Posting WB6JDH's 30 meter mods for the 40A. These are more comprehensive than W0CH's mods and will be the ones we pursue on our builds. Stay tuned.
30 Meter Mods Per WB6JDHRegarding the oscillator circuit for U4, Dick was concerned because the drive to the base of Q7, the final transistor, was too low at 2.5 Vpp. At first, he thought it might be due to the crystal activity of X6 hanging off of pin 6. However, fishing through his stack of 8 MHz crystals and finding a substitution with more poop did not alter the drive that much.
(1) Install 8 MHz crystals for X1-X6.
(2) Matched X1-X4.
(3) T1 is wound as follows:
(4) T2 is wound as follows:
(5) T3 is wound as follows:
(6) C6 is 22 pF NPO
(7) C9-C13 are 470 monos or 470 C0G or 470 NP0 if they fit
(8) C14 is 47 pF NP0
(9) C35 is around 151 pF (see notes below)
(10) C38 is 68 pF NP0
(11) C52 and C53 are 1500 Poly
(12) L4 is 10 uH
(13) L5 is 15 uH
(14) L6 is 24T
(15) L9 is 62T
(16) RFC1 = 15 uH
(17) R12 is 18-22 Ohms
(18) AGC Mute Circuit
R4 is 4.7 MOhm
R8 can be jumpered
Sanity Check Test Measurements
(1) U2 measurement:
Pin 6 = 0.80 Vpp
(2) U4 measurements (on transmit):
Pin 4 = 0.15 Vpp
Pin 6 = 0.30 Vpp
(3) Q5 Measurements (on transmit):
Source = 1.0 Vpp
(4) Q7 Measurements (on transmit):
Base 2.5 Vpp
(5) Q8 Measurements:
Source = 3.4 Vpp
R23 on C7 side = 0.8 Vpp
Rxvfo label = 0.8 Vpp
Instead he proposed dropping the value of C35 hanging off of U4 pin 7 with something a little less than the 270 pF original. Dick tried a 150 pF NPO cap and reports that the power output for his 30m version was raised to just a shade under two watts. Recommend you try a temporary insertion until an optimum value is found. But be careful, however, as too low a value will swamp U4's oscillator.
Important Building Caveat
Q6 is the 2N222A driver and it requires a ferrite bead on the base. This is specified in the original NorCal 40A assembly manual as:
FERRITE BEAD, 0.146" O.D., 0.138" long #64 material
The original assembly instructions tell you to "[after installing the ferrite bead on the base] push Q6 down onto the PC board so that its leads are as short as possible". Follow these instructions but be sure that the bead DOES NOT TOUCH the other two leads of Q6 or you will be in a world of hurt.
When I was doing the final tests on this, a mysterious 2.80 VDC voltage showed up on the collector of Q4 (the 8V TX keying circuit) and in the part of the circuit involving U4 (the transmit mixer) and Q5 (the buffer). This voltage turned on transmit and resulted in two tones (signal generator input and transmit side tone) and other general havoc. After trying to isolate Q4 from any other external influences like removing it (Q4), U4, and Q5, C29, and C30, the mysterious voltage still persisted.
Finally, I surmised that it must be Q6 and it was wicked out causing the voltage to disappear. After Q6 was correctly reinstalled with the ferrite bead insulated from the other two leads and all of the removed components were replaced things more or less returned to normal. Seems that some ferrite beads are made of materials that are conductive and caused a voltage to appear on the base of Q6 that leaked back into the transmit mixer circuit and played the hell described above.
..oh! And this note: Joe, W2KJ, sent me a PM and said correctly pointed out that the TO-18 (metal can) 2N2222A has a different pinout than the PN2222A. The former (looking at it with the metal tab on the left) is EBC left-to-right while the TO-92 PN2222A (looking at the flat side) is CBE left to right. But hey, PNP BJTs are kinda the same as MPF 109s or J310s, right?
Bottom line, I replaced Q6 with a PN2222A in lieu of the TO-18 2N2222A and that seems to work fine.