I'm still scratching my head a bit about HOW the resistor prescribed by Edgardo, LU1AR, cured the 250 kHz oscillation problem that was plaguing the JBOT amplifier in my 20 meter DSB rig. Earlier I'd posted an excerpt from a CQ article in which Doug DeMaw talks about swamping and Q killing. Last week I got a very thoughtful e-mail from esteemed homebrewer Todd, VE7BPO. Here is an excerpt :
Thoughts and Considerations
Let’s discuss squashing low frequency oscillations in a QRP transmitter; say at 200 KHz or so. A low value resistor across the coil (12t -- FT37-43) often works well to stop these.
Oscillations come from the transistor: gain versus frequency isn’t linear, nor is impedance at transistor ports. We’ll often add negative feedback and such to stabilize an amplifier towards unconditional status. In my Tx circuit that oscillated, no feedback was applied.
In the case of an inductor wound on a FT37-43 or FT50-43, the Q is already low (say 8- 15 or so). Obviously a resistor in parallel with such a coil is not going to lower Q since Q is already quite low. That R will reduce the inductor impedance and thus may serve to decrease the low frequency gain of the RF amplifier to stop any low frequency oscillations. This might not work so well with a way-high fT transistor where decoupling might be hampered if UHF oscillations are singing.
Doug DeMaw often referred to the parallel resistor as a Q-killer. If we examine the equations describing parallel, or series resonant circuits -- if the Q of a tank is high enough, we can practically ignore the effect of resistance at resonance. Conversely if we add a resistance and make it high enough, we might even obliterate the resonant frequency or “kill the Q”. Engineers have long placed an R into a parallel-tuned circuit to drop Q and stop oscillations — they refer to it as damping. 1 example might be in old TV sets where a variable resistance was added to peaking coils to prevent a tank from ringing at a frequency determined by the coil L and distributed C. This applies to higher Q inductors and not our FT37-43 inductor.
Decoupling
Our teacher, Wes, teaches us in EMRFD that coupling often occurs along the DC power supply lines. Further, he’s taught us to decouple AC by placing high impedance in this path. Often the impedance is a low-pass filter with series element(s) of a high Z and shunt element(s) with a low Z. The filter must present a simple short circuit (or perhaps just a resistance) at low frequency so DC flows to the amplifier.
Final
Oscillations should likely be identified and treated according their frequency. This topic looks advanced and all RLC networks deserve more attention from us.
Todd, VE7BPO --- Feb 27, 2013
Our book: "SolderSmoke -- Global Adventures in Wireless Electronics"
http://soldersmoke.com/book.htm
Our coffee mugs, T-Shirts, bumper stickers:
http://www.cafepress.com/SolderSmoke
Our Book Store:
http://astore.amazon.com/contracross-20
