[Amps] PA efficiency

[Jim Thomson]

From: David Kirkby <drkirkby at gmail.com>
Subject: Re: [Amps] PA efficiency
To: "Carey Lockhart, KC5GTT" <kc5gtt at gmail.com>
Cc: amps at contesting.com
Message-ID: <BANLkTik7gt5eZmp_R9Pp3DiTQveq8FrCbg at mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

On 18 May 2011 13:20, Carey Lockhart, KC5GTT <kc5gtt at gmail.com> wrote:
> i was affraid of that. 58.3%. my gs31 2 meter amplifier i use for a standby
> isnt much more efficent than my gs-35 that just croaked. i have been using
> these yu1aw designs. have yet to get the rated operating paramters listed. i
> dont even have a clue where to start. i am wondering if there is a more
> proven design better suited to my skill level. any suggestions?
>
> Carey, kc5gtt

It's totally pointless stating the efficiency to 3 significant figures
(58.3%). Whilst if you use a DVM you can read DC voltages and currents
to well under 1%, you wont be able to measure the RF power to anything
like that accuracy. So your efficiency is probably somewhere between
48% and 68%. So worrying about the odd 1 percent is totally pointless
unless you use professional, laboratory quality test equipment to
measure the RF power.

In fact, I doubt your high voltage probe is sufficiently to measure
the voltage to better than 1%.

Dave

##  I'm sure it's  very easy to calculate eff a lot better than the..."48% - 68%"
range you mentioned.  If you were really adamant about it, you could use a 
calibrated fluke or hp dvm... and measure the loaded B+  across  each electrolytic
in any HV supply.  And ditto with measuring plate current.. [ just insert the dvm in series
with existing plate current meter]. 

##  as for the wattmeter  end of it... the bird is spec'd  at  +/- 5%  of full scale.....
anywhere on the scale.   Which implies that unless the meter is nailerd right to the
top end... you won't get  +/- 5 %.     IE:  try using a 5 kw slug  to meausre a 625 w 
signal from an hf amp.   It reads 700 w.   Switch to a 2500 w  slug... and it now reads
675 w.   Switch to 1 kw slug..and it now reads just 625 w.   Which one do you think
is the most accurate ?   A 1 kw slug implies +/- 5%  of 1 kw =  50 watts. 

##  So if the wattmeter with the 1 kw slug is used to measure a 100 w  xcvr, the
meter could read  say 100w.  Now the real power could be 150w..and the meter is
reading 50 w low.   or the real power could be just 50w... and the meter  is reading 50w
high.  That's a huge margin of error..and obviously way more than  +/-5%.   Point is..if
the bird meter isn't in the top  end of it's scale... [ ie: use the smallest slug possible]....
then your readings  will be  out to lunch. 

##  A few yrs back... I ordered up several NEW  2.5 kw HF slugs  from Coaxial Dynamics.
I shoveled them into the CD 83 wattmeter.. hot.. with amp running a cxr..... and resulting
readings  were close... but still  all over the map..with some slugs high..and some low. 
Which one is correct... who knows?     Put a heat gun or hair dryer  near a bird slug..and you
can watch the readings  change 10%  in front of your eyes.  Take a reading when room temp
in the shack is cold..like  60 deg F..then do it again  when shack temp is 85 deg F.... and you
will be in for an eye opener.   Same deal in a mobile set up..in summer time... with car windows 
rolled up..and sitting in the hot sun. 

## I gave up on the bird/cd meter's..and now use the array solution's  'power-master' wattmeter.
It's not affected by heat, and if you  don't like the calibration, you can change it in software....in
+/-  1%  increments..from   a  max of  +15%.....down to a low of -15%.    They calibrate each coupler by 
hand..and provide individual calibration number's  to use  for HF...  one for  fwd..and a 2nd calibration
for rvs.   Then  2 x more just for the 6m   band.    Same deal for their vhf/uhf couplers.   That's as good
as it gets..and good enough for me.   The other problem with the bird/cd  units is severe lack of resolution...
esp at the upper end of each scale. 

##  You could also use  a CALIBRATED scope  across a  dummy load... and measure the rms voltage.  Then
use V squared /  50 ohms.   

##  either way.... you can get something a bit more accurate than  just.. 48-68%.  If I built an amp that I thought
had 68% eff.... then found out it was really only 48%... I'd go nuts.  Not good enough. 

later.. Jim  VE7RF