Added LIN OUT measurements
Locked
 
|
Within the last few minutes, I've added static harmonic plots at LIN
OUT = 2, 5 and 10 in addition to the GIF animation from 2 to 100 at http://www.cliftonlaboratories.com/elecraft_k3_receive_audio.htm The data shows a considerable increase in odd order harmonic distortion between LIN OUT = 5 and 10. Looking at the animation, however, should convince the viewer that the 3rd harmonic is consistently down 45 dB over a wide range of LIN OUT settings (10 to 100.) Jack _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
|
|
Jack, Could you do me a favour please? I'm interested in what happens on the headphone output if you re-run the tests using a 2.7 or 2.8kHz filter with the audio tone set low enough in frequency so that the 3rd harmonic lies within the filter passband (i.e. below 900Hz). I use the Heil phones which have a 200 ohm impedance so if you could use that as the load that would be great. Thanks (if you can do this) Regards, Paul M1PAF |
|
|
On Wed, 3 Sep 2008 14:00:27 -0700 (PDT), Paul Fletcher wrote:
Paul, I's clear from your email that you don't understand how a ham receiver works. See comments interspersed. >Could you do me a favour please? I'm interested in what happens on the >headphone output if you re-run the tests using a 2.7 or 2.8kHz filter with >the audio tone set low enough in frequency so that the 3rd harmonic lies >within the filter passband (i.e. below 900Hz). First, the 2.7 or 2.8 kHz filters are operating at 8.8 MHz, not at audio. They are 2.7 kHz or 2.8 kHz WIDE at the 8.8 MHz IF. Thus, they have NO effect on the measurement of audio distortion in the manner that Jack is doing it. >I use the Heil phones which have a 200 ohm impedance so if you could use that >a the load that would be great. Another misconception. Headphones and loudspeakers do not have a constant impedance -- indeed, their impedance isn't even resistive! Rather, their impedance varies widely over the audio spectrum. International stardards define the impedance of loudspeakers and headphones as the MINIMUM value of the impedance within their operating bandwidth. See my tutorial about this in http://audiosystemsgroup.com/RFI-Ham.pdf Now, the load impedance IS relevant to distortion measurement -- in general, amplifiers produce the greatest distortion when they are providing the most current, so a lower impedance load would draw more current for the same output voltage. BUT -- different headphones have different voltage sensitivities. That is, it takes more voltage to get the same loudness in one headphone than another. AND, a lower impedance headset will take less voltage to get the same power. See how interwined all this stuff is? The closest that Jack could come to honoring your request is to measure the distortion when actually driving a set of those headphones with a specified output voltage over a range of frequencies (because Z, and thus I, is a variable). 73, Jim Brown K9YC _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
|
|
Jim,
Maybe I wasn't clear what I was asking. I was asking for a carrier to be injected that produced an audio tone of less than 900Hz so that the third harmonic of that was within the passband. I'm interested in the DSP and if that is introducing the effect I'm hearing. As for the difference between impedance and resistance your point is duly noted and the only way this could be achieved is to use a set of Heil cans as the load (or something with very similar characteristics). 73 Paul. |
|
|
On Wed, 3 Sep 2008 21:37:00 -0700 (PDT), Paul Fletcher wrote:
>I'm interested in the DSP and if >that is introducing the effect I'm hearing. The DSP is not operating at baseband audio either -- it is another IF just above the audio passband. The only way it's going to generate an audio product as distortion is the IM of two or more RF signals. 73, Jim K9YC _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
|
|
Jim, I completely understand the K3 architecture and that the DSP operates at a low IF. DSP is still software though and can still have bugs - no aspertions being cast just a simple statement. This can produce weird effects and analogies to analogue techiniques are only really applicable when describing DSP topologies - implementation is a whole different ball game. As I'm sure you are aware from your audio work (very interesting articles by the way) digital processing of any kind of signal introduces a raft of new pitfalls such as Nyquist aliasing, clock jitter, differing data formats etc etc. I see strange things happening that I cannot attribute to non-linear effects of the AF stage. Let me explain (maybe my fourier