KX3 12kHz audio spike
Locked
 
|
While doing some tests tonight to compare the white noise of my Icom
IC-7000 to the white noise of the KX3, both having the AF gain turned all the way down, I noticed something a little unusual in the KX3's spectrum: http://imgur.com/0QF9vXh There is quite a spike of noise at 12kHz (well, relative to everything else around it anyway), with harmonics at 24, 36, and 48 kHz. The latter three of course are inaudible to humans. I found that even without the considerable amplification that I was using to get the levels high enough to analyze, I was able to hear the 12kHz tone in my headphones with the AF gain turned up higher than 30 or so. The sound is present even with everything else disconnected from the unit, and it begins immediately after it is powered up, before it begins receiving. I could see it possibly giving someone fatigue or a headache after a while, even if not consciously aware of it. Has anyone else experienced this, or does anyone have the test equipment to see whether this is present for others and not just me? Nick -- *N6OL* Saying something doesn't make it true. Belief in something doesn't make it real. And if you have to lie to support a position, that position is not worth supporting. ______________________________________________________________ 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 |
|
|
Hello Nick,
I would like to do similar test for my K3 and KX3. Could you please advise how I can do that? TNX & 73, Johnny VR2XMC 寄件人︰ Nicklas Johnson <[hidden email]> 收件人︰ elecraft <[hidden email]> 傳送日期︰ 2013年08月6日 (週二) 1:30 PM 主題︰ [Elecraft] KX3 12kHz audio spike While doing some tests tonight to compare the white noise of my Icom IC-7000 to the white noise of the KX3, both having the AF gain turned all the way down, I noticed something a little unusual in the KX3's spectrum: http://imgur.com/0QF9vXh There is quite a spike of noise at 12kHz (well, relative to everything else around it anyway), with harmonics at 24, 36, and 48 kHz. The latter three of course are inaudible to humans. I found that even without the considerable amplification that I was using to get the levels high enough to analyze, I was able to hear the 12kHz tone in my headphones with the AF gain turned up higher than 30 or so. The sound is present even with everything else disconnected from the unit, and it begins immediately after it is powered up, before it begins receiving. I could see it possibly giving someone fatigue or a headache after a while, even if not consciously aware of it. Has anyone else experienced this, or does anyone have the test equipment to see whether this is present for others and not just me? Nick -- *N6OL* Saying something doesn't make it true. Belief in something doesn't make it real. And if you have to lie to support a position, that position is not worth supporting. ______________________________________________________________ ______________________________________________________________ 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 Nicklas Johnson
This is most likely power supply ripple from the variable loading of the
DSP, which occurs at a 12 kHz rate. You may be able to reduce it by setting MENU:RX ISO to ON. This increases average current in part by preventing the DSP from "sleeping" between tasks, thus smoothing out some of the power supply load. It is always best to use headphones or external speakers that have limited response above 5 kHz with the KX3. The radio will not pass audio information above this frequency, so the only thing wide response headphones or speakers will provide is additional noise. This is particularly true if your ears haven't "aged" to provide some low-pass function for you :-) 73, Lyle KK7P > While doing some tests tonight to compare the white noise of my Icom > IC-7000 to the white noise of the KX3, both having the AF gain turned all > the way down, I noticed... ______________________________________________________________ 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 |
|
|
Hello Lyle,
