Topics

#audio #audio


 

An RPi3 plus an older draws hat is running with nwdr14 and adc gain set to 0 dB. Audio knob from the radio is set to give an average level of 50 in received direwolf packets.

An RPi4 plus the new draws hat with all else equal, I have to either crank the audio knob up a quarter turn or leave it and adjust adc gain to +7 dB in alsamixer to get the same average level 50 in direwolf.

Both sets of hardware were updated through piwiz and git pull on the n7nix dir.

direwolf is decoding equally well in either case so this may be a non-issue from a practical perspective. But should this be a concern? If not, I’ll happily move on.

Thanks for any insight. I will not be able to respond with any requested console output until this evening. There are some snd_card_* differences between the boards from showudrc if that could be relevant.

BTW, thank you Basil for your scripts. My students studied fm side band theory and ultimately used measure_deviate along with gnu radio and rtf-sdr to put it into practice. The project was well received.


 

Hi David,

There should be no difference in Audio Input between the revisions.

6dB sounds suspiciously like the resistor settings. Run Basil’s Script and let us know.

Bryan


 

Thanks Bryan. I just tried new DRAWS with old Pi and same deal with needing to raise the gain. I've attached the two showudrc outputs. Host:husky is the one I have to turn up the gain.

David


From: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io> on behalf of Bryan Hoyer <bhhoyer@...>
Sent: Wednesday, December 11, 2019 4:17 PM
To: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io>
Subject: Re: [draws and udrc] #audio
 
Hi David,

There should be no difference in Audio Input between the revisions.

6dB sounds suspiciously like the resistor settings. Run Basil’s Script and let us know.

Bryan



Basil Gunn
 

David,

Trying to understand differences between the two units you are
using.

Things to try for more data points.
1. swap RPi SD cards
2. swap DRAWS hat

One unit has a power-down switch?
dtoverlay=gpio-shutdown,gpio_pin=6,gpio_pull=up

What radio (make, model) are you using?

You set the lowest level (PTM_P1) for the DAC PowerTune control (DAC
Playback PT) which drops the LO Drive output voltage down to 330mv
P-P. What happens if you use PTM_P2?

See comment below.

David Lowry <huskyrunnr@...> writes:

An RPi3 plus an older draws hat is running with nwdr14 and adc gain
set to 0 dB. Audio knob from the radio is set to give an average level
of 50 in received direwolf packets.

An RPi4 plus the new draws hat with all else equal, I have to either
crank the audio knob up a quarter turn or leave it and adjust adc gain
to +7 dB in alsamixer to get the same average level 50 in direwolf.

Both sets of hardware were updated through piwiz and git pull on the
n7nix dir.

direwolf is decoding equally well in either case so this may be a
non-issue from a practical perspective. But should this be a concern?
If not, I’ll happily move on.

Thanks for any insight. I will not be able to respond with any
requested console output until this evening. There are some snd_card_*
differences between the boards from showudrc if that could be
relevant.

BTW, thank you Basil for your scripts. My students studied fm side
band theory and ultimately used measure_deviate along with gnu radio
and rtf-sdr to put it into practice. The project was well received.
Thanks and happy to hear that you found the script useful. Appreciate
the feedback.
/Basil n7nix


 

Thanks Basil,

 

“Things to try for more data points.
1. swap RPi SD cards
2. swap DRAWS hat”

I’ll try these things.

“One unit has a power-down switch?
dtoverlay=gpio-shutdown,gpio_pin=6,gpio_pull=up”

Correct, on my old hardware, I loaded nwdr14 and then added my embellishments. No embellishments with the new hardware.

“What radio (make, model) are you using?”

Yaesu FT-2980.

“You set the lowest level (PTM_P1) for the DAC PowerTune control (DAC
Playback PT) which drops the LO Drive output voltage down to 330mv
P-P. What happens if you use PTM_P2?”

For the following, showudrc attached, I used only the RPi3, a single sd card with new nwdr14 image updated by piwiz and git pull on n7nix directory. The mic gain on the radio was set to 5. I swapped DRAWS hats. I still get a large difference in sensitivity to Audio In from the radio. For Audio Out, LO was -6 dB in all cases. I nulled the carrier for both hats at PT_P1, PT_P2, and PT_P3. I used

measure_deviate.sh –f 1042 –c left

PT_P3:

Old DRAWS; PCM_-35.5 dB, New DRAWS; PCM_-13.5 dB

PT_P2:

Old DRAWS; PCM_-30.5 dB, New DRAWS; PCM_-9 dB

PT_P1

Old DRAWS; PCM_-21 dB, New DRAWS; PCM_1 dB

I still worry about my DIY cable. I had problems making it correctly, got what I thought was correct and everything worked fine. Then, the new board came and the different audio levels blew my confidence. The Audio In level for the KPC3+ is closer to the old draws board. But I made that cable too.

