HV gain stage with fewest number of transistors

[alan]

Hello people:

Looking to build a high voltage amp for as cheap as possible. So I
figured that if I use a decent op-amp, and then some transistors as an
output stage in a composite amplifier configuration, that it would be a
lot cheaper than the $150+ HV op-amps.

So I guess the transistor gain stage would have these characteristics:
Gain of >40x (have to look into some composite amp stuff to figure out
the actual gain to use)
Bipolar to +/-200V
Doesn't need to output more than a few mA
speed of maybe 1kHz
I guess it doesn't have to be super-linear coz the feedback from the
input stage will fix that.
umm, I can't think of any other characteristics

If anybody can point me to some links for such a circuit, or even to a
complete amp design and let me know just what portion I can excise, that
would be cool.

 

[Winfield Hill]

alan wrote...

Looking to build a high voltage amp for as cheap as possible. So I
figured that if I use a decent op-amp, and then some transistors as an
output stage in a composite amplifier configuration, that it would be
a lot cheaper than the $150+ HV op-amps.

I think Apex PA42 plastic-SIP amp is more affordable, about $45.


--
Thanks,
- Win

 

[Daniel Haude]

On Wed, 10 Nov 2004 03:47:38 -0800,
alan <no-longer-valid@yahoo.com> wrote
in Msg. <cmsv71$9d2$1@news.Stanford.EDU>

Looking to build a high voltage amp for as cheap as possible.

Noisy or quiet?

How about AoE, p. 169? Can't beat the parts count...

--Daniel

--
"With me is nothing wrong! And with you?" (from r.a.m.p)

 

[John Larkin]

On Wed, 10 Nov 2004 03:47:38 -0800, alan <no-longer-valid@yahoo.com>
wrote:

Hello people:

Looking to build a high voltage amp for as cheap as possible. So I
figured that if I use a decent op-amp, and then some transistors as an
output stage in a composite amplifier configuration, that it would be a
lot cheaper than the $150+ HV op-amps.

So I guess the transistor gain stage would have these characteristics:
Gain of >40x (have to look into some composite amp stuff to figure out
the actual gain to use)
Bipolar to +/-200V
Doesn't need to output more than a few mA
speed of maybe 1kHz
I guess it doesn't have to be super-linear coz the feedback from the
input stage will fix that.
umm, I can't think of any other characteristics

If anybody can point me to some links for such a circuit, or even to a
complete amp design and let me know just what portion I can excise, that
would be cool.

I posted a schematic to abse a while back. The composite opamp had one
small lv opamp driving two optoisolators, with the opto output
transistors in a totem pole output stage. It uses a Fairchild
optoisolator that's good for 400 volts on the phototransistor side. So
the whole thing reduces to three small surface-mount parts and a few
passives. I used this to apply bias voltages to some electrodes in a
particle detector.

I could post it again if this sounds interesting.

What's your application?

John

 

[Winfield Hill]

Daniel Haude wrote...

alan wrote

Looking to build a high voltage amp for as cheap as possible.

How about AoE, p. 169? Can't beat the parts count...

That only does unipolar HV, the O.P. wanted bipolar. The
additional parts in the drawing in my post provide bipolar.


--
Thanks,
- Win

 

[Winfield Hill]

Jim Thompson wrote...

Winfield Hill wrote:

Given that 1200V MOSFETs are readily available, it's easy to
make up to +/-600V dc-coupled amplifiers, using a totem-pole
output stage like we show in AoE, fig. 3.75, but with an added
negative rail and level-shift stage. Here's the basic idea:

. basic high-voltage MOSFET dc amplifier
. by Winfield Hill
.
. + supply rail -----------------+-------,
. | |
. R11 |
. | D Q4
. +--+--G
. C2 | | S
. ,--R4--||-, | C |
. IN | | | B--+ Q5
. o--R3--+--|+\ | | E |
. | >---+--R5--, X | R10
. ,--|-/ | | | | out
. | E | '----+-+--R12--o
. | gnd -+- B Q1 | D1 |
. | | C +--|<|----+
. | | | +---R13---+
. | | | | |
. +---------- | -- | ------- | -+--R2--+
. | | | | | |
. R1 R13 | C3 | '--||--+
. | | E ,-||--+ C1 |
. gnd +- B | | |
. | Q2 C R8 | R9
. | | | D Q3 |
. R14 +---++--G C4
. | | | S |
. | R6 C | gnd
. | | B--+
. | | E | Add a zener across
. | | | R7 each FET gate!
. | | | |
. - supply rail --+----+----+----'

The circuit is inexpensive and works well at low frequencies but
it can be further enhanced for a -3dB bandwidth over 100kHz.

