K3 and USB Audio Codec

[oe9fwv]

This is a question to the development team: are there any thoughts or
possibilities to add a USB port to the K3 or P3 which could introduce a
USB "soundcard" to the K3 like it is in the IC-7200 or the upcoming TS-
590s (?)

73! de Werner OE9FWV

--
I suffer occasional delusions of adequacy.


Email powered by Pegasus Mail free at <http://www.pmail.com>
Homepage: <http://www.qsl.net/oe9fwv/>
Fone [phone]
Fax [phone]
Mobile [phone]
Fax-Email Gateway [phone]


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Brett Howard]

No the K3 already has transformer isolated line in and line out
connections.  These make for very simple connections to onboard sound
cards and if you need an outboard sound card with just two 3.5mm audio
ports these can be had for less than 20 bucks online.  

Keeping things in this fashion keeps the higher noise digital equipment
outside of your highly sensitive receiver.

~Brett (N7MG)

On Sat, 2010-07-10 at 14:50 +0200, Dr. Werner Furlan wrote:
> This is a question to the development team: are there any thoughts or
> possibilities to add a USB port to the K3 or P3 which could introduce a
> USB "soundcard" to the K3 like it is in the IC-7200 or the upcoming TS-
> 590s (?)
>
> 73! de Werner OE9FWV
>


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[oe9fwv]

I don't trust the quality of an external soundcard of this price. It might be sufficient for PSK31 but probably not for Winmor. The Signalink USB is good, but not for 20 bucks.
I have not heard of a problem with the DSP in the IC-7200 or 7600.
From what I have heard this is not a real soundcard but the IF passband after the DSP is sent to the audio codec via USB.
It would make it very easy using the K3 for digital soundcard modes without another external device.

73! de Werner OE9FWV

[Brett Howard]

I've got a sound good sound card that I just installed into the computer
and it already is very easy to use digital modes with the K3 without an
external device.

~Brett (N7MG)

On Sat, 2010-07-10 at 12:48 -0700, oe9fwv wrote:

> I don't trust the quality of an external soundcard of this price. It might be
> sufficient for PSK31 but probably not for Winmor. The Signalink USB is good,
> but not for 20 bucks.
> I have not heard of a problem with the DSP in the IC-7200 or 7600.
> >From what I have heard this is not a real soundcard but the IF passband
> after the DSP is sent to the audio codec via USB.
> It would make it very easy using the K3 for digital soundcard modes without
> another external device.
>
> 73! de Werner OE9FWV
>
>
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Joe Subich, W4TV-4]

 > I don't trust the quality of an external soundcard of this price.
 > It might be sufficient for PSK31 but probably not for Winmor. The
 > Signalink USB is good, but not for 20 bucks.

The USB Audio CODEC integrated circuit in the IC-7000 or IC-7600 has
the same specifications as the USB Audio CODEC in Signalink or a $15
USB headset adapter.  There is no benefit to putting audio to digital
conversion in the radio and bringing the audio out using a driver
dependent protocol.

An integrated "USB audio CODEC" is nothing more than a gimmick that
provides no performance benefit while limiting user flexibility/choice,
increasing cost and decreasing reliability.

If one were going to provide digital output, it would be far better
to stream the raw in I/Q data from the DSP's ADC via firewire rather
than push it through two additional A/D and D/A steps and reduce it
to a single channel of audio.

73,

    ... Joe, W4TV


On 7/10/2010 3:48 PM, oe9fwv wrote:

>
> I don't trust the quality of an external soundcard of this price. It might be
> sufficient for PSK31 but probably not for Winmor. The Signalink USB is good,
> but not for 20 bucks.
> I have not heard of a problem with the DSP in the IC-7200 or 7600.
>> From what I have heard this is not a real soundcard but the IF passband
> after the DSP is sent to the audio codec via USB.
> It would make it very easy using the K3 for digital soundcard modes without
> another external device.
>
> 73! de Werner OE9FWV
>
>
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Ross Primrose N4RP]

On 7/10/10 3:48 PM, oe9fwv wrote:
> I don't trust the quality of an external soundcard of this price. It might be
> sufficient for PSK31 but probably not for Winmor. The Signalink USB is good,
> but not for 20 bucks.
>    
The SignalLink USB is just a $20 USB sound card with a VOX circuit added
anyway.... See W4TV's post titled "Re: [Elecraft] AFSK - K3 - SignaLink
USB" from 01JUL10....

73, Ross N4RP

--
FCC Section 97.313(a) “At all times, an amateur station must use the minimum transmitter power necessary to carry out the desired communications.”