David



From: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io> on behalf of Basil Gunn <basil@...>
Sent: Thursday, December 12, 2019 10:32 AM
To: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io>
Subject: Re: [draws and udrc] #audio
 

David,

Trying to understand differences between the two units you are
using.

Things to try for more data points.
1. swap RPi SD cards
2. swap DRAWS hat

One unit has a power-down switch?
dtoverlay=gpio-shutdown,gpio_pin=6,gpio_pull=up

What radio (make, model) are you using?

You set the lowest level (PTM_P1) for the DAC PowerTune control (DAC
Playback PT) which drops the LO Drive output voltage down to 330mv
P-P. What happens if you use PTM_P2?

See comment below.

David Lowry <huskyrunnr@...> writes:

> An RPi3 plus an older draws hat is running with nwdr14 and adc gain
> set to 0 dB. Audio knob from the radio is set to give an average level
> of 50 in received direwolf packets.
>
> An RPi4 plus the new draws hat with all else equal, I have to either
> crank the audio knob up a quarter turn or leave it and adjust adc gain
> to +7 dB in alsamixer to get the same average level 50 in direwolf.
>
> Both sets of hardware were updated through piwiz and git pull on the
> n7nix dir.
>
> direwolf is decoding equally well in either case so this may be a
> non-issue from a practical perspective. But should this be a concern?
> If not, I’ll happily move on.
>
> Thanks for any insight. I will not be able to respond with any
> requested console output until this evening. There are some snd_card_*
> differences between the boards from showudrc if that could be
> relevant.
>
> BTW, thank you Basil for your scripts. My students studied fm side
> band theory and ultimately used measure_deviate along with gnu radio
> and rtf-sdr to put it into practice. The project was well received.

Thanks and happy to hear that you found the script useful. Appreciate
the feedback.
/Basil n7nix




 

OK, I attached the wrong file. Sorry. Here is the showudrc for the powertune/single-Pi/single-sd card experiment.


From: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io> on behalf of David Lowry <huskyrunnr@...>
Sent: Thursday, December 12, 2019 6:22 PM
To: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io>
Subject: Re: [draws and udrc] #audio
 

Thanks Basil,

 

“Things to try for more data points.
1. swap RPi SD cards
2. swap DRAWS hat”

I’ll try these things.

“One unit has a power-down switch?
dtoverlay=gpio-shutdown,gpio_pin=6,gpio_pull=up”

Correct, on my old hardware, I loaded nwdr14 and then added my embellishments. No embellishments with the new hardware.

“What radio (make, model) are you using?”

Yaesu FT-2980.

“You set the lowest level (PTM_P1) for the DAC PowerTune control (DAC
Playback PT) which drops the LO Drive output voltage down to 330mv
P-P. What happens if you use PTM_P2?”

For the following, showudrc attached, I used only the RPi3, a single sd card with new nwdr14 image updated by piwiz and git pull on n7nix directory. The mic gain on the radio was set to 5. I swapped DRAWS hats. I still get a large difference in sensitivity to Audio In from the radio. For Audio Out, LO was -6 dB in all cases. I nulled the carrier for both hats at PT_P1, PT_P2, and PT_P3. I used

measure_deviate.sh –f 1042 –c left

PT_P3:

Old DRAWS; PCM_-35.5 dB, New DRAWS; PCM_-13.5 dB

PT_P2:

Old DRAWS; PCM_-30.5 dB, New DRAWS; PCM_-9 dB

PT_P1

Old DRAWS; PCM_-21 dB, New DRAWS; PCM_1 dB

I still worry about my DIY cable. I had problems making it correctly, got what I thought was correct and everything worked fine. Then, the new board came and the different audio levels blew my confidence. The Audio In level for the KPC3+ is closer to the old draws board. But I made that cable too.

David



From: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io> on behalf of Basil Gunn <basil@...>
Sent: Thursday, December 12, 2019 10:32 AM
To: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io>
Subject: Re: [draws and udrc] #audio
 

David,

Trying to understand differences between the two units you are
using.

