Power Supply Design

Looking for some guidance or ideas. I have all these great components
sitting around, and they're just begging to be put together.
I have a rewound MOT which right now has 60 turns. I have a capacitor
bank rated for 90VDC (I know I'm pushing it with the 60 turn
transformer, I will probably pull out some windings for a better soa)
I have a heatsink for 12 TO-3 transistors, right now they are
2n3055's. What I am driving for is an adjustable supply over the full
range, possibly current limiting but not necessary. Are there any op
amps that have high supply voltage, say 50V? Are there other high
voltage regulators that I can drive the transistor bank with? Is
there a better, cheap transistor than the 3055? Will 3055's handle an
80VDC load?

I'm just looking for some ideas. The design is really pretty much
open to change. I just happened to pick up some surplus stuff, and
always wanted a large DC power supply. I'm shooting for around
0-50VDC, and would love as much as 20A, but I realize that this will
probably be more around 8A at full voltage.

Anyway, has anyone had good success building large DC supplies, away
from using a 317 and a big parrallel bank? Any ideas would be
appreciated.

Thanks for reading,
Steve Kamego

 

On 4 Apr 2005 21:11:08 -0700, bigcat@meeow.co.uk wrote:

steve wrote:

Looking for some guidance or ideas. I have all these great
components
sitting around, and they're just begging to be put together.
I have a rewound MOT which right now has 60 turns.

You havent told us the one piece info that needs 2b known: how have you
rewound the MOT? '60 turns' isnt even a transformer, its an inductor.

When rewinding you will need to put many mroe turns on the primary,
probably 4x as many if you want it to run continuous.


Anyway, has anyone had good success building large DC supplies, away
from using a 317 and a big parrallel bank? Any ideas would

Thats how they used to be done, parallel linear and discrete. 50v x 20A
= 1kW, you'll only get that from a MOT for max 15 mins from cold with
fan cooling. Use 2-4 MOTs if you want it continuous. Teh advantage with
multiple MOTs is you dont need to rewind the primaries, just run them
in series.


Depending on your app, you might find tap changing a whole lot easier
than a tr regulator. Easy to implement when winding your own secondary.


NT

Sorry about not being specific. I re-wound the secondary coil.
Primary is still intact. I haden't thought of using multiple MOT's,
I'll take that into consideration. By tap changing, do you just mean
changing the tap that the input is on to change the output voltage?
And, what do you mean by tr regulator?

Thanks for the reply, I appreciate it!
Steve

 

[Fritz Schlunder]

<steve> wrote in message news:2rf0511ma668t4n3hokcibtf71etj96gkb@4ax.com...

Looking for some guidance or ideas. I have all these great components
sitting around, and they're just begging to be put together.
I have a rewound MOT which right now has 60 turns. I have a capacitor
bank rated for 90VDC (I know I'm pushing it with the 60 turn
transformer, I will probably pull out some windings for a better soa)
I have a heatsink for 12 TO-3 transistors, right now they are
2n3055's. What I am driving for is an adjustable supply over the full
range, possibly current limiting but not necessary. Are there any op
amps that have high supply voltage, say 50V? Are there other high
voltage regulators that I can drive the transistor bank with? Is
there a better, cheap transistor than the 3055? Will 3055's handle an
80VDC load?

The good news is there are high voltage op amps rated for 50V or more that
would make excellent building blocks for this type of supply. One
possibility might be the LM675.

http://www.national.com/ds/LM/LM675.pdf

They are available from Digikey although they aren't exactly free. The
LM675 has an internal overtemperature protection feature which might be very
handy in helping protect your main transistors from excessive temperature if
you mount them all on the same heatsink. Be sure to realize the case of the
2N3055 device is connected to the collector (so presumably to the +Vin
supply), while the case of the LM675 is connected to -Vee (or in your case
presumably ground), so someone needs to be electrically isolated from the
heatsink.