As you can see the circuit includes short-circuit protection. My
finished designs generally have more parts to add features such as
HV-output monitoring, reduced cross-over distortion (at point X),
increased slew rate, improved power-supply rejection, etc.

Looks like a MOSFET twist on my MC1554 design, for which I won a
coffee bet with Tom Frederiksen... he didn't think I could do an
all_NPN design.

It's often called a totem-pole output stage, and I first ran into
it in the late 60s in DTL logic gates. When did you do the mc1554?

Problem: Loop gain is higher when upper half of output stage is
active than with lower half.

Yes, and the pulldown gain depends on the load. This all forces
an overly conservative compensation.

I bet I can guess method at "X" ;-)

...Jim Thompson

Yep, there are several good candidates working towards class AB.
A big problem is one doesn't know much about each MOSFET's linear
gate operating voltage... For one series of these HV amplifiers
I even designed a little simple floating servo system.


--
Thanks,
- Win

 

[Ken Smith]

In article <cmsv71$9d2$1@news.Stanford.EDU>,
alan <no-longer-valid@yahoo.com> wrote:

Hello people:

Looking to build a high voltage amp for as cheap as possible. So I
figured that if I use a decent op-amp, and then some transistors as an
output stage in a composite amplifier configuration, that it would be a
lot cheaper than the $150+ HV op-amps.

So I guess the transistor gain stage would have these characteristics:
Gain of >40x (have to look into some composite amp stuff to figure out
the actual gain to use)
Bipolar to +/-200V
Doesn't need to output more than a few mA
speed of maybe 1kHz
I guess it doesn't have to be super-linear coz the feedback from the
input stage will fix that.
umm, I can't think of any other characteristics

Who about this for cheap? (or did you mean low cost?)




From op-amp ------\/\/\---\!
!------ GND
/! +300V
! !
\ \
/ / 5K
\ \ 100W
! !
! !!--+------- OUTPUT
+----+-----------!
! ! !!--- BIG MOSFET
/ /---/ ! on BIG SINK
\ ^ 15V /
/ ! \
! ! /
-----+----+----------------
-300V
--
--
kensmith@rahul.net forging knowledge

 

[Ken Smith]

In article <cmt6ht02p7u@drn.newsguy.com>,
Winfield Hill <hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote:
[...]

. basic high-voltage MOSFET dc amplifier
. by Winfield Hill
With changes by Ken Smith:
.
. + supply rail -----------------+-------,
R11B !
. Are you +------ | -------
. sure? (+/-) R11 | !
. ! | D Q4 --- added
. V +--+--G --- Cnoise
. C2 | | S !
. ,--R4--||-, | C | GND
. IN | | | B--+ Q5
. o--R3--+--|+\ | | E |
. | >---+--R5--, X | R10
. ,--|-/ | | | | out
. | E | '----+-+--R12----o
. | gnd -+- B Q1 | D1 |
. | | C +--|<|----+
. | | | +---R13---+
. | | | | |
. +---------- | -- | ------- | -+--R2--+
. | | | | | |
. R1 R13 | C3 | '--||--+
. | | E ,-||--+ C1 |
. gnd +- B | | |
. | Q2 C R8 | R9
. | | | D Q3 |
. R14 +---++--G C4
. | | | S |
. | R6 C | gnd
. | | B--+
. | | E | Add a zener across
. | | | R7 each FET gate!
. | | | |
. - supply rail --+----+----+----'

The circuit is inexpensive and works well at low frequencies but
it can be further enhanced for a -3dB bandwidth over 100kHz.

As you can see the circuit includes short-circuit protection. My
finished designs generally have more parts to add features such as
HV-output monitoring, reduced cross-over distortion (at point X),
increased slew rate, improved power-supply rejection, etc.


