DSP
Locked
 
|
I note that Wayne makes a point of the fact that the DSP in the KX3 is
32-bit, floating point. This buzzphrase is also used by other manufacturers to tout their DSPs, but I can't find any similar description of the DSP hardware in the K3. To raise again a point that was mentioned yesterday: if there's a potential for getting more DSP horsepower in the K3, I'm sure that many K3 owners would happily pay for a retrofit. Tony KT0NY ______________________________________________________________ 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 |
|
|
>From the datasheet for the DSP chip used in the K3:
The TMS320VC33 DSP is a 32-bit, floating-point processor manufactured in 0.18-Tm four-level-metal CMOS (TImeline) technology. The TMS320VC33 is part of the TMS320C3x generation of DSPs from Texas Instruments. The K3's DSP -is- 32 bit floating point. As Wayne has said, the K3 ain't dead yet, and will continue to be developed. Notice the decreasing hair in recent pics of Lyle. 8-) Bruce, N1RX ______________________________________________________________ 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 |
|
|
Thanks, Bruce. I didn't look hard enuff.
Tony KT0NY On Sat, May 21, 2011 at 8:55 AM, Bruce Beford <[hidden email]>wrote: > >From the datasheet for the DSP chip used in the K3: > > The TMS320VC33 DSP... > > ______________________________________________________________ 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 |
|
|
In reply to this post by Bruce Beford-2
Nice overview of the chip at http://focus.ti.com/lit/an/spra526/spra526.pdf
It's not a NR chip. It's a CPU with embedded memory and gizmos designed with digital signal processing in mind. You could use it in a model airplane if you wanted to, or construct a mini PC out of it. So it allows inclusion of appropriate K3 firmware development in a tiny box. The big improvement that is pertinent to a mini rig in the new chip appears to be CMOS and lower power, and more so because the processor is touted at lower power at higher speeds. Run it at less than WFO and the power draw is even less. I remember that there was a serious current addition (from viewpoint of battery operation) when they added DSP to the K2. If one was going to make a serious quantum jump in DSP with higher-end processing than you see in a K3, most certainly a jump in component price-point would be involved. A QRP miniature, with the intense price competition, is NOT the place to start. What you are seeing in the KX3, with that 32 bit processor, is prior high end stuff coming down enough in cost and battery drain to put all of the K3's high end processing, and fairly mature DSP program code, in a mini. That's the opposite of what Tony seemed to be worried about in earlier posts. AND, all of this seems to suggest that a K4 is going to be SMALLER than a K3, not larger. Bring it on. 73, Guy. On Sat, May 21, 2011 at 9:55 AM, Bruce Beford <[hidden email]> wrote: > >From the datasheet for the DSP chip used in the K3: > > The TMS320VC33 DSP is a 32-bit, floating-point processor manufactured in > 0.18-Tm four-level-metal CMOS > (TImeline) technology. The TMS320VC33 is part of the TMS320C3x generation of > DSPs from Texas > Instruments. > > The K3's DSP -is- 32 bit floating point. As Wayne has said, the K3 ain't > dead yet, and will continue to be developed. Notice the decreasing hair in > recent pics of Lyle. 8-) > > Bruce, N1RX > > > ______________________________________________________________ > 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 > 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 |
|
|
On Sat, May 21, 2011 at 10:01 AM, Guy Olinger K2AV <[hidden email]>wrote:
> ....That's the opposite of what Tony seemed to be worried about... Yeah, my interest is just in having the most powerful DSP possible, whatever the current draw and/or cost. I have modest antennas; there's a lot of noise in the ether here; and the propagation from here to the countries I still need is such that the signals, if and when there are signals, are extremely weak and covered up by buzzing, humming, grinding racket. I realize that there's a Shannon limit to what a DSP can do, but I'm always hoping for more and more magic. Tony KT0NY ______________________________________________________________ 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 |
|
|
In reply to this post by Tony Estep
CONTENTS DELETED
The author has deleted this message.
