Adjustment of the modified phase inverter bias correction circuits for V12 and V13 (Dave’s circuits - presented
here) requires matching two resistances for each tube. For example, for V13 the combined series resistance of R109+R112 (which is measured
with the receiver OFF between the (+) side of 91C and pin 1 of V13) is adjusted to match the resistance of the nominally 47K resistor R116B (which is measured
with the receiver OFF between the (+) side of the 22 uF cap and pin 3 of V13). The wipers of the adjustment pots R108 and R109 connect directly to 91C(+). Either wiper is a more accessible place for the connection than the cap itself. For more info, see the sticky Larry provided on
PI-circuit-mods for several models.
I never built or used a
Wheatstone bridge (outside using a VTVM). I remember reading it’s a good way to compare resistances, so for fun I set out to use a bridge circuit to match the resistances for the phase inverter adjustment. On the left below is a drawing of Dave’s P.I.
circuit mod (Copyright D Gillespie Designs / DGD). The center figure illustrates the connections for the Wheatstone bridge adjustment method.
It's convenient to have stackable banana plug leads for the DMM and the little rig shown in the upper photo. It has two banana plugs linked by two matched resistors in series. I used 100K 0.1% resistors; 47Ks would give about 10% better sensitivity. Everything was hooked up as shown in the drawing and photos (
receiver OFF), and the adjustment requires simply using the pot to null the DMM voltage. That’s it.
With stock pots, getting the response down to 0.5 mV is easy; hitting 0V exactly is difficult. Potentials of 100-200 uV can be reached with a little patience. Sensitivity was measured by adjusting to null (within 400 uV) and then recording the voltage increase when 4.7M, 25M, and 100M resistors are added parallel to 33K resistor R112. The experiment shows there is high sensitivity: A 230-ohm mismatch out of 47K gives an 8.5 mV meter reading. A difference of 45 ohms (0.1%) is easily detected – it gives a 1.9 mV potential on the DMM. A 10-ohm resistance difference gave 0.5 mV on the DMM. Those numbers are very close to what
theory predicts. The precision of the resistance measurements at 47K prevents making a good comparison of the two adjustment methods. They are the same within the precision limits.
Until told otherwise I’ll doubt a 50-ohm (0.1%) error in adjustment would cause distortion I could hear. Could it even be measured? This is not presented to be a better way to do the adjustment* – for me it was just something new to try, and simply fun. I hoped you liked reading about it and that you’ll tell me if I am mistaken about anything here. -d
*I'm not convinced about that.
*PRO: The method reduces variables because it eliminates having to change connections in the course of the adjustment process.
*CON: The rig and battery are added gear. There’s a learning curve.