Well, 24 days into building a radio and I’m finally starting on something that works at radio frequency. It’s been about three weeks of lockdown and four weeks of isolation and I have to say, the project remains a great distraction from the news.
True to the “build a stage and test it” philosophy, I’ve tackled the BFO and product detector, and finally achieved success, although perhaps a qualified success.
Not much going on there, right? You’d think so, but as always, a few challenges arose.
First, I have to say, I think building with tubes is a bit easier today, given the relatively smaller size of the components. This makes it easier to keep leads short and still have room to work. It’s hard to imagine you are looking at five hours of work in the picture below, but here it is….
The BFO wiring is the half a 12AX7 at the bottom of the picture. It’s easy to differentiate from the audio amp because it is fed with regulated high voltage rather than unregulated, and I’ve color-coded unregulated HV as red, and regulated HV as white and orange.
This may seem neurotic, but I’ve found I need to do everything I can to keep things straight, and color-coding the wiring is yet another help. Filaments are the traditional green, anything associated with primary AC transformer leads and switching is black, and AVC/bias lines are blue. RF connections are via miniature copper braid coax.
After carefully checking my wiring I confidently fired up the radio, but the BFO laughed at me, hence the BFO ho ho. After rechecking the wiring, this time tracing over the stage wiring as recorded in my lab notebook, I realized I missed the B+ connection to the bottom of the BFO coil. Ya!
After adding that wire, I confidently fired up the radio, and the BFO laughed at me again. At this point, I could draw the circuit in my sleep. I’d tested every component, and was left with two possibilities. Either the Handbook schematic had an error – remember this happened already with the filament wiring – or I failed to solder something properly.
Life has taught me that when I think someone else has done something wrong, the first thing to do is make sure I didn’t do something wrong, and sure enough, life was right again.
In the picture, there is a blue resistor just to the right of the oscilloscope probe. I had cleverly passed the resistor through the solder lug to ground it, and then used the rest of the resistor lead to go back to the tube socket to ground an adjacent pin. This plan only works if you solder the resistor at the solder lug. In my defense, the resistor body obscured the view of the solder connection to the lug, and I could not see that I’d botched the solder until I moved the resistor to look. A few seconds of heat and – third time’s the charm, I got this.
Or more accurately, a little less than half of that. With the regulated B+ on the oscillator, I saw only about 125 millivolts of RF. The 300 millivolts pictures is with full B+ applied.
I did that little experiment because I was concerned that 125 millivolts of BFO was not enough. I don’t know that for sure. I’m used to solid state receivers, with diode rings used for product detectors, and they require a few volts to work right.
But I’ve never used the simple two diode product detector shown in the JMD so the jury is out on that. A later handbook suggests that the important point is to keep the BFO voltage at 10 to 20 times the signal level. As I haven’t wired and tested the AVC yet, I’m not sure what maximum signal level to expect at the detector, so the jury will have to remain out for a while.
I finished up the product detector wiring, applied a signal generator to the product detector, and I easily heard -50 dBm or so properly detected, and the detector didn’t distort noticeably until about 0 dBm. So things may be working.
Next step will be to wire the IF amp and filter stage to see what happens when a signal is fed into the crystal filter. Tune in tomorrow.