Efficiency of the TS-590S(G) Speech Processor
Ever since the TS-590S and TS-590SG has been sold, people are complaining about the speech processor. The “punch” is missing and the “talk power” is weak, they say. I must that admit, that you will never get an ICOM-like brute force modulation out of the TS-590S, but the speech processor is exactly doing what it was made for.
The human voice has a peak to average power ratio (PAR) of about 15 dB, which means the average power is 15 dB below the peaks. Please read that again: The average power is 15 dB below the peaks. Using a transmitter with ALC, the PAR can be reduced to something about 10 dB, which means the average power is 10 dB below the peaks. This would show about 10 watts on a decent power meter in average mode. Sounds not too much, but that’s how the human voice modulates. For details please read this bulletin of the Collins Radio Company from 1960.
With some additional speech processing, this 10 dB PAR can be reduced even more to something about 6-8 dB. This shows about 20 watts on the power meter in average mode. This PAR is known to be adaquate, even for DX pileups or in contests, independent of the output power. If you want to be louder, get a bigger antenna or buy an amplifier, but please do not reduce the PAR even more.
Really strong compressed audio shows you even more, maybe 30-40 watts. But that sounds really really awful, at least to me. It also reduces the intelligibility of the spoken words. Please check this RadCom article Improving the intelligibility of SSB transmissions of Martin G8JNJ, published February 2009, for more information.
I did some measurements on my TS-590S, using a power attenuator and an oscilloscope. I made screenshots of the spoken words “Alpha”, “Monday” and “Kenwood” with and without speech processor. You can clearly see the efficiency of the processing. The peaks are always 100 watts, as you can see from the CW waveform at the bottom. This has been measured and verified with a calibrated Rohde & Schwarz power meter.
The word “Alpha” with speech processor OFF
The word “Alpha” with speech processor ON
The word “Monday” with speech processor OFF
The word “Monday” with speech processor ON
The word “Kenwood” with speech processor OFF
The word “Kenwood” with speech processor ON>
CW waveform (100 W, 1 ms rise time)