Warning, lengthly, boring and technically redundant post to follow. If you don't have a TS-830S or have a problem with one on the bench this will probably put you to sleep.
Audio input picked up by the microphone comes to the IF unit and is amplified by Q19-21 (attention to the R110-R111-R112/C 102-C103-C104 network) The input adapts to any microphone impedance of from 500 ohms to 50k ohms. This amplified output is converted to a DSB signal by the balanced modulator D29-32 (1N60 x 4) passes through buffer amplifier Q22 (2SK19) a 455 ceramic filter and the output appears as a SSB signal. The signal passes either a buffer amplifier Q23 (2SC1815) or the speech processor consisting of Q24, 26 and 27. Then the first transmit mixer Q28 (3SK73) converts this to an 8.83 MHz SSB signal. From the service manual .
The R110-R111-R112/C 102-C103-C104 network may lend itself to 500-50K ohm but I have yet to find a better solution than the MC-50. The audio processing is done at RF rather than AF frequencies and if used with some restraint ie, very rarely bumping into ALC action, then it works amazingly well.
Lack of crispness in transmit audio. Cause: The RF-bypass capacitor, at the collector of the 1st audio amplifier, is so large that it bypasses the higher audio frequencies along with the RF. {The amount of treble-rolloff in a stock radio is about 4.5db at 2800Hz.} Fix: On the IF Unit, at Q19, change C106 from .015µF to approx. 0.001µF. Q19 can be found directly behind the filter space for the larger of the two, optional CW-filters.~ Rich Measures.
As Rich describes I ran into this with my 830. In general , use the compressor/processor but much past 1:00 is getting close to distortion products.
The part of the Kenwood alignment instructions to deep-six is the first step, where you adjust the processor oscillator to 451.4 KHz. Instead, inject a 0.3mv 1 KHz audio signal into the mic input, and adjust TC4 (the oscillator frequency control) for maximum signal at TP2. You have to get the audio into the filter passband first, where you can see it, and in this rig, and probably many others, 451.4 KHz didn't cut it - There was no signal making it through at all.
Once you have 1 KHz solidly on the scope, go ahead and peak the 455 KHz transformers as the alignment procedure dictates. Then change the audio injection frequency to 300 Hz and adjust TC4 for -6db, or half the voltage. That places the audio signal correctly in the passband of the ceramic filters. You will have to go back & forth a few times between 1 KHz and 300 Hz, until 300 Hz is exactly half the voltage of the 1 Khz signal measured at TP2 - That's the -6db point, as Kenwood intended. Lastly, adjust VR7 so there's no difference in TP2 signal level when you turn the processor on and off.~Rich Measures (again...)
I've only had to deal with this on my personal radio and that because I was a bit of a lysdexic dumbass in the first place. This one bit of information was invaluable for this one incident. "451.4 KHz didn't cut it - There was no signal making it through at all."
Thanks to Rich Measures and the original service manual. Huge Schematics available on request .
kop