--
Thanks,
- Win

--
--
kensmith@rahul.net forging knowledge

 

[Tony Williams]

In article <61h4p05gprne8eifnmrugqnsqkmd7acn03@4ax.com>,
John Larkin wrote:

I posted a schematic to abse a while back. The composite opamp
had one small lv opamp driving two optoisolators, with the opto
output transistors in a totem pole output stage. It uses a
Fairchild optoisolator that's good for 400 volts on the
phototransistor side. So the whole thing reduces to three small
surface-mount parts and a few passives. I used this to apply bias
voltages to some electrodes in a particle detector.

That was a good post, a totally original idea that would
never have dawned on me. Deceptively simple as well.

--
Tony Williams.

 

[John Larkin]

On Thu, 11 Nov 2004 11:08:45 +0000 (GMT), Tony Williams
<tonyw@ledelec.demon.co.uk> wrote:

In article <61h4p05gprne8eifnmrugqnsqkmd7acn03@4ax.com>,
John Larkin wrote:

I posted a schematic to abse a while back. The composite opamp
had one small lv opamp driving two optoisolators, with the opto
output transistors in a totem pole output stage. It uses a
Fairchild optoisolator that's good for 400 volts on the
phototransistor side. So the whole thing reduces to three small
surface-mount parts and a few passives. I used this to apply bias
voltages to some electrodes in a particle detector.

That was a good post, a totally original idea that would
never have dawned on me. Deceptively simple as well.

Actually, I should have got brave and made it full class-B. That would
result in a +-200v opamp with 6 parts.

John

 

[John Larkin]

On Wed, 10 Nov 2004 03:47:38 -0800, alan <no-longer-valid@yahoo.com>
wrote:

Hello people:

Looking to build a high voltage amp for as cheap as possible. So I
figured that if I use a decent op-amp, and then some transistors as an
output stage in a composite amplifier configuration, that it would be a
lot cheaper than the $150+ HV op-amps.

Another strange idea posted to abse; class AB opto-coupled HV
composite opamp.

John

 

[Winfield Hill]

Ken Smith wrote...

Winfield Hill rote:
[...]
. basic high-voltage MOSFET dc amplifier
. by Winfield Hill
With changes by Ken Smith:
.
. + supply rail -----------------+-------,
R11B !
. Are you +------ | -------
. sure? (+/-) R11 | !
. ! | D Q4 --- added
. V +--+--G --- Cnoise
. C2 | | S !
. ,--R4--||-, | C | GND
. IN | | | B--+ Q5
. o--R3--+--|+\ | | E |
. | >---+--R5--, X | R10
. ,--|-/ | | | | out
. | E | '----+-+--R12----o
. | gnd -+- B Q1 | D1 |
. | | C +--|<|----+
. | | | +---R13---+
. | | | | |
. +---------- | -- | ------- | -+--R2--+
. | | | | | |
. R1 R13 | C3 | '--||--+
. | | E ,-||--+ C1 |
. gnd +- B | | |
. | Q2 C R8 | R9
. | | | D Q3 |
. R14 +---++--G C4
. | | | S |
. | R6 C | gnd
. | | B--+
. | | E | Add a zener across
. | | | R7 each FET gate!
. | | | |
. - supply rail --+----+----+----'

Are you sure? (+/-)

Right. Silly typo, already corrected an earlier post in this
thread.

added Cnoise...

Much better, splitting R11 and adding a C_noise / C_bootstrap cap.
This effectively creates an ac pullup current source. The R6 / R7
pulldown current-limiting scheme is also used, saving a transistor.

.. basic high-voltage MOSFET dc amplifier
.. by Winfield Hill
..
.. + supply rail -----------------+-------,
.. +600V | |
.. R11B |
.. | |
.. +------ | -----,
.. | | |
.. R11A | |
.. | D |
.. +--+--G Q4 | C_bootstrap
.. C2 | | S _|_
.. ,--R4--||-, | C | ---
.. IN | __ | | B--+ Q5 |
.. o--R3--+--|- \ | | E | |
.. | >--+--R5--, X | R10 |
.. ,--|+_/ | Q1 | | | | out
.. | E | '----+-+----+---R12---o
.. | gnd -+- B | D1 |
.. | | C +--|<|----+
.. | R13 | +---R15---+
.. | | | | |
.. +---------- | -- | ------- | -+--R2--+
.. | | | Q2 | '--||--+
.. R1 | E ,-||-+ C1 |
.. | +- B | C3 | R9
.. gnd | C R8 | |
.. | | | D Q3 C4
.. R14 +----+--G |
.. | | S gnd
.. | R6 |
.. | | R7 Add a VGS zener
.. -600V | | | on each MOSFET!
.. - supply rail --+----+---------'

Furthermore, Q2 and R13 R14 can be eliminated if V- is below -350V.