The bad news is 2N3055 devices are crappy and are only rated for 60V
maximum.

http://www.onsemi.com/pub/Collateral/2N3055-D.PDF

Technically with a 90V input you could handle an 80V load since only 10V
would be across the transistors, but in that case you could not output 0V
since that would apply 90V to the devices. They would naturally be toasted
in short order. For a practical powersupply you will have to take off
transformer turns until you limit yourself to under 60V peak on the input of
your capacitors.

You will want to use emitter resistors to insure good current balancing with
the devices.

I would strongly strongly strongly suggest integrating an adjustable current
limit into the supply. Current limiting is so valuable, both as a feature
and for protecting the supply itself from self destructing. I can't stress
how great current limiting is. Definitely use it. It isn't too hard to
implement, and anything of this power level needs it.

I'm just looking for some ideas. The design is really pretty much
open to change. I just happened to pick up some surplus stuff, and
always wanted a large DC power supply. I'm shooting for around
0-50VDC, and would love as much as 20A, but I realize that this will
probably be more around 8A at full voltage.

Indeed 20A might be a bit greedy unless you are willing to go to great
lengths to achieve that much. Although the 2N3055 is specified for an
absolute maximum power dissipation of 115W each, that number is quite
fanciful since it requires the case temeperature to remain at 25 deg. C.
Something of around or less than 50W each is more realistic assuming a very
good heatsink with robust forced air cooling. Assuming 12 devices total,
this suggest a power dissipation of 600W or less. Assuming the output is at
0V with input at 50V, then maybe a current of up to 12A might not be too
unrealistically attainable.

Something else to keep in mind assuming you live in North America or other
low voltage country... Standard wall outlets are rated for no more than
15A, but standard derating procedure for safety and reliability reasons
suggest not to draw more than say 12A from the outlet. Capacitive input
filters fed through a rectifier from a sinusoidal voltage source have bad
power factor. Something around 0.6 might be typical, though it could be
lower for higher capacitance with small series resistance. So with an
output power of 600W (or just waste of that much) you could be using up a
substantial amount of the outlet's capability. I've tripped my 15A circuit
breaker before with a 500W load at low power factor (+maybe 350W of other
low power factor equipment on the circuit including my computer).

Anyway, has anyone had good success building large DC supplies, away
from using a 317 and a big parrallel bank? Any ideas would be
appreciated.

Thanks for reading,
Steve Kamego

 

On Fri, 8 Apr 2005 00:08:52 -0700, "Fritz Schlunder" <me@privacy.net>
wrote:

steve> wrote in message news:2rf0511ma668t4n3hokcibtf71etj96gkb@4ax.com...
Looking for some guidance or ideas. I have all these great components
sitting around, and they're just begging to be put together.
I have a rewound MOT which right now has 60 turns. I have a capacitor
bank rated for 90VDC (I know I'm pushing it with the 60 turn
transformer, I will probably pull out some windings for a better soa)
I have a heatsink for 12 TO-3 transistors, right now they are
2n3055's. What I am driving for is an adjustable supply over the full
range, possibly current limiting but not necessary. Are there any op
amps that have high supply voltage, say 50V? Are there other high
voltage regulators that I can drive the transistor bank with? Is
there a better, cheap transistor than the 3055? Will 3055's handle an
80VDC load?


The good news is there are high voltage op amps rated for 50V or more that
would make excellent building blocks for this type of supply. One
possibility might be the LM675.

http://www.national.com/ds/LM/LM675.pdf

They are available from Digikey although they aren't exactly free. The
LM675 has an internal overtemperature protection feature which might be very
handy in helping protect your main transistors from excessive temperature if
you mount them all on the same heatsink. Be sure to realize the case of the
2N3055 device is connected to the collector (so presumably to the +Vin
supply), while the case of the LM675 is connected to -Vee (or in your case
presumably ground), so someone needs to be electrically isolated from the
heatsink.