I really like your last sentence. I can still hear somewhere around 12Khz or higher. This hearing ability is good for Hifi audio but bad for listening to radio. My ears are very sensitive to DSP artifacts. This is the reason why I found the DSP board upgrade for earlier version K3 was necessary. TNX & 73, Johnny VR2XMC 寄件人︰ Lyle Johnson <[hidden email]> 收件人︰ elecraft <[hidden email]> 傳送日期︰ 2013年08月6日 (週二) 8:51 PM 主題︰ Re: [Elecraft] KX3 12kHz audio spike This is most likely power supply ripple from the variable loading of the DSP, which occurs at a 12 kHz rate. You may be able to reduce it by setting MENU:RX ISO to ON. This increases average current in part by preventing the DSP from "sleeping" between tasks, thus smoothing out some of the power supply load. It is always best to use headphones or external speakers that have limited response above 5 kHz with the KX3. The radio will not pass audio information above this frequency, so the only thing wide response headphones or speakers will provide is additional noise. This is particularly true if your ears haven't "aged" to provide some low-pass function for you :-) 73, Lyle KK7P > While doing some tests tonight to compare the white noise of my Icom > IC-7000 to the white noise of the KX3, both having the AF gain turned all > the way down, I noticed... ______________________________________________________________ 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 Lyle Johnson
Lyle: I gave setting RX ISO to ON a try, and though I can see some small
changes in the spectrum analyzer, it doesn't seem to change the 12 kHz spike or its harmonics. Granted, my hearing is abnormally good for my age. I can still hear up to about 18 kHz, which is pretty unusual for an almost-38-year-old I'm told. Flyback transformers in old TVs and monitors are the worst thing in the world. I wonder if connecting a 100uF capacitor across the headphone output would be about right to attenuate noise above 5kHz. Johnny: I did this test by routing the headphone output of the KX3 into the input of a Roland Quad Capture, which is effectively just a really high-quality sound card, turned the gain on the Quad Capture up about halfway, and sampled the audio at 192kHz, 32-bit mono into Sound Forge. Then I ran Sound Forge's spectrum analysis tool on the captured audio. You could probably perform a similar test using an ordinary sound card to see whether you have the spike at 12kHz, but you might not see the ones at 24, 36, and 48kHz if your sound card won't sample above 48000 Hz or won't provide enough gain. Free tools like Audacity also provide spectrum analysis if you don't already have something like Sound Forge. Thanks for the input; it's nice to know at least that I'm not crazy ;-) 73, Nick On 6 August 2013 05:51, Lyle Johnson <[hidden email]> wrote: > This is most likely power supply ripple from the variable loading of the > DSP, which occurs at a 12 kHz rate. > > You may be able to reduce it by setting MENU:RX ISO to ON. This increases > average current in part by preventing the DSP from "sleeping" between > tasks, thus smoothing out some of the power supply load. > > It is always best to use headphones or external speakers that have limited > response above 5 kHz with the KX3. The radio will not pass audio > information above this frequency, so the only thing wide response > headphones or speakers will provide is additional noise. This is > particularly true if your ears haven't "aged" to provide some low-pass > function for you :-) > > 73, > > Lyle KK7P > > While doing some tests tonight to compare the white noise of my Icom >> IC-7000 to the white noise of the KX3, both having the AF gain turned all >> the way down, I noticed... >> > ______________________________**______________________________**__ > Elecraft mailing list > Home: http://mailman.qth.net/**mailman/listinfo/elecraft<http://mailman.qth.net/mailman/listinfo/elecraft> > Help: http://mailman.qth.net/mmfaq.**htm<http://mailman.qth.net/mmfaq.htm> > Post: mailto:[hidden email].**net <[hidden email]> > > This list hosted by: http://www.qsl.net > Please help support this email list: http://www.qsl.net/donate.html > -- *N6OL* Saying something doesn't make it true. Belief in something doesn't make it real. And if you have to lie to support a position, that position is not worth supporting. ______________________________________________________________ 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 |
|
|
Nick,
I think Lyle meant to type RX I/Q, not RX ISO. Turning on the RX I/Q outputs (which is accompanied by the DSP load leveling) will reduce the 12 kHz spur by 15 to 20 dB typically. Doing this also increases overall supply current by about 10 mA. When looking at these spurs on an audio spectrum analyzer that uses a very narrow resolution bandwidth, they're quite prominent. But they are at or below the noise in a 12 kHz (or greater) bandwidth. For quite a few of us we would never know they were there without the instrumentation. I guess good hearing can be a blessing or a curse, depending on what you're listening to... :-) 73, Rich AC7MA On 08/06/2013 07:41 AM, Nicklas Johnson wrote: > Lyle: I gave setting RX ISO to ON a try, and though I can see some small > changes in the spectrum analyzer, it doesn't seem to change the 12 kHz > spike or its harmonics. > > Granted, my hearing is abnormally good for my