Interesting GaN power part: A3G26D055N

[Phil Hobbs]

125V VDSmax, asymmetric dual depletion-mode GaN FET in common-source
connection.

It's characterized for 100 MHz to 2690 MHz, which probably means no
internal matching network, so it might be good for fast switching.

Weirdly it has no I_Dmax rating.

$22 at Richardson

Cheers

Phil Hobbs


--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

[jla...@highlandsniptechnology.com]

On Wed, 15 Sep 2021 11:55:52 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>125V VDSmax, asymmetric dual depletion-mode GaN FET in common-source
>connection.
>
>It's characterized for 100 MHz to 2690 MHz, which probably means no
>internal matching network, so it might be good for fast switching.
>
>Weirdly it has no I_Dmax rating.
>
>$22 at Richardson
>
>Cheers
>
>Phil Hobbs

Interesting, but the specs are terrible. No DC curves, no Idss, no
Imax, no capacitances, and a diagram that pretends that it's two
amplifiers.

I wonder where the 100 MHz limit comes from.

Biasing is the usual RF nonsense: play with it until it works.



--

Father Brown's figure remained quite dark and still;
but in that instant he had lost his head. His head was
always most valuable when he had lost it.

[Joe Gwinn]

On Wed, 15 Sep 2021 09:29:01 -0700, jla...@highlandsniptechnology.com
wrote:

>On Wed, 15 Sep 2021 11:55:52 -0400, Phil Hobbs
><pcdhSpamM...@electrooptical.net> wrote:
>
>>125V VDSmax, asymmetric dual depletion-mode GaN FET in common-source
>>connection.
>>
>>It's characterized for 100 MHz to 2690 MHz, which probably means no
>>internal matching network, so it might be good for fast switching.
>>
>>Weirdly it has no I_Dmax rating.
>>
>>$22 at Richardson
>>
>>Cheers
>>
>>Phil Hobbs
>
>Interesting, but the specs are terrible. No DC curves, no Idss, no
>Imax, no capacitances, and a diagram that pretends that it's two
>amplifiers.
>
>I wonder where the 100 MHz limit comes from.
>
>Biasing is the usual RF nonsense: play with it until it works.

The key is that it is intended for implementation of a Doherty
Amplifier, which are widely used in linear RF final amplifier stages
because it's far more efficient with spikey data comm waveforms (which
look like Gaussian noise) than class A or B.

.<https://en.wikipedia.org/wiki/Doherty_amplifier>


Joe Gwinn

[Jan Panteltje]

On a sunny day (Wed, 15 Sep 2021 13:16:27 -0400) it happened Joe Gwinn
<joeg...@comcast.net> wrote in <r9a4kglapomav14rd...@4ax.com>:

https://www.microwaves101.com/encyclopedias/doherty-amplifiers
scroll down for nice board!

[John Larkin]

On Wed, 15 Sep 2021 13:16:27 -0400, Joe Gwinn <joeg...@comcast.net>

Doherty amps use DC too. They need to be biased.

I'm not a fan of s-params, but they aren't on the data sheet either.
Doherty amps are pretty nonlinear, so a Spice model would be nice.

RF people seem to be allergic to Spice.

[Phil Hobbs]

jla...@highlandsniptechnology.com wrote:
> On Wed, 15 Sep 2021 11:55:52 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> 125V VDSmax, asymmetric dual depletion-mode GaN FET in common-source
>> connection.
>>
>> It's characterized for 100 MHz to 2690 MHz, which probably means no
>> internal matching network, so it might be good for fast switching.
>>
>> Weirdly it has no I_Dmax rating.
>>
>> $22 at Richardson

>

> Interesting, but the specs are terrible. No DC curves, no Idss, no
> Imax, no capacitances, and a diagram that pretends that it's two
> amplifiers.
>
> I wonder where the 100 MHz limit comes from.
>
> Biasing is the usual RF nonsense: play with it until it works.

It is two amplifiers--it's intended for a boosted class-B application,
sort of like a Class G. The amps are in fact dual depletion FETs with
different characteristics wired common-source. (The pad is the source
contact.)

[Gerhard Hoffmann]

Am 15.09.21 um 20:33 schrieb John Larkin:

> Doherty amps use DC too. They need to be biased.

> I'm not a fan of s-params, but they aren't on the data sheet either.
> Doherty amps are pretty nonlinear, so a Spice model would be nice.

> RF people seem to be allergic to Spice.

Power-up/down sequence is described in detail.

Spice means "Simulation Program with Integrated Circuit Emphasis"

When you do boards, then things like transmission lines already require
hacks to work at all. Pin diodes do not, since SPICE does not know the
concept of carrier lifetime.

On page 10/11 they say they have a s2p file. These are 2 port
s-parameters in Touchstone format, probably multiple by operating point.

Input return loss of these devices is probably negative at low
frequencies. IE you put 1 mW in and 100 mW come back. That is no
fun to stabilize and do something you want at the same time.

We once tried this with early Agilent phemts for use at 144 MHz.
When we had that thing stable, noise factor etc was uninteresting.
On 432 MHz all it took was some inductive source degeneration.

Maybe they'll have an ADS device kit, but buying Keysight Advanced
Design System with the proper options costs an arm and a leg.


Gerhard

[Tom Gardner]

On 15/09/21 19:33, John Larkin wrote:
> I'm not a fan of s-params, but they aren't on the data sheet either.
> Doherty amps are pretty nonlinear, so a Spice model would be nice.
>
> RF people seem to be allergic to Spice.

They work in the frequency domain, not the time domain, and use
different types of simulator to assess the non-linear effects,
e.g. Harmonic Balance Simulators

From http://literature.cdn.keysight.com/litweb/pdf/ads2004a/pdf/adshbapp.pdf
Harmonic balance is a highly accurate frequency-domain analysis
technique for obtaining the steady state solution of nonlinear
circuits and systems. It is usually the method of choice for
simulating analog RF and microwave problems that are most
naturally handled in the frequency domain. Once the steady
state solution is calculated, the harmonic balance simulator
can be used to do the following.
1. Compute quantities such as third-order intercept (TOI) points,
total harmonic distortion (THD), and inter-modulation distortion
components.
2. Perform power amplifier load-pull contour analyses.
3. Perform nonlinear noise analyses.
The harmonic balance method assumes that the input stimulus
consists of a few steady-state sinusoids. Therefore the solution
is a sum of steady state sinusoids that includes the input
frequencies in addition to any significant harmonics or mixing
terms.

[John Larkin]

On Wed, 15 Sep 2021 21:24:47 +0200, Gerhard Hoffmann <dk...@arcor.de>
wrote:

>Am 15.09.21 um 20:33 schrieb John Larkin:
>
>
>> Doherty amps use DC too. They need to be biased.
>
>> I'm not a fan of s-params, but they aren't on the data sheet either.
>> Doherty amps are pretty nonlinear, so a Spice model would be nice.
>
>> RF people seem to be allergic to Spice.
>
>Power-up/down sequence is described in detail.
>
>Spice means "Simulation Program with Integrated Circuit Emphasis"
>
>When you do boards, then things like transmission lines already require
>hacks to work at all. Pin diodes do not, since SPICE does not know the
>concept of carrier lifetime.

LT Spice seems to do a reasonable job with reverse recovery on various
diodes, if maybe a bit rectangular about the snap-off. It doesn't have
any PINS in its library.

Spice has lossy and lossless tx lines. I use them a lot.

