Two series pass devices?...

[piglet]

Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

<https://www.youtube.com/watch?v=FKC6lnTKH-c>

.... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

piglet

 

[Ricky]

On Wednesday, August 30, 2023 at 5:11:38 PM UTC-4, piglet wrote:

Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

He has a forum. You could just go there and ask him. I\'m sure he would love to discuss it with you.

--

Rick C.

- Get 1,000 miles of free Supercharging
- Tesla referral code - https://ts.la/richard11209

 

[Don]

piglet wrote:

Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

A pertinent previous post:

<https://www.eevblog.com/forum/projects/usupply-usb-rev-b-update/>

contains a link to this schematic:

<https://eevblog.com/files/uSupplyUSBrevB.pdf>

Does your SED post pertain to pass transistors Q5 and Q6 shown in the
schematic?

Danke,

--
Don, KB7RPU, https://www.qsl.net/kb7rpu
There was a young lady named Bright Whose speed was far faster than light;
She set out one day In a relative way And returned on the previous night.

 

[Fred Bloggs]

On Wednesday, August 30, 2023 at 5:11:38 PM UTC-4, piglet wrote:

Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

They use similar in HP benchtops. It\'s necessary if you want a precision programmable current limit. Until it limits, it\'s just a saturated FET putting most of the voltage on the voltage regulation FET.

Diode D4 in the voltage regulator does not inspire confidence.

Note on schematic says D4 prevents interference by input diodes ( of error opamp ) on the voltage measurement, meaning at R22+ R23 junction. If I could read the pn of the op amp it would be clear, but it\'s not readable as is. Maybe it\'s some quasi-Norton amp. Whatever it is the input is acting on about a diode drop below the voltage at external (+) input.

Schematic symbols of resistors look like inductors. FETs look like RJ-45 connectors.

piglet

 

[Don]

Nevermind. The Youtube video finally downloaded. The pass transistors
you ask about are Q1 and Q2.

Danke,

--
Don, KB7RPU, https://www.qsl.net/kb7rpu
There was a young lady named Bright Whose speed was far faster than light;
She set out one day In a relative way And returned on the previous night.

 

[John Larkin]

On Wed, 30 Aug 2023 22:11:28 +0100, piglet <erichpwagner@hotmail.com>
wrote:

Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

piglet

That\'s the craziest design I\'ve ever seen. It must be a deliberate
parody.

 

[Phil Hobbs]

John Larkin <jlarkin@highlandSNIPMEtechnology.com> wrote:

On Wed, 30 Aug 2023 22:11:28 +0100, piglet <erichpwagner@hotmail.com
wrote:

Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

piglet

That\'s the craziest design I\'ve ever seen. It must be a deliberate
parody.

Disclaimer: I find EEVblog practically unwatchable, so I haven’t looked at
the schematic.

One reason to separate the current limit and voltage regulator functions is
to sharpen up the current limiter by putting it inside the voltage
regulator loop.

I commonly do that with LND150 depletion FETs in HV bias supplies for APDs
and SiPMs. They look like ~1k resistors below about 1.5 mA, which is enough
to hurt the gain linearity, but the feedback loop effectively knocks that
down to way under an ohm.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC /
Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics

 

[John Larkin]

On Thu, 31 Aug 2023 02:23:59 -0000 (UTC), Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

John Larkin <jlarkin@highlandSNIPMEtechnology.com> wrote:
On Wed, 30 Aug 2023 22:11:28 +0100, piglet <erichpwagner@hotmail.com
wrote:

Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

piglet

That\'s the craziest design I\'ve ever seen. It must be a deliberate
parody.



Disclaimer: I find EEVblog practically unwatchable, so I haven’t looked at
the schematic.

It\'s a small lab power supply powered by USB-C. It has several
microprocessors, layers of isolators and switching regs and linear
regs, a custom enclosure with custom-machined connectors, a custom
planar transformer, a custom heat sink, a zillon layer PCB, and what
must be tens of KLOCs. It\'s apparently been in design for about 5
years.

Squeaky talks about it in an almost bearable tone.

Really, it\'s funny enough to have a look at. Skip the blather and look
at the schematic pages.

(I had hernia surgery today and am still goofy from the drugs, so this
sort of nonsense fit right in.)