age. I can still hear up to > about 18 kHz, which is pretty unusual for an almost-38-year-old I'm told. > Flyback transformers in old TVs and monitors are the worst thing in the > world. > > I wonder if connecting a 100uF capacitor across the headphone output would > be about right to attenuate noise above 5kHz. > > Johnny: I did this test by routing the headphone output of the KX3 into the > input of a Roland Quad Capture, which is effectively just a really > high-quality sound card, turned the gain on the Quad Capture up about > halfway, and sampled the audio at 192kHz, 32-bit mono into Sound Forge. > Then I ran Sound Forge's spectrum analysis tool on the captured audio. > You could probably perform a similar test using an ordinary sound card to > see whether you have the spike at 12kHz, but you might not see the ones at > 24, 36, and 48kHz if your sound card won't sample above 48000 Hz or won't > provide enough gain. Free tools like Audacity also provide spectrum > analysis if you don't already have something like Sound Forge. > > Thanks for the input; it's nice to know at least that I'm not crazy ;-) > > 73, > > Nick > > > > On 6 August 2013 05:51, Lyle Johnson <[hidden email]> wrote: > >> This is most likely power supply ripple from the variable loading of the >> DSP, which occurs at a 12 kHz rate. >> >> You may be able to reduce it by setting MENU:RX ISO to ON. This increases >> average current in part by preventing the DSP from "sleeping" between >> tasks, thus smoothing out some of the power supply load. >> >> It is always best to use headphones or external speakers that have limited >> response above 5 kHz with the KX3. The radio will not pass audio >> information above this frequency, so the only thing wide response >> headphones or speakers will provide is additional noise. This is >> particularly true if your ears haven't "aged" to provide some low-pass >> function for you :-) >> >> 73, >> >> Lyle KK7P >> >> While doing some tests tonight to compare the white noise of my Icom >>> IC-7000 to the white noise of the KX3, both having the AF gain turned all >>> the way down, I noticed... >>> >> ______________________________**______________________________**__ >> Elecraft mailing list >> Home: http://mailman.qth.net/**mailman/listinfo/elecraft<http://mailman.qth.net/mailman/listinfo/elecraft> >> Help: http://mailman.qth.net/mmfaq.**htm<http://mailman.qth.net/mmfaq.htm> >> Post: mailto:[hidden email].**net <[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 Lyle Johnson
On 8/6/2013 5:51 AM, Lyle Johnson wrote:
> It is always best to use headphones or external speakers that have > limited response above 5 kHz with the KX3. The radio will not pass > audio information above this frequency, so the only thing wide > response headphones or speakers will provide is additional noise. Horsepucky. This is nothing more than an excuse for poor design. There are several possible solutions, perhaps in combination, but using lousy headphones or speakers is not one of them -- headphones and speakers with poor HF response are also likely to have very bumpy amplitude and phase response in midrange. First, the audio bandwidth ought to be limited in the radio -- a few poles of low pass around 6 kHz would make a nice dent at 12 kHz. Second, if there's that much ripple, it sure sounds to me like the power supply is either inadequately filtered or poorly regulated, or there's insufficient decoupling somewhere in the audio chain. Or perhaps even a circuit layout issue. On several occasions, I've urged the Elecraft engineering team to attend one of Henry Ott's excellent EMC workshops. Henry talks about keeping track of where the current is flowing -- ALL of the current, not just the "intended" current, and he talks about the invisible schematic hidden behind the "ground" symbol, which is one of the most common ways we lose track of the return current. If, for example, the current associated with that 12 kHz clock happens to share a return current path with an audio gain stage, it gets added to the audio. It's the same sort of mechanism that if it's at the junction of the box and the outside world we call a "Pin One Problem." 73, Jim K9YC ______________________________________________________________ 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 |
|
|
Ahh, that's why I love this reflector. One always has a chance to learn new technical terms On 8/6/2013 12:46 PM, Jim Brown wrote: > > Horsepucky. . . . . . ______________________________________________________________ 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 Jim Brown-10