--
Thanks,
- Win

 

[alan]

Thanks for the sugestions, everybody. It looks like getting an Apex
PA97 is the easiest thing to do.

 

[Winfield Hill]

alan wrote...

Thanks for the sugestions, everybody. It looks like getting
an Apex PA97 is the easiest thing to do.

One thing to watch out for, the PA97 doesn't have any current
limiting on its output. It also doesn't have class-AB biasing
for its output transistors, and therefore it has high crossover
distortion if used for high-frequency ac outputs. If you need
faster response, consider pre-biasing the output with a constant
current to one rail, e.g., with a resistor pulling down to the
-Vs rail, or use my favorite LND150 as a 500V 1mA current sink.


--
Thanks,
- Win

 

[Winfield Hill]

alan wrote...

Looking to build a high voltage amp for as cheap as possible. So I
figured that if I use a decent op-amp, and then some transistors as an
output stage in a composite amplifier configuration, that it would be a
lot cheaper than the $150+ HV op-amps.

So I guess the transistor gain stage would have these characteristics:
Gain of >40x (have to look into some composite amp stuff to figure out
the actual gain to use)
Bipolar to +/-200V
Doesn't need to output more than a few mA
speed of maybe 1kHz
I guess it doesn't have to be super-linear coz the feedback from the
input stage will fix that.
umm, I can't think of any other characteristics

If anybody can point me to some links for such a circuit, or even to
a complete amp design and let me know just what portion I can excise,
that would be cool.

Here's a copy of one of my posts on the subject.

Given that 1200V MOSFETs are readily available, it's easy to
make up to +/-600V dc-coupled amplifiers, using a totem-pole
output stage like we show in AoE, fig. 3.75, but with an added
negative rail and level-shift stage. Here's the basic idea:

.. basic high-voltage MOSFET dc amplifier
.. by Winfield Hill
..
.. + supply rail -----------------+-------,
.. | |
.. R11 |
.. | D Q4
.. +--+--G
.. C2 | | S
.. ,--R4--||-, | C |
.. IN | | | B--+ Q5
.. o--R3--+--|+\ | | E |
.. | >---+--R5--, X | R10
.. ,--|-/ | | | | out
.. | E | '----+-+--R12--o
.. | gnd -+- B Q1 | D1 |
.. | | C +--|<|----+
.. | | | +---R13---+
.. | | | | |
.. +---------- | -- | ------- | -+--R2--+
.. | | | | | |
.. R1 R13 | C3 | '--||--+
.. | | E ,-||--+ C1 |
.. gnd +- B | | |
.. | Q2 C R8 | R9
.. | | | D Q3 |
.. R14 +---++--G C4
.. | | | S |
.. | R6 C | gnd
.. | | B--+
.. | | E | Add a zener across
.. | | | R7 each FET gate!
.. | | | |
.. - supply rail --+----+----+----'

The circuit is inexpensive and works well at low frequencies but
it can be further enhanced for a -3dB bandwidth over 100kHz.

As you can see the circuit includes short-circuit protection. My
finished designs generally have more parts to add features such as
HV-output monitoring, reduced cross-over distortion (at point X),
increased slew rate, improved power-supply rejection, etc.


--
Thanks,
- Win

 

[Ken Smith]

In article <cn169402hv4@drn.newsguy.com>,
Winfield Hill <hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote:
[.. my added Cnoise ..]

Much better, splitting R11 and adding a C_noise / C_bootstrap cap.
This effectively creates an ac pullup current source. The R6 / R7
pulldown current-limiting scheme is also used, saving a transistor.

