[Top] [All Lists]

[AMPS] G2DAF Circuit

To: <>
Subject: [AMPS] G2DAF Circuit
From: (measures)
Date: Thu, 26 Oct 2000 12:31:47 -0700
>> Although the load would change with drive, I am not convinced that is as
>> big a problem as it might first appear.  When drive is less, the output is
>> less. Buckshot usually occurs at peaks, as amplifiers flatten out.  What
>"Buckshot" (IMD products) can appear any level transition during 
>the envelope change, but it is most harmful if it occurs during the 
>periods where envelope power has the highest time-averaged power.
>I'd rather have someone operating next to me with occasional peak 
>clipping than non-linearity at about 50% of the envelope peaks or 
>less, because the splatter is more constant! Speech processors 
>actually reduce peak-clipping in the amplifier stages. The reason 
>they have a reputation for increasing splatter is they increase the 
>time the envelope stays at higher power levels. 
>If you run a processor on a system with clean amplifiers, the 
>system actually gets less objectionable because occasional peaks 
>and overshoots are removed.
>If the PA or amplifier stages have non-linearity at medium to high 
>power levels, a processor will make the system sound worse on 
>adjacent channels.
>The G2DAF system would be the last system I'd run with 
>processing of any type.         
>> is more troubling is whether the exciter is properly loaded on both half
>> cycles at peak power.  That is the reason for using a tuned input in the
>> grounded grid circuit. 
>If the G2DAF system is working, it probably is working because 
>the screen circuit ISN'T doing anything at all. 
>Every bit of power applied to the screen that supposedly increases 
>output comes from the exciter. The cathode would have to be 
>"robbed" of a significant amount of drive power to supply power to 
>the screen.
>A simple "triode" connection of the tetrode, with attention to voltage 
>division between screen and control grids, would have less 
>distortion and just as much gain.
>The DAF system is similar to the "forward-acting speech 
>processor" in the ARRL Handbook that samples input signal, 
>rectifies it, and uses it to limit gain. That circuit doesn't work at all 
>like the article claims, has poor envelope characteristics, yet 
>people swear by it.
>A two-tone test will not show up many types of splatter-causing 
>problems. It certainly doesn't show most power supply regulation  
>problems! The screen grid in the DAF is poorly regulated! 
Amen to that.  

>While a two-tone test is one of the least effective ways to tell if a 
>PA is clean, many people use something even less reliable. They 
>depend on on-the-air testimonials from a few people with S7 noise 
One of the better on the air testimonials was given to Roy Neal, K6DUE, 
NBC's science advisor.  //   Roy was driving a 50w max input solid state 
amp. with 100w of SSB.  The amp. was clearly being overdriven.  One of 
Roy's "nice guy" fans reported that he lost Roy's signal completely by 
tuning off frequency  500 cycles. 
>One thing I commonly have heard is an operator getting a 
>complaint often asks another random person (who often has a S-5 
>or stronger background noise) how his "signal sounds". Of course 
>it sounds fine to someone who doesn't know what to look for and 
>who might even have a high background noise, especially when the 
>other person wants to be a "nice guy".

-  Rich..., 805.386.3734,  

FAQ on WWW:     
Administrative requests:

<Prev in Thread] Current Thread [Next in Thread>