View Single Post
Old 10th Jun 2008, 10:49 am   #7
paolo67
Retired Dormant Member
 
Join Date: Jun 2008
Location: Milan, Italy
Posts: 15
Default Re: Racal RA17L BFO behaviour

Thank you Al.

Here how I progressed: I replaced the L! I dug in the junk box (there's probably more value there than anywhere else...) and I found a coil with approximately the same inductance as the original (about 18mH), same size and a tap! How lucky I am I thought.

The tap of the original is, based on the resistance, at 90% of the coil while the one of the replacement is at 50%. Also, the original coil is completely impregnated of some white insulation material (while the replacement is not). Another difference is the core: the original has a fixed core, the replacement is on air, although I could insert a core but it would significantly increase its inductance (8x).

The harmonics are gone (40 dB below the 100kHz signal) and waveforms look quite good at any point, especially at G1 it is nearly a perfect sinusoid.

Here the voltages I measured (the anodic voltage is 227V, the valve is the original CV4014):

Anode 222 Vdc
Screen grid 222 Vdc
Cathode 0 Vdc 500 Vac (pp)
Control grid ? Vdc (unmeasurable with my equipment) 600 Vac (pp)

So the A-K current is then (227-222)/5.7k = 29uA. According to your figures, it should be (210-180)/4.7k = 6.4mA.

The output after C215 is approx. 2.7 Vpp when the BFO is in circuit. I tried several values for C215 with no significant change. This is a direct effect of the small A-K current I think.

Most important, the BFO works quite poorly: although I can tweak the trimmer to have 100kHz, it is impossible to decode any SSB signal, just no way.

I've been surprised to find such a high voltage in the tank. It would explain why the original coil is so accurately insulated - to prevent arcing.

Now, my guess is that the position of the tap is making the difference here. However, it is an oscillator, probably one of the most critical kind of circuits, and it is far beyond my knowledge to re-calculate proper values for the components.

A poor man idea would be to try a lower value for R125 (or R216?) to get less negative bias at G1. However, I'm worried of the already very high voltages there...

Paolo
paolo67 is offline