|
|
|
Wayne was talking about an optional "roofing filter that is in the
audio range ahead of the DAC" in the YouTube video. So it is seriously down-converted before it goes digital. It's not a Flex 5000. 73, Guy. On Sat, May 21, 2011 at 4:05 PM, Ron D'Eau Claire <[hidden email]> wrote: > And I'd not be at all surprised to see such upgrades available at some > future time if the improvement is significant in the K3 application and > there is sufficient demand. But that may not be the case. > > The K3's receiver is a superhet (i.e. converts the input signal to a fixed > intermediate frequency where it is amplified, filtered then demodulated). A > successful superhet receiver is a carefully balanced system from the antenna > input to the audio or data output. One component that might make a huge > improvement in one superhet design may have little or no benefit in another > superhet design. In the 90 years since Col. Armstrong developed the superhet > receiver platform all of the improvements have been in the various > components, not in the underlying design. And that continues. > > However, a full SDR receiver may not use the superhet architecture at all. > An SDR receiver typically converts the RF signal at the antenna input > directly into digital data and from that point on it's all crunching bits > and bytes until, if an analog output such as audio to speaker or phones is > needed, the digital data is used to produce an analog audio signal. I got > the impression from the video with Wayne that the KX3 *may* use a full SDR > receiver. > > Whether or not the KX3 uses a full SDR or retains a superhet architecture, > it will have quite different demands on the components than the K3, > including the DSP system. > > I'm sure there will be a LOT of test data comparing the KX3 and K3 receiver > as well as endless anecdotal comparisons shared here on the reflector by > those with both units. I'm looking forward to them. > > Ron AC7AC > > > -----Original Message----- > I note that Wayne makes a point of the fact that the DSP in the KX3 is > 32-bit, floating point. This buzzphrase is also used by other manufacturers > to tout their DSPs, but I can't find any similar description of the DSP > hardware in the K3. To raise again a point that was mentioned yesterday: if > there's a potential for getting more DSP horsepower in the K3, I'm sure that > many K3 owners would happily pay for a retrofit. > > Tony KT0NY > > ______________________________________________________________ > 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 > 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 |
|
|
Perhaps we just need to take a breather on this thing. This general discussion (no fingers specifically pointed) is getting to be as ridiculous as the end of world. The prototype about which we know very little is already being reengineered by people who know even less, and somehow this has morphed into a "K3 is dead, and has to be brought up to spec and maybe NR will be better and where is the K3 upgrade" discussion. Because a better (?) DSP will solve every problem. Right. The radio is at least 6 months away, and we already have a Yahoo "users group", where the speculation and reengineering will no doubt take on a life of it's own.
Good grief ... zip it back up, take a cold shower, and let's see -- rationally -- what the radio is and is not. And LET ELECRAFT ENGINEER IT!!!! Sent from my iPhone On May 21, 2011, at 4:30 PM, Guy Olinger K2AV <[hidden email]> wrote: > Wayne was talking about an optional "roofing filter that is in the > audio range ahead of the DAC" in the YouTube video. So it is > seriously down-converted before it goes digital. It's not a Flex > 5000. 73, Guy. > > On Sat, May 21, 2011 at 4:05 PM, Ron D'Eau Claire <[hidden email]> wrote: >> And I'd not be at all surprised to see such upgrades available at some >> future time if the improvement is significant in the K3 application and >> there is sufficient demand. But that may not be the case. >> >> The K3's receiver is a superhet (i.e. converts the input signal to a fixed >> intermediate frequency where it is amplified, filtered then demodulated). A >> successful superhet receiver is a carefully balanced system from the antenna >> input to the audio or data output. One component that might make a huge >> improvement in one superhet design may have little or no benefit in another >> superhet design. In the 90 years since Col. Armstrong developed the superhet >> receiver platform all of the improvements have been in the various >> components, not in the underlying design. And that continues. >> >> However, a full SDR receiver may not use the superhet architecture at all. >> An SDR receiver typically converts the RF signal at the antenna input >> directly into digital data and from that point on it's all crunching bits >> and bytes until, if an analog output such as audio to speaker or phones is >> needed, the digital data is used to produce an analog audio signal. I got >> the impression from the video with Wayne that the KX3 *may* use a full SDR >> receiver. >> >> Whether or not the KX3 uses