One reason to separate the current limit and voltage regulator functions is
to sharpen up the current limiter by putting it inside the voltage
regulator loop.

I commonly do that with LND150 depletion FETs in HV bias supplies for APDs
and SiPMs. They look like ~1k resistors below about 1.5 mA, which is enough
to hurt the gain linearity, but the feedback loop effectively knocks that
down to way under an ohm.

Cheers

Phil Hobbs

 

[piglet]

On 30/08/2023 22:40, Ricky wrote:

On Wednesday, August 30, 2023 at 5:11:38 PM UTC-4, piglet wrote:
Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

He has a forum. You could just go there and ask him. I\'m sure he would love to discuss it with you.

Thanks, good call, I\'d completely forgotten!

piglet

 

[piglet]

On 30/08/2023 23:50, Fred Bloggs wrote:

On Wednesday, August 30, 2023 at 5:11:38 PM UTC-4, piglet wrote:
Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

They use similar in HP benchtops. It\'s necessary if you want a precision programmable current limit. Until it limits, it\'s just a saturated FET putting most of the voltage on the voltage regulation FET.


Diode D4 in the voltage regulator does not inspire confidence.

Note on schematic says D4 prevents interference by input diodes ( of error opamp ) on the voltage measurement, meaning at R22+ R23 junction. If I could read the pn of the op amp it would be clear, but it\'s not readable as is. Maybe it\'s some quasi-Norton amp. Whatever it is the input is acting on about a diode drop below the voltage at external (+) input.

Schematic symbols of resistors look like inductors. FETs look like RJ-45 connectors.


piglet

The opamp is OPA2180 zero drift/auto zero type with inverse parallel
input protection diodes. D4 looks mega kludgy - if it were mine I\'d
investigate using a different opamp or taking the error signal from the
voltage follower used for metering. Still I guess in real life the
voltage drop error across D4 at nanoamp opamp input bias current is near
negligible?

piglet

 

[piglet]

On 31/08/2023 04:36, John Larkin wrote:

(I had hernia surgery today and am still goofy from the drugs, so this
sort of nonsense fit right in.)

Glad to have supplied some light relief!

piglet

 

[piglet]

On 30/08/2023 23:55, Don wrote:

Nevermind. The Youtube video finally downloaded. The pass transistors
you ask about are Q1 and Q2.

Danke,

Yes. I am used to seeing the outputs of voltage and current error amps
orr-ed together to control just one series pass device.

piglet

 

[Fred Bloggs]

On Thursday, August 31, 2023 at 7:03:29 AM UTC-4, piglet wrote:

On 30/08/2023 23:50, Fred Bloggs wrote:
On Wednesday, August 30, 2023 at 5:11:38 PM UTC-4, piglet wrote:
Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

They use similar in HP benchtops. It\'s necessary if you want a precision programmable current limit. Until it limits, it\'s just a saturated FET putting most of the voltage on the voltage regulation FET.


Diode D4 in the voltage regulator does not inspire confidence.

Note on schematic says D4 prevents interference by input diodes ( of error opamp ) on the voltage measurement, meaning at R22+ R23 junction. If I could read the pn of the op amp it would be clear, but it\'s not readable as is. Maybe it\'s some quasi-Norton amp. Whatever it is the input is acting on about a diode drop below the voltage at external (+) input.

Schematic symbols of resistors look like inductors. FETs look like RJ-45 connectors.


piglet
The opamp is OPA2180 zero drift/auto zero type with inverse parallel
input protection diodes. D4 looks mega kludgy - if it were mine I\'d
investigate using a different opamp or taking the error signal from the
voltage follower used for metering. Still I guess in real life the
voltage drop error across D4 at nanoamp opamp input bias current is near
negligible?

None of what he has makes any sense, neither D4 nor that 2kR resistor. What\'s probably going on is he\'s hitting the Vset input with too large a step from his D/A, and that\'s causing the differential clamp diodes to conduct and produce screwy outputs until the loop recovers. He should limit his steps to say 50mV and allow settling time of 5x of what he\'s pre-measured before applying another one. Get rid of D4 and the 2k.

piglet

 

[Don]

piglet wrote:

Ricky wrote:
piglet wrote:
Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

Diode D4 in the voltage regulator does not inspire confidence.

He has a forum. You could just go there and ask him. I\'m sure he
would love to discuss it with you.

