinking in developing a pre-amp board 2.5W or 5W out to control a PA. The PA iam using is 100 W and contains LPF filters. I think that makes a nice little transceiver.

Hope this helps and gives you an idea about the development. 

73 Johan
PA3GSB


Op woensdag 25 november 2020 om 21:29:10 UTC+1 schreef pete...@gmail.com:

Pierre Martel

unread,
Nov 25, 2020, 5:01:54 PM11/25/20
to Radioberry
Wow! lots of changes! 

I will have some time to work on the project with the snow already starting to drop and the virus keeping us even more inside. (been working from home since mid mars and I feel like a bear inside a cage from time to time So I need new stuff to do. 

That and a multiband, multi digital mode transceiver and the corresponding repeater but only for UHF, will make the long winter evening less difficult to pass.. 

A BIG thank you Johan for the hard work! 

Pierre
VE2PF

--
You received this message because you are subscribed to the Google Groups "Radioberry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to radioberry+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/radioberry/c31e1b4a-e997-44c1-8b3a-e91ae7126816n%40googlegroups.com.

Tokio

unread,
Nov 25, 2020, 10:45:42 PM11/25/20
to Radioberry
Hi Johan,

While waiting my Rpi-4B, I'm getting info about how to set up Radioberry with Rpi-4B.

The first thing to do is to set up Rpi-4B.
You mentioned "Linux version 5" in a message to Pierre.
However on the wiki, Raspberry pi Operations System (Raspibian) should be installed.
What do you mean by "Linux version5"? Is there a need to modify Raspibian on Rpi-4B?

73,
Tokio
ja1cca


2020年11月26日木曜日 6:24:38 UTC+9 Radioberry:

pa3gsb

unread,
Nov 26, 2020, 2:53:40 AM11/26/20
to Radioberry
Hi Tokio

With linux 5  iam pointing to raspberry pi OS. No specials are required ; just the plain raspberry pi image.

The wiki is uptodate here.

73 Johan
PA3GSB

Op donderdag 26 november 2020 om 04:45:42 UTC+1 schreef Tokio:

Tokio

unread,
Nov 26, 2020, 4:32:21 AM11/26/20
to Radioberry
Hi Johan,

Thank you for the clarification.
I follow the wiki:

73,
Tokio ja1cca


2020年11月26日木曜日 16:53:40 UTC+9 pa3gsb:

Tokio

unread,
Nov 29, 2020, 4:49:12 PM11/29/20
to Radioberry
Hi Johan,

Rpi4 arrived here. The radioberry with Rpi4 is good for RX. However, it doesn't work for TX due to many spurs on TX.

As to Radioberry gateware/driver/firmware, I installed them subsequently including piHPSDR following the URL

The result is as follows (please look at the attached picture):
(1) Many TX spurs. But on the screen of PiHPSDR, the transmission seems to be normal.
(2) Sometime with this firmware, the ADC output is opposite, that means LSB becomes USB. e.g., on 80-meter band. This is observed only by receiving.
(3) Receiving is good especially with SDR Console V3. This client software is very beautiful.

The gateware 72.3 would be the for HL2. I'm obliged to think that there is an anomaly in driver/firmware.
FYI, the above result is the same for a couple of friends of mine using Radioberry with Rpi4.

73,
Tokio
ja1cca
PNG78.png
IMG_0080.jpeg

Radioberry

unread,
Nov 30, 2020, 6:28:50 AM11/30/20
to Radioberry


Hi Tokio,

i have tesed the latest release.

Hereby the spectrum:


spectrum_72.3.jpg



Please check your configuration:


See check section.


You are running via WIFI?   can you repeat the measurement using the local host (127.0.0.1)  connection?


Also the change of LSB to USB or otherwise are not happening.

Seems a local problem.

Johan
PA3GSB


Op zondag 29 november 2020 om 22:49:12 UTC+1 schreef Tokio:

Tokio

unread,
Nov 30, 2020, 8:17:18 AM11/30/20
to Radioberry
Hi Johan,

Thank you for your spectrum analyzer screenshot.
This makes me get reassured.

I did the test again. Still no joy.
The sideband change occurred (please look at the SDR Console screenshot). With this, the noise level became very high.
I did a power cycle of Rpi4. Now this time , this sideband change disappeared. The noise level returned to the normal.