[John Larkin]

On Wed, 15 Sep 2021 15:21:55 -0400, Phil Hobbs

<pcdhSpamM...@electrooptical.net> wrote:

>jla...@highlandsniptechnology.com wrote:
>> On Wed, 15 Sep 2021 11:55:52 -0400, Phil Hobbs
>> <pcdhSpamM...@electrooptical.net> wrote:
>>
>>> 125V VDSmax, asymmetric dual depletion-mode GaN FET in common-source
>>> connection.
>>>
>>> It's characterized for 100 MHz to 2690 MHz, which probably means no
>>> internal matching network, so it might be good for fast switching.
>>>
>>> Weirdly it has no I_Dmax rating.
>>>
>>> $22 at Richardson
>
>>
>> Interesting, but the specs are terrible. No DC curves, no Idss, no
>> Imax, no capacitances, and a diagram that pretends that it's two
>> amplifiers.
>>
>> I wonder where the 100 MHz limit comes from.
>>
>> Biasing is the usual RF nonsense: play with it until it works.
>
>
>It is two amplifiers--it's intended for a boosted class-B application,
>sort of like a Class G. The amps are in fact dual depletion FETs with
>different characteristics wired common-source. (The pad is the source
>contact.)
>
>Cheers
>
>Phil Hobbs

Given a depletion fet, I would have expected Idss to be supplied.

[Phil Hobbs]

Gerhard Hoffmann wrote:
> Am 15.09.21 um 20:33 schrieb John Larkin:
>
>
>> Doherty amps use DC too. They need to be biased.
>
>> I'm not a fan of s-params, but they aren't on the data sheet
>> either. Doherty amps are pretty nonlinear, so a Spice model would
>> be nice.
>
>> RF people seem to be allergic to Spice.
>
> Power-up/down sequence is described in detail.
>
> Spice means "Simulation Program with Integrated Circuit Emphasis"
>
> When you do boards, then things like transmission lines already
> require hacks to work at all. Pin diodes do not, since SPICE does not
> know the concept of carrier lifetime.
>
> On page 10/11 they say they have a s2p file. These are 2 port
> s-parameters in Touchstone format, probably multiple by operating
> point.
>
> Input return loss of these devices is probably negative at low
> frequencies.

With a conjugate match, probably so, but not in a 50-ohm system surely?
The abs max gate current is only 10 mA or something.

> IE you put 1 mW in and 100 mW come back. That is no fun to stabilize
> and do something you want at the same time.

A 2N3904 emitter follower will oscillate if you run it hot enough and
connect the base to AC ground. (An RC + emitter follower was a popular
power-on reset circuit in the early personal computer days, till folks
found out how badly it oscillated.)

>
> We once tried this with early Agilent phemts for use at 144 MHz. When
> we had that thing stable, noise factor etc was uninteresting. On 432
> MHz all it took was some inductive source degeneration.
>
> Maybe they'll have an ADS device kit, but buying Keysight Advanced
> Design System with the proper options costs an arm and a leg.

Well, it occurred to me that a 125-V, 8W FET that had 18 dB gain at 3
GHz and worked down to below 100 MHz might be quite an interesting pulse
output stage. ;)

[Phil Hobbs]

One would, wouldn't one. Or at least I_Dmax, but no.

Hey, it's only 21 bucks, cut them some slack. ;)

[John Larkin]

On Wed, 15 Sep 2021 18:58:59 -0400, Phil Hobbs

I suppose that the RF boys want to sell billions of parts, and assume
they will do a lot of hand-holding for a few big buyers.

They don't want our kind.

[Phil Hobbs]

Get in touch with your inner grad student. ;)

[John Larkin]

That bit me once, using a 2N2219 as the logic reset on a military
project. We didn't know what the heck was going on.

My latest instant-on Colpitts oscillator, with a SAV541,
has a ** 499 ohm ** series gate resistor.

>
>>
>> We once tried this with early Agilent phemts for use at 144 MHz. When
>> we had that thing stable, noise factor etc was uninteresting. On 432
>> MHz all it took was some inductive source degeneration.
>>
>> Maybe they'll have an ADS device kit, but buying Keysight Advanced
>> Design System with the proper options costs an arm and a leg.
>
>Well, it occurred to me that a 125-V, 8W FET that had 18 dB gain at 3
>GHz and worked down to below 100 MHz might be quite an interesting pulse
>output stage. ;)

The EPC GaN fets are nice and dirt cheap.

https://www.dropbox.com/s/izvf0thpe5ddty9/T577B_Vp50.JPG?raw=1

50 volts, 50 ohms, totem pole.

https://www.dropbox.com/s/t1whjfnrtkn2ucx/T577s_P500D.jpg?raw=1

[Phil Hobbs]

Love the AliExpress blob-top construction. ;)

(Yes, I know why you're doing it, but no, I couldn't pass that one up.)

[John Larkin]

On Wed, 15 Sep 2021 20:15:49 -0400, Phil Hobbs

The EPC parts need physical protection. We had adventures finding a
glob-top that didn't eventually break the BGAs. We had failures that I
hoped weren't a circuit design problem; they weren't. The epoxy was
cracking them.

The stuff we're using now is a medium-hard-set 1-part thing that cures
at 85c.

[Phil Hobbs]

How did that wind up working?

>
>>
>>>
>>> We once tried this with early Agilent phemts for use at 144 MHz.
>>> When we had that thing stable, noise factor etc was
>>> uninteresting. On 432 MHz all it took was some inductive source
>>> degeneration.
>>>
>>> Maybe they'll have an ADS device kit, but buying Keysight
>>> Advanced Design System with the proper options costs an arm and a
>>> leg.
>>
>> Well, it occurred to me that a 125-V, 8W FET that had 18 dB gain at
>> 3 GHz and worked down to below 100 MHz might be quite an
>> interesting pulse output stage. ;)
>
>
> The EPC GaN fets are nice and dirt cheap.
>
> https://www.dropbox.com/s/izvf0thpe5ddty9/T577B_Vp50.JPG?raw=1
>
> 50 volts, 50 ohms, totem pole.
>
> https://www.dropbox.com/s/t1whjfnrtkn2ucx/T577s_P500D.jpg?raw=1

I see the P500 is on your web site--it's nice looking, and a very decent
advance, especially in the trigger jitter department. Factor-of-2
improvements don't come easily once you're down in the 20-ps
neighbourhood, especially when vendors pull the rug out by discontinuing
all their pHEMT parts. :(

(I still secretly like VFDs though.) ;)

[Phil Hobbs]

Too-hard epoxy is a famous problem in optics and optomechanics as
well--at low temperatures it's liable to rip chunks out of lenses,
squash glass bead thermistors so they become intermittent, and destroy
things like pots and electrolytic caps.

For electronics potting, a bottom layer of silicone helps a lot, but a
glob-topped BGA is more like a lens assembly--I can believe that soft
epoxy is the ticket.
h

[Gerhard Hoffmann]

Am 16.09.21 um 00:43 schrieb Phil Hobbs:

> Gerhard Hoffmann wrote:
>> Am 15.09.21 um 20:33 schrieb John Larkin:
>>
>>
>>> Doherty amps use DC too. They need to be biased.
>>
>>> I'm not a fan of s-params, but they aren't on the data sheet
>>> either. Doherty amps are pretty nonlinear, so a Spice model would
>>> be nice.
>>
>>> RF people seem to be allergic to Spice.
>>
>> Power-up/down sequence is described in detail.
>>
>> Spice means "Simulation Program with Integrated Circuit Emphasis"
>>
>> When you do boards, then things like transmission lines already
>> require hacks to work at all. Pin diodes do not, since SPICE does not
>> know the concept of carrier lifetime.
>>
>> On page 10/11 they say they have a s2p file. These are 2 port
>> s-parameters in Touchstone format, probably multiple by operating
>> point.
>>
>> Input return loss of these devices is probably negative at low
>> frequencies.
>
> With a conjugate match, probably so, but not in a 50-ohm system surely?

Neg. return loss = S11 means that in the 50 Ohm system there comes more
out than gets in.

