Maker Pro
Maker Pro

Class AB amplifier

S

Simon Burley

Jan 1, 1970
0
Hi,

I'm looking for a push-pull output stage design for a single +500v rail
supply. It would need to drive no more than 1mA.

I guess that complementary design with bipolar PNP transistors are out,
since I can't find any PNP devices that have a reasonable margin.

What's the best way to go? A quasi-complementary output stage using only
NPN transistors, or design using MOSFETs? And can someone provide a
reference or url to a sample design or either transistor type?

Thanks
Simon
 
W

Winfield Hill

Jan 1, 1970
0
Simon said:
I'm looking for a push-pull output stage design for a single +500v rail
supply. It would need to drive no more than 1mA.

I guess that complementary design with bipolar PNP transistors are out,
since I can't find any PNP devices that have a reasonable margin.

What's the best way to go? A quasi-complementary output stage using only
NPN transistors, or design using MOSFETs? And can someone provide a
reference or url to a sample design or either transistor type?

There's a circuit in the Art of Electronics, page 169.
If you don't have a copy, that's one of the pages that
Google has scanned, just type "low-power piezo driver"
into the search box. Pages 168-171 are available. As
you can see, two high-voltage power MOSFETs are used,
one to pull up and the other to pull down, via a diode.
That kind of output stage is called a totem pole, and
was used in TTL logic gates designed in the late 60s.

We used BUZ-50B power MOSFETs in our original design,
but you can substitute an IRFBG20 or another HV part.
Infineon's SPA02N80C3 is a low-cost 800V part in stock
at DigiKey, with a low gate capacitance, Ciss = 290pF.
Its Coss is much higher than the IRFBG20, 130 vs 52pF.
 
E

Eeyore

Jan 1, 1970
0
Simon said:
Hi,

I'm looking for a push-pull output stage design for a single +500v rail
supply. It would need to drive no more than 1mA.

I guess that complementary design with bipolar PNP transistors are out,
since I can't find any PNP devices that have a reasonable margin.

What's the best way to go? A quasi-complementary output stage using only
NPN transistors, or design using MOSFETs? And can someone provide a
reference or url to a sample design or either transistor type?

Well .... you could use a couple of devices in cascode. MSA42 and 92 have a
300V rating.

Graham
 
S

Simon Burley

Jan 1, 1970
0
Winfield said:
There's a circuit in the Art of Electronics, page 169.
If you don't have a copy, that's one of the pages that
Google has scanned...

Many thanks, just what I was looking for.
I will see if I can get a copy from Amazon, looks like a very useful book.

Simon
 
M

Meat Plow

Jan 1, 1970
0
Hi,

I'm looking for a push-pull output stage design for a single +500v rail
supply. It would need to drive no more than 1mA.

I guess that complementary design with bipolar PNP transistors are out,
since I can't find any PNP devices that have a reasonable margin.

What's the best way to go? A quasi-complementary output stage using only
NPN transistors, or design using MOSFETs? And can someone provide a
reference or url to a sample design or either transistor type?

Thanks
Simon

Why 500 volts?
 
J

John Larkin

Jan 1, 1970
0
Hi,

I'm looking for a push-pull output stage design for a single +500v rail
supply. It would need to drive no more than 1mA.

I guess that complementary design with bipolar PNP transistors are out,
since I can't find any PNP devices that have a reasonable margin.

What's the best way to go? A quasi-complementary output stage using only
NPN transistors, or design using MOSFETs? And can someone provide a
reference or url to a sample design or either transistor type?

Thanks
Simon


I posted my optocoupler-based hv amp to a.b.s.e. a while back. It
takes about 6 or so parts. But the highest-voltage optos I know of are
400 volt parts.

Cascoding low-voltage optocouplers with depletion-mode mosfets should
work fine.

John
 
J

Jim Thompson

Jan 1, 1970
0
I posted my optocoupler-based hv amp to a.b.s.e. a while back. It
takes about 6 or so parts. But the highest-voltage optos I know of are
400 volt parts.

Cascoding low-voltage optocouplers with depletion-mode mosfets should
work fine.

John

At 1mA load who needs push-pull.

Personally, I'd use a TUBE, driven from the cathode with solid state
whatever. UNKILLABLE ;-)

...Jim Thompson
 
J

John Larkin

Jan 1, 1970
0
At 1mA load who needs push-pull.


What, single-ended class A with a resistive pullup? Do the math.

Personally, I'd use a TUBE, driven from the cathode with solid state
whatever. UNKILLABLE ;-)

Well, at 5 KV maybe.

John
 
T

Tim Williams

Jan 1, 1970
0
John Larkin said:
What, single-ended class A with a resistive pullup? Do the math.

Duh, invoke a +1000V supply. At 1mA, it's *practically* static
electricity, no one will notice. ;-)

Tim
 
J

Jim Thompson

Jan 1, 1970
0
What, single-ended class A with a resistive pullup? Do the math.

So what's a few watts amongst friends ?:)
Well, at 5 KV maybe.

John

I've actually done it at 30KV... almost 50 years ago, when there were
damn few semiconductors ;-)

...Jim Thompson
 
B

BobW

Jan 1, 1970
0
Simon Burley said:
Many thanks, just what I was looking for.
I will see if I can get a copy from Amazon, looks like a very useful book.

Simon

You'll see that it's an expensive book, however, the old saying "you get
what you pay for" absolutely applies here.

If you do get it, perhaps you can coerce Mr. Hill into autographing it for
you.

Bob
 
B

BobW

Jan 1, 1970
0
Jim Thompson said:
At 1mA load who needs push-pull.

Personally, I'd use a TUBE, driven from the cathode with solid state
whatever. UNKILLABLE ;-)

...Jim Thompson


Yep. I agree.

Back in 1980, I built a fully complimentary bridged AB 400V amplifier (using
two NPNs in series and two PNP's in series [the inside guys were slaves]) to
directly drive electret headphones.

It worked extrememly well, but when I needed to build another one I didn't
want all that complexity. The second one was a pair of NPNs with resistor
pullups. The resistor value was selected low enough to give me adequate high
frequency response (into the capacitive load of the headphones). It did get
hotter than the first one, but WTF? It still works after about 20 years of
use.

Bob
 
J

John Larkin

Jan 1, 1970
0
So what's a few watts amongst friends ?:)


I've actually done it at 30KV... almost 50 years ago, when there were
damn few semiconductors ;-)

...Jim Thompson


I've used a 1B3 HV diode as an active regulator element, with the
control being the filament voltage.

John
 
J

Jim Thompson

Jan 1, 1970
0
I've used a 1B3 HV diode as an active regulator element, with the
control being the filament voltage.

John

Ooooooh :-(

...Jim Thompson
 
F

Fred Bartoli

Jan 1, 1970
0
Jim Thompson a écrit :
Ooooooh :-(

On a friend's toobs amplifier which had some pb with push-pull quiescent
current balancing and really huge primary inductance, I had him doing
that within a FB loop.
(he was crazy about having an opamp 'in the signal path' despite it
wouldn't really have been with a more classical solution. You know that
kind of guy).
 
J

John Larkin

Jan 1, 1970
0
You don't need the negative supply (or opto-coupler) unless you need
to get really close to ground.

...Jim Thompson


I was thinking of an opamp with +-2000 volt supply rails.

There's a tube, still in use by Kepco or somebody, that's a nice 8KV
pentode. And the old color teevees had a shunt regulator tube that was
good to 25 kv or something absurd.

John
 
Top