Replacing PIN Diodes on the TS-940
The often mentioned but never really published modification to install Pin Diodes in the TS-940 is a blanket swap out, replaceing switching diodes for both the Band Pass Filter and the IF Filters, a total of 44 diodes, 20 of which are on the IF Unit. Basically, the guy who developed the modification followed the convention that if switching diodes are the source of IMD or elevated noise floor in the receive path, they must also be relevant in the transmit path as well. Changing the filter switching diodes in the transmit IF path on the TS-940 is a bit dubious, especially as the bias properties of these diodes were not evaluated to determine if the replacement diodes would perform optimally. Apparently, on-line comments about this upgrade have all been positive, but also include some mixed praise for transmit purity when using a transverter. That is a bit of misnomer, the true cure for Phase Noise, which was highly evident in FM mode when using a transverter, was actually accomplished by modifications to the PLL Unit. Since the TS-930 does not include the FM mode, PLL Phase Noise, if it were present, is really not relevant. Although there might be some possible PLL Phase Noise creeping into the transmit signal for FSK, the modulation frequencies are so low in the audio spectrum that I cannot imagine how anyone would measure it, let alone hear it.
While it is true that both Kenwood radios were designed using conventional wisdom engineering practices driven by cost containment, hence the use of diode switches. Very efficient, small footprint, low cost. The result was selection of devices that were readily available but less than optimal.
The behavior of switching diodes at RF frequencies were not as well understood by the commodity Armature Radio engineers, but were of major concern for high end test instrument designers at companies like Tektronix, and Hewlett Packard, who researched them thoroughly, and in the case of HP, went so far as to manufacture their own Pin diodes with very specific performance criteria specifically for the reasons of limiting self-generated noise at extreme low level signal amplitudes, highly desirable in test equipment far more sensitive than radios.
The switching diodes in the Band Pass Filter on both the TS-930 and TS-940 are directly in the path of a tide of very low level wide band RF signals, as low as .012 uV, injected from the antenna. The more efficient the antenna, the more there are. The switching diodes job is to direct to RF into the appropriate bandwidth limiting filter stage for the desired frequency band. Normally, most of this broadband “noise” is isolated by the BPF for all but the selected band, and the only interference of note will be in band, most often caused by strong adjacent signals. However, the barrage of wide band RF continues to hammer the diodes, and if they are not selective enough or biased properly, their junctions will absorb the out-of-band RF and oscillate at its frequency and at harmonic offsets, classic IMD. Still, the diode junction noise is the source of the overall received noise noted when there is no antenna attached, and limiting it with better diodes can result in lowering that noise floor by as much as 3 dB.
The transmitter is a whole other story. Diode switches are utilized in the transmit path specifically to select the desired IF filters for operating mode, CW, SSB, AM, etc.;. The RF signals that pass through them are typically scaled in millivolts, much stronger than any received signal at the antenna port. IMD products in the transmit path are the result of local circuit design, and self-oscillation of the diodes in the path is of minimal concern, especially since the purity of the transmit signal is relatively clean, and gets much cleaner once it has been passed through the filters. . What should be noted is where a diode is inserted in the RF path, its behavior can and will have an effect on the purity of RF that passes through it. However, that depends greatly on the bias state of the diode and its relative resistance/capacitance while passing the RF signal.
IF Filter switching is significantly different in the 930 and 940. On your TS-930, the diodes switches active in the transmit IF path are selecting the filters, and taking note of the INRAD upgrades, I did not feel it would be appropriate to mess with that, especially since the transmit signal path is shown to be extremely clean at the drive level used by the PA.
Sorry for the tardy reply.
The primary reason for replacing the diodes in the BPF section of the TS-930 is to eliminate a major source of receiver noise. There are 20 diodes in the input BPF stage of the early model TS-930, 22 in the late model. For some unfathomable reason, Kenwood used 1S2588 diodes for this application, which are actually Varicap diodes, which change their series resistance/capacitance with the amplitude of the signal that passes through them. This generates a lot of PN junction noise. Replacing them with BA-479G's, which are a true PIN diode, will lower the receiver noise floor by at least 10 dB.
Unlike the TS-940, the TS-930 does not use diode switching in the transmit path.
Even in the TS-940, PN Junction Noise is not really an issue in the transmit path anyway, the effect is so minor it would not be noticed in the output signal, so there is no real justification to arbitrarily replacing the original diodes used there.