Thanks, good call, I\'d completely forgotten!

Yes, Ricky\'s suggestion is indeed excellent. And what do you know?
Someone at the forum already asked and Dave doesn\'t remember:

<https://www.eevblog.com/forum/blog/eevblog-1561-usupply-usb-power-supply-part-20/msg5033464/#msg5033464>

Regardless, it\'s my intention to find the time to digest the video
first, then post questions to Dave at the forum.

Danke,

--
Don, KB7RPU, https://www.qsl.net/kb7rpu
There was a young lady named Bright Whose speed was far faster than light;
She set out one day In a relative way And returned on the previous night.

 

[John Larkin]

On Thu, 31 Aug 2023 12:03:20 +0100, piglet <erichpwagner@hotmail.com>
wrote:

On 30/08/2023 23:50, Fred Bloggs wrote:
On Wednesday, August 30, 2023 at 5:11:38?PM UTC-4, piglet wrote:
Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

They use similar in HP benchtops. It\'s necessary if you want a precision programmable current limit. Until it limits, it\'s just a saturated FET putting most of the voltage on the voltage regulation FET.


Diode D4 in the voltage regulator does not inspire confidence.

Note on schematic says D4 prevents interference by input diodes ( of error opamp ) on the voltage measurement, meaning at R22+ R23 junction. If I could read the pn of the op amp it would be clear, but it\'s not readable as is. Maybe it\'s some quasi-Norton amp. Whatever it is the input is acting on about a diode drop below the voltage at external (+) input.

Schematic symbols of resistors look like inductors. FETs look like RJ-45 connectors.


piglet

The opamp is OPA2180 zero drift/auto zero type with inverse parallel
input protection diodes. D4 looks mega kludgy - if it were mine I\'d
investigate using a different opamp or taking the error signal from the
voltage follower used for metering. Still I guess in real life the
voltage drop error across D4 at nanoamp opamp input bias current is near
negligible?

piglet

Using a chopamp in that kluge is yet another absurdity.

 

[whit3rd]

On Thursday, August 31, 2023 at 9:56:46 AM UTC-7, John Larkin wrote:

On Thu, 31 Aug 2023 12:03:20 +0100, piglet <erichp...@hotmail.com
wrote:

The opamp is OPA2180 zero drift/auto zero type with inverse parallel
input protection diodes. D4 looks mega kludgy - if it were mine I\'d
investigate using a different opamp or taking the error signal from the
voltage follower used for metering. Still I guess in real life the
voltage drop error across D4 at nanoamp opamp input bias current is near
negligible?

Using a chopamp in that kluge is yet another absurdity.

I must agree; there\'s THREE power pass transistors, one for the switch mode coarse regulator, one
for current, one for voltage, and you could easily limit the current at the switch mode node (there\'s
a switch-current-sense resistor there anyhow). The design has all the hallmarks of being
an exercise in using medium-scale-integration blocks, each with a single function.
That\'s how you end up needing two microprocessors to obey two instructions: supply-this-voltage
except limit-to-this-current.

One nice item, though, is worth mentioning: it displays BOTH the target-or-actual output voltage and output current
at the same time. One of my bench supplies has one-analog-meter-does-both; simpler, but a pain to use.

 

[Lasse Langwadt Christensen]

fredag den 1. september 2023 kl. 03.55.10 UTC+2 skrev whit3rd:

On Thursday, August 31, 2023 at 9:56:46 AM UTC-7, John Larkin wrote:
On Thu, 31 Aug 2023 12:03:20 +0100, piglet <erichp...@hotmail.com
wrote:

The opamp is OPA2180 zero drift/auto zero type with inverse parallel
input protection diodes. D4 looks mega kludgy - if it were mine I\'d
investigate using a different opamp or taking the error signal from the
voltage follower used for metering. Still I guess in real life the
voltage drop error across D4 at nanoamp opamp input bias current is near
negligible?
Using a chopamp in that kluge is yet another absurdity.
I must agree; there\'s THREE power pass transistors, one for the switch mode coarse regulator, one
for current, one for voltage, and you could easily limit the current at the switch mode node (there\'s
a switch-current-sense resistor there anyhow). The design has all the hallmarks of being
an exercise in using medium-scale-integration blocks, each with a single function.
That\'s how you end up needing two microprocessors to obey two instructions: supply-this-voltage
except limit-to-this-current.