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Julian, G4ILO]

Ross Primrose wrote

The SignalLink USB is just a $20 USB sound card with a VOX circuit added
anyway.... See W4TV's post titled "Re: [Elecraft] AFSK - K3 - SignaLink
USB" from 01JUL10....

I have just spent most of the day trying to set up one of my other radios to receive HF APRS on 300baud packet using a "$20 USB sound card" and the AGW Packet Engine soundcard modem.

I had my K3 connected to the PC internal sound card and my FT-817 connected to the $20 USB card. I actually tried two USB devices. One was a real cheapie, more like $10. The other should have been quite high performance as it quotes a 48kHz sample rate and has two SP-DIF sockets on the front in addition to the usual array of speaker outputs.

All of the packets copied were received by the K3 / internal soundcard. None were copied using the USB soundcard. To make sure it wasn't the radio causing the difference I swapped the audio cables over so the K3 was feeding the cheap soundcard and the FT-817 the internal one. All the packets copied then came from the FT-817.

It was curious as I set up both radios / soundcards with simultaneously running copies of Fldigi. The waterfalls looked identical and PSK31 and PSK63 were copied using both setups. But 300baud packet using the AGWPE was a complete dead loss.

I think there are various factors that could be at work here. I'm sure there are people more knowledgeable than me on this reflector so I stand to be corrected but I think it's possible that USB sound cards work at a fixed sample rate and perform resampling in the drivers which may result in the loss of information critical to decoding. Internal sound cards actually sample at the rate selected by the software. So USB sound cards may work fine with some less critical modes or with applications that use the same sample rate as the sound card, but with others they perform poorly.

Therefore I think internal sound cards are always going to perform better, though the difference may only be noticeable with critical modes. Unfortunately they are getting expensive as bog standard sound cards are now obsolete, so you have to pay for a lot of special effects and other fancy stuff that you don't need.

As a related aside, a few weeks ago I recorded signals bounced off the moon from Arecibo from outdoors using a home made yagi and my FT-817. My recording device was a netbook. The signals were extremely weak but barely copyable. When I played back the WAV files on the shack PC using a USB sound card as the audio all I heard was noise. I then "borrowed" the internal sound card from the K3 to use for audio playback and heard the signals exactly as I did originally. Proof positive for me that USB sound cards do deliver a poor quality signal.

[Kok Chen]

On Jul 10, 2010, at 7/10    2:57 PM, Joe Subich, W4TV wrote:

> If one were going to provide digital output, it would be far better
> to stream the raw in I/Q data from the DSP's ADC via firewire rather
> than push it through two additional A/D and D/A steps and reduce it
> to a single channel of audio.

FireWire is not necessarily a great solution either.  It will  
definitely require special drivers at the computer end (viz, the  
Edirol FA-66 that is common in ham SDR circles).

One good connection that I have come across is S/PDIF that is  
available on the Icom IC-7800.  It allows a connection between the  
radio and the computer that is completely non-galvanic.   No ground  
loops, no RFI, down to -144 dB.

S/PDIF provides an audio path with up to 20 bits of resolution (about  
120 dB, with 1 bit being 6.02 dB), with 24 bits being an option by  
using the extra 4 steering bits as data.

The standard sampling rate for S/PDIF is 48,000 samples per second  
stereo, which would support almost 50 kHz of bandwidth on an baseband  
I/Q channel.  S/PDIF provides up to about 30 feet separation between  
equipment.

Both Firewire and S/PDIF lacks a standard way of passing control/
status signals. The Flex-5000 (which uses FireWire) used to hack into  
the MIDI (Musical Instrument Digital Interface) channel for control/
status; but I have not been following that development and it might  
have since changed.  The Icom IC-7800 goes through its regular CAT  
path (RS-232 or CI-V, you get to choose, I believe).

In any case, if you want a rig like the K3 to support 123 dB of  
dynamic range I/Q output, it will not come cheap.  The Asahi AK-5394A  
is probably the most expensive component in the front end of the  
Flex-5000, and you have to be very careful with the circuit layout to  
get a -125 dB noise floor when mixing analog and digital components.

We are definitely not talking about 16-bit codecs like the one in the  
SignaLink USB -- which by the way, is a TI/Burr-Brown PCM-2902, and it  
is not the $20 that people have been mentioning, but is $5.85 at  
DigiKey :-).

