wrote:
Hi all and thanks again for all the responses.

Well I had another 10 minutes to spare again today so took a different
secondary tapping to try to get more voltage headroom. Following that
tweak, I now have 25.5VDC available across the large electrolytics.
This is obviously plenty enough for 13.8VDC regulated; but is it now
*too* much to drop efficiently?
Thanks,
p.

It's more than ideal but there is no such thing as "too much" if you've
got an adequate supply of heat sinks and circulating air :-).

Is 25.5VDC measured at your "full load" current or at "no load"? If
that's what it is like under a typical load then you'll get about 50%
efficiency from a linear supply.

A switching step-down converter is possible but if that were an option
you wouldn't be where you are today.

The no-load voltage is irrelevant for power dissipation because with no
current there's no heat.

Tim.

Yes sorry, Tim. I should have mentioned that the 25.5VDC is the
open-circuit voltage across the caps.
I assume this new drop is do-able because this PSU in its original
configuration provided 24VDC regulated from 40VDC unreg across the
caps. That seemed like a hell of a drop without generating heaps of
heat, but I freely admit my recollection of linear PSU design (like a
lot of other electronics-related stuff) is very, very hazy.

wrote:
Yes sorry, Tim. I should have mentioned that the 25.5VDC is the
open-circuit voltage across the caps.
I assume this new drop is do-able because this PSU in its original
configuration provided 24VDC regulated from 40VDC unreg across the
caps. That seemed like a hell of a drop without generating heaps of
heat, but I freely admit my recollection of linear PSU design (like a
lot of other electronics-related stuff) is very, very hazy.

A MUCH more relevant number would be the average DC voltage and minimum
DC voltage across the caps under pretty much full load.

A good dummy load for an unregulated supply is auto bulbs/headlamps in
series/parallel as appropriate.

The average DC voltage under full load is very relevant to overall heat
dissipation.

The minimum DC voltage (that is, put it on a scope and observe the
bottom of the ripple) under full load is very relevant to making sure
you have enough headroom for your regulator.

If you don't have a scope you can guess what the ripple is from
average DC and rectifier type (half or full wave). Guess won't be far
off from reality unless you've got abysmal ESR caps or really really
sucky transformer.

As to headroom: Typically for a high-current supply on a 723 chip,
there would be a "secondary" supply at higher DC voltage that helped
you develop drive to the pass transistor. Without this you need several
extra volts headroom.

So far you haven't given us an awful lot to go on :-).

Tim.

Paul,

If you didn't understand where I was going, Tim's is a good summary of my
original post.

73, Steve, K9DCI



"Tim Shoppa" wrote in message
oups.com...
wrote:
Yes sorry, Tim. I should have mentioned that the 25.5VDC is the
open-circuit voltage across the caps.
I assume this new drop is do-able because this PSU in its original
configuration provided 24VDC regulated from 40VDC unreg across the
caps. That seemed like a hell of a drop without generating heaps of
heat, but I freely admit my recollection of linear PSU design (like a
lot of other electronics-related stuff) is very, very hazy.

A MUCH more relevant number would be the average DC voltage and minimum
DC voltage across the caps under pretty much full load.

A good dummy load for an unregulated supply is auto bulbs/headlamps in
series/parallel as appropriate.

The average DC voltage under full load is very relevant to overall heat
dissipation.

The minimum DC voltage (that is, put it on a scope and observe the
bottom of the ripple) under full load is very relevant to making sure
you have enough headroom for your regulator.

If you don't have a scope you can guess what the ripple is from
average DC and rectifier type (half or full wave). Guess won't be far
off from reality unless you've got abysmal ESR caps or really really
sucky transformer.

As to headroom: Typically for a high-current supply on a 723 chip,
there would be a "secondary" supply at higher DC voltage that helped
you develop drive to the pass transistor. Without this you need several
extra volts headroom.

So far you haven't given us an awful lot to go on :-).

Tim.

Okay Tim, thanks. I'm not short of test equipment; only the savvy to
use much of it. ;-)
Seriously, in this instance I can easily measure the ripple nadir under
full load - if I can source some headlamps from someplace. This may
take a few days so I'll report back in due course.