The result of "sudo modinfo radioberry" is different from the github you mentioned.

pi@raspberrypi:~ $ sudo modinfo radioberry
filename:       /lib/modules/5.4.72-v7l+/kernel/drivers/sdr/radioberry.ko
version:        0.6
license:        GPL
description:    Radioberry SDR device driver. (rpi-4)
author:         Johan Maas - pa3...@gmail.com
srcversion:     89D65EA9BD663FEC435EEE3
alias:          of:N*T*Csdr,radioberryC*
alias:          of:N*T*Csdr,radioberry
depends:
name:           radioberry
vermagic:       5.4.72-v7l+ SMP mod_unload modversions ARMv7 p2v8

===== Github Info Quote ======

$ sudo modinfo radioberry

filename: /lib/modules/5.4.51-v7l+/kernel/drivers/sdr/radioberry.ko

version: 0.7

license: GPL

description: Radioberry SDR device driver. (rpi-4)

author: Johan Maas - pa3...@gmail.com

srcversion: 753F1653130B45CCCC5FE2A

alias: of:N*T*Csdr,radioberryC*

alias: of:N*T*Csdr,radioberry

depends:

name: radioberry

vermagic: 5.4.51-v7l+ SMP mod_unload modversions ARMv7 p2v8

======= Github info Unquote ====

No WiFi used for the test.
I still have many spurs when TX, and piHPSDR screen also shows high-level spurs.

Is there any advice to clear this trouble?

73,
Tokio
ja1cca


2020年11月30日月曜日 20:28:50 UTC+9 Radioberry:
sudo radioberry.PNG
radioberry.PNG
IMG_0088.jpeg
USB.PNG

Tokio

unread,
Nov 30, 2020, 8:32:47 AM11/30/20
to Radioberry
Johan,

I forgot to mention that Tune button of piHPSDR doesn't disable RX. I can hear the receiving signal. I suppose this receiving signals send as TX signals.

Tokio

2020年11月30日月曜日 22:17:18 UTC+9 Tokio:

Radioberry

unread,
Nov 30, 2020, 9:21:07 AM11/30/20
to Radioberry

Tokio

pihpsdr has a duplex checkbox somewhere. switch it of to make it simplex.

Please update all the components

cd /tmp
wget https://github.com/pa3gsb/Radioberry-2.x/releases/download/2020.11.14/radioberry_install.sh
sudo chmod +x radioberry_install.sh
./radioberry_install.sh 

your driver and firmware are not the latest software components

The timing in your logging seems fine. 

The path for TX is from RPI to Radiobery using 5 pins... maybe there is something wrong?

It is possible to test the radioberry in transmit mode without using this signal path from rpi - rb.


Please use the key input pin 14 and look to your signal.... this signal is generated inside the FPGA and should produce a good signal.

Hope this helps to make clear where the problem is.

I assume some soldering problem.

73 Johan
PA3GSB
Op maandag 30 november 2020 om 14:32:47 UTC+1 schreef Tokio:

Tokio

unread,
Nov 30, 2020, 11:04:54 PM11/30/20
to Radioberry
Hi Johan,

I really appreciate your advice, and would like to use Radioberry with Rpi-4 also for TX.

I would like to have further your advice.

(1) As to soldering of my Radioberry, this board works fine for both RX and TX with Rpi-3. So, this is not the problem.

(2) I updated Firmware. Now I get the result of "sudo modinfo radioberry," exactly the same as your Github.

(3) Your advice below is the next thing to do.
=== Quote ===
It is possible to test the radioberry in transmit mode without using this signal path from rpi - rb.
Please use the key input pin 14 and look to your signal.... this signal is generated inside the FPGA and should produce a good signal.
=== Unquote ===

I don't understand how to generate a signal from FPGA by " from rpi - rb". Pin 14  of CN-301 "key input".
Could you elaborate this procedure?

73,
Tokio
ja1cca

2020年11月30日月曜日 23:21:07 UTC+9 Radioberry:

Radioberry

unread,
Dec 1, 2020, 5:57:01 AM12/1/20
to Radioberry
Tokio,


Your jump to conclusion about the soldering is not right. The  rpi-3 tx is done by SPI and the rpi-4 uses as said 5 pins.

