Amps
[Top] [All Lists]

Re: [Amps] HV switch-mode power supply

To: Amps reflector <amps@contesting.com>
Subject: Re: [Amps] HV switch-mode power supply
From: John Wilson <amps@grebe.plus.com>
Date: Sat, 23 May 2009 17:05:03 +0100
List-post: <amps@contesting.com">mailto:amps@contesting.com>
All,
Here's my take on this, particularly from a VHF contests point of view. 
We need good HV SMPS because;

1.  Must be lighter, and hopefully smaller and more efficient than 
equivalent linear supplies, - lugging heavy PSUs to contest sites is a 
(back) pain.

2.  It will cope with poor quality power supplies - we currently have to 
hire generators for VHF portable contests that are vastly overrated to 
cope with the huge swing, poor power factor, and distorted current 
waveform from linear PSU's and still give a reasonably constant voltage.

3. In the denser parts of Europe, received signals in a VHF contests on 
a good site can be enormous and so signal quality is very important. The 
supply therefore needs to be regulated .
a)  To cope with smaller poorly regulated generators as above,
b)  To allow the most linear operation of the amplifier to reduce 
splatter.  Tuning a linear amp and throttling it back to give good 
linearity needs care even on a decent mains supply. On a mutliband 
contest site with poor quality power whose voltage is varying with the 
demands of the other stations, it's a nightmare.

4. It needs power factor correction.  Again to allow smaller, cheaper, 
more efficient (hence lower CO2), generators to be used.

5. It needs low ripple and spurious on its o/p.  Any switching ripple on 
the high voltage lines will modulate the signal. 100 / 120 Hz ripple on 
a linear supply is not too serious as it only needs to be say 40 dB 
below peak speech level i.e 1% and its harmonics will get lost in the 
intermod / splatter.  However beyond say 10 kHz from carrier, outside 
the splatter zone, it becomes very serious .  A close by station may 
well be receiving -20 dBm signals.  His noise level in 3 kHz is likely 
to be in the -130 to -140 dBm region depending on band thus the ripple 
needs to be better than about -110 dB down (say <3 mV).  Otherwise local 
fixed stations will call you and tell you they can hear you at say 75 
kHz intervals up the band.  I think this is achievable but it will need 
careful design.  There is also the need to consider induced ripple into 
other nearby equipment, supplies and cables so well screened enclosures 
and good mains filtering are important.

6. It must be very quiet on receive.  Lower band (6,4 and 2m) VHF 
contests and DX are a demanding environment.  The best contacts are 
often at signal levels at or below the noise floor thus radiated and 
induced spurs which may be acceptable on transmit may well not be on 
receive.  The logging computer is even a problem and I've recently had 
to put its plastic brick PSU into a metal box with further filtering.  
Modern Fast switching PSU devices produce spurs up to unbelievably high 
frequencies which can be radiated out of the box  all over, especially 
through fan and meter openings.   The current answer to this, on 12 V 30 
A transceiver SMPS PSUs, of putting a switching frequency tweak control 
is not the answer.  It may be useful on HF but by 50 MHz the oscillator 
harmonics are so wide they are a continuous mush.   Thus it needs to 
shut down on receive, preferably completely.

7. Design
a. Try and design the circuit so that it is easily understandable or at 
least layout the circuit diagram so that its easier to follow, I think 
this point should not be overlooked as it seems very difficult, even for 
those of us with some circuit design experience, to fathom the operation 
of most modern SMPSs.  This will help immensely with testing and fault 
finding. 
b. It needs to use components that are readily obtainable world wide and 
ICs which are widely used and unlikely to go obsolete quickly.
c.  It should be easily adaptable for different B+ voltage and current 
requirements and maybe even provide a screen supply for 4CX users.

So get to it folks, a good proven design will be much appreciated.

73 John G3UUT  (G4ZAP/P group)

_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

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