>  The abs max gate current is only 10 mA or something.
>
>> IE you put 1 mW in and 100 mW come back. That is no fun to stabilize
>> and do something you want at the same time.
>
> A 2N3904 emitter follower will oscillate if you run it hot enough and
> connect the base to AC ground.  (An RC + emitter follower was a popular
> power-on reset circuit in the early personal computer days, till folks
> found out how badly it oscillated.)

Connecting the base to AC ground is not enough. It takes an inductance.
If you measure into the base/gate of a capacitively loaded follower with
a network analyzer, you see a smallish capacitance with a negative
resistor in series. It takes some inductor to create a resonance
for the oscillation frequency.
And the gate stopper resistor may need to be surprisingly large to
overcompensate the negative resistance for sure. Bad for noise
performance.

(Note to selves: what is the resulting Vnoise of +1KOhm and -1KOhm
in series in Spice? Is it ~0 by optimizing the resistors away?)


------
That is my pet problem: A usual low noise amplifier for baseband
has a FET with some small source resistor, optional cascode and
then an op amp with feedback into the source.

With the feedback closed, the FET is no longer CS because the
source follows the gate quite precisely. It is off just by ~loop gain.
The FET thinks ist works as a follower. The 0.2 Ohm source resistor
gives a wrong impression.

The low drain impedance of the cascode only completes the world as
perceived by the input FET: I'm a follower!

But the feedback is late through the op amp. The source voltage
is like being capacitively loaded. ---> the Fet generates negative
Rin like cap. loaded followers like to do.

If you spice the circuit in AOE3, page 152 and plot
re(Vgenerator/Igenerator), you'll see the negative real part of
the input impedance over some frequency range.
Vgenerator is the voltage source at the input.

I did not find a good solution. Replacing the op amp with a VCVS
works nicely but it is hard to solder. A THS4303 3 GHz op amp is
fast enough, but has an astronomic 1/f noise that dominates despite
the fist stage.

Compensating everything to death works, but I want at least 1 MHz BW.
I finally left the loop open. That is surprisingly stable over temp.
I'm less than happy because I like to enforce the operating conditions.

Any ideas?
------


> Cheers

Gerhard

BTW: Where is Win? Last time he suddenly disappeared was during
the preparation of AOE3. Can we expect a new book? I hope he's well!

[jla...@highlandsniptechnology.com]

On Wed, 15 Sep 2021 20:50:52 -0400, Phil Hobbs

Very well. Tempco is around 5 PPM/K and jitter is around 5 ps RMS per
microsecond before we phase-lock it. One part in 200K is good in this
game.

I gave up on the driven guard... too many side effects. I just cut out
the planes below the critical stuff to reduce the FR4 capacitance.

We are shipping these now... finally.

http://www.highlandtechnology.com/DSS/P500DS.shtml

>>
>>>
>>>>
>>>> We once tried this with early Agilent phemts for use at 144 MHz.
>>>> When we had that thing stable, noise factor etc was
>>>> uninteresting. On 432 MHz all it took was some inductive source
>>>> degeneration.
>>>>
>>>> Maybe they'll have an ADS device kit, but buying Keysight
>>>> Advanced Design System with the proper options costs an arm and a
>>>> leg.
>>>
>>> Well, it occurred to me that a 125-V, 8W FET that had 18 dB gain at
>>> 3 GHz and worked down to below 100 MHz might be quite an
>>> interesting pulse output stage. ;)
>>
>>
>> The EPC GaN fets are nice and dirt cheap.
>>
>> https://www.dropbox.com/s/izvf0thpe5ddty9/T577B_Vp50.JPG?raw=1
>>
>> 50 volts, 50 ohms, totem pole.
>>
>> https://www.dropbox.com/s/t1whjfnrtkn2ucx/T577s_P500D.jpg?raw=1
>
>I see the P500 is on your web site--it's nice looking, and a very decent
>advance, especially in the trigger jitter department. Factor-of-2
>improvements don't come easily once you're down in the 20-ps
>neighbourhood, especially when vendors pull the rug out by discontinuing
>all their pHEMT parts. :(

Actually, comparing typicals the old P400 was a bit better. When we
get some time we plan to find out why.

This one can install a new set of delay and width values every
trigger. So it can do time sweeps and such. And you can fiddle the
spinner knob, to change a delay or a width, and it never misses a
trigger or gives weird pulses.

>
>(I still secretly like VFDs though.) ;)

Terrible thing to confess, in public.

>
>Cheers
>
>Phil Hobbs

[jla...@highlandsniptechnology.com]

On Thu, 16 Sep 2021 03:08:30 +0200, Gerhard Hoffmann <dk...@arcor.de>
wrote:

On leaded parts, the leads and wire bonds are enough.

Power mosfets like to oscillate too.

[Steve Wilson]

Phil Hobbs <pcdhSpamM...@electrooptical.net> wrote:

[...]

> A 2N3904 emitter follower will oscillate if you run it hot enough and
> connect the base to AC ground. (An RC + emitter follower was a popular
> power-on reset circuit in the early personal computer days, till folks
> found out how badly it oscillated.)

Just about any small-signal bipolar will oscillate in that configuration.
Another example is a capacitance multiplier, or any circuit with a long
wire connected to the base. It makes a classic Colpitts, which is a very
vigorous oscillator. It needs some inductance in the base, usually supplied
by trace inductance and the impedance of the capacitor. The feedback is
supplied by stray trace capacitance, the capacitance from base to emittor,
and any capacitance from emitter to ground. The circuit can also oscillate
in a cascode configuration with a common-emitter stage feeding into a
common-base stage, where the feedback is a complex picture involving both
transistors.

The cure, as is well known, is a small resistance in the base right at the
base connection, or a ferrite bead. The bead is preferred when you need the
absolute lowest input noise.

There are a number of ways to detect the oscillation. A scope probe or
small screwdriver on a sensitive node may kill the oscillation. A small
coil of wire driving a 1N34 point contact germanium diode may generate
enough voltage to detect the oscillation. A spectrum analyzer with enough
bandwidth can easily detect oscillations, especially when there are more
than one. Waving a hand near the circuit may change a voltage or current
enough to show the circuit is oscillating. The circuit may break into
oscillation at some point in the operating cycle and cause distortion in
the output.

A sensible precaution is to assume any wide bandwidth circuit is going to
oscillate and to add something in the base during layout. That way you are
not caught with the problem of how to add it after production has started.

Of course, in this day and age, I'm merely preaching to the choir. However,
it is still surprising to see a post where someone has neglected to add
some base impedance and has no room in a tight layout to add it later.

[...]

> Cheers
>
> Phil Hobbs




--
The best ideas occur in the theta state. - sw

[Phil Hobbs]

I'm pretty sure you can get negative Rin due to device behaviour alone
on account of having two lags in the loop. I'm cranking away on an
expert report for a Friday deadline just now, but I'll go look it up in
Hollister over the weekend if I remember.

> And the gate stopper resistor may need to be surprisingly large to
> overcompensate the negative resistance for sure. Bad for noise
> performance.
>
> (Note to selves: what is the resulting Vnoise of +1KOhm and -1KOhm in
> series in Spice? Is it ~0 by optimizing the resistors away?)

Unfortunately, despite the square root negative resistors don't have
imaginary noise. ;)

> ------ That is my pet problem: A usual low noise amplifier for
> baseband has a FET with some small source resistor, optional cascode
> and then an op amp with feedback into the source.
>
> With the feedback closed, the FET is no longer CS because the source
> follows the gate quite precisely. It is off just by ~loop gain. The
> FET thinks ist works as a follower. The 0.2 Ohm source resistor gives
> a wrong impression.
>
> The low drain impedance of the cascode only completes the world as
> perceived by the input FET: I'm a follower!

I normally use local feedback around the FET, either with a fast PNP
wraparound plus a fixed tail current source, or else the NPN version,
where the drain of the FET feeds back to the base of the tail current
source (the White cathode follower idea).