And also the duplex mode is different!

But anyway:

Use the pin 14 of the RB connector and use a wire to ground (keying);  the gateware will go into tx mode and wil generate a signal.... 

Hope this helps... to figure out what is wrong!

Johan
PA3GSB

Op dinsdag 1 december 2020 om 05:04:54 UTC+1 schreef Tokio:

Tokio

unread,
Dec 1, 2020, 6:38:29 AM12/1/20
to Radioberry
Johan,

Thank you very much for your comments.

I didn't know that the connection between FPGA and Raspberry pi is different from Rpi-3 to Rpi-4. Now it is understood.

This afternoon, I checked the test you recommended pin 14 of CN301 is grounded. I confirmed the clean signal on my Spectrum Analyzer. Please look at the attached picture. When pin 14 is grounded, piHPSDR is running. The frequency on piHPSDR is 7,100 KHz/LSB, and the output carrier is measured 7,120 KHz, 20 KHz up, whose level is around 17 dB. (PA setting is set 38.8 in piHPSDR). 

(1) How do we understand this 20 KHz difference?

From the above result, I think that my Radioberry FPGA and ADC are working properly. Then the problem resides between FPGA and Rpi-4.

(2)To locate the problem, how do we proceed as the next step?

Your help would be much needed.

73,
Tokio
ja1cca

2020年12月1日火曜日 19:57:01 UTC+9 Radioberry:
PNG84.png

Radioberry

unread,
Dec 1, 2020, 1:54:03 PM12/1/20
to Radioberry

Tokio

Not too much time today....

Nice your output is there with the FPGA generated signal!  


First check your soldering:



do you have gpio modules you are loading  in the  /boot/config.txt

like:    gpio=4-13,16-27=ip,pu


As a first quick change : I have doubled the FIFO tx buffer size.  The radioberry gateware for the CL016 devices is added.

Copy this to the /lib/firmware folder.

sudo sytemctl stop radioberry (stop firmware)
sudo modprobe -r radioberry  (to unload the driver) 
sudo modprobe radioberry (to load the driver)
sudo systemctl start radioberry (start firmware)

start pihpsdr.

You are lucky to learn a lot! When your radio is working the fun is over....

Have fun

Johan
PA3GSB


Op dinsdag 1 december 2020 om 12:38:29 UTC+1 schreef Tokio:
radioberry.rbf

Tokio

unread,
Dec 1, 2020, 7:48:23 PM12/1/20
to Radioberry
Johan,

Thank you very much for your action.

(1) I checked soldering of from txclk to txdata[3]. The conductivity test by tester is OK.

(2) Adding "gpio gpio=4-13,16-27=ip,pu" in /boot/config.txt causes the high-level noise on RX.
     To which pins should the gpio set to input and pull-up be set?

(3) My FPGA is CL025.
Could you prepare rbf for CL025?

I appreciate your help.

73,
Tokio
ja1cca



2020年12月2日水曜日 3:54:03 UTC+9 Radioberry:

pa3gsb

unread,
Dec 2, 2020, 9:34:10 AM12/2/20
to Radioberry
Tokio

Some misunderstanding about the /boot/config.txt; the intention was only to ask if there where some gpio module settings in your file? Not to add the line!

But do i understand it right that the setting is actually  influencing the RX ?

The RPI is a general purpose system... so different settings can influence the working of the radio... so i am searching for these interfering settings.


OK ... CL025 ; attached the larger FIFO for tx. 

Copy this to the /lib/firmware folder.

sudo sytemctl stop radioberry (stop firmware)
sudo modprobe -r radioberry  (to unload the driver) 
sudo modprobe radioberry (to load the driver)
sudo systemctl start radioberry (start firmware)

start pihpsdr.

This is a quick try....... 

I am not able to reproduce your problem.... so it is a little hard to figure out what is going wrong ; i will think about some logging to get an idea where things are missing.

Please check this and running localhost...

73 Johan
PA3GSB



Op woensdag 2 december 2020 om 01:48:23 UTC+1 schreef Tokio:
radioberry.rbf

pa3gsb

unread,
Dec 2, 2020, 1:52:55 PM12/2/20
to Radioberry
Tokio

Please tune in CW mode.