transform theory is a bit rusty - not used it for 20 years or so). As far as I can remember when a sine wave starts to clip odd order harmonics are introduced at varying phase and amplitude in relation to the fundamental, depending on the level of clipping. So far so good. Now, I have decoded a harmonically related PSK31 signal (i.e. a "ghost") with no problem whatsoever. This implies that the phase relationships are as per the "real" signal. Also the bandwidth of the ghost signal is exactly the same as the "real" signal with Lin out set at 4 (yes I realise this is not the phones output but trust me the ghost signal appears on the phones output as well). The thing that puzzles me is this. When a PSK signal is idling it is two tones a set frequency apart. For the purposes of this excercise let's say the tones are 150Hz and 200Hz. The third harmonic of each tone is 450Hz and 600Hz respectively so the ghost signal should have a bandwidth of 150Hz but it doesn't - it's still 50Hz (50Hz may not be the real spacing of the but you get my point). If my reasoning is off beam I would welcome an explaination (still learning something new each day). Regards, Paul |
RE: Added LIN OUT measurements
|
|
In reply to this post by Paul Fletcher
> Maybe I wasn't clear what I was asking. I was asking for a > carrier to be injected that produced an audio tone of less > than 900Hz so that the third harmonic of that was within the > passband. I'm interested in the DSP and if that is > introducing the effect I'm hearing As has been stated a dozen times in a dozen different ways, what you are hearing is simply the effect of overdriving the audio amplifiers. I posted the following table showing harmonic distortion products for the speaker amplifier the other day: > The following table shows harmonic levels relative to a 500 Hz > reference tone at the given voltage across an 8 Ohm speaker. > 0 dB is 1V Peak (.7V RMS). +10 dB represents 2.25V RMS (signal > + distortion). +13 dB is 3.2V RMS (the speaker amplifier is in > the compression region at the threshold of clipping) and +15 > dB is 4V RMS (the speaker amplifier is in hard clipping). > > Reference 2nd 3rd 4th 5th 6th 7th 8th 9th > ============================================================ > +15 dB -37 -15 -45 -26 -60 -43 -51 -37 > +13 dB -49 -31 -51 -35 -57 -39 -63 -49 > +10 dB -70 -68 -95 -67 -85 -69 -85 -73 > 0 dB -66 -74 -70 -77 -73 -76 -77 -80 > -10 dB -54 -66 -63 -90 -73 -82 -80 -87 > -20 dB -50 -78 -68 -74 -73 -74 -74 -78 > > In order to drive the audio amplifier hard enough to create > the distortion, it was necessary to disable the AGC or set > AGC SLP to low values (reduce the AGC reduction for strong > signals). Until the level is increased to the point that the amplifier is in compression and/or clipping, harmonic levels are extremely low. Here is a similar set of data points for the headphone output: 8 Ohm (old Yaesu) headset Reference 2nd 3rd 4th 5th 6th 7th 8th 9th ============================================================ -6 dB -60 -36 -74 -37 -73 -39 -75 -44 -7 dB -60 -69 -79 -75 -85 -77 -87 -81 -10 dB -61 -69 -82 -60 -73 -66 -87 -80 -20 dB -72 -60 -73 -66 -74 -70 -75 -74 -30 dB -61 -56 -63 -64 -65 -67 -66 nf 32 Ohm (computer) headset Reference 2nd 3rd 4th 5th 6th 7th 8th 9th ============================================================ +1 dB -62 -20 -68 -28 -82 -50 -70 -46 0 dB -62 -73 -89 -79 -90 -84 -94 -86 -10 dB -70 -66 -80 -73 -85 -80 -86 -81 -20 dB -72 -60 -73 -66 -74 -70 -75 -74 -30 dB -52 -52 -52 nf -54 -60 -55 nf 200 Ohm (Heil Proset) headset Reference 2nd 3rd 4th 5th 6th 7th 8th 9th ============================================================ +4 dB -64 -48 -90 -47 -84 -50 -80 -53 +3 dB -65 -79 -95 -86 -98 -93 -97 -93 0 dB -68 -79 -82 -84 -86 -80 -84 -83 -10 dB -76 -75 -82 -86 -84 nf -86 nf -20 dB -72 -80 -73 nf -75 nf -76 nf 0 dB is 770 mV RMS (1V Peak) measured for the reference tone across the headphone element - total audio level will be somewhat higher with distortion/noise. Note: in every case harmonic levels are extremely low, generally down more than 60 dB right to the point that the headphone amplifier goes into compression/clipping. "nf" indicares harmonic product not measurable (in the noise floor). The table shows the harmonic levels at maximum "linear" output and again with the reference tone raised just 1 dB above the clipping point. Now my ears are not "young" and have some loss from more than 30 years around control room monitors that were always "too loud" and the many hours around the big pumps/blowers in the TV transmitter sites. Still, with all of the tested headphones the audio level is nearly painful at least 3 dB before the amplifier went into clipping and the harmonic products exceed -60 dB. Based on these and previous tests, anyone who can not get adequate headphone levels without distortion should be look- ing for more efficient headphones and/or loudspeakers. 