Notes on DK92/CV-5172 (1AC6) Conversion from UK

With most work by James Duckworth

First Attempt (single component & works but...)

Observed problems on plugging DK92 staight into G500, H500 and 600's are..

(1) alignment shift due to greater grid capacitance which is correctable on all bands up to 19m, but needs approx 1.5 turns unwound on 16m band osc coil for correct alignment.However, If you don't mind the stations shifted right up to the top of the band, it may be ok with the osc coil core simply removed.

(2)Instability on the SW bands arising from the lower voltage/higher current operating design of the oscillator anode, sorry Plate!..anyway pin 3. (My DK-92 data alleged 30v only,not 60! {60v is the listed maximum, 30v is the recommended operating point - app}). This is most severe on the high gain front end 600's ..i.e. all SW bands, and hardly present at all on the G500. The H500's go unstable first from 6-8mhz on the 4-8mhz band, then the 25m band. the 19m band seems the most stable for some reason.

My first pass to fix this was to do almost exactly what you suggest in the treatise, i.e. the logical step of reducing VP, by making a potential divider with the 3.3k resistor. This starts to clear the instability but kills off the oscillator on 16m. It is possible to cure the bands on a discrete basis.i.e the 6-8mhz instability can be fixed by soldering a 10kresistor accross the osc trimmer on top of the coil tower which is easily accessible. HOWEVER..the final solution which fixes all bands reliably and I think repeatably with no loss of performance following realignment, is to solder a 5.6kohm resistor across windings EF ** of the common oscillator coil secondary (L3 on G/H 500's, L4 on 600's). This provides a critical amount of AC damping and low impedance grid drive on all bands without affecting DC conditions (5.6k shunting 3ohms).

It is however a critical optimum...a couple of K higher and instability starts to return. A couple of K lower and a form of AF instability sets in.

It took 2 full days of non stop experiments! so be happy mister..this is it!. Seriously, If an easier alternative can be found , I want to know, but its only one resistor and the 16m out of band effect is an unavoidable consequence of the DK92 extra capacity.

Note: Since then we have found that the 5.6k resistor works best with the "small plate" versions of the tube (Scoba & Heintz and Kaufmann) while a lower value (3.7k) is better for the "large plate" versions.(app)

First Design Modified G-500 schematic

Second Trial (three components but much more satisfactory)

While the first attempt was a "quick and dirty" mechanism, for the second trial a more indepth analysis of the circuit and the tubes were made. First, the differences in recommended components was noted: for the 1L6 a grid leak of 200k was recommended by General Electric in Essential Charactoristics. Zenith began with a 180k for the G-500 then reduced to 100k for all later models. For the 1AC6 a much lower grid leak is recommended: 27k. In experiments 15k seems to work well.

Second, the oscillator plate voltage of 88v (reduced to 78v in A-600) is clearly in excess of the 60v max rating for the 1AC6/DK92. At the same time reducing to the recommended 30v would bias the oscillator plate below the screens (45v) and this would not be good. A 15k resistor in series from pin F of the oscillator coil to pin 3 of the tube was found to reduce the effective voltage to 55-60v.

Finally a decoupling capacitor was inserted between pin 3 and pin 1. Thusfar this seems to provide stable operation on all bands with both large and small plate tubes.

Second Design Modified G-500 Schematic

Yet to come

The final problem to address is the band shift - while correctable through alignment on all but the 16m band, this highest range is shifted beyond the range of even the substitution of a brass slug (#10-32) for the ferrite core in the oscillator coil. This is probably through the altered time constant of the circuit induced by both the change in tube structure and the oscillator grid circuit. When this occured, Zenith usually modified the coil and unwinding a turn and a half from the 16m oscillator coil has this effect.

However it is noted that when Zenith changed the grid leak resistor from 180k to 100k for the H-500, the coupling capacitor was also changed from 75mmf to 250mmf. This larger value has shown less alignment drift than the smaller unit. However when the value was increased to about 350 mmf, sensitivity decreased so for this conversion on a G-500 only, the capacitor (C-12 on drawing) must be changed to 250mmf, H-5 and all -600 already have this value (though should be checked).