How does this look for you?

73 Johan

Op woensdag 2 december 2020 om 15:34:10 UTC+1 schreef pa3gsb:

Tokio Endo

unread,
Dec 3, 2020, 5:30:52 AM12/3/20
to pa3gsb, Radioberry
Johan,

Thank you for your advice.

When CW/AM mode selected, the TX output is very clean as CN301-pin14 grounded. 

However, for LSB/USB, TX output has many spurs. That is the same as Two-tone test.

Could you advise me how I can use LSB/USB with Rpi-4?

As to the frequency, it’s good when RBW of spectrum analyzer is adequate.

73,
Tokio
ja1cca



2020/12/03 3:52、pa3gsb <pa3...@gmail.com>のメール:

-- 
You received this message because you are subscribed to the Google Groups "Radioberry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to radioberry+...@googlegroups.com.

pa3gsb

unread,
Dec 3, 2020, 5:49:28 AM12/3/20
to Radioberry
Tokio

Can you tune in SSB mode in for instance Powersdr?

Is it an pihpsdr problem or a radioberry problem.


What happens when using the microphone in lsb or usb mode?

73 Johan
PA3GSB

Op donderdag 3 december 2020 om 11:30:52 UTC+1 schreef Tokio:

Tokio

unread,
Dec 3, 2020, 6:50:24 AM12/3/20
to Radioberry
Johan,

With PowerSDRmRX 3.4.9, I have the same spurs on TX output in LSB/USB, and also CWL !
With piHPSDR and mic connected, it is the same. Many spurs on Tx output.

73,
Tokio


2020年12月3日木曜日 19:49:28 UTC+9 pa3gsb:

pa3gsb

unread,
Dec 4, 2020, 2:21:23 AM12/4/20
to Radioberry
Tokio

I think i have found the problem.... takes some time to improve.... 

Come back to you later!

73 Johan
PA3GSB

Op donderdag 3 december 2020 om 12:50:24 UTC+1 schreef Tokio:

Tokio Endo

unread,
Dec 4, 2020, 2:54:51 AM12/4/20
to pa3gsb, Radioberry
Johan,

Thank you very much for your update.

Take your time. It’s a hobby.

Once you modify the code, I’m in a position to do a test anytime.

I really appreciate your work.

73,
Tokio
ja1cca

2020/12/04 16:21、pa3gsb <pa3...@gmail.com>のメール:

You received this message because you are subscribed to a topic in the Google Groups "Radioberry" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/radioberry/HK1xxmC1tDc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to radioberry+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/radioberry/07219795-487d-4735-9bbb-c842b0096732n%40googlegroups.com.

Radioberry

unread,
Dec 6, 2020, 6:12:02 AM12/6/20
to Radioberry
Tokio

Hope the following is also working for you:

I have tested this using SPARK SDR and pihpsdr!

Some background info:

The TX FIFO buffer logic of the HL-2 gateware looks like ; when FIFO is full it is not allowing 
to push the data into the FIFO; skipping the sample.

I have tested this in the initial developement... i did not see this happening; but this is the 
problem you are seeing when TX-ing. I did decrease the buffer size and got the same behaviour.

In the RB the behaviour is different than for the HL-2; it gives distortion.

For a CW tune signal the signal send to the RB is constant giving no problems... but a tune signal in LSB is
a repeating sequence of samples...removing a sample because the FIFO is full resorts in the problem.


In the RPI-3 implementation the TX data is send via the SPI bus. After each sample transferring to
the RB the number of present samples in the FIFO is returned.

If the #samples is near to a full TX FIFO buffer the firmware is waiting for some time before sending new samples.


This concept is used in the new developement in a slightly different implementation.


To make it work you need to install using the following script:

cd /tmp
sudo chmod +x radioberry_install.sh
./radioberry_install.sh


INSTALL:
- gateware:  CL016 or CL025   : 72.4 gateware
- firmware: Build version: rpi-4-2020-12-05
- driver: MODULE_VERSION("0.8");

shutdown and reboot!

Hope this helps.

