Maker Pro
Maker Pro

High current power supply?

A

aran

Jan 1, 1970
0
I will design an AC/DC power-supply that can give up to 100A with a voltage
range 2V-24V. Almost transient- and noise-free DC output is demanded. Would
like to switch between current-, voltage- or power (with maximum current
and/or voltage) regulation. I will also pulse the output, i.e. use the PS as
pulse generator (rectangular pulse shape).



I would like to hear opinion of the people out there for best design
principals.





I believe the switch-suppliers have higher noise than linear suppliers.
Using bipolar transistor(s) will give a low efficiency (as heat losses). May
be IGBT transistors used together with an integrated driver circuit?



I appreciate any suggestion, literature, link, etc.
 
R

Rene Tschaggelar

Jan 1, 1970
0
aran said:
I will design an AC/DC power-supply that can give up to 100A with a voltage
range 2V-24V. Almost transient- and noise-free DC output is demanded. Would
like to switch between current-, voltage- or power (with maximum current
and/or voltage) regulation. I will also pulse the output, i.e. use the PS as
pulse generator (rectangular pulse shape).

What does almost transient- and noise free mean ?
200mVpp, 20mVpp, 2mVpp, 200uVpp ?
This determines whether a switcher, a thyristor regulated or a linear
supply is required.

In case of a linear supply, can you switch the transformer fast enough
or do you intend watercooling ?
I believe the switch-suppliers have higher noise than linear suppliers.
Using bipolar transistor(s) will give a low efficiency (as heat losses).

Fet heat losses are equally high in a linear supply.


Rene
 
A

aran

Jan 1, 1970
0
Well, under 20mVpp will be good enough, but must be satisfied even several
kV transients in the 220V-net (or other sources). Is the solution SRC's
then? Do you mean SRC at the primary side or?

No water cooling is intended, will avoid if not absolute necessary.

MOSFETs has very law drain resistance, a few miliohm. Are there another
factors for the losses?

Thank you for your responsd
 
J

Jamie

Jan 1, 1970
0
better get your self a good size transformer of around 30 volts on the
sec side or so and some large SCR/Triac's ect..
 
K

Ken Smith

Jan 1, 1970
0
MOSFETs has very law drain resistance, a few miliohm. Are there another
factors for the losses?[/QUOTE]

You bet:

The time it takes for a MOSFET to go from off to fully on is not quite
zero. During that time there are largish losses in the MOSFET.

No inductor is perfect

No capacitor is perfect

No diode in perfect.
 
W

Winfield Hill

Jan 1, 1970
0
Ken Smith wrote...
aran wrote: [..]
MOSFETs has very law drain resistance, a few miliohm. Are there
another factors for the losses?

You bet:

The time it takes for a MOSFET to go from off to fully on is not
quite zero. During that time there are largish losses in the MOSFET.

Actually, when driven with a modern driver, power MOSFETs switch
ON rather quickly, minimizing those switching losses. Furthermore
in many designs the current is low at switch-on. It's the turnoff
that's a big problem. Most of the FET's structure, those portions
close to the gate metalization, can be made to switch off pretty
fast, but a portion continues to conduct and given the high drain
currents at switch off, can create considerable FET dissipation.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)
 
C

CFoley1064

Jan 1, 1970
0
Subject: High current power supply?
From: "aran" [email protected]
Date: 7/19/2004 7:38 AM Central Daylight Time
Message-id: <[email protected]>

I will design an AC/DC power-supply that can give up to 100A with a voltage
range 2V-24V. Almost transient- and noise-free DC output is demanded. Would
like to switch between current-, voltage- or power (with maximum current
and/or voltage) regulation. I will also pulse the output, i.e. use the PS as
pulse generator (rectangular pulse shape).

Hmmn. If I were in your shoes, I would get the manual for the old HP6259B and
6456B high current DC power supplies, and possibly rent one to open up and take
a look at. If you need low noise, you're going to need a linear, and you
should look at the SCR pre-regulation setup of the 6259B as the way to go. It
does CV/CC, too. You'll also need a small forklift to move the HP. I've found
it's best just to keep them on the pallet they're shipped on.

Remember you're going to require a 240VAC three phase transformer with all that
power, so you'll want to get the manual for the threee phase version. For the
pulse generation, you'll have to work up a post-power supply series high-side
switch.

Keep the manual with you, take it to the water closet, sleep with it under your
pillow, and try to grok the fullness. Once you're there, make sure you have a
long heart-to-heart with the transformer manufacturer you choose. Let them
know what you're using it for, and make sure they understand the potential
problems. If they don't or they say it's not relevant, talk to someone else.
You'll have to get the transformer custom-manufactured, which ain't cheap, and
will certainly have lead time. Also, get a corruption of high throughput 240V
fans if you're not cooling with water -- you'll need them all.

Apart from that, it looks like a straightforward project. So's the high side
switch.

Oh, yes. This little project has the genuine possibility of being fatal. Do
make sure you use good engineering practice, and make sure somebody else is in
the room at all times during development to disconnect you from the contraption
if you happen to close a circuit.

You were born for this. I particularly like your initial statement "I will
design..." That kind of take-no-prisoners, failure-is-not-an-option attitude
is good to see in engineers. Of course, there are other ways of looking at
this...

Good luck, and please be careful
Chris
 
R

Rene Tschaggelar

Jan 1, 1970
0
aran said:
Well, under 20mVpp will be good enough, but must be satisfied even several
kV transients in the 220V-net (or other sources). Is the solution SRC's
then? Do you mean SRC at the primary side or?

No water cooling is intended, will avoid if not absolute necessary.

MOSFETs has very law drain resistance, a few miliohm. Are there another
factors for the losses?


Well, in a linear supply, the FET is not in the mOhm state, but somewhat
above to get rid of several volts. Then the losses are apparent.

Without water cooling you won't be able to have a linear supply that
delivers 100A at 2V, because that'd dissipate 100A times 22V or 2.2kW.

So that leaves a polyphase switcher or a thyristor regulated supply.
Lownoise switchers are somewhat tricky.

Rene
 
K

Ken Smith

Jan 1, 1970
0
Winfield Hill said:
Actually, when driven with a modern driver, power MOSFETs switch
ON rather quickly, minimizing those switching losses. Furthermore
in many designs the current is low at switch-on. It's the turnoff
that's a big problem. Most of the FET's structure, those portions
close to the gate metalization, can be made to switch off pretty
fast, but a portion continues to conduct and given the high drain
currents at switch off, can create considerable FET dissipation.

Also:

In some designs the current is also zero or near zero at the off edge
too. Psudo-resonant designs have lower switching losses because of this.

If there is much winding capacitance in the transformer or slowish silicon
diodes are used on the output side, the current at the turn on edge can
exceed the current at the turn off. This is one of the ways a bad
transformer design raises losses.
 
Top