[Elecraft] K3 Harmonic Distortion

[Joe Subich, W4TV]

VP8NO writes: 

> Perhaps someone suitably qualified and tooled up might like to try and
> [repeat] and comment upon the view presented by Rob Sherwood at the 
> Dayton convention.  According to him the K3 audio amp stage has 
> serious shortcomings.

I had to listen to the audio of Sherwood's talk before looking 
in the right area.  First, Rob doesn't say the K3 audio has 
serious shortcomings, he complains about distortion on the 
speaker.  Sherwood also admits that using a high impedance 
speaker (e.g., amplified computer speakers) or high impedance 
headphones resolves the issue!

My measurements did not show that distortion but I was measuring 
at the headphone output with a high impedance load - the lowest 
impedance headphones I use are 32 Ohms.  As Jack Smith points 
out, the headphones and the speakers have separate audio 
amplifiers but share a common DAC.

Since Sherwood's comments were based on the speaker output and 
a LOW impedance load, I dragged out an old Optimus XTX25 speaker 
- an unpowered 8 Ohm speaker I've used with other radios for many 
years.  With the 8 Ohm speaker is was easy to generate harmonic 
distortion - particularly if the K3 were set to SPKRS = 2 with 
only one speaker connected - simply by turning up the volume 
until the amplifier was forced into clipping/saturation!  

With the 8 Ohm speaker, 1 to 2V peak represents a comfortable 
to loud listening level for me.  At those levels, the harmonics 
are all down more than 65 dB and any other distortion is very 
acceptable.  

The following table shows harmonic levels relative to a 500 Hz 
reference tone at the given voltage across the speaker.  0 dB 
is 1V Peak (.707V 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).  

Again, as in the case of the Line Out, if the audio gain,  
AGC and RF gain are operated to keep the speaker amplifier 
out of the clipping range - or away from the compression 
region just before clipping - the distortion products are 
entirely acceptable.  For those with noisy shacks, impaired 
hearing or using the internal speaker it may be difficult 
to find a comfortable listening level without entering the 
compression region.  

For those who wish to operate without AGC and with maximum 
RF gain, I would suggest setting the AF gain to a moderate 
level and using powered (computer) speakers or a 100 watt 
per channel stereo amplifier capable of outputs greater 
than the 1 watt per channel at .1% THD+N specified for the 
LM4950 with an 8 Ohm load!  Note, simply going from 8 to 
16 Ohm speakers will improve THD by 8 dB. 

73, 

   ... Joe, W4TV