73
Chen, W7AY




______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[oe9fwv]

hi Joe,

I am a bit late for a reply, but anyway.
The way the codec is implemented is (as far I understand it) different and
more straitforward. There is no "soundcard" in the radio, but the data
stream is directly derived from the IF-DSP.
My friend Gert, OE3ZK described it as follows:

Es wird kein eigener (billig) SoundChip eingesetzt. Die USB Audio CODEC
Aufbereitung erfolgt direkt aus der ZF-DSP der Transceiver. Abgesehen von
den Quarzungenauigkeiten der billig Sound-Chips wird  DSP nicht
kaskadiert, das ist der wesentliche Qualitätsunterschied.

I translate:
There is no (cheap) Soundchip in the radio. The USB Audio Codec is
directly created out of the IF-DSP of the transceiver. There are no
inaccuracies of the cheap soundchip quartz and there is no cascading of
DSP's, that is a significant difference in quality.

I have a block diagram of the IC-7200 on my website, you can see it here:
http://oe9fwv.shacknet.nu/oe9fwv/bilder/IC-7200%20USB.jpg

This is an interesting approach in my opinion and has two major
advantages. Better quality of the signal for soundcard modes independent
from external soundcards, no additional cables and devices for mobile use.
A radio and the computer.
Thats all. Remote control included.

73! de Werner OE9FWV


Joe Subich, W4TV-4 [via Elecraft] schrieb am 13 Jul 2010 um 9:17:

>
>
> > I don't trust the quality of an external soundcard of this price.
> > It might be sufficient for PSK31 but probably not for Winmor. The
> > Signalink USB is good, but not for 20 bucks.
>
> The USB Audio CODEC integrated circuit in the IC-7000 or IC-7600 has
> the same specifications as the USB Audio CODEC in Signalink or a $15
> USB headset adapter. There is no benefit to putting audio to digital
> conversion in the radio and bringing the audio out using a driver
> dependent protocol.
>
> An integrated "USB audio CODEC" is nothing more than a gimmick that
> provides no performance benefit while limiting user flexibility/choice,
> increasing cost and decreasing reliability.
>
> If one were going to provide digital output, it would be far better
> to stream the raw in I/Q data from the DSP's ADC via firewire rather
> than push it through two additional A/D and D/A steps and reduce it
> to a single channel of audio.
>
> 73,
>
>   ... Joe, W4TV
>
>
> On 7/10/2010 3:48 PM, oe9fwv wrote:
>
> >
> > I don't trust the quality of an external soundcard of this price. It
> might be
> > sufficient for PSK31 but probably not for Winmor. The Signalink USB is
> good,
> > but not for 20 bucks.
> > I have not heard of a problem with the DSP in the IC-7200 or 7600.
> >> From what I have heard this is not a real soundcard but the IF passband
> > after the DSP is sent to the audio codec via USB.
> > It would make it very easy using the K3 for digital soundcard modes
> without
> > another external device.
> >
> > 73! de Werner OE9FWV

--
There is so much sand in Northern Africa, that if we spread it out, it
would completely cover the Sahara Desert.


Email powered by Pegasus Mail free at <http://www.pmail.com>
Homepage: <http://www.qsl.net/oe9fwv/>
Fone [phone]
Fax [phone]
Mobile [phone]
Fax-Email Gateway [phone]


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Joe Subich, W4TV-4]

Werner,

 > There is no (cheap) Soundchip in the radio. The USB Audio Codec is
 > directly created out of the IF-DSP of the transceiver. There are no
 > inaccuracies of the cheap soundchip quartz and there is no cascading
 > of DSP's, that is a significant difference in quality.

The block diagram tells a completely different story.  It shows the
USB sound chip IC3202 is a PCM2901 with AAFO and UMOD.  PCM2901 is
from the same family of USB CODECs used in many of the amateur
"soundcard interfaces" - e.g., RigExpert, Navigator, Signalink USB,
etc.  The connections - AAFO and UMOD - are AUDIO.  AAFO is ANALOG
demodulator audio (the output of D/A IC1692 following the DSP) and
UMOD is an AUDIO input to the transmit DSP modulator (IC1661).

The performance of "USB Audio CODEC" in the 7200 is (and can be)
no different than that of any other well designed PCM2901 circuit
driven by the "demodulator" output on the ACC jack in any of the
other Icom transceivers.

 > This is an interesting approach in my opinion and has two major
 > advantages. Better quality of the signal for soundcard modes
 > independent from external soundcards, no additional cables and
 > devices for mobile use.