Bootstrapping the drain of the FET helps a fair amount. It's possible
to sneak a BFP640 bootstrap inside the local feedback loop (PNP
wraparound or White-style follower) if you're careful, but things get
less stable as you do that. It's also possible to bootstrap the whole
FET + PNP local loop.

>
> But the feedback is late through the op amp. The source voltage is
> like being capacitively loaded. ---> the Fet generates negative Rin
> like cap. loaded followers like to do.

I generally rely on the local feedback for the high-frequency
performance. One of my bootstrap follower gizmos has a gain of 0.9997
near DC, and > 0.999 out to a megahertz or so. (I have no idea how to
measure that directly--those numbers come from looking at bandwidth as a
function of the bootstrapped capacitance.

The reason I care about getting so close to A_V = 1.0000 is the
time-domain performance. If the FET stage has 0.8 nV noise in 1 Hz and
the op amp has 12 or so (LM6171, a fave), then there's no big noise
advantage to gains much above 1-(0.5 0.8 / 12) = 0.97. However, any
bootstrap produces a closely-spaced pole/zero pair that don't exactly
cancel, and that causes settling whoopdedoos at late times. From that
point of view, 0.9997 is 100 times better than 0.97. (Win and I had a
disagreement about this long ago in this very boutique.)

> If you spice the circuit in AOE3, page 152 and plot
> re(Vgenerator/Igenerator), you'll see the negative real part of the
> input impedance over some frequency range. Vgenerator is the voltage
> source at the input.
>
> I did not find a good solution. Replacing the op amp with a VCVS
> works nicely but it is hard to solder. A THS4303 3 GHz op amp is fast
> enough, but has an astronomic 1/f noise that dominates despite the
> fist stage.

The SiGe:C BJTs have pretty low 1/f corners as well as high betas, so
they may be a win at very low frequency. I measured the DC paramaeters
of a few BFP640H and BFP780 devices on my HP 4145B awhile back, and
found betas around 200-300 with AC Early voltages near 250V. (The DC
Early voltage measurements in the datasheets are corrupted by the
effects of device heating--some even show negative collector resistances!)

Beta holds up very well at low collector current in these gizmos, so
maybe you can find a bias setting that gets you enough bandwidth and low
enough noise. The BFP780 is a medium-power 20 GHz transistor, so R_bb'
should be small.

The noise temperature of an ideal BJT emitter is T_J/2, so its 1-Hz
voltage noise ought to be 1 nV at a collector current of about 100 uA.
For the BFP650FHs I measured, that corresponds to a base current of
right around 315 nA.

> Compensating everything to death works, but I want at least 1 MHz
> BW. I finally left the loop open. That is surprisingly stable over
> temp. I'm less than happy because I like to enforce the operating
> conditions.
>
> Any ideas?

A few. ;)

One approach is to use a chopamp to keep the bias conditions constant.
You do have to watch out for its noise, of course, but a lot of the time
you can synchronize it with the measurement.

> BTW: Where is Win? Last time he suddenly disappeared was during the
> preparation of AOE3. Can we expect a new book? I hope he's well!

Rob Legg said some months ago that Win was well, but busy.

We seem to have lost a lot of the positive-SNR contributors lately--Win,
James Arthur, Jim Thompson (whose SNR was somewhat variable) and Spehro
come to mind, and of course Joerg is spending his time working off all
those beer calories on his mountain bike. Woodgate apparently got
hacked off about the incivility here and sloped off to the LTspice group
some years ago. I even miss Miso, believe it or not.

At this point SED seems to be all corona, all the time. A pity.

Cheers

Phil

[Jan Panteltje]

On a sunny day (Wed, 15 Sep 2021 20:57:07 -0400) it happened Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote in
<shu4p4$rj2$1...@gioia.aioe.org>:

>
>Too-hard epoxy is a famous problem in optics and optomechanics as
>well--at low temperatures it's liable to rip chunks out of lenses,
>squash glass bead thermistors so they become intermittent, and destroy
>things like pots and electrolytic caps.
>
>For electronics potting, a bottom layer of silicone helps a lot, but a
>glob-topped BGA is more like a lens assembly--I can believe that soft
>epoxy is the ticket.

I did some with hotglue.
Was not RF though, no idea, but it stays a bit softlike.

[Dmitriy Pshonkin]

> I normally use local feedback around the FET, either with a fast PNP
> wraparound plus a fixed tail current source, or else the NPN version,
> where the drain of the FET feeds back to the base of the tail current
> source (the White cathode follower idea).

Unipolar power supply ada4807 + cph3910,
the output of the op-amp is connected to the source of jfet, the voltage at this pin is 0.7 Volts in operating mode.
The name of the bootstrap circuit for a photodiode as a voltage follower is not entirely correct - the signal source is connected to the gate-source and the circuit works as a cascade with a common source and corresponding amplification.

https://ixbt.photo/?id=photo:1441128

[Phil Hobbs]

For a one-off, I'd do that in a heartbeat. In production, something
with zero volatiles and an actual relevant datasheet is comforting. ;)

[Phil Hobbs]

Sure JFET follower + op amp is a useful configuration, especially since
it allows you to use some hairy-chested CFA to get better slew rate
without trashing the noise too badly. Most of my high-performance
preamps are for photodiodes of one kind or another, which is why I was
talking about bootstraps specifically.

Now that op amps all have horrible input capacitance (despite what their
lying sleazebag datasheets will tell you), I often build bootstraps that
way too. Really helps the bandwidth and e_N*C noise at high Z, at the
price of probably worse 1/f noise and definitely worse offset and drift.
It gets really old when your bootstrap is beautiful but the op amp
trips over its own big feet. ;)

Back in the BF862 days I used to servo I_D to exactly I_DSS using an
auxiliary op amp, which got rid of the drift very neatly. You can do
reasonably well with a CPH3910 because its zero-drift current is some
reasonable value like 15 mA.

Putting the bootstrap inside the loop also lets you use amps such as the
very nice LM6171, which is a CFA with a built-in buffer driving the
noninverting input. I have no idea how they managed to get such nice
symmetrical inputs doing that--good CMR, good Zin, and so on--but the
price you pay is 10-nV noise. (In 1 Hz, natch.)

Ten nanovolts is no big worry in at TIA with > 20k feedback resistance,
as long as it isn't getting multiplied by the e_N*C mechanism, so the
combination is a very good one. It's especially useful in optical
stuff, where you often don't know what sort of nasty fast laser pulse is
going into that poor photodiode.

That puts a big premium on high slew rate, even if the application
itself isn't that wideband--once the amp loses control of its input, the
measurement goes straight into the tank.

Cheers

Phil Hobbs

[Phil Hobbs]

Gerhard Hoffmann wrote:
> Am 16.09.21 um 00:43 schrieb Phil Hobbs:
>> Gerhard Hoffmann wrote:
>>> Am 15.09.21 um 20:33 schrieb John Larkin:
>>>
>>>
>>>> Doherty amps use DC too. They need to be biased.
>>>
>>>> I'm not a fan of s-params, but they aren't on the data sheet
>>>> either. Doherty amps are pretty nonlinear, so a Spice model would
>>>> be nice.
>>>
>>>> RF people seem to be allergic to Spice.
>>>
>>> Power-up/down sequence is described in detail.
>>>
>>> Spice means "Simulation Program with Integrated Circuit Emphasis"
>>>
>>> When you do boards, then things like transmission lines already
>>> require hacks to work at all. Pin diodes do not, since SPICE does not
>>> know the concept of carrier lifetime.
>>>
>>> On page 10/11 they say they have a s2p file. These are 2 port
>>> s-parameters in Touchstone format, probably multiple by operating
>>> point.
>>>
>>> Input return loss of these devices is probably negative at low
>>> frequencies.
>>
>> With a conjugate match, probably so, but not in a 50-ohm system surely?
>
> Neg. return loss = S11 means that in the 50 Ohm system there comes more
> out than gets in.

