TF3MA sidetone mod

I have not been satisfied with the sound of the current side-tone in my K2. I have done the side-tone mod, using a 47mH choke across Rp5 and 47k in series with C24.
Looking at the modified circuit, it is a High-Pass filter with some peaking in the 600Hz region and the choke does not attenuate the higher order harmonics of the side-tone. The higher order harmonics n>5 are non musical and make the tone sound harsh. The higher order harmonics fall in the 2 kHz and above region where the ear is most sensitive, so you have to get rid of those higher harmonics, if you want a soft sidetone. I find it easier to zero-beat a station if the sidetone is a pure sinus, if the 2nd and 3rd harmonics are -30dB down rel to the fundamental, I find it difficult.
Why not make a Low-Pass filter using the 47mH choke (mine has an internal resistance of Rs=180 Ohm)?
I have the 47kOhm in series with C24 as the original mod suggests. It is there to smooth the transision of the square wave, so leave it in.
I moved the choke and put it in series with C33 the 2.2uF cap and put an unpolarized 2.2uF cap across Rp5 pins 7 and 10, making the circuit into a genuine LP filter. Rs, the internal resistance of the choke controls the Q of the filter, higher Rs results in lower Q. For lower Q, resistance (100 to 200 Ohms) can be inserted in series with the choke. I do not recommend using chokes with higher Rs than 200 Ohms.
This made a huge improvement of the side-tone sound for me, now it is soft and sweet :-).

Installing the mod, my way

To make the mod I had to cut the trace from C33 leading to U10, I cut it next to C33 and inserted the choke across the gap, soldered it to the C33 pad and the pad for C35, this is done on the underside of the Control Board. Then solder a small 2.2uF cap (unpolarized is better than electrolytic) between pins 7 and 10 on Rp5.

No trace cutting needed, insert the choke in series with C33

Jerry, WA2DKG reported that he did not cut the trace but inserted the choke in series with C33, just remove C33 and insert the choke in one hole and a new C33 (it needs longer leads!) in the other hole observing the polarity of the cap and join the two, he laid C33 flat against the board. Tnx Jerry.

Measurements of the new sidetone harmonics


Here are measurements of the side-tone harmonics before and after. Insertion loss of my filter is on the order of 5dB. Higher Q coils make the insertion loss lower. 
Side-tone at F0=600Hz, reference level 0dB. Before is the sidetone with
K2 sidetone mod, after is the sidetone with the TF3MA mod.

                Before          After

600Hz           F0=0dB          F0=0db
1200Hz          F1=-16dB        F1=-42dB
1800Hz          F2=-27dB        F2=-42dB
2400Hz          F3=-54dB        F3=-70dB
3000Hz          F4=-47dB        F4=-70dB

Here is the spectrum of a 500Hz sidetone on my K2 #2024. Sorry I did not take a graph of the "before" spectrum.

Measurements done with the baudline program running on Linux and plotted using Gnuplot_
If you want a more harsh note, then use a 1uF or less for the cap across Rp5. This will make the insertion loss lower and also make the cutoff freq of the filter higher.

Frequency response of the new filter, from Spice3

The frequency response of the filter was calculated using spice3 program running on Linux. The red curve is my current filter. I have plotted the freq response for several combinations of L and C.

How does it sound?

If you can play ogg vorbis file here is the live sound of the 500Hz sidetone on my K2 on 20m band. TF3MA sound (sid.ogg 75kB)
73 TF3MA
----
Last modified: Mán Jan 6 10:12:35 GMT 2003