no it is not, one mcu is for an isolated USB interface, the other one for running the display/keypad/menu/supply

 

[Jasen Betts]

On 2023-09-03, Jan Panteltje <alien@comet.invalid> wrote:

On a sunny day (Sat, 02 Sep 2023 13:12:42 -0700) it happened John Larkin
jlarkin@highlandSNIPMEtechnology.com> wrote in
uh57fihhi4d8hc8u1k4o18o2qf79crilf9@4ax.com>:

On Sat, 2 Sep 2023 19:25:50 -0000 (UTC), Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

John Larkin <jlarkin@highlandSNIPMEtechnology.com> wrote:
On Thu, 31 Aug 2023 12:05:31 +0100, piglet <erichpwagner@hotmail.com
wrote:

On 31/08/2023 04:36, John Larkin wrote:

(I had hernia surgery today and am still goofy from the drugs, so this
sort of nonsense fit right in.)



Glad to have supplied some light relief!

piglet


Good quote, appropriate to that power supply:


It was the embodiment of the idea that perfection is not when there is
nothing left to add but when there is nothing left to take away.

Keith Houston in Empire of the Sum



As the wise man said,

“All programs have bugs, and all programs can be made smaller. Therefore
all programs can be reduced to a single incorrect instruction.“

Cheers

Phil Hobbs

Phil Hobbs

PDP-11 had the RESET instruction. As in full hardware powerup-type
reset. Saved wear and tear on the power switch.

It also has the Land Mine instruction,

MOV -(PC), -(PC)

which copied itself one word below, and executed from there.

In linux systems you just type, as root, \'poweroff\' in a terminal.

As to starting anew with nothing:
s u d o rm -rf /*
remove spaces in \'sudo\' (added those just to prevent poor creatures from doing a cut and paste)
as it will erase everything on your system.

I tried

rm -rf /

yesterday, rm refused to comply I had to say

rm --no-preserve-root -rf /

And yes, as I am always root, I once typed rm * -rf by accident while in /
was quick to stop with ctrlC but hat to put back some X libraries...

Do not try this at home.. ;-)

I was at work.

--
Jasen.
🇺🇦 Слава Україні

 

[Jan Panteltje]

On a sunny day (Tue, 5 Sep 2023 04:28:59 -0000 (UTC)) it happened Jasen Betts
<usenet@revmaps.no-ip.org> wrote in <ud6aqb$2r1$3@gonzo.revmaps.no-ip.org>:

On 2023-09-03, Jan Panteltje <alien@comet.invalid> wrote:
On a sunny day (Sat, 02 Sep 2023 13:12:42 -0700) it happened John Larkin
jlarkin@highlandSNIPMEtechnology.com> wrote in
uh57fihhi4d8hc8u1k4o18o2qf79crilf9@4ax.com>:

On Sat, 2 Sep 2023 19:25:50 -0000 (UTC), Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

John Larkin <jlarkin@highlandSNIPMEtechnology.com> wrote:
On Thu, 31 Aug 2023 12:05:31 +0100, piglet <erichpwagner@hotmail.com
wrote:

On 31/08/2023 04:36, John Larkin wrote:

(I had hernia surgery today and am still goofy from the drugs, so this
sort of nonsense fit right in.)



Glad to have supplied some light relief!

piglet


Good quote, appropriate to that power supply:


It was the embodiment of the idea that perfection is not when there is
nothing left to add but when there is nothing left to take away.

Keith Houston in Empire of the Sum



As the wise man said,

“All programs have bugs, and all programs can be made smaller. Therefore
all programs can be reduced to a single incorrect instruction.“

Cheers

Phil Hobbs

Phil Hobbs

PDP-11 had the RESET instruction. As in full hardware powerup-type
reset. Saved wear and tear on the power switch.

It also has the Land Mine instruction,

MOV -(PC), -(PC)

which copied itself one word below, and executed from there.

In linux systems you just type, as root, \'poweroff\' in a terminal.

As to starting anew with nothing:
s u d o rm -rf /*
remove spaces in \'sudo\' (added those just to prevent poor creatures from doing a cut and paste)
as it will erase everything on your system.