The only "advantage" is the lack of a separate box.  The design
is no different than the IC-7000 and a product like the microHAM
USB Interface III.  However, IC-7200 lacks support for FSK, pFSK
(fldigi), audio keyed CW, a computer CW interface and audio pots
to allow optimizing drive levels to take full advantage of the
PCM2901 dynamic range (the PCM2901 has *NO* input gain control).

73,

    ... Joe, W4TV

On 8/17/2010 2:21 AM, Dr. Werner Furlan wrote:

> hi Joe,
>
> I am a bit late for a reply, but anyway.
> The way the codec is implemented is (as far I understand it) different and
> more straitforward. There is no "soundcard" in the radio, but the data
> stream is directly derived from the IF-DSP.
> My friend Gert, OE3ZK described it as follows:
>
> Es wird kein eigener (billig) SoundChip eingesetzt. Die USB Audio CODEC
> Aufbereitung erfolgt direkt aus der ZF-DSP der Transceiver. Abgesehen von
> den Quarzungenauigkeiten der billig Sound-Chips wird  DSP nicht
> kaskadiert, das ist der wesentliche Qualitätsunterschied.
>
> I translate:
> There is no (cheap) Soundchip in the radio. The USB Audio Codec is
> directly created out of the IF-DSP of the transceiver. There are no
> inaccuracies of the cheap soundchip quartz and there is no cascading of
> DSP's, that is a significant difference in quality.
>
> I have a block diagram of the IC-7200 on my website, you can see it here:
> http://oe9fwv.shacknet.nu/oe9fwv/bilder/IC-7200%20USB.jpg
>
> This is an interesting approach in my opinion and has two major
> advantages. Better quality of the signal for soundcard modes independent
> from external soundcards, no additional cables and devices for mobile use.
> A radio and the computer.
> Thats all. Remote control included.
>
> 73! de Werner OE9FWV
>
>
> Joe Subich, W4TV-4 [via Elecraft] schrieb am 13 Jul 2010 um 9:17:
>
>>
>>
>>> I don't trust the quality of an external soundcard of this price.
>>> It might be sufficient for PSK31 but probably not for Winmor. The
>>> Signalink USB is good, but not for 20 bucks.
>>
>> The USB Audio CODEC integrated circuit in the IC-7000 or IC-7600 has
>> the same specifications as the USB Audio CODEC in Signalink or a $15
>> USB headset adapter. There is no benefit to putting audio to digital
>> conversion in the radio and bringing the audio out using a driver
>> dependent protocol.
>>
>> An integrated "USB audio CODEC" is nothing more than a gimmick that
>> provides no performance benefit while limiting user flexibility/choice,
>> increasing cost and decreasing reliability.
>>
>> If one were going to provide digital output, it would be far better
>> to stream the raw in I/Q data from the DSP's ADC via firewire rather
>> than push it through two additional A/D and D/A steps and reduce it
>> to a single channel of audio.
>>
>> 73,
>>
>>    ... Joe, W4TV
>>
>>
>> On 7/10/2010 3:48 PM, oe9fwv wrote:
>>
>>>
>>> I don't trust the quality of an external soundcard of this price. It
>> might be
>>> sufficient for PSK31 but probably not for Winmor. The Signalink USB is
>> good,
>>> but not for 20 bucks.
>>> I have not heard of a problem with the DSP in the IC-7200 or 7600.
>>>>  From what I have heard this is not a real soundcard but the IF passband
>>> after the DSP is sent to the audio codec via USB.
>>> It would make it very easy using the K3 for digital soundcard modes
>> without
>>> another external device.
>>>
>>> 73! de Werner OE9FWV
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Leigh L. Klotz Jr WA5ZNU]

Kok Chen wrote

One good connection that I have come across is S/PDIF that is  
available on the Icom IC-7800.  It allows a connection between the  
radio and the computer that is completely non-galvanic.   No ground  
loops, no RFI, down to -144 dB.

S/PDIF provides an audio path with up to 20 bits of resolution (about  
120 dB, with 1 bit being 6.02 dB), with 24 bits being an option by  
using the extra 4 steering bits as data.

The standard sampling rate for S/PDIF is 48,000 samples per second  
stereo, which would support almost 50 kHz of bandwidth on an baseband  
I/Q channel.  S/PDIF provides up to about 30 feet separation between  
equipment.

I built an amplitude-modulated IR fiber link using cheap parts and got poor results due to fluorescent light leakage into the jacketed cables.  

I've tried hard to find available chipsets for S/P DIF but come up short.  A well-shielded box or internal optical connection for radio would be a nice thing to have for ham equipment.

Leigh/WA5ZNU