[Amps] LDMOS availability

[Carl]

----- Original Message ----- 
From: "Jim Thomson" <jim.thom at telus.net>
To: <amps at contesting.com>
Sent: Saturday, June 10, 2017 1:48 AM
Subject: [Amps] LDMOS availability


> Date: Fri, 9 Jun 2017 22:02:53 -0600
> From: Jim Garland <4cx250b at miamioh.edu>
> To: amps group <amps at contesting.com>
> Subject: Re: [Amps] LDMOS availability
>
> <There are two issues, here, one being the definition of Class C, and the
> <other being linearity. I'm inclined to agree with Bill on both. It's not
> hard to imagine modulating the bias, so that on voice peaks, for
> example, the amp moves in the direction of Class B, but regresses to
> Class C at low drive levels.In that case, the class of operation varies
> with the drive. That said, Class C is by definition non-and that occurs
> whenever the device doesn't conduct over a full RF cycle.
>
> I have a homebrew amp with two-tiered bias. On SSB, in pauses in speech,
> the amp is biased at a very low resting current (approaching Class C) to
> cut down on heat, but on voice peaks where the linearity is needed, it
> switches to Class B. One has to be careful playing that trick, however,
> to avoid switching transients.
>
> 73,
>
> Jim W8ZR
>
> ## Another amp config that switches about, is the infamous G2DAF circuit.
> I have a pair of  4-400s in that config.  Done right, they operate just 
> fine,  but the
> component values have to be chosen very very carefully..esp the screen 
> caps.  Nobody
> has ever complained of  imd...and cant tell the difference between the 
> G2DAF amp, and the
> GG amps..when I toggle between them...while listening off freq.   A lot of 
> G2DAF amps are not
> configured correctly, and I dont recommend  the design  for any new 
> construction projects.
> I have seen the G2DAF config used on 4-400, 4-1000,  4CX-1000, 4CX-250B.
> They essentially switch from class C-B-AB..and back again  from AB-B-C.
>
> ##  I looked into the  two-tiered bias scheme back in the 70s. When it 1st 
> came out
> in QST, its function was to cut off idle current  between dots and dashes 
> for qsk cw.
>
> ##  To avoid the dreaded switching transients, esp on ssb,  the 
> three-tiered bias scheme
> was introduced.  Cut off on RX, the  bias V is high...like 40-60 vdc. 
> With PTT activated,
> the  bias V is reduced, like 15-20 vdc, so idle current is real low,  like 
> 20-40 ma.   With RF
> detected, bias V toggles to its normal linear state, like  5-8 vdc,  such 
> that normal idle current occurs.
> While talking on ssb, then the bias is just toggling between  say 15-20 
> vdc...and 5-8 vdc.
> With the two-tiered  scheme, the bias is toggling between say  40 - 60 
> vdc and    5-8 vdc.
> That takes time..and the amp ends  up being non linear while the drive 
> level increases..esp
> at low levels.  If you have any amount of back ground noise, their will be 
> enough RF from  the xcvr,
> even with just a few milliwatts, to toggle the bias,  defeating the 
> purpose of the bi or ...tri-state bias scheme.
> A simple noise gate..or downward expander, would solve that issue. A noise 
> gate is just a high ratio
> downward expander.
>
> ##  Even with the three – tiered scheme, its still not fast enough to 
> toggle between syllables, so normal
> say 5-8 vdc bias stays on for the duration of the word phrases. It will 
> however, toggle between words..
> sometimes.
>
> ##  I played  with it..and finally gave up.   Where it  sorta  works good 
> is if using PTT-Footswitch SSB.
> For VOX  ssb,  it does not buy you anything..... except between the last 
> word spoken....and when the vox drops out.
> Which is of such short duration, with  the typ 300-500 msec  vox delay.
>
> ###   For  CW use, I found that its just as easy to simply increase the 
> bias Voltage a bunch, such that idle current
> is really low, like  15-40 ma.   Then you get the heat reduction between 
> dots + dashes.  Idle power is WAY down, heat
> is WAY down.   Tank eff increases a little bit,  1-3 %,  further reducing 
> dissipation when driven full bore.  On any GG
> amp, you cant cut off the idle current, and run it into class B  or  Class 
> C.....other wise you will get key clix.   It has to
> draw a little bit of idle current, hence the  15-40 ma.   15-40 ma is a 
> helluva lot better  vs  150-220 ma,  heat wise.
> 14-40 ma also works good on FM.
>
> Jim   VE7RF


**  CQ ran an article around 1949-50 using a sliding bias scheme on a 304TL 
which ran gobs of idle current in AB. IIRC it was only partially successfull 
at well less than full bore. Full details escape me and I dont have time to 
dig in to the library.

Carl 


---
This email has been checked for viruses by AVG.
http://www.avg.com