73, ... Joe, W4TV > -----Original Message----- > From: [hidden email] > [mailto:[hidden email]] On Behalf Of Paul Fletcher > Sent: Thursday, September 04, 2008 12:37 AM > To: [hidden email] > Subject: Re: [Elecraft] Added LIN OUT measurements > > > > Jim, > > Maybe I wasn't clear what I was asking. I was asking for a > carrier to be injected that produced an audio tone of less > than 900Hz so that the third harmonic of that was within the > passband. I'm interested in the DSP and if that is > introducing the effect I'm hearing. As for the difference > between impedance and resistance your point is duly noted and > the only way this could be achieved is to use a set of Heil > cans as the load (or something with very similar characteristics). > > 73 Paul. _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
Headphone Output - SSB Mode - Single Tone
|
|
In reply to this post by Paul Fletcher
At the request of Paul, M1PAF, I've measured the harmonic distortion
seen at the headphone output in a K3 operating with with 2.7 KHz bandwidth USB mode with a single tone at 500 Hz offset. This allows up to the fifth harmonic to be within the DSP passband. The details are at http://www.cliftonlaboratories.com/elecraft_k3_receive_audio.htm. Audio level is set for a comfortable listening level with a Heil Pro Set, corresponding to about 30 mV RMS. The short story is 2nd harmonic is at least 80 dB down and the 3rd harmonic is 76 dB down. Fourth and fifth harmonics are in the noise and not visible. An artifact at 3.9 KHz is also visible, although quite far down compared with the desired signal. Jack K8ZOA www.cliftonlaboratories.com _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
Re: Headphone Output - SSB Mode - Single Tone
|
|
Thanks Jack. I'll try to repeat you test conditions here to compare results. 73 Paul. |
RE: Headphone Output - SSB Mode - Single Tone
|
|
In reply to this post by Jack Smith-6
Jack, > The short story is 2nd harmonic is at least 80 dB down and the > 3rd harmonic is 76 dB down. Fourth and fifth harmonics are in > the noise and not visible. An artifact at 3.9 KHz is also visible, > although quite far down compared with the desired signal. That is remarkably close to what I measured with my K3 in CW with a Heil Proset as the load. I posted those results on Thursday. I've run them again in USB with a 500 Hz reference tone. DSP is set to 2800, center frequency is 1500 Hz (default). Signal source is an XG-2 (50 uV/-73 dBm) and ACG is Slow. 200 Ohm (Heil Proset) headset 500 Hz ---------------- Harmonic ------------------- Reference 2nd 3rd 4th 5th 6th 7th 8th 9th ============================================================ +4 dB -64 -45 -79 -47 -81 -49 -78 -51 +3 dB -66 -78 nf -83 nf -92 -98 -93 0 dB -68 -79 nf -82 nf -89 -95 -93 -10 dB -76 -75 -80 -82 -84 nf -86 nf -20 dB -72 nf -73 nf -75 nf -76 nf -30 dB -64 nf -62 nf -65 nf -66 nf 0 dB is .7 V RMS (1V Peak) measured for the reference tone across the headphone element - total audio level will be somewhat higher with distortion/noise. "nf" indicates the harmonic product not measurable (in the noise floor). Note: at +4 dB, the audio amplifier is in clipping and there are additional IMD products approximately +/- 200 Hz from the reference and each of the odd harmonics. 73, ... Joe, W4TV > -----Original Message----- > From: [hidden email] > [mailto:[hidden email]] On Behalf Of Jack Smith > Sent: Saturday, September 06, 2008 2:14 PM > To: Paul Fletcher > Cc: [hidden email] > Subject: [Elecraft] Headphone Output - SSB Mode - Single Tone > > > At the request of Paul, M1PAF, I've measured the harmonic distortion > seen at the headphone output in a K3 operating with with 2.7 KHz > bandwidth USB mode with a single tone at 500 Hz offset. This > allows up > to the fifth harmonic to be within the DSP passband. > > The details are at > http://www.cliftonlaboratories.com/elecraft_k3_receive_audio.htm. > > Audio level is set for a comfortable listening level with a Heil > Pro Set, corresponding to about 30 mV RMS. > > The short story is 2nd harmonic is at least 80 dB down and the 3rd > harmonic is 76 dB down. Fourth and fifth harmonics are in the > noise and not visible. An artifact at 3.9 KHz is also visible, > although quite far down compared with the desired signal. > > Jack K8ZOA > www.cliftonlaboratories.com > > > > _______________________________________________ > Elecraft mailing list > Post to: [hidden email] > You must be a subscriber to post to the list. > Subscriber Info (Addr. Change, sub, unsub etc.): > http://mailman.qth.net/mailman/listinfo/elecraft > > Help: http://mailman.qth.net/subscribers.htm > Elecraft web page: http://www.elecraft.com _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
RE: Headphone Output - SSB Mode - Single Tone
|
|
Roughly, What AF settings did you need to get 1V peak i.e. gain hi or lo and what front panel AF gain? When I run the tests here I'll set the output using a scope but just curious. 73 Paul |
RE: Headphone Output - SSB Mode - Single Tone
|
|
> What AF settings did you need to get 1V peak i.e. gain hi or > lo and what front panel AF gain? CONFIG: AF GAIN = HI -30 dBV at 9:00 -20 dBV at 10:00 -10 dBV at 12:00 0 dBV at 2:30 +3 dBV at 3:30 +4 dBV at maximum 1V peak an excessively high headphone level. Even with my noise abused ears, -30 to -20 dBV is a comfortable listening level with the Heil ProSet (200 Ohm transducers) and -10 dBV is almost painful. At 0 dBV (1V peak), I can hear the ProSet across the room. 73, ... Joe, W4TV > -----Original Message----- > From: [hidden email] > [mailto:[hidden email]] On Behalf Of Paul Fletcher > Sent: Sunday, September 07, 2008 1:50 AM > To: [hidden email] > Subject: RE: [Elecraft] Headphone Output - SSB Mode - Single Tone > > > > > > Joe Subich, W4TV-3 wrote: > > > > DSP is set to 2800, center frequency is 1500 Hz (default). Signal > > source is an XG-2 (50 uV/-73 dBm) and ACG is Slow. > > > > Roughly, > > What AF settings did you need to get 1V peak i.e. gain hi or > lo and what > front panel AF gain? When I run the tests here I'll set the > output using a > scope but just curious. > > 73 Paul > > -- > View this message in context: > http://n2.nabble.com/Added-LIN-OUT-measurements-tp834406p950890.html > Sent from the Elecraft mailing list archive at Nabble.com. > > _______________________________________________ > Elecraft mailing list > Post to: [hidden email] > You must be a subscriber to post to the list. > Subscriber Info (Addr. Change, sub, unsub etc.): > http://mailman.qth.net/mailman/listinfo/elecraft > > Help: http://mailman.qth.net/subscribers.htm > Elecraft web page: http://www.elecraft.com _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
Re: Headphone Output - SSB Mode - Single Tone
|
|
-30 dBV is the level I found comfortable listening with my Heil Pro Set
as well. Jack Joe Subich, W4TV wrote: >> What AF settings did you need to get 1V peak i.e. gain hi or >> lo and what front panel AF gain? >> > > CONFIG: AF GAIN = HI > > -30 dBV at 9:00 > -20 dBV at 10:00 > -10 dBV at 12:00 > 0 dBV at 2:30 > +3 dBV at 3:30 > +4 dBV at maximum > > 1V peak an excessively high headphone level. Even with my > noise abused ears, -30 to -20 dBV is a comfortable listening > level with the Heil ProSet (200 Ohm transducers) and -10 dBV > is almost painful. At 0 dBV (1V peak), I can hear the ProSet > across the room. > > 73, > > ... Joe, W4TV > > > > >> -----Original Message----- >> From: [hidden email] >> [mailto:[hidden email]] On Behalf Of Paul Fletcher >> Sent: Sunday, September 07, 2008 1:50 AM >> To: [hidden email] >> Subject: RE: [Elecraft] Headphone Output - SSB Mode - Single Tone >> >> >> >> >> >> Joe Subich, W4TV-3 wrote: >> >>> DSP is set to 2800, center frequency is 1500 Hz (default). Signal >>> source is an XG-2 (50 uV/-73 dBm) and ACG is Slow. >>> >>> >> Roughly, >> >> What AF settings did you need to get 1V peak i.e. gain hi or >> lo and what >> front panel AF gain? When I run the tests here I'll set the >> output using a >> scope but just curious. >> >> 73 Paul >> >> -- >> View this message in context: >> http://n2.nabble.com/Added-LIN-OUT-measurements-tp834406p950890.html >> Sent from the Elecraft mailing list archive at Nabble.com. >> >> _______________________________________________ >> Elecraft mailing list >> Post to: [hidden email] >> You must be a subscriber to post to the list. >> Subscriber Info (Addr. Change, sub, unsub etc.): >> http://mailman.qth.net/mailman/listinfo/elecraft >> >> Help: http://mailman.qth.net/subscribers.htm >> Elecraft web page: http://www.elecraft.com >> > > _______________________________________________ > Elecraft mailing list > Post to: [hidden email] > You must be a subscriber to post to the list. > Subscriber Info (Addr. Change, sub, unsub etc.): > http://mailman.qth.net/mailman/listinfo/elecraft > > Help: http://mailman.qth.net/subscribers.htm > Elecraft web page: http://www.elecraft.com > > Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
«
Return to Elecraft
|
1 view|%1 views
| Free forum by Nabble | Edit this page |