a full SDR or retains a superhet architecture, >> it will have quite different demands on the components than the K3, >> including the DSP system. >> >> I'm sure there will be a LOT of test data comparing the KX3 and K3 receiver >> as well as endless anecdotal comparisons shared here on the reflector by >> those with both units. I'm looking forward to them. >> >> Ron AC7AC >> >> >> -----Original Message----- >> I note that Wayne makes a point of the fact that the DSP in the KX3 is >> 32-bit, floating point. This buzzphrase is also used by other manufacturers >> to tout their DSPs, but I can't find any similar description of the DSP >> hardware in the K3. To raise again a point that was mentioned yesterday: if >> there's a potential for getting more DSP horsepower in the K3, I'm sure that >> many K3 owners would happily pay for a retrofit. >> >> Tony KT0NY >> >> ______________________________________________________________ >> 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 >> > ______________________________________________________________ > 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 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 |
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1 Yep, good advice. On 5/21/2011 4:48 PM, Grant Youngman wrote: Perhaps we just need to take a breather on this thing. This general discussion (no fingers specifically pointed) is getting to be as ridiculous as the end of world. The prototype about which we know very little is already being reengineered by people who know even less, and somehow this has morphed into a "K3 is dead, and has to be brought up to spec and maybe NR will be better and where is the K3 upgrade" discussion. Because a better (?) DSP will solve every problem. Right. The radio is at least 6 months away, and we already have a Yahoo "users group", where the speculation and reengineering will no doubt take on a life of it's own. Good grief ... zip it back up, take a cold shower, and let's see -- rationally -- what the radio is and is not. And LET ELECRAFT ENGINEER IT!!!! - -- R. Kevin Stover ACØH -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJN2DdbAAoJEAu8BkU0m9MnyRgIAKTEcxT16nmI6kkYhUVnKdLN FP9/4ZbX6aQ9l4nhURpLLvzBLE3Yc0oT7Fb6MHRUA5+1OFVnAHaZzXW8pxdiZ8eg aa4tqaUI2wANjVXrq9Aa+6IbdgvnKcysLMljVy8rKmYhnRxON7YDWEpq+9jOz5uW QQ0Ws7q3v+ZeTZBfAAlt3bvtNtjQsHQ7S8zigbysPUhAZACE+aQuld7NNRLK9IJb d3VPhYcMRAYu4/6Nii22s5QpqNd44dNGRH9qouJidox4snZZv0OEo/0HnUar4C/1 N7/AVNeWvzEoAs12pmq/u3Od/iFPvbIpuUBYdaZ6sMqiYRf92K4u5c/uPAUEmok= =1UGo -----END PGP SIGNATURE----- ______________________________________________________________ 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 |
|
|
In reply to this post by Guy, K2AV
Guy, the Flex is down converted as well to 11kHz before the A/D.
Jim, W4ATK On May 21, 2011, at 4:30 PM, Guy Olinger K2AV wrote: > Wayne was talking about an optional "roofing filter that is in the > audio range ahead of the DAC" in the YouTube video. So it is > seriously down-converted before it goes digital. It's not a Flex > 5000. 73, Guy. > > On Sat, May 21, 2011 at 4:05 PM, Ron D'Eau Claire <[hidden email]> wrote: >> And I'd not be at all surprised to see such upgrades available at some >> future time if the improvement is significant in the K3 application and >> there is sufficient demand. But that may not be the case. >> >> The K3's receiver is a superhet (i.e. converts the input signal to a fixed >> intermediate frequency where it is amplified, filtered then demodulated). A >> successful superhet receiver is a carefully balanced system from the antenna >> input to the audio or data output. One component that might make a huge >> improvement in one superhet design may have little or no benefit in another >> superhet design. In the 90 years since Col. Armstrong developed the superhet >> receiver platform all of the improvements have been in the various >> components, not in the underlying design. And that continues. >> >> However, a full SDR receiver may not use the superhet architecture at all. >> An SDR receiver typically converts the RF signal at the antenna input >> directly into digital data and from that point on it's all crunching bits >> and bytes until, if an analog output such as audio to speaker or phones is >> needed, the digital data is used to produce an analog audio signal. I got >> the impression from the video with Wayne that the KX3 *may* use a full SDR >> receiver. >> >> Whether or not the KX3 uses a full SDR or retains a superhet architecture, >> it will have quite different demands on the components than the K3, >> including the DSP system. >> >> I'm sure there will be a LOT of test data comparing the KX3 and K3 receiver >> as well as endless anecdotal comparisons shared here on the reflector by >> those with both units. I'm looking forward to them. >> >> Ron AC7AC >> >> >> -----Original Message----- >> I note that Wayne makes a point of the fact that the DSP in the KX3 is >> 32-bit, floating point. This buzzphrase is also used by other manufacturers >> to tout their DSPs, but I can't find any similar description of the DSP >> hardware in the K3. To raise again a point that was mentioned yesterday: if >> there's a potential for getting more DSP horsepower in the K3, I'm sure that >> many K3 owners would happily pay for a retrofit. >> >> Tony KT0NY >> >> ______________________________________________________________ >> 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 >> > ______________________________________________________________ > 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 JIM ROGERS, W4ATK [hidden email] http://web.me.com/jimrogers_w4atk K3/100 P3 K2/10 ______________________________________________________________ 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 |