I had started to draw exactly that and then stopped for some reason
<smoke comes from ears as memory accessed>
The circuit already has a big imbalance between the pull down gain and the
pull up gain. The bootstrap increases the pull up gain. This is the one
that is already high. I had assumed that the combination of R11 and R15
had been selected to reduce this enough to prevent oscillation on the up
swing.

I'd also like to add a suggestion about the current limitters.

To source
! To drain
! !
! \
! / To gate
! \ R201 !
! D1 ! !/
+---S<-----/\/\/\----+------!
! R202 !\
/ !
\ !
/ !
\ R sense !
! !
+-----------------------------
!

The schottky diode D1 is optional. It reduces the size of Rsense.
R201 and R202 work to add a foldback to the current limit. This can allow
you to let the current limit be quite a bit higher without putting the
MOSFETs at risk.

Also:

The op-amp is running on +/- supplies. I worry a bit about the
emitter base junction of Q1 during power up.

You may want to add a resistor in series with C1, Imagine that the output
is swung up to the rail and then someone suddenly shorts it to ground.
The op-amp would be the point of failure. If you protect it, the whole
circuit would be able to survive random shorting.


The supplies are large enough and the value of R5 high enough that the
emitter diode of Q1's non-linearity doesn't get into things much. If the
op-amp section included a RC of equal TC to the R4,C2 pair, the op-amp
can be made to act more like a current source to Q1. In this circuit, the
output non-linearities are going to swamp any improvement that you get.

D1 should be a fast recovery device or Schottky. I have used a circuit
very like this and discovered that in my case, the diode was important to
the distortion (Funny squiggle on rising edge). The circuit was not audio
but was providing a large swing "logical" signal that triggered a
responder. You triggered the responder by pulsing its power wire down and
back up by about 5V. The responder's maker didn't suggest how ugly the
signal could be and still work so I needed to make a fairly clean looking
thing just to avoid having to redesign. The circuit regulated and pulsed
the power line.

In my version, only the pull up had a real current limit circuit. The
pull down just relied on a source resistor to keep things within reason.

. basic high-voltage MOSFET dc amplifier
. by Winfield Hill
.
. + supply rail -----------------+-------,
. +600V | |
. R11B |
. | |
. +------ | -----,
. | | |
. R11A | |
. | D |
. +--+--G Q4 | C_bootstrap
. C2 | | S _|_
. ,--R4--||-, | C | ---
. IN | __ | | B--+ Q5 |
. o--R3--+--|- \ | | E | |
. | >--+--R5--, X | R10 |
. ,--|+_/ | Q1 | | | | out
. | E | '----+-+----+---R12---o
. | gnd -+- B | D1 |
. | | C +--|<|----+
. | R13 | +---R15---+
. | | | | |
. +---------- | -- | ------- | -+--R2--+
. | | | Q2 | '--||--+
. R1 | E ,-||-+ C1 |
. | +- B | C3 | R9
. gnd | C R8 | |
. | | | D Q3 C4
. R14 +----+--G |
. | | S gnd
. | R6 |
. | | R7 Add a VGS zener
. -600V | | | on each MOSFET!
. - supply rail --+----+---------'

Furthermore, Q2 and R13 R14 can be eliminated if V- is below -350V.


--
Thanks,
- Win

--
--
kensmith@rahul.net forging knowledge

 

[Ken Smith]

In article <cmsv71$9d2$1@news.Stanford.EDU>,
alan <no-longer-valid@yahoo.com> wrote:

Hello people:

Looking to build a high voltage amp for as cheap as possible. So I
figured that if I use a decent op-amp, and then some transistors as an
output stage in a composite amplifier configuration, that it would be a
lot cheaper than the $150+ HV op-amps.

So I guess the transistor gain stage would have these characteristics:
Gain of >40x (have to look into some composite amp stuff to figure out
the actual gain to use)
Bipolar to +/-200V
Doesn't need to output more than a few mA
speed of maybe 1kHz
I guess it doesn't have to be super-linear coz the feedback from the
input stage will fix that.
umm, I can't think of any other characteristics

If anybody can point me to some links for such a circuit, or even to a
complete amp design and let me know just what portion I can excise, that
would be cool.

How about an LT1206 and a step up transformer?


--
--
kensmith@rahul.net forging knowledge