That is absolutely a real effect. Back in the mid-70's I was a product engineer responsible trying to sell tuning diodes to the TV tuner industry. UHF tuners in the U.S. back then were still almost all mechanical, and even though the rest of the world had pretty much already converted to electronic tuning it was a tough sell. Those RF engineers back then were VERY conservative but they really, really knew their stuff. I remember a couple of them giving me a mini-primer on how they designed a mechanical tuner, and it was fascinating. Coupling from the oscillator to the mixer was partially affected by current flow in the very heavy stamped chassis (i.e., ground) partition, and they described how if they needed more coupling they just punched a hole here and there to direct the current where it would have more effect. It might be easy to dismiss such effects as being at VHF frequencies, but that would be a mistake. Currents most certainly do not flow uniformly in ground planes, and coupling effects can be real whenever significant gain is involved. 73, Dave AB7E On 8/6/2013 1:46 PM, Jim Brown wrote: > > Or perhaps even a circuit layout issue. On several occasions, I've > urged the Elecraft engineering team to attend one of Henry Ott's > excellent EMC workshops. Henry talks about keeping track of where the > current is flowing -- ALL of the current, not just the "intended" > current, and he talks about the invisible schematic hidden behind the > "ground" symbol, which is one of the most common ways we lose track of > the return current. If, for example, the current associated with that > 12 kHz clock happens to share a return current path with an audio gain > stage, it gets added to the audio. It's the same sort of mechanism > that if it's at the junction of the box and the outside world we call > a "Pin One Problem." > > 73, Jim K9YC ______________________________________________________________ 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 8/6/2013 8:49 PM, David Gilbert wrote:
> Currents most certainly do not flow uniformly in ground planes, and > coupling effects can be real whenever significant gain is involved. Actually, what happens is that a trace above an ideal "ground plane" (the ground layer) forms a transmission line, with the return current flowing in the ground plane directly below he trace -- UNLESS the ground plane below the trace is broken, for example, by the circuit board layout guy realizing he left something off the main layer, and putting on the ground layer instead. In which case there is no longer a transmission line, the return current flows wherever it can. inductance is added to the path, which causes magnetic coupling to other circuits, and it forms an antenna, so it can radiate, into other circuitry, or outside the box, or both. So the RF current in a ground plane is VERY well behaved and predictable. The problems arise when the PC layout guy doesn't understand the function of the ground plane, and that it CANNOT be interrupted without consequences. Like zipper noise on tuning in the K3. 73, Jim K9YC ______________________________________________________________ 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 |
|
|
I didn't say that RF current in a ground plane wasn't "well-behaved", whatever that means. I said that it wasn't uniform, as in uniform density. It's not. 73, Dave AB7E On 8/7/2013 2:10 AM, Jim Brown wrote: > On 8/6/2013 8:49 PM, David Gilbert wrote: >> Currents most certainly do not flow uniformly in ground planes, and >> coupling effects can be real whenever significant gain is involved. > > Actually, what happens is that a trace above an ideal "ground plane" > (the ground layer) forms a transmission line, with the return current > flowing in the ground plane directly below he trace -- UNLESS the > ground plane below the trace is broken, for example, by the circuit > board layout guy realizing he left something off the main layer, and > putting on the ground layer instead. In which case there is no longer > a transmission line, the return current flows wherever it can. > inductance is added to the path, which causes magnetic coupling to > other circuits, and it forms an antenna, so it can radiate, into other > circuitry, or outside the box, or both. > > So the RF current in a ground plane is VERY well behaved and > predictable. The problems arise when the PC layout guy doesn't > understand the function of the ground plane, and that it CANNOT be > interrupted without consequences. Like zipper noise on tuning in the K3. > > 73, Jim K9YC > ______________________________________________________________ > 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 |
|
|
I decided to build a thing to work around the problem when using