The bad news is 2N3055 devices are crappy and are only rated for 60V
maximum.

http://www.onsemi.com/pub/Collateral/2N3055-D.PDF

Technically with a 90V input you could handle an 80V load since only 10V
would be across the transistors, but in that case you could not output 0V
since that would apply 90V to the devices. They would naturally be toasted
in short order. For a practical powersupply you will have to take off
transformer turns until you limit yourself to under 60V peak on the input of
your capacitors.

You will want to use emitter resistors to insure good current balancing with
the devices.

I would strongly strongly strongly suggest integrating an adjustable current
limit into the supply. Current limiting is so valuable, both as a feature
and for protecting the supply itself from self destructing. I can't stress
how great current limiting is. Definitely use it. It isn't too hard to
implement, and anything of this power level needs it.


I'm just looking for some ideas. The design is really pretty much
open to change. I just happened to pick up some surplus stuff, and
always wanted a large DC power supply. I'm shooting for around
0-50VDC, and would love as much as 20A, but I realize that this will
probably be more around 8A at full voltage.


Indeed 20A might be a bit greedy unless you are willing to go to great
lengths to achieve that much. Although the 2N3055 is specified for an
absolute maximum power dissipation of 115W each, that number is quite
fanciful since it requires the case temeperature to remain at 25 deg. C.
Something of around or less than 50W each is more realistic assuming a very
good heatsink with robust forced air cooling. Assuming 12 devices total,
this suggest a power dissipation of 600W or less. Assuming the output is at
0V with input at 50V, then maybe a current of up to 12A might not be too
unrealistically attainable.

Something else to keep in mind assuming you live in North America or other
low voltage country... Standard wall outlets are rated for no more than
15A, but standard derating procedure for safety and reliability reasons
suggest not to draw more than say 12A from the outlet. Capacitive input
filters fed through a rectifier from a sinusoidal voltage source have bad
power factor. Something around 0.6 might be typical, though it could be
lower for higher capacitance with small series resistance. So with an
output power of 600W (or just waste of that much) you could be using up a
substantial amount of the outlet's capability. I've tripped my 15A circuit
breaker before with a 500W load at low power factor (+maybe 350W of other
low power factor equipment on the circuit including my computer).


Anyway, has anyone had good success building large DC supplies, away
from using a 317 and a big parrallel bank? Any ideas would be
appreciated.

Thanks for reading,
Steve Kamego

Thanks for the feedback. I will probably start seriously running
though the design process in a few weeks. I will definantely keep the
plug load capacity in mind, I live in an apartment, and I definantely
don't want to cause any trouble. I can easily pull turns of wire from
the transformer, I originally wound it with as many turns as I could
fit inside it, so I wouldn't mind bringing it down to a rectified
output around 50 or 60 VDC. I also will, as NT said, have the
un-regulated power avaliable for those times that power of that nature
is necessary. I appreciate all the help, maybe I'll finally get this
supply build (I've been working on it here and there for about two
years now).

Thanks Again,
Steve Kamego

 

[John Crighton]

On 12 Apr 2005 17:39:09 -0700, bigcat@meeow.co.uk wrote:
<snip>

I like variacs. With a bit extra ingenuity the OP could turn the MOT
into a variac with linear slider. Now that would be a lot more use than
a 20V 8A psu - and would leave all those components left over for
another psu.

NT

OK you have got me interested. I am all ears.
How would you turn a MOT into a variac?
Regards,
John Crighton
Sydney

 

[Adrian Tuddenham]

<bigcat@meeow.co.uk> wrote:

[variac]

Now make a sliding knob with a wide carbon brush on the bottom, make it
run along this track. The brush must be wide so as not to produce much
turn shorting current. A slotted brush or 2 brushes would be better.

The carbon of a variac brush is fairly high resistance so that it can
operate in contact with more than one turn at a time without giving rise
to huge circulating currents. This allows smoother voltage control
without breaks between the steps.


--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk