[Elecraft] K3 Harmonic Distortion

[Jack Smith]

Working through the K3's schematic this morning, it seems that both the 
headphone output port and the LIN OUT port are driven by similar 
devices, a Texas Instruments DAC with integrated headphone amplifier. 
(Two DACs, one for LIN OUT and one for headphones and speaker output.)

As near as I can see from tracing from one sheet to another, U29, a 
TLV320ACC23 DAC with integrated headphone amplifier drives the TTC-108 
LIN OUT transformers directly. Headphone and speaker audio, in contrast, 
are derived from a second TLV320 DAC driving a LM4950 amplifier for 
speaker output and an LM4811 for headphone output.

I've ordered a couple of TTC-108 transformers today and will run some 
experiments with them over  the next week or so. I've looked at similar 
transformers developed for modem and telephone line coupling, but not 
the TTC-108 specifically, when trying to find a cheaper output 
transformer to use with the Softrock Lite receivers. There, the emphasis 
was more on high frequency response, up to 100 KHz, but distortion was 
also a major issue. The solution of choice in my opinion for the 
Softrock receivers turns out to be a MILSPEC Triad transformer, SP-70, 
with excellent frequency response and quite low distortion. In single 
lot prices, an SD-70 costs about four times as much as a TTC-108.

Jack



Leigh L. Klotz, Jr WA5ZNU wrote:
> It would be nice to see this problem with the the K3IO board addressed 
> at some point.  It may be, as Jim K9YC has hinted, related to core 
> saturation from low frequency components that could attenuated before 
> reaching the board.  Or it may be that the transformers that fit in 
> the space are as big as they can be, and a third-party outboard audio 
> amplifier and isolation board is necessary, bypassing the transformers 
> in the K3.
> But whatever the solution is, I'm looking forward to it, as using PSK 
> with a waterfall display on the K3 requires me to ride the AGC, LIN 
> OUT, and filters in a way that I didn't on the K2.
>
> Leigh/WA5ZNU
>
>> I've found that the LIN OUT port has typical third harmonic levels 
>> around -45 dB from the fundamental, with the LIN OUT level set at 10, 
>> the factory default. The headphone port, under the same conditions, 
>> will show the third harmonic down about 70 dB or so at levels 
>> appropriate for my Sony high sensitivity headphones. Cranking the 
>> audio level up will cause higher distortion on the headphone port, 
>> but only at levels that are impossible to listen to as they are so 
>> loud as to  be unbearable.
>>
>> I've also seen some non-harmonic artifacts, particularly when very 
>> strong signals are pumped into the K3's antenna port.
>>
>> Jack K8ZOA
>>
>> Julian, G4ILO wrote:
>>> I think it was me who used the term "ghost" signals when referring 
>>> to the
>>> harmonically related CW traces I could see on the CW Skimmer display. I
>>> probably would have been completely unaware of it if I had never 
>>> tried this
>>> program. It was only obvious as I was listening to CW with a narrow 
>>> filter
>>> and the "ghost" traces were appearing in the part of the waterfall 
>>> that you
>>> would be expect to be dark because it was outside of the passband. I 
>>> would
>>> agree with the suggestion that the harmonics are at a low level 
>>> relative to
>>> the main signal because they are masked by band noise when you open the
>>> filter up. I don't think there is much risk of calling someone on 
>>> the third
>>> harmonic by mistake. I don't know what the voltage level of the line 
>>> output
>>> is but the traces disappear if you turn the level down to 2, which
>>> unfortunately is too low for most sound card programs.
>>>
>>> I don't know if those same harmonics are present on the speaker or 
>>> headphone
>>> outputs, as distinct from the computer audio feed. I can only say 
>>> that I
>>> find the K3 audio exceptionally pleasant to listen to. I much prefer 
>>> it to
>>> the K2, or the FT-817, which are the only other radios I have. The only
>>> other radio I have had that sounded as clean and low-distortion as 
>>> the K3
>>> was an IC-756PRO that I had many years ago.
>>>
>>> If the harmonics are down by the kind of amounts that others have 
>>> measured
>>> then the distortion figure is still within the specification for 
>>> what after
>>> all is not meant to be an audiophile quality hi-fi output stage.
>>>
>>> -----
>>> Julian, G4ILO. K2 #392  K3 #222.
>>> http://www.g4ilo.com/ G4ILO's Shack   http://www.ham-directory.com/ Ham
>>> Directory    http://www.g4ilo.com/kcomm.html KComm for Elecraft K2 
>>> and K3   
>>
>
>