headphones. I did find that turning I/Q on reduced the spike significantly, but didn't eliminate it entirely. I found that a pair of .47uF capacitors and a 3mH choke did a nice job of reducing the noise over 6kHz though: http://i.imgur.com/bAiquCc.png I'd never built a pi-network before, so this was a decent learning exercise too. The reality matched up with the math quite well, I'm pleased to say. The spur at 12kHz went from about -59dB to -78dB, almost exactly 18dB (as before, I'm also adding a lot of gain in the capture device to emphasize the problem; in practice, the spur isn't anywhere near that loud). It takes the spur at 12kHz down about 18dB, which is enough to get it well under the noise threshold for normal listening conditions for my ears. Of course it would be much better if the 12kHz spur weren't there in the first place so it wasn't necessary to try to filter it out after the fact. One thing that's kind of interesting in the spectrum analysis is that filtering above 6kHz looks like it emphasized the spectrum below 6kHz. I suspect that this is not actually what happened (given that the filter is all passive components), but rather the filtering either slightly changed the results of the FFT calculations, or possibly the responsiveness of the audio capture device. I'd be interested to hear others' theories on this unexpected result, too. Nick On 7 August 2013 10:25, David Gilbert <[hidden email]> wrote: > > I didn't say that RF current in a ground plane wasn't "well-behaved", > whatever that means. I said that it wasn't uniform, as in uniform density. > It's not. > > 73, > Dave AB7E > > > > On 8/7/2013 2:10 AM, Jim Brown wrote: > >> On 8/6/2013 8:49 PM, David Gilbert wrote: >> >>> Currents most certainly do not flow uniformly in ground planes, and >>> coupling effects can be real whenever significant gain is involved. >>> >> >> Actually, what happens is that a trace above an ideal "ground plane" (the >> ground layer) forms a transmission line, with the return current flowing >> in the ground plane directly below he trace -- UNLESS the ground plane >> below the trace is broken, for example, by the circuit board layout guy >> realizing he left something off the main layer, and putting on the ground >> layer instead. In which case there is no longer a transmission line, the >> return current flows wherever it can. inductance is added to the path, >> which causes magnetic coupling to other circuits, and it forms an antenna, >> so it can radiate, into other circuitry, or outside the box, or both. >> >> So the RF current in a ground plane is VERY well behaved and predictable. >> The problems arise when the PC layout guy doesn't understand the function >> of the ground plane, and that it CANNOT be interrupted without >> consequences. Like zipper noise on tuning in the K3. >> >> 73, Jim K9YC >> ______________________________**______________________________**__ >> Elecraft mailing list >> Home: http://mailman.qth.net/**mailman/listinfo/elecraft<http://mailman.qth.net/mailman/listinfo/elecraft> >> Help: http://mailman.qth.net/mmfaq.**htm<http://mailman.qth.net/mmfaq.htm> >> Post: mailto:[hidden email].**net <[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<http://mailman.qth.net/mailman/listinfo/elecraft> > Help: http://mailman.qth.net/mmfaq.**htm<http://mailman.qth.net/mmfaq.htm> > Post: mailto:[hidden email].**net <[hidden email]> > > This list hosted by: http://www.qsl.net > Please help support this email list: http://www.qsl.net/donate.html > -- *N6OL* Saying something doesn't make it true. Belief in something doesn't make it real. And if you have to lie to support a position, that position is not worth supporting. ______________________________________________________________ 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 |
|
|
Haha glad it helped you, but it wasn't originally my idea. Credit goes to whomever had suggested it to me here on the list. Turning IQ on will draw around 10ma more power by my reading, but it's worth it not to hear that godawful squeal, IMHO.
Nick (I am still using my 3-pole LC filter when I use headphones) Jessie Oberreuter <[hidden email]> wrote: > >On Thu, 8 Aug 2013, Nicklas Johnson wrote: > >> I did find that turning I/Q on reduced the spike significantly, but >> didn't eliminate it entirely. > > Holy Cow! That did it! Rich! I wonder if your dev unit had the rx >I/Q turned on! Nicklas, THANK YOU! That takes it down to the point >where >I can no longer hear it! There's still general high-freq fatigue with >the >ear-buds, but that's hardly unique to the KX3, and I don't generally >use >ear-buds with radios anyway. > > Nicklas, you just made my day! 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 Elecraft
|
1 view|%1 views
| Free forum by Nabble | Edit this page |