Things to try for more data points.
1. swap RPi SD cards
2. swap DRAWS hat

One unit has a power-down switch?
dtoverlay=gpio-shutdown,gpio_pin=6,gpio_pull=up

What radio (make, model) are you using?

You set the lowest level (PTM_P1) for the DAC PowerTune control (DAC
Playback PT) which drops the LO Drive output voltage down to 330mv
P-P. What happens if you use PTM_P2?

See comment below.

David Lowry <huskyrunnr@...> writes:

> An RPi3 plus an older draws hat is running with nwdr14 and adc gain
> set to 0 dB. Audio knob from the radio is set to give an average level
> of 50 in received direwolf packets.
>
> An RPi4 plus the new draws hat with all else equal, I have to either
> crank the audio knob up a quarter turn or leave it and adjust adc gain
> to +7 dB in alsamixer to get the same average level 50 in direwolf.
>
> Both sets of hardware were updated through piwiz and git pull on the
> n7nix dir.
>
> direwolf is decoding equally well in either case so this may be a
> non-issue from a practical perspective. But should this be a concern?
> If not, I’ll happily move on.
>
> Thanks for any insight. I will not be able to respond with any
> requested console output until this evening. There are some snd_card_*
> differences between the boards from showudrc if that could be
> relevant.
>
> BTW, thank you Basil for your scripts. My students studied fm side
> band theory and ultimately used measure_deviate along with gnu radio
> and rtf-sdr to put it into practice. The project was well received.

Thanks and happy to hear that you found the script useful. Appreciate
the feedback.
/Basil n7nix




 

Well, I must have done something to the first board during my cable fiasco a year ago.

I’ve had a tm-v71a since the first of last summer, new-in-box until this morning. I set it up. This eliminates my mad cable skills and gives me some known ALSA parameters from setalsa-tmv71a.sh. PCM for the old hat needs to be set much lower. PCM for the new hat is much closer to the script values, not sure what deviation the script is shooting for. Audio In looks right for the new board as well.

Bryan, is it possible to fubar that ti chip in a way that increases sensitivity? Because that is what it appears I have done. In any case, both hats work reliably on both xmit and rev.

What deviation should I shoot for on the tm-v71a?


 

Rcv, not rev. Thanks spellcheck.


Basil Gunn
 

What deviation should I shoot for on the tm-v71a?
From page 83 of the Kenwood TM-V71a manual
Select 1200 or 9600 bps for the data transfer rate, depending on your TNC.
1200 bps: Transmit data input (PKD) sensitivity is 40 mV p-p , input impedance is 10 kΩ.
9600 bps: Transmit data input (PKD) sensitivity is 2 V p-p , input impedance is 10 kΩ.

I use the 9600 bps connection even with 1200 baud data which just selects the
Kenwood discriminator output (PR9 Output 9600 (bps)) ie. pin 4 on the
mDin6 connector is selected by ALSA setting IN1.

Then you can control your packet data speed with a single setting in
/etc/direwolf.conf MODEM 1200 or MODEM 9600. Make sure your Kenwood
Menu 518 is set to DATSPD 9600.

/Basil n7nix

David Lowry <huskyrunnr@...> writes:

Well, I must have done something to the first board during my cable fiasco a year ago.

I’ve had a tm-v71a since the first of last summer, new-in-box until this morning. I set it up. This eliminates my mad cable skills and gives me some known ALSA parameters from setalsa-tmv71a.sh. PCM for the old hat needs to be set much lower. PCM for the new hat is much closer to the script values, not sure what deviation the script is shooting for. Audio In looks right for the new board as well.

Bryan, is it possible to fubar that ti chip in a way that increases sensitivity? Because that is what it appears I have done. In any case, both hats work reliably on both xmit and rev.

What deviation should I shoot for on the tm-v71a?


 

"I use the 9600 bps connection even with 1200 baud data which just selects the
Kenwood discriminator output (PR9 Output 9600 (bps)) ie. pin 4 on the
mDin6 connector is selected by ALSA setting IN1."

Thank you, this works well with my old board. For the setalsa-tmv71a script, is the switch file placed in /etc/ax25 manually or does some other script do that? 

I am still having issues though. With the new board, alsa parameters from setalsa-tmv71a run with the discriminator, I barely get any modulation at all, judging from the spectrum of measure_deviate.sh -f 1042 -c left .

I looked for a spot on the board to measure V p-p and I noticed that the 100 Ohm resistors from the low pass audio filters, the ones where C17 and C20 are removed, are labeled differently between boards. On my old board, they are marked 101, which I am guessing is 10*10^1 Ohms. On my new board, they are marked 104, which I am guessing is 10*10^4 Ohms.