Sure. But when the feedback is via some very high-Z RC, which it is in
a GaN FET, the power coming out the input depends on the termination.

One of the good things about GaN FETs in general is their tiny C_DG. I
haven't used them so far, but pHEMTs are electrically similar and are
really amazingly stable, so by analogy I'd expect GaN parts to be like
that too.

>>   The abs max gate current is only 10 mA or something.
>>
>>> IE you put 1 mW in and 100 mW come back. That is no fun to stabilize
>>> and do something you want at the same time.

Do you have experience with HFETs showing that behaviour? The thing
only has 20 dB of gain at high frequency, so it would have to be wasting
it all driving its own input!

Cheers

Phil Hobbs

(Omitted this point in my earlier reply.)

[Phil Hobbs]

[jla...@highlandsniptechnology.com]

On Thu, 16 Sep 2021 08:58:34 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>Jan Panteltje wrote:
>> On a sunny day (Wed, 15 Sep 2021 20:57:07 -0400) it happened Phil Hobbs
>> <pcdhSpamM...@electrooptical.net> wrote in
>> <shu4p4$rj2$1...@gioia.aioe.org>:
>>
>>>
>>> Too-hard epoxy is a famous problem in optics and optomechanics as
>>> well--at low temperatures it's liable to rip chunks out of lenses,
>>> squash glass bead thermistors so they become intermittent, and destroy
>>> things like pots and electrolytic caps.
>>>
>>> For electronics potting, a bottom layer of silicone helps a lot, but a
>>> glob-topped BGA is more like a lens assembly--I can believe that soft
>>> epoxy is the ticket.
>>
>> I did some with hotglue.
>> Was not RF though, no idea, but it stays a bit softlike.
>
>For a one-off, I'd do that in a heartbeat. In production, something
>with zero volatiles and an actual relevant datasheet is comforting. ;)
>
>Cheers
>
>Phil Hobbs

And easy to dispense without long strings.

[John Larkin]

Here's an xray of a SAV541. The wire bonds are long.

https://www.dropbox.com/s/qttp09vpw3kcee3/SAV541_Xray.jpg?raw=1

The EPC BGA parts don't have wirebonds, but are too big for really
fast stuff.



If a man will begin with certainties, he shall end with doubts,
but if he will be content to begin with doubts he shall end in certainties.
Francis Bacon

[Gerhard Hoffmann]

Am 16.09.21 um 15:27 schrieb Phil Hobbs:

> Gerhard Hoffmann wrote:
>> Am 16.09.21 um 00:43 schrieb Phil Hobbs:

>>>> On page 10/11 they say they have a s2p file. These are 2 port
>>>> s-parameters in Touchstone format, probably multiple by operating
>>>> point.
>>>>
>>>> Input return loss of these devices is probably negative at low
>>>> frequencies.
>>>
>>> With a conjugate match, probably so, but not in a 50-ohm system surely?
>>
>> Neg. return loss = S11 means that in the 50 Ohm system there comes more
>> out than gets in.
>
> Sure.  But when the feedback is via some very high-Z RC, which it is in
> a GaN FET, the power coming out the input depends on the termination.

The power comes probably from the source via Cgss which is large.
And for oscillation you don't need much. Just more than you did
put in.

> One of the good things about GaN FETs in general is their tiny C_DG.  I
> haven't used them so far, but pHEMTs are electrically similar and are
> really amazingly stable, so by analogy I'd expect GaN parts to be like
> that too.
>>>   The abs max gate current is only 10 mA or something.
>>>
>>>> IE you put 1 mW in and 100 mW come back. That is no fun to stabilize
>>>> and do something you want at the same time.
>
> Do you have experience with HFETs showing that behaviour?  The thing
> only has 20 dB of gain at high frequency, so it would have to be wasting
> it all driving its own input!

Not with PHEMTs, but with ordinary FETs and BJTs.
Well 100 mW is probably exaggerated with 2V/10 mA on the drain or so.
:-) At low enough frequencies, the gain is breathtaking.

This here is is S11 of 1 or 2 lame IF3602. The Circuit is much like
the one from AOE3.

<
https://www.flickr.com/photos/137684711@N07/34701106245/in/album-72157662535945536/
>

In the middle of the biggest non-dotted circle where the 1 can be seen
is the point "50 Ohms real". Where the 2 is: 100 Ohm real.
Where the 0.5 is: 25 Ohm real. On the left side at the 0 ( non-dotted)
is the short; on the right end is infinity. Everything is scaled to
50 Ohms by convention, but it would work for 75 Ohms just as well.
Everything on the horizontal axis is a real resistance.

The points in the upper half have a inductance in series, the points
in the lower half have a capacitor in series.
If you go far away from 50 Ohms you leave the sweet spot of a network
analyzer. Then you are better served by an impedance analyzer.

Left from 0 and the rest outside the unit circle is the area of negative
resistance. That can only happen with active networks since someone
has to deliver the energy.

The trace of the input S11 of the IF3602 amplifier does exactly that.
The sweep starts at 150 KHz (with full specs @ 300 KHz- 8 GHz) and
at Marker 1 we are already out of the passive circle.

The markers can compute the impedance from S11 (pronounced S-one-one)
and we see -144 Ohms and 1.3 nF at 250 KHz. Without that feature it
would be food for the pocket calculator.

(In my case, Go41C for Android. The man who was my boss for the
diploma thesis was slightly shocked, had a déjà vue on Friday evening
in the bar seeing my cell phone. We are all HP-41 fans. :-))

Has anybody measured the voltage noise and 1/f corner for EPC2018 &
friends?

> Cheers
Gerhard

[Gerhard Hoffmann]

Am 16.09.21 um 07:01 schrieb Phil Hobbs:

> We seem to have lost a lot of the positive-SNR contributors lately--Win,
> James Arthur, Jim Thompson (whose SNR was somewhat variable) and Spehro
> come to mind, and of course Joerg is spending his time working off all
> those beer calories on his mountain bike.  Woodgate apparently got
> hacked off about the incivility here and sloped off to the LTspice group
> some years ago.  I even miss Miso, believe it or not.

And Fred Bertoli, he's badly missed, too.

Joerg is often on de.sci.electronis, probably homesick, but he would
never ever admit that. ;-)

There are others who used to read s.e.d. but never came out of the
closet. The father of the AD797 comes to mind. I know him from some
audio BBS. If that is better than here, we have seriously lost.

cheers, Gerhard

[Phil Hobbs]

You've just gotten spoiled. Back in the DMOS days you'd just have
poured on the coal and powered through. ;)

Cheers

Phil Hobbs(Also spoiled)

[John Larkin]

On Thu, 16 Sep 2021 18:27:19 -0400, Phil Hobbs

The four source bonds are interesting. The slab is not ground!

--

[Phil Hobbs]

There's no substrate conduction, which is one reason pHEMTs are so fast.
The channel is a 2-dimensional electron gas (2DEG), which greatly
reduces scattering and so greatly increases mobility. There's no
diffusion transport involved.

Pretty cute.

Cheers

Phil Hobbs

[Phil Hobbs]

We haven't heard much from George Herold lately either. C'mon back
guys, all is forgiven. ;)

Cheers

Phil Hobbs

[jla...@highlandsniptechnology.com]

Does anyone here know about wirebonds? Can anyone estimate the
inductances? I could add them to the Spice model.

[Dmitriy Pshonkin]

> The markers can compute the impedance from S11 (pronounced S-one-one)
> and we see -144 Ohms and 1.3 nF at 250 KHz. Without that feature it
> would be food for the pocket calculator.

This is not a problem if the signal source is a 2000-3000 pF photodiode with an internal series resistance of ten ohms.

