Steve Nosko wrote:
(snip)
FYI:
On one I built, I wound a few extra turns on the transformer and added some
diodes to provide the driver collector voltage. This was a 5V 30A supply
and I was trying to minimize loss. When TTL was in vogue...

That is a very good idea. You can do something like that very cheaply
by just providing a small extra filter cap that is fed by two extra
diodes, to make a positive supply that doesn't have the full ripple
sag of the big caps.

I would build the thing with a 15 or 16 volt transformer instead of
the 18 volt one specified, unless you have lots of trouble with low
line voltage.

If you do use an 18 volt one, you can lower the peak currents and cool
the transformer, capacitors and output transistors off by putting a
big resistor in series with the transformer primary, such that at full
load you just barely have enough DC to keep the regulator
functioning. You still get the heat, but it is dumped into a
resistor, instead of those other components. It also reduces the
current thump when you turn the thing on and have to both charge the
caps and handle transformer core saturation.

I think you can also improve the load transient response by putting a
10 ohm resistor across the output transistors, base to emitter, and a
100 ohm resistor base to emitter on the driver transistor. I haven't
calculated the closed loop frequency response of this design.

--
John Popelish

Wonderful, I might learn something here.
Here are some more details about the existing parts.


I use an existing center tapped transformer which gives twice 10.7 V. I also
have 4 caps of 167 000mF at 7.5 V (which are currently hooked up for 15V max
voltage) and two fat diodes. The transformer is probably good for 25 to 30A
( it has a 3.5 X 4.5 X 6 inch core).

If the 15 V I get right now at my filter stage is barely enough for decent
regulation I could double that (and also use a bridge rectifier) but then I
would have to regulate that higher voltage down to the 13.8V I will most
often use.

I looked at the design http://www.rason.org/Projects/powsupply/powsupply.htm
that Tracy had suggested and it looks manageable to me (the Astron design
someone suggested have additional windings on their transformers which I do
not have on mine) but I do not have the expertise to change parts in the
design to accommodate the higher voltage and current I could get out of my
transformer. I am not sure if I only have to change the current limiting
resisitor at the 723 as well as the resistors which determine the output
voltage or if more elaborate changes involving the 2N3055's are necessary.

regards Uwe







in article , John Popelish at
wrote on 2/2/04 7:02 PM:

Steve Nosko wrote:
(snip)
FYI:
On one I built, I wound a few extra turns on the transformer and added some
diodes to provide the driver collector voltage. This was a 5V 30A supply
and I was trying to minimize loss. When TTL was in vogue...

That is a very good idea. You can do something like that very cheaply
by just providing a small extra filter cap that is fed by two extra
diodes, to make a positive supply that doesn't have the full ripple
sag of the big caps.

I would build the thing with a 15 or 16 volt transformer instead of
the 18 volt one specified, unless you have lots of trouble with low
line voltage.

If you do use an 18 volt one, you can lower the peak currents and cool
the transformer, capacitors and output transistors off by putting a
big resistor in series with the transformer primary, such that at full
load you just barely have enough DC to keep the regulator
functioning. You still get the heat, but it is dumped into a
resistor, instead of those other components. It also reduces the
current thump when you turn the thing on and have to both charge the
caps and handle transformer core saturation.

I think you can also improve the load transient response by putting a
10 ohm resistor across the output transistors, base to emitter, and a
100 ohm resistor base to emitter on the driver transistor. I haven't
calculated the closed loop frequency response of this design.

Uwe Langmesser wrote in message ...
Wonderful, I might learn something here.
Here are some more details about the existing parts.


I use an existing center tapped transformer which gives twice 10.7 V. I also
have 4 caps of 167 000mF at 7.5 V (which are currently hooked up for 15V max
voltage) and two fat diodes. The transformer is probably good for 25 to 30A
( it has a 3.5 X 4.5 X 6 inch core).

If the 15 V I get right now at my filter stage is barely enough for decent
regulation I could double that (and also use a bridge rectifier) but then I
would have to regulate that higher voltage down to the 13.8V I will most
often use.

I looked at the design http://www.rason.org/Projects/powsupply/powsupply.htm
that Tracy had suggested and it looks manageable to me (the Astron design
someone suggested have additional windings on their transformers which I do
not have on mine) but I do not have the expertise to change parts in the
design to accommodate the higher voltage and current I could get out of my
transformer. I am not sure if I only have to change the current limiting
resisitor at the 723 as well as the resistors which determine the output
voltage or if more elaborate changes involving the 2N3055's are necessary.

This transformer is so close to the desired 13.8 volt output that I
would be tempted to use it unregulated. But you might want ot hang a
big 12 volt, 5 watt zener and a 10 ohm 2 watt resistor (in series)
across it to provide a minimum load. It will have some ripple
voltage, but it should do for most loads designed to operate from an
automotive charging system (which also has considerable variation).

If you want a regulated output, this transformer is not a very good
fit, unless you go with a switching regulator.

"John Popelish" wrote in message
...
Steve Nosko wrote:
(snip)
FYI:
On one I built, I wound a few extra turns on the transformer and added
some
diodes to provide the driver collector voltage. ...

That is a very good idea. You can do something like that very cheaply
by just providing a small extra filter cap that is fed by two extra
diodes, to make a positive supply that doesn't have the full ripple
sag of the big caps.

FYI: It was a 28 V supply when I started. I re-wound the secondary, I think
it was #10 wire, so a few extra turns of smaller wire (added to each end of
the secondary, as I recall) was a treat at the end.

...lower the peak currents and cool
the transformer, capacitors and output transistors off by putting a
big resistor in series with the transformer primary, such that at full
load you just barely have enough DC to keep the regulator
functioning. You still get the heat, but it is dumped into a
resistor, instead of those other components.

This is interesting. I have to think about it. It'll slow the filter
cap charge time and lower and spread out the peak current.
T=RC and all that....

It also reduces the
current thump

I like that.

I think you can also improve the load transient response by putting a
10 ohm resistor across the output transistors, base to emitter, and a
100 ohm resistor base to emitter on the driver transistor. I haven't
calculated the closed loop frequency response of this design.
John Popelish

Yikes! Shades of my final year in control theory. I don't think I
even bothered to look at this for the supply above.

--
Steve N, K,9;d, c. i My email has no u's.