73 Johan
PA3GSB

Op vrijdag 4 december 2020 om 08:54:51 UTC+1 schreef Tokio:

Tokio Endo

unread,
Dec 7, 2020, 12:25:02 AM12/7/20
to Radioberry
Johan,

Thank you very much for your work to identify the problem including the remedy.

I installed the modified gateware/driver/firmware. The client software is piHPSDR.

Just after the installation, piHPSDR works fine when Tune in LSB as well as CWL. I put the microphone and the output of TX of Radioberry is good. 

However, when I change PA setting of Menu from 38.8 to around 55 to check the linearity of TX out in conjunction with PA setting, and  I change LSB, CWL, LSB, CWL, and so on. Then I got the spurs shown in the attached picture.

This anomaly is cleared after power cycle of Rpi-4. But when I do again the above operation, the spurs occur again. The spurs are very different from the ones I had previously.

There would be a bug in driver/firmware.

73,
Tokio
ja1cca






2020/12/06 20:12、Radioberry <radio...@googlegroups.com>のメール:

pa3gsb

unread,
Dec 7, 2020, 5:26:38 AM12/7/20
to Radioberry
Tokio

Tnx. For the feedback.

Your actions using pihpsdr are not send to the radioberry. 

The driver slider will set the gain of the tx path; this info is send to the RB.

Do you have the SPI module loaded? please uncheck this in raspiconfig.

No time at this moment to reproduce your finding....

Is this behaviour also there using powersdr... which is also having these pa settings?

73 Johan

PA3GSB

Op maandag 7 december 2020 om 06:25:02 UTC+1 schreef Tokio:

Tokio

unread,
Dec 7, 2020, 6:10:20 AM12/7/20
to Radioberry
Johan,

Congratulations!

[piHPSDR]
I've disabled SPI that had been "Enabled" when I did the test.
No spurs when changing the PA setting values as well as Mode changes(Though you explained this mode change info is not sent to Radioberry).
I set Rasberry pi Interface as the following wiki:
I think I2C is no longer needed to be enabled either when we are using Rpi-4 with Radioberry. Is this right?

[PowerSDRmRx]
No spurs when changing the PA settings.

Regarding the interface setting of Raspberry Pi 4,  I hope the above wiki should be updated.

Thank you again for your work to remedy the problem.

73,
Tokio
ja1cca


2020年12月7日月曜日 19:26:38 UTC+9 pa3gsb:

pa3gsb

unread,
Dec 7, 2020, 9:11:37 AM12/7/20
to Radioberry
Tokio

Tnx nice that it is working for you as well; nice step forward to make the radioberry  software more mature!

Iam using one of the chip select signals of the SPI bus to inform about tx fifo full..... SPI driver seems to interfere!

I will update the wiki.

Also the release page needs some attention... iam willing to make a test release (using all the latest git updates)  and 'stable' releases. 


The i2c is not coupled to the FPGA; no interference  you can keep this one checked; people using i2c filtering (supported by the firmware) have to check it.


One nice thing for performance improvement is to bring the TX stream back to the SPI bus and using the 4 data tx signals for rx... which improves the max rx througput.


Have fun,

73 Johan
PA3GSB




Op maandag 7 december 2020 om 12:10:20 UTC+1 schreef Tokio:

Tokio Endo

unread,
Dec 7, 2020, 7:26:29 PM12/7/20
to pa3gsb, Radioberry
Johan,

Thank you very much for your technical explanation data exchange between Rpi-4 and Radioberry.
It helps us to understand the interface.

As to I2C, I know someone makes use of the control of peripherals related to the radio.

I would like to mention that occasionally uploading the gateware to Radioberry FPGA may not work properly, then I need to power cycle the Rpi-4, of course, Radioberry.

In addition to the above observation, I would like you to revise the overview especially about the interface to elaborate on that:

Then, it would be a very nice technical information in order to make use of your Radioberry.

There are a number of Radioberry enthusiasts here in Japanese language environment, and one of them let me know Radioberry several months ago.

I’m confident that Radioberry will let us begin the journey of the experimental SDR project.

I really appreciate your work.

73,
Tokio
ja1cca


2020/12/07 23:11、pa3gsb <pa3...@gmail.com>のメール:

Reply all
Reply to author
Forward
0 new messages