I saw old circuit solutions in which the input capacitance jfet was compensated by the upper transistor in the cascode through its own drain-gate capacitance.
They did not just keep the source-drain voltage constant, but with overcompensation. But this will not reduce the noise ((
In your circuit, the IF3602 drain is fixed by an NPN transistor to GND on HF, and if you tie it to the jfet source, how will the input impedance change?

> Has anybody measured the voltage noise and 1/f corner for EPC2018 &
> friends?

I am very curious how does the EPC2038 work in the input stage of an amplifier?))

[Gerhard Hoffmann]

Am 17.09.21 um 11:02 schrieb Dmitriy Pshonkin:

That was nothing with photo diodes, just an amplifier for measuring
voltage noise.

cheers, Gerhard

[Phil Hobbs]

Straight wires are about 20 nH per inch. There's not enough inductive
coupling there to make that vary much.

Cheers

Phil Hobbs

[Steve Wilson]

Gerhard Hoffmann <dk...@arcor.de> wrote:

> Am 17.09.21 um 11:02 schrieb Dmitriy Pshonkin:

[...]

>>> Has anybody measured the voltage noise and 1/f corner for EPC2018 &
>>> friends?
>>
>> I am very curious how does the EPC2038 work in the input stage of an
>> amplifier?))
>>
>
> That was nothing with photo diodes, just an amplifier for measuring
> voltage noise.
>
> cheers, Gerhard

1. Can you tell us the input voltage noise and corner frequency of the
EPC2038?

2. How do you couple the input signal to the gate at +2.5V? See

Figure 2: Transfer Characteristics, page 2

https://datasheet.octopart.com/EPC2038-EPC-datasheet-138895975.pdf

3. What is the bandwidth of the amplifier?

4. What is the input capacitance?

[Gerhard Hoffmann]

Am 17.09.21 um 15:35 schrieb Steve Wilson:

> Gerhard Hoffmann <dk...@arcor.de> wrote:
>
>> Am 17.09.21 um 11:02 schrieb Dmitriy Pshonkin:
>
> [...]
>
>>>> Has anybody measured the voltage noise and 1/f corner for EPC2018 &
>>>> friends?
>>>
>>> I am very curious how does the EPC2038 work in the input stage of an
>>> amplifier?))
>>>
>>
>> That was nothing with photo diodes, just an amplifier for measuring
>> voltage noise.
>>
>> cheers, Gerhard
>
> 1. Can you tell us the input voltage noise and corner frequency of the
> EPC2038?

I don't know. Above I have asked if someone has done it :-)

All I have done with it up to now is making a test circuit as a chopper
and I have verified that I can solder it to the board. That is not
important, just curiosity. Soldering was easy with hot air.

> 2. How do you couple the input signal to the gate at +2.5V? See

The chopper has the drain at the input, vaguely as a voltage divider.
Gate voltage is the LVCmos switching frequency. 5V CMOS would be
quite close to killing it.
A chopper may profit from low charge injection.

As an amplifier, one could use the circuit that Win used for
measuring the transistor noise tables in AOE3. It has a NPN
to servo the gate bias. But if you have interest in 1/f and
below, you may not be able to tell the difference between
servo effects and input noise.

Gerhard

[jla...@highlandsniptechnology.com]

On Fri, 17 Sep 2021 16:22:29 +0200, Gerhard Hoffmann <dk...@arcor.de>

wrote:

>Am 17.09.21 um 15:35 schrieb Steve Wilson:
>> Gerhard Hoffmann <dk...@arcor.de> wrote:
>>
>>> Am 17.09.21 um 11:02 schrieb Dmitriy Pshonkin:
>>
>> [...]
>>
>>>>> Has anybody measured the voltage noise and 1/f corner for EPC2018 &
>>>>> friends?
>>>>
>>>> I am very curious how does the EPC2038 work in the input stage of an
>>>> amplifier?))
>>>>
>>>
>>> That was nothing with photo diodes, just an amplifier for measuring
>>> voltage noise.
>>>
>>> cheers, Gerhard
>>
>> 1. Can you tell us the input voltage noise and corner frequency of the
>> EPC2038?
>
>I don't know. Above I have asked if someone has done it :-)

I use them and have done some off-label testing. But I switch them
hard and don't care about noise.

The gates are kind of strange. At high gate voltages, there seems to
be some longterm damage mechanism slowly increasing gate current.

I'd guess that they are noisy, but numbers would be interesting.

>
>All I have done with it up to now is making a test circuit as a chopper
>and I have verified that I can solder it to the board. That is not
>important, just curiosity. Soldering was easy with hot air.
>
>
>> 2. How do you couple the input signal to the gate at +2.5V? See
>
>The chopper has the drain at the input, vaguely as a voltage divider.
>Gate voltage is the LVCmos switching frequency. 5V CMOS would be
>quite close to killing it.
>A chopper may profit from low charge injection.
>
>As an amplifier, one could use the circuit that Win used for
>measuring the transistor noise tables in AOE3. It has a NPN
>to servo the gate bias. But if you have interest in 1/f and
>below, you may not be able to tell the difference between
>servo effects and input noise.
>
>Gerhard

[Steve Wilson]

OK, Thanks.

ISTR Win used a bunch of bipolars to get the noise down to pV/root(Hz) for
ribbon microphones. This required an extremely large input coupling
capacitor to prevent shorting out the base bias and still pass the low
frequencies needed.

I am working on a different approach that DC-couples the input signal to an
op amp, or several in parallel for low noise. There is no couping capacitor
needed, and the op amp eliminates the risk of blowing out the base due to
the high surge current from a power supply. The capacitors in the feedback
network need to be insulated or covered to minimize low frequency voltage
shift due to random air currents affecting the tempco of the capacitors.
Thanks to P. Hobbs for this observation.

I am also working on high attenuation ripple filters for the op amp power
supplies. I'd appreciate your comments.

Here is the ASC file for the input stage:

Version 4
SHEET 1 1888 724
WIRE 544 32 480 32
WIRE 672 32 608 32
WIRE 480 112 480 32
WIRE 528 112 480 112
WIRE 672 112 672 32
WIRE 672 112 608 112
WIRE 720 160 576 160
WIRE 864 160 720 160
WIRE 576 176 576 160
WIRE 720 176 720 160
WIRE 864 176 864 160
WIRE 480 192 480 112
WIRE 512 192 480 192
WIRE 544 192 512 192
WIRE 640 208 608 208
WIRE 672 208 672 112
WIRE 672 208 640 208
WIRE 400 224 384 224
WIRE 544 224 400 224
WIRE 384 240 384 224
WIRE 480 256 480 192
WIRE 576 256 576 240
WIRE 784 256 576 256
WIRE 784 272 784 256
WIRE 864 272 864 256
WIRE 384 336 384 320
WIRE 480 368 480 336
WIRE 528 368 480 368
WIRE 576 368 528 368
WIRE 672 368 576 368
WIRE 720 368 720 240
WIRE 720 368 672 368
WIRE 784 368 784 352
WIRE 480 384 480 368
WIRE 576 384 576 368
WIRE 672 384 672 368
WIRE 480 464 480 448
WIRE 576 464 576 448
WIRE 672 480 672 464
FLAG 384 336 0
FLAG 400 224 Vin
FLAG 640 208 U1O
FLAG 480 464 0
FLAG 528 368 R1C1
FLAG 784 368 0
FLAG 864 272 0
FLAG 512 192 U1N
FLAG 576 464 0
FLAG 672 480 0
SYMBOL voltage 384 224 R0
WINDOW 0 33 22 Left 2
WINDOW 3 37 77 Left 2
WINDOW 123 35 103 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value 10
SYMATTR Value2 AC 1
SYMBOL opamps\\lt1028 576 144 R0
SYMATTR InstName U1
SYMBOL res 624 96 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R2
SYMATTR Value 4950
SYMBOL res 464 240 R0
SYMATTR InstName R1
SYMATTR Value 50
SYMBOL cap 464 384 R0
SYMATTR InstName C1
SYMATTR Value 10mf
SYMATTR SpiceLine Rser=25m Lser=2n
SYMBOL voltage 784 368 M180
WINDOW 0 33 22 Left 2
WINDOW 3 33 44 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V2
SYMATTR Value 15
SYMBOL voltage 864 160 R0
WINDOW 0 33 22 Left 2
WINDOW 3 33 44 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V3
SYMATTR Value 15
SYMBOL cap 608 48 M270
WINDOW 0 32 32 VTop 2
WINDOW 3 0 32 VBottom 2
SYMATTR InstName C2
SYMATTR Value 100p
SYMBOL cap 560 384 R0
SYMATTR InstName C3
SYMATTR Value 10mf
SYMATTR SpiceLine Rser=25m Lser=2n
SYMBOL res 656 368 R0
SYMATTR InstName Rleak
SYMATTR Value 1e6
SYMBOL diode 704 240 M180
WINDOW 0 24 64 Left 2
WINDOW 3 24 0 Left 2
SYMATTR InstName D1
SYMATTR Value 1N4001
TEXT 368 -96 Left 2 ;'Measure Power Supply, Zener, LED Noise
TEXT 368 -64 Left 2 !.ac dec 500 1m 1e8\n;.tran 2

