The 6KZ8 Triode-Pentode (R-4 Crystal Oscillator/Premixer)

The 6KZ8 triode-pentode is just a 6GH8 with a different pinout, so if push comes to shove, you can rewire your R-4's 6KZ8 socket for a 6JN8, which differs from the 6KZ8 only in that its triode's plate and cathode pins are transposed relative to the KZ8. (It's almost funny how manufacturers have fiddled with rebasing the 6GH8 for various purposes. General Electric's 8102, used in Progress Line equipment, is a "12 volt" 6GH8 variant that includes yet another variation on triode pinout, the purpose of doing which was done, I guess, to achieve a [temporary] sole-source advantage as the Prog Line was promoted and marketed.)

Just Say No to 6HS6 Scalpers

Multiple Drake products of the 1960s used the 6HS6 pentode for received-signal and oscillator-system mixing. In the early 2020s, new 6HS6s cost over $20 on Ebay as a result of scalpers charging what the traffic will bear. There is no reason Drake 4-line owners should obsess about 6HS6 replacement; the 6AH6 is essentially the same tube. (Don't get stuck on the slight difference in their specified transconductance [g_m]); tube-to-tube variation within a type, and across both of those types, will include both tubes' published g_m values—and of course in practice g_m depends on exactly the operating conditions set up by associated resistors.) That said, I strongly recommend having a look at the 6DK6/8136, which in Drake applications is equivalent to or better than the 6AH6/6HS6 in every respect aside from requiring less heater power and having swapped G3 and K connections. I've rewired the first mixer in my R-4A to use a 6DK6, and it works just fine. (Resistance must be added across the 6DK6 heater to make the replacement draw 0.45 A. If you sub a 6AH6 or resistance-bridged 6DK6/8136 for a 6HS6, also verify that voltage division across the set's seriesed-heater tubes is within spec for all tubes involved.)

So why did Drake use the 6HS6 instead of the 6AH6 if the 6HS6 being "hotter" (higher transconductance) is not the reason? We may need to look no further than cost; in the 1963 Allied Radio catalog, the net price of the 6AH6 is $2.35 and the 6HS6 is $1.25. (Meanwhile, the at-least-as-good 6DK6 was selling for $1.18...) As a possible technical justification for the choice, the 6HS6 has a controlled heater-warmup-time characteristic ("11 seconds," a standard specification) and the 6AH6 doesn't. This difference can be quite important in seriesed-paralleled heater arrangements such as those used in all R-4-series receivers except the R-4C, as (in a two-seriesed-same-heater-volage-tubes example), the tube that comes to temperature first ends up overheated while its partner remains underheated. (The first time I series-connected the heaters of two 6AG7 pentodes across 12.6 V, after a minute's warmup the heater voltage of one was 8 and the heater voltage of the other was 4.6!) Cross-paralleling seriesed-heater tubes tends to equalize such unequal heating, but using tubes with controlled heater warmup times can help considerably in keeping heater voltages within spec as tubes age and are replaced over the life of a set.

Your S-Meter as Questionable Diagnostics Tool

In Jason Buchanan's Drake tips page at http://n1su.com/Drake/tubes.html we read of the 6KZ8 that "gain/transconductance can vary as much as 1 S unit on your meter out of a sleeve of five tubes." Considering that the R-4's 6KZ8 participates in setting the set's antenna-jack-to-detector gain by supplying heterodyning energy to the 6HS6 first mixer, this is a significant finding that may call for a different tube or realignment (of associated transformers) or both, including consideration of a 6GH8 or 6JN8 as replacement.

S-meter changes on replacing other key tubes in the set may require further investigation and/or alignment. As a result of the general AGC and S-meter-drive subsystem design in Drake receivers, the fact that a given radio's S-meter may read more or less (on, one hopes, a test signal of repeatable strength) after replacing one or more tubes that participate in the set's AGC/S-meter subsystems cannot be assumed to indicate a change in sensitivity (ability to distinguish signals from noise) or even overall antenna-jack-to-detector gain. This is so because the AGC/S-meter circuitry in Drake 2- and 4-line receivers and transceivers use (a) dc amplification in implementing AGC and (b) a dc bridge circuit for S-meter drive. Merely changing the gain of the AGC system's dc amplifier (in the 4-line, a 12AV6; in the 2-line, a 6BF6) can change a set's S-meter deflection for a given signal with no change in the set's sensitivity or antenna-jack-to-detector gain. Likewise, because the plate currents of a set's mixer tube(s) participate in balancing the S-meter bridge, changes in that balance as a result of tube replacement or substitution can change S-meter indication without necessarily changing the set's sensitivity and/or antenna-jack-to-detector gain. Changing a tube or tubes that participate in the AVC amplifier and/or S-meter bridge requires that the bridge be re-zeroed at least, and preferably rebalanced and recalibrated per Drake's instructions—if one cannot break free of the misconception that how high a relative, non-absolute S-meter reads actually matters.