Audio In is still a bit low on this new board as well.

Could you folks confirm whether or not  the "104" resistor could account for my output sound level discrepancies between my boards? Should I try to jump those "104" resistors?

I've attached a photo of the area where C17 and C20 are removed. The two "104" resistors are the ones in between where those C's were.

David


From: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io> on behalf of Basil Gunn <basil@...>
Sent: Friday, December 13, 2019 10:30 AM
To: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io>
Subject: Re: [draws and udrc] #audio
 

> What deviation should I shoot for on the tm-v71a?

From page 83 of the Kenwood TM-V71a manual
Select 1200 or 9600 bps for the data transfer rate, depending on your TNC.
1200 bps: Transmit data input (PKD) sensitivity is 40 mV p-p , input impedance is 10 kΩ.
9600 bps: Transmit data input (PKD) sensitivity is 2 V p-p , input impedance is 10 kΩ.

I use the 9600 bps connection even with 1200 baud data which just selects the
Kenwood discriminator output (PR9 Output 9600 (bps)) ie. pin 4 on the
mDin6 connector is selected by ALSA setting IN1.

Then you can control your packet data speed with a single setting in
/etc/direwolf.conf MODEM 1200 or MODEM 9600.  Make sure your Kenwood
Menu 518 is set to DATSPD 9600.

/Basil n7nix

David Lowry <huskyrunnr@...> writes:

> Well, I must have done something to the first board during my cable fiasco a year ago.
>
> I’ve had a tm-v71a since the first of last summer, new-in-box until this morning. I set it up. This eliminates my mad cable skills and gives me some known ALSA parameters from setalsa-tmv71a.sh. PCM for the old hat needs to be set much lower. PCM for the new hat is much closer to the script values, not sure what deviation the script is shooting for. Audio In looks right for the new board as well.
>
> Bryan, is it possible to fubar that ti chip in a way that increases sensitivity? Because that is what it appears I have done. In any case, both hats work reliably on both xmit and rev.
>
> What deviation should I shoot for on the tm-v71a?
>
>




 

I had a closer look at the boards. Everywhere there is a 101 on the old board, there is a 104 on the new board. Even on what must be audio in. Is this part of the revision?


 

This morning I went ahead and measured and then jumped R20 with a wire and left R26 alone. I ran the setalsa-tmv71a.sh with the switch file in place. I ran measure_deviate.sh -f 1042 on first the left then the right mDin6. The radio was set to dataspd 9600. The left port is now modulating the carrier with as much deviation as my old hat (about 1800 Hz deviation is all I can get in this mode). The right port is barely deviating. Low res pics attached. Higher res upon request.

Help... I do not have the experience to replace these six "104" resistors, if indeed they should all be replaced.

David Lowry


From: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io> on behalf of David Lowry <huskyrunnr@...>
Sent: Friday, December 13, 2019 8:41 PM
To: udrc@nw-digital-radio.groups.io <udrc@nw-digital-radio.groups.io>
Subject: Re: [draws and udrc] #audio
 
I had a closer look at the boards. Everywhere there is a 101 on the old board, there is a 104 on the new board. Even on what must be audio in. Is this part of the revision?


Basil Gunn
 

"I use the 9600 bps connection even with 1200 baud data which just selects the
Kenwood discriminator output (PR9 Output 9600 (bps)) ie. pin 4 on the
mDin6 connector is selected by ALSA setting IN1."

Thank you, this works well with my old board. For the setalsa-tmv71a
script, is the switch file placed in /etc/ax25 manually or does some
other script do that?
To run 1200 baud packet using the discrimintor on the TM-V71a edit
setalsa-tmv71a.sh
MODE_9600_ENABLE=true
and run the script.

I am still having issues though. With the new board, alsa parameters
from setalsa-tmv71a run with the discriminator, I barely get any
modulation at all, judging from the spectrum of measure_deviate.sh -f
1042 -c left .

I looked for a spot on the board to measure V p-p and I noticed that
the 100 Ohm resistors from the low pass audio filters, the ones where
C17 and C20 are removed, are labeled differently between boards. On my
old board, they are marked 101, which I am guessing is 10*10^1
Ohms. On my new board, they are marked 104, which I am guessing is
10*10^4 Ohms.

Audio In is still a bit low on this new board as well.