Here is the frequency response:

[AC Analysis]
{
Npanes: 1
{
traces: 1 {524290,0,"V(u1o)"}
X: ('M',1,0.001,0,1e+008)
Y[0]: (' ',0,0.125892541179417,6,125.892541179417)
Y[1]: (' ',0,-280,40,80)
Volts: (' ',0,0,3,9.99,0.002,10.012)
Log: 1 2 0
GridStyle: 1
PltMag: 1
}
}
[Transient Analysis]
{
Npanes: 1
{
traces: 3 {524290,0,"V(m1s)"} {589828,0,"V(m1g)"} {524291,0,"V(vdc)"}
X: (' ',1,0,0.2,2)
Y[0]: (' ',1,-0.9,0.9,9)
Y[1]: (' ',0,1e+308,30,-1e+308)
Volts: (' ',0,0,0,-0.9,0.9,9)
Log: 0 0 0
GridStyle: 1
PltMag: 1
PltPhi: 1 0
}
}

Here are different versions of the power supply ripple filters.

Version 4
SHEET 1 1888 1296
WIRE -48 48 -64 48
WIRE 16 48 -48 48
WIRE 144 48 16 48
WIRE 288 48 240 48
WIRE 320 48 288 48
WIRE 448 48 320 48
WIRE 576 48 448 48
WIRE 768 48 672 48
WIRE 864 48 768 48
WIRE 896 48 864 48
WIRE -64 64 -64 48
WIRE 768 64 768 48
WIRE 896 64 896 48
WIRE 288 80 288 48
WIRE 16 128 16 48
WIRE 32 128 16 128
WIRE 160 128 112 128
WIRE 192 128 192 112
WIRE 192 128 160 128
WIRE 448 128 448 48
WIRE 464 128 448 128
WIRE 592 128 544 128
WIRE 624 128 624 112
WIRE 624 128 592 128
WIRE 192 144 192 128
WIRE 624 144 624 128
WIRE 768 144 768 128
WIRE -64 160 -64 144
WIRE 288 160 288 144
WIRE 896 160 896 144
WIRE 192 224 192 208
WIRE 624 224 624 208
WIRE -16 256 -64 256
WIRE 144 256 -16 256
WIRE 288 256 240 256
WIRE 320 256 288 256
WIRE 448 256 320 256
WIRE 576 256 448 256
WIRE 768 256 672 256
WIRE 864 256 768 256
WIRE 896 256 864 256
WIRE -64 272 -64 256
WIRE 768 272 768 256
WIRE 896 272 896 256
WIRE 288 288 288 256
WIRE -16 336 -16 256
WIRE 32 336 -16 336
WIRE 160 336 112 336
WIRE 192 336 192 320
WIRE 192 336 160 336
WIRE 448 336 448 256
WIRE 464 336 448 336
WIRE 592 336 544 336
WIRE 624 336 624 320
WIRE 624 336 592 336
WIRE 192 352 192 336
WIRE 624 352 624 336
WIRE 768 352 768 336
WIRE -64 368 -64 352
WIRE 288 368 288 352
WIRE 896 368 896 352
WIRE 192 432 192 416
WIRE 624 432 624 416
WIRE 16 464 16 128
WIRE 144 464 16 464
WIRE 288 464 240 464
WIRE 368 464 288 464
WIRE 400 464 368 464
WIRE 448 464 400 464
WIRE 576 464 448 464
WIRE 736 464 672 464
WIRE 768 464 736 464
WIRE 880 464 768 464
WIRE 896 464 880 464
WIRE 768 480 768 464
WIRE 896 480 896 464
WIRE 368 496 368 464
WIRE 16 544 16 464
WIRE 32 544 16 544
WIRE 160 544 112 544
WIRE 192 544 192 528
WIRE 192 544 160 544
WIRE 208 544 192 544
WIRE 288 544 288 464
WIRE 288 544 272 544
WIRE 448 544 448 464
WIRE 464 544 448 544
WIRE 592 544 544 544
WIRE 624 544 624 528
WIRE 624 544 592 544
WIRE 656 544 624 544
WIRE 736 544 736 464
WIRE 736 544 720 544
WIRE 768 560 768 544
WIRE 368 576 368 560
WIRE 896 576 896 560
WIRE 16 656 16 544
WIRE 144 656 16 656
WIRE 304 656 240 656
WIRE 368 656 304 656
WIRE 400 656 368 656
WIRE 448 656 400 656
WIRE 576 656 448 656
WIRE 736 656 672 656
WIRE 768 656 736 656
WIRE 880 656 768 656
WIRE 896 656 880 656
WIRE 768 672 768 656
WIRE 896 672 896 656
WIRE 368 688 368 656
WIRE 16 736 16 656
WIRE 48 736 16 736
WIRE 160 736 112 736
WIRE 192 736 192 720
WIRE 192 736 160 736
WIRE 208 736 192 736
WIRE 304 736 304 656
WIRE 304 736 288 736
WIRE 448 736 448 656
WIRE 480 736 448 736
WIRE 592 736 544 736
WIRE 624 736 624 720
WIRE 624 736 592 736
WIRE 640 736 624 736
WIRE 736 736 736 656
WIRE 736 736 720 736
WIRE 768 752 768 736
WIRE 368 768 368 752
WIRE 896 768 896 752
WIRE 16 864 16 736
WIRE 144 864 16 864
WIRE 272 864 240 864
WIRE 288 864 272 864
WIRE 352 864 288 864
WIRE 448 864 352 864
WIRE 576 864 448 864
WIRE 720 864 672 864
WIRE 784 864 720 864
WIRE 864 864 784 864
WIRE 896 864 864 864
WIRE 896 880 896 864
WIRE 352 896 352 864
WIRE 784 896 784 864
WIRE 192 944 192 928
WIRE 288 944 288 864
WIRE 624 944 624 928
WIRE 720 944 720 864
WIRE 352 976 352 960
WIRE 784 976 784 960
WIRE 896 976 896 960
WIRE 16 1040 16 864
WIRE 32 1040 16 1040
WIRE 208 1040 112 1040
WIRE 240 1040 240 1008
WIRE 240 1040 208 1040
WIRE 448 1040 448 864
WIRE 464 1040 448 1040
WIRE 640 1040 544 1040
WIRE 672 1040 672 1008
WIRE 672 1040 640 1040
WIRE 240 1056 240 1040
WIRE 672 1056 672 1040
WIRE 240 1136 240 1120
WIRE 672 1136 672 1120
FLAG -64 160 0
FLAG -48 48 VCC
FLAG 192 224 0
FLAG 320 48 Q1E
FLAG 288 160 0
FLAG 160 128 Q1B
FLAG 592 128 Q2B
FLAG 768 144 0
FLAG 864 48 Classic
FLAG 624 224 0
FLAG 896 160 0
FLAG 400 464 Q7E
FLAG 368 576 0
FLAG 160 544 Q7B
FLAG 592 544 Q8B
FLAG 768 560 0
FLAG 880 464 NPN
FLAG 896 576 0
FLAG 352 976 0
FLAG 240 1136 0
FLAG 272 864 Q3C
FLAG 208 1040 Q4B
FLAG 784 976 0
FLAG 672 1136 0
FLAG 864 864 Sziklai
FLAG 640 1040 Q6B
FLAG 896 976 0
FLAG 400 656 Q9C
FLAG 368 768 0
FLAG 160 736 Q9B
FLAG 592 736 Q10B
FLAG 768 752 0
FLAG 880 656 PNP
FLAG 896 768 0
FLAG 192 432 0
FLAG 320 256 Q11E
FLAG 288 368 0
FLAG 160 336 Q11B
FLAG 592 336 Q12B
FLAG 768 352 0
FLAG 864 256 NegVolt
FLAG 624 432 0
FLAG 896 368 0
FLAG -64 368 0
SYMBOL voltage -64 48 R0
WINDOW 0 33 22 Left 2
WINDOW 3 37 77 Left 2
WINDOW 123 35 103 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value 15
SYMATTR Value2 AC 1
SYMBOL npn 144 112 R270
WINDOW 0 43 35 VRight 2
WINDOW 3 65 27 VRight 2
SYMATTR InstName Q1
SYMATTR Value 2N3904
SYMBOL res 128 112 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R1
SYMATTR Value 10k
SYMBOL cap 176 144 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C1
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 272 80 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C2
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL npn 576 112 R270
WINDOW 0 43 35 VRight 2
WINDOW 3 69 25 VRight 2
SYMATTR InstName Q2
SYMATTR Value 2N3904
SYMBOL res 560 112 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R2
SYMATTR Value 10k
SYMBOL cap 608 144 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C3