|
|
In reply to this post by Grant Youngman
I thought the purpose of these groups was endless speculation and the creation of pent up demand. Now you want to ruin that :-)
--- On Sat, 5/21/11, Grant Youngman <[hidden email]> wrote: > Perhaps we just need to take a > breather on this thing. This general discussion (no fingers > specifically pointed) is getting to be as ridiculous as the > end of world. The prototype about which we know very > little is already being reengineered by people who know even > less, and somehow this has morphed into a "K3 is dead, and > has to be brought up to spec and maybe NR will be better and > where is the K3 upgrade" discussion. Because a better > (?) DSP will solve every problem. Right. The radio is at > least 6 months away, and we already have a Yahoo "users > group", where the speculation and reengineering will no > doubt take on a life of it's own. > > Good grief ... zip it back up, take a cold shower, and > let's see -- rationally -- what the radio is and is not. And > LET ELECRAFT ENGINEER IT!!!! > ______________________________________________________________ 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 |
|
|
One thing the Elecrew does very well is keeping overlap out of its product
lineup. The fact that there are no discontinued transceivers in the lineup supports that. While some overlap is inevitable, each transceiver seems to be geared to serving a certain audience with certain needs and operating habits. While the KX3 is appears to be highly adaptable, it's not a K3, or a K2, or a reason to bring the others in line with its capabilities. My main problem isn't "how do I get a KX3" - but rather, "how do I get a KX3 and a K3 and a P3 and a KPA500 and a K1 and ...." --Andrew, NV1B > somehow this has morphed into a "K3 is dead, and > > has to be brought up to spec and maybe NR will be better and > > where is the K3 upgrade" discussion. ______________________________________________________________ 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 |
|
|
In reply to this post by Tony Estep
Interesting. Can you explain how exactly you define "DSP horsepower"? Is
there evidence that K3 does not have enough "DSP horsepower"? Are you sure the current K3 DSP software used full potential of the TI DSP chip? If we talking about processing, somewhere I heard that KX3 use 24 bit sigma-delta ADC. Now that is something new with potential. If I remember correctly, the K3 uses 16 bit codec. I am a little skeptical that 24 bit SD will give measurable advantage over 16 bit ADC in ham radio application. I guess at some point it will be measured. 73, Igor, N1YX -----Original Message----- From: [hidden email] [mailto:[hidden email]] On Behalf Of Tony Estep Sent: Saturday, May 21, 2011 9:43 AM To: Elecraft Subject: [Elecraft] DSP I note that Wayne makes a point of the fact that the DSP in the KX3 is 32-bit, floating point. This buzzphrase is also used by other manufacturers to tout their DSPs, but I can't find any similar description of the DSP hardware in the K3. To raise again a point that was mentioned yesterday: if there's a potential for getting more DSP horsepower in the K3, I'm sure that many K3 owners would happily pay for a retrofit. Tony KT0NY ______________________________________________________________ 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 ______________________________________________________________ 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 |
|
|
In reply to this post by Tony Estep
According to the K3 Owner's Manual, the K3 uses a 24-bit ADC. From the
schematic, it appears to be the Burr-Brown PCM1804DB, a very nice part. Bob NW8L -----Original Message----- Interesting. Can you explain how exactly you define "DSP horsepower"? Is there evidence that K3 does not have enough "DSP horsepower"? Are you sure the current K3 DSP software used full potential of the TI DSP chip? If we talking about processing, somewhere I heard that KX3 use 24 bit sigma-delta ADC. Now that is something new with potential. If I remember correctly, the K3 uses 16 bit codec. I am a little skeptical that 24 bit SD will give measurable advantage over 16 bit ADC in ham radio application. I guess at some point it will be measured. 73, Igor, N1YX ______________________________________________________________ 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 |
«
Return to [KX3]
|
1 view|%1 views
| Free forum by Nabble | Edit this page |