"Bob Henderson" <[EMAIL PROTECTED]> made an utterance to the drakelist gang ---------------------------------------------------------------------- Many thanks to those who have written off reflector expressing interest and offering encouragement.
The primary mystery is why the 6EJ7 1st mixer in this particular R-4C has such an abysmal IMD3 performance. The mixer circuit isn't complex and I've been through it a dozen times looking for faulty components but find none. Without any specific knowledge of what the injection levels to the 6EJ7 ought to be, I've had little choice but to follow my instincts and aim to increase the ratio of premix to wanted signals. Having found the 6BA6 RF amp guilty of producing IMD3 products only 60dB or so down and having found that increasing the screen voltage to 100V improved this to around 80dB, Steve, VK6VZ told me that Tom, W8JI recommends increasing the screen volts to 130V. This I have done and so far as I can determine, IMD3 products at the output of the RF amp are now >90dB down. That has to be good. However, the big problem is the 6EJ7 1st mixer and of course, the increased output from the RF amp, consequent upon the increase in screen volts, exacerbates that. So I've reduced the gain of the RF amp to around 12-15dB through the introduction of a 2k2 resistor and 50nF bypass cap into the cathode feed. This allows me to leave the AGC1 feed to the RF amp unmolested. Next, I set about increasing premix injection. Basically I want to double this from around 2V p.p to around 4V p.p. It wouldn't be any use messing about with the premix circuit itself, as I need the same injection from the T-4XC premix chain. It certainly doesn't make much sense to have to modify the T-4XC premix chain too. I decided use of a 2:1 UNUN was worth a try. I wound 7 trifilar turns on a T37-2 core (I think T37-7 would be better, but I don't have one) and connected them appropriately. I disconnected the premix injection from the 10pF ceramic cap on g1 of the 6EJ7. I then connected the premix to the input side of the 1:2 UNUN and attached the output to the 10pF cap. Maybe the UNUN provides an unwelcome bit of loading but the premix level still increased to a more desirable 3.5V p.p. My fears about the T37-2 core not being good enough for 10m where the premix injection is around 35 MHz appear unfounded. I still get a good 3.5V p.p. there. So where has all of this got me? Well, with two 30mV signals at the input to the RF amp I now get an output of around 380mV, which feeds into the 6EJ7. The premix injection is now 3.5V giving me an injection ratio of >9:1. That feels a lot better to me than the <2:1 I had before I made any changes. So what about IMD3 response? Well that has improved from around 45dB to around 70dB. Now 70dB is somewhat short of impressive but nonetheless, it's a sackload more than 45dB :-)) This 70dB figure holds good for 60 kHz spacing and for 2 kHz spacing. (Sherwood CF600/6 installed). Remaining concerns? Despite the significant change in injection ratio at the input to the 1st mixer, examination of the IMD3 products at its output confirm that this mixer is still the fundamental constraint on overall IMD3 performance. I'm still scratching my head over why this mixer in its standard config should be so bad. Though at least I can now listen usefully to 40m in the early evening while I do so:-) 73 Bob, 5B4AGN ---------------------------------------------------------------------- On Behalf of "Bob Henderson" <[EMAIL PROTECTED]> Submissions: drakelist@www.zerobeat.net Unsubscribe: [EMAIL PROTECTED] - unsubscribe drakelist in body Hopelessly Lost: [EMAIL PROTECTED] - help in body of message Zerobeat Web Page: http://www.zerobeat.net Brought to you courtesy of TLCHost.net http://www.tlchost.net/ ----------------------------------------------------------------------
