1
\$\begingroup\$

In the old days with vacuum tubes for power amplifiers we carefully tuned the output stage, dipped the plate current to get maximum efficiency.

Looking at todays transistor PA circuits I don't see a lot of effort to tune the final stage. There is an output network and some ratio of turns used to help match, but no effort is made to carefully "tune/dip" it. So what is different about solid state devices that careful tuning is not required?

\$\endgroup\$
2
  • \$\begingroup\$ What kind of power output levels are you talking about from the PA? \$\endgroup\$
    – SteveSh
    Jan 28 at 2:30
  • \$\begingroup\$ Note that RF transistors are driven with a VSWR circuit protection very efficient. Note also that voltage power supplies are "big" for tubes ( some 250 to 1000 V) and "low" for transistors. \$\endgroup\$
    – Antonio51
    Jan 28 at 18:09

1 Answer 1

Reset to default
0
\$\begingroup\$
  • Cost of variable capacitors (two are required for tube PA in the common PI-network) is quite high.
  • Since collector impedance of transistor PA is very low compared with plate impedance of tubes, very large-value variable capacitors would be needed for transistor PA collector filter. Rotating plate contacts would have to handle high current.
  • Some folks like to twiddle knobs, but the more-modern sentiment seems to favour no-tune filter networks. These kind of filters have multiple stages to achieve harmonic suppression - far more difficult to twiddle than the 2-capacitor PI network having Q of ten or twenty.
  • Tubes could handle the mis-match while the user twidded. Transistors are more fussy.
  • To aid tuning the high-Q PI filter, a current indicator is needed (to see how the two capacitors peak and dip plate current). This is another expensive component.
\$\endgroup\$
4
  • \$\begingroup\$ That all makes sense. However, I had a different aspect to the question. Let me clarify. I am tinkering with building a small PA amplifier with BJTs. I'm now at the stage of designing the components for the final stage. So....what I'm wondering is there an equivalent to "dipping" when monitoring the collector current while connected to a good 50 ohm load. or how do I tweek those components to optimize the output. \$\endgroup\$
    – user3799318
    Jan 29 at 2:06
  • \$\begingroup\$ for reference we are talking about a PA between 1-10 watts in the 20mhz range \$\endgroup\$
    – user3799318
    Jan 29 at 2:10
  • \$\begingroup\$ By asking "how do I optimize the output", I'd look at two key measurements: (1) harmonic rejection - determine how attenuated 2nd, 3rd harmonic compared with 20 MHz fundamental...required for legal use on-air. (2) PA efficiency - compare DC power input with RF power output. Anything less than 50% is unacceptable. You might be able to approach 90% if you measure only DC average current to PA collector or drain, and don't include DC current to input bias or driver stage(s). \$\endgroup\$
    – glen_geek
    Jan 29 at 16:22
  • \$\begingroup\$ yep, that's in the plan. Its just not easy to tweak things as the turns ratios are controlled with wire on torroids so you have to pull out the torroids and change the wires. No simple variable cap to twist to attempt to "dip". maybe its not as hard as I think. \$\endgroup\$
    – user3799318
    Jan 29 at 21:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.