Could you folks confirm whether or not the "104" resistor could
account for my output sound level discrepancies between my boards?
Should I try to jump those "104" resistors?
Bryan should get back to you on that.

I've attached a photo of the area where C17 and C20 are removed. The
two "104" resistors are the ones in between where those C's were.

David


 

Hi David,

I will look into this. The last run of DRAWS did not have any value changes in the audio section. You may have found another assembly error.

Thanks,
Bryan K7UDR

On Dec 14, 2019, at 6:13 PM, Basil Gunn <basil@...> wrote:


"I use the 9600 bps connection even with 1200 baud data which just selects the
Kenwood discriminator output (PR9 Output 9600 (bps)) ie. pin 4 on the
mDin6 connector is selected by ALSA setting IN1."

Thank you, this works well with my old board. For the setalsa-tmv71a
script, is the switch file placed in /etc/ax25 manually or does some
other script do that?

To run 1200 baud packet using the discrimintor on the TM-V71a edit
setalsa-tmv71a.sh
MODE_9600_ENABLE=true
and run the script.

I am still having issues though. With the new board, alsa parameters
from setalsa-tmv71a run with the discriminator, I barely get any
modulation at all, judging from the spectrum of measure_deviate.sh -f
1042 -c left .

I looked for a spot on the board to measure V p-p and I noticed that
the 100 Ohm resistors from the low pass audio filters, the ones where
C17 and C20 are removed, are labeled differently between boards. On my
old board, they are marked 101, which I am guessing is 10*10^1
Ohms. On my new board, they are marked 104, which I am guessing is
10*10^4 Ohms.

Audio In is still a bit low on this new board as well.

Could you folks confirm whether or not the "104" resistor could
account for my output sound level discrepancies between my boards?
Should I try to jump those "104" resistors?

Bryan should get back to you on that.

I've attached a photo of the area where C17 and C20 are removed. The
two "104" resistors are the ones in between where those C's were.

David



 

Thanks Bryan! No rush, my old board will keep me busy.

David


 

David, you are right!

100ohm Resistors we’re incorrectly Stuffed as 100k creating a 20dB attenuator.

I should have caught it when we had the cap problem.

With the new run, 8 Components were stuffed incorrectly:

C17, C20 (0603) should be 0.1uF but 10uF was stuffed

Easiest Fix is to Remove

R20, R21, R26, R27, R39, R40 (0603) 100ohms but 100kohms was stuffed

Easiest Fix is to solder 100ohms on top (in parallel)

Assembly has current stock in for re-work.

We’ll make it right for boards in the field, whatever it takes.

I’m out of the office until January 6th.

Thanks for your patience,
Bryan K7UDR


 

Rework Directions/Image


John McRae
 

I just purchased one a week or so ago Bryan,

I am wondering wether it was "corrected" before it was posted to me here in Australia.

If not, I will source the correct components and repopulate the pcb.

MERRY XMAS from Australia.

73 de VK5PO

On 23/12/2019 7:03 am, Bryan Hoyer wrote:
Rework Directions/Image

-- 
MY GMAIL desktop Email accoount,
JF McRae, VK5PO/VK5SIX


 

Good deal Bryan, thanks for the instructions. I’m going to try the piggyback. Always wanted to learn smd. I’ll practice on scrap. Have a good break. David


Ruben .
 

Hi All,

If we are affected by the error, will we be notified? I’ve had mine about a month.

I’ve configured the software, but have not yet transmitted, everything works on receive. Getting antenna and go-box built to enable portable operation. The Pi-DRAWS case would be real helpful.

Also installing Auto-Access Point and WinLink PAT configuration by KM4ACK.

I can break them, but I can’t fix them.

Thanks and 73,

Ruben
WA2NBL

On Dec 22, 2019, at 12:53 PM, Bryan Hoyer <@K7UDR> wrote:

David, you are right!

100ohm Resistors we’re incorrectly Stuffed as 100k creating a 20dB attenuator.

I should have caught it when we had the cap problem.

With the new run, 8 Components were stuffed incorrectly:

C17, C20 (0603) should be 0.1uF but 10uF was stuffed

Easiest Fix is to Remove

R20, R21, R26, R27, R39, R40 (0603) 100ohms but 100kohms was stuffed

Easiest Fix is to solder 100ohms on top (in parallel)

Assembly has current stock in for re-work.

We’ll make it right for boards in the field, whatever it takes.

I’m out of the office until January 6th.

Thanks for your patience,
Bryan K7UDR