I tried

rm -rf /

yesterday, rm refused to comply I had to say

rm --no-preserve-root -rf /

Modern versions of Linux keep amazing me!
That \'rm\' I did was maybe 15 or 20 years ago.

There is so much crap build into today\'s Linux distros.. utilities
I find for example Ubuntu, that I just installed on my laptop,
a disaster.
Not only does it not allow root to run X programs, it also does
sneaky automatic updates, and if you switch that off, then it has something called \'snap\'
that will do updates by itself.
Debian is more sane...

My thought was, after fighting 2 days with Ubuntu , that Linux is dead.
There are fortunately other distros..
The other news I did read is that Linux wants to stop supporting reiserfs in the kernel.
Now Reiserfs is a magnitude more reliable than what distros come with these days
I have it on several systems and many backups ..
Never a problem, ext4 has a lot of strange problems.
So, Unix, the old Unix , .. it was good.
I have SLS Linux from 1998 on a floppy somewhere..


Kiddies now add their own take on it while not understanding the basics IMO.
Same for a lot of other software and libraries.
So a good clean Unix must be somewhere to be found.

Do they drop reiserfs because the poor guy is in jail?
Politics into technology? What about people\'s backups?
Sell more storage media? Bloat sells, the more stupid mods to Linux the more
updates needed the more bandwidth needed so you have to buy into 5G and fiber...
As you can see from this reply I am all for going back to smoke signals.
Works even from the moon to earth if the smoke clouds are big enough.

And yes, as I am always root, I once typed rm * -rf by accident while in /
was quick to stop with ctrlC but hat to put back some X libraries...

Do not try this at home.. ;-)

I was at work.

Yea, sorry ;-)

 

[piglet]

On 31/08/2023 3:31 pm, Fred Bloggs wrote:

On Thursday, August 31, 2023 at 7:03:29 AM UTC-4, piglet wrote:
On 30/08/2023 23:50, Fred Bloggs wrote:
On Wednesday, August 30, 2023 at 5:11:38 PM UTC-4, piglet wrote:
Pardon a post about electronic design. In a recent Dave Jones EEVBlog
video nr 1561 ...

https://www.youtube.com/watch?v=FKC6lnTKH-c

... he shows his USB powered variable bench power supply. At about 28:10
the regulator schematic is shown. I am baffled why he uses two series
pass transistors; one for current, one for voltage. I cannot recall ever
seen the functionality split into two power transistors in series
before. Is that laziness or brilliance? Does anyone have ideas about the
design reasoning?

They use similar in HP benchtops. It\'s necessary if you want a precision programmable current limit. Until it limits, it\'s just a saturated FET putting most of the voltage on the voltage regulation FET.


Diode D4 in the voltage regulator does not inspire confidence.

Note on schematic says D4 prevents interference by input diodes ( of error opamp ) on the voltage measurement, meaning at R22+ R23 junction. If I could read the pn of the op amp it would be clear, but it\'s not readable as is. Maybe it\'s some quasi-Norton amp. Whatever it is the input is acting on about a diode drop below the voltage at external (+) input.

Schematic symbols of resistors look like inductors. FETs look like RJ-45 connectors.


piglet
The opamp is OPA2180 zero drift/auto zero type with inverse parallel
input protection diodes. D4 looks mega kludgy - if it were mine I\'d
investigate using a different opamp or taking the error signal from the
voltage follower used for metering. Still I guess in real life the
voltage drop error across D4 at nanoamp opamp input bias current is near
negligible?

None of what he has makes any sense, neither D4 nor that 2kR resistor. What\'s probably going on is he\'s hitting the Vset input with too large a step from his D/A, and that\'s causing the differential clamp diodes to conduct and produce screwy outputs until the loop recovers. He should limit his steps to say 50mV and allow settling time of 5x of what he\'s pre-measured before applying another one. Get rid of D4 and the 2k.



piglet

I don\'t think it is do with settling time or steps - simply that when
the output is shorted (or otherwise much lower than the set output
voltage) then the op-amp clamp diodes will let Vset from the DAC flow
into the voltage divider measurement node and cause the wrong output
voltage to be displayed. What is the 2kR resistor you refer to? I see
R23 which is 10Megaohm.

The D4 kludge is truly appalling.

piglet