[hpsdr] Demeter: PSU's

[Tony Langdon]

At 06:35 AM 6/25/2007, Andreas Troschka wrote:

>Let me spend my 50Cents on this topic too.
> >From the reading of the various posts showing pros and cons of 
> linear and switching power supply technologies I could notice that 
> one important point has been disregarded, EMI.

I disagree, it was heavily discussed a few weeks ago.

>This behaviour causes strong RF harmonics generation. Efficiently 
>shielding of such a circuit is far from being such simple as it may seem.

But it CAN be done, and it can be quite effective.

>I'm not sustaining the idea it is impossible to use some switching 
>circuits in the power supply without getting significant noise 
>increase. It is my conviction it's not worth the effort.

It depends what your needs are.  As a data point, I myself have run a 
switchmode supply for 3 years.  This supply is a 40A unit, running at 
around 14.3V DC (I tweaked it up because I have a standby power 
system on it that had a Shottky diode in the supply path).

As for noise, the supply is extremely quiet.  These particular units 
were obtained cheaply from a local electronics outlet, and a couple 
of local hams put them under the microscope, as they had the exact 
same concerns expressed here.  Both physical examinations and various 
tests were performed.

Physically, at least 5 RFI suppression networks were identified (as 
you say, it takes a lot of effort to do properly).

As for on air testing, everyone (including myself) who has tested 
these supplies has found the spurious outputs to be well below the 
local noise floor on any band.  For the purpose of my tests, HF was 
the primary focus (as that's where I expected the worst 
problems).  The spurs are only detectable (and weakly at that) if no 
antenna is used.  I cartainly was unable to see any difference 
between the linear supply and the switchmode supply on receive.  On 
Tx however, there is a world of difference.  The linear supply gets 
extremely hot under high duty cycle, even down to 5A load (it's 
capable of 20A).  The switchmode supply handles over 30A and stays 
cool as a cucumber.  In fact, the main limitation I've seen is I need 
to use heavier cabling. :-)

I want to get a second switcher for the second shack, so I can crank 
the power up on some other links. :)

As one of my requirements is for high duty cycle operations (comes 
from running VoIP gateways), the switchmode has the greatest overall 
advantage.  Also, because space is limited, the lower weight and lack 
of a 50 Hz transformer (and its associated magnetic field) are also 
of advantage to me.

Horses for courses.  Every choice is a compromise, and my experience 
still leaves me favouring a switchmode supply, even after considering 
the downsides (EMI, etc).  Yes, it takes effort to keep the RFI/EMI 
from getting out of the supply, but there are some talented people 
here, and the effort is worth it.

Probably a bigger issue is the safety aspect of building the supply, 
and from that PoV, it might be better looking at a choice of linear 
supply for the low current components (Atlas and everything connected 
to it except the PA), or a DC-DC converter for portable/mobile/shack 
supply operation.  That also allows us to choose what route we want 
to go.  I would consider a low current linear supply for everything 
except the PA.  For serious power, the economics and weight don't work for me.

73 de VK3JED
http://vkradio.com


 1182719123.0