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 752 64 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C4
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 880 48 R0
SYMATTR InstName R3
SYMATTR Value 1k
SYMBOL res 128 528 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R4
SYMATTR Value 10k
SYMBOL cap 272 528 R90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C5
SYMATTR Value 1000uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 352 496 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C6
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 560 528 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R5
SYMATTR Value 10k
SYMBOL cap 752 480 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C7
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 880 464 R0
SYMATTR InstName R6
SYMATTR Value 1k
SYMBOL pnp 240 928 M270
WINDOW 0 60 59 VLeft 2
WINDOW 3 97 68 VLeft 2
SYMATTR InstName Q3
SYMATTR Value 2N3906
SYMBOL npn 192 1008 R270
WINDOW 0 43 35 VRight 2
WINDOW 3 80 25 VRight 2
SYMATTR InstName Q4
SYMATTR Value 2N3904
SYMBOL res 128 1024 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R7
SYMATTR Value 47k
SYMBOL cap 224 1056 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C8
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 336 896 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C9
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL pnp 672 928 M270
WINDOW 0 60 59 VLeft 2
WINDOW 3 97 68 VLeft 2
SYMATTR InstName Q5
SYMATTR Value 2N3906
SYMBOL npn 624 1008 R270
WINDOW 0 43 35 VRight 2
WINDOW 3 80 25 VRight 2
SYMATTR InstName Q6
SYMATTR Value 2N3904
SYMBOL res 560 1024 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R8
SYMATTR Value 47k
SYMBOL cap 656 1056 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C10
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 768 896 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C11
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 880 864 R0
SYMATTR InstName R9
SYMATTR Value 1k
SYMBOL npn 144 528 R270
WINDOW 0 43 35 VRight 2
WINDOW 3 64 27 VRight 2
SYMATTR InstName Q7
SYMATTR Value 2N3904
SYMBOL npn 576 528 R270
WINDOW 0 43 35 VRight 2
WINDOW 3 63 29 VRight 2
SYMATTR InstName Q8
SYMATTR Value 2N3904
SYMBOL cap 720 528 R90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C12
SYMATTR Value 1000uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 304 720 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R10
SYMATTR Value 10k
SYMBOL cap 112 720 R90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C13
SYMATTR Value 1000uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 352 688 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C14
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 736 720 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R11
SYMATTR Value 10k
SYMBOL cap 752 672 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C15
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 880 656 R0
SYMATTR InstName R12
SYMATTR Value 1k
SYMBOL cap 544 720 R90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C16
SYMATTR Value 1000uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL pnp 240 720 M270
WINDOW 0 60 59 VLeft 2
WINDOW 3 97 68 VLeft 2
SYMATTR InstName Q9
SYMATTR Value 2N3906
SYMBOL pnp 672 720 M270
WINDOW 0 60 59 VLeft 2
WINDOW 3 97 68 VLeft 2
SYMATTR InstName Q10
SYMATTR Value 2N3906
SYMBOL cap 176 352 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C17
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 272 288 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C18
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 560 320 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R13
SYMATTR Value 10k
SYMBOL cap 608 352 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C19
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL cap 752 272 R0
WINDOW 0 34 15 Left 2
WINDOW 3 34 52 Left 2
SYMATTR InstName C20
SYMATTR Value 470uf
SYMATTR SpiceLine Rser=17m Lser=10n
SYMBOL res 880 256 R0
SYMATTR InstName R14
SYMATTR Value 1k
SYMBOL pnp 144 320 R270
WINDOW 0 40 38 VRight 2
WINDOW 3 60 26 VRight 2
SYMATTR InstName Q11
SYMATTR Value 2N3906
SYMBOL pnp 576 320 R270
WINDOW 0 40 38 VRight 2
WINDOW 3 60 26 VRight 2
SYMATTR InstName Q12
SYMATTR Value 2N3906
SYMBOL voltage -64 368 R180
WINDOW 0 33 22 Left 2
WINDOW 3 37 77 Left 2
WINDOW 123 35 103 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V2
SYMATTR Value 15
SYMATTR Value2 AC 1
SYMBOL res 128 320 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R15
SYMATTR Value 10k
TEXT 200 -96 Left 2 ;'Low Noise Ripple Filter for Opamps
TEXT 200 -64 Left 2 !.ac dec 500 1m 1e6
TEXT 120 1216 Left 2 ;Capacitors are Kemet Solid Polymer, 470uF 25V, 17mohm
ESR

Here's the PLT files. Watch the wrap:

[AC Analysis]
{
Npanes: 1
{
traces: 5 {524290,0,"V(classic)"} {524291,0,"V(negvolt)"}
{524292,0,"V(npn)"} {524293,0,"V(pnp)"} {524294,0,"V(sziklai)"}
X: ('M',0,0.001,0,1e+006)
Y[0]: (' ',0,1e-012,20,10)
Y[1]: (' ',0,-400,80,480)
Volts: (' ',0,0,3,5.864,0.001,5.876)
Log: 1 2 0
GridStyle: 1
PltMag: 1

[Gerhard Hoffmann]

Am 18.09.21 um 01:19 schrieb Steve Wilson:

> ISTR Win used a bunch of bipolars to get the noise down to pV/root(Hz) for
> ribbon microphones. This required an extremely large input coupling
> capacitor to prevent shorting out the base bias and still pass the low
> frequencies needed.

He has a differential version that does not require the huge input
capacitor. I have built the single-ended version:
<
https://www.flickr.com/photos/137684711@N07/44311358915/in/album-72157662535945536/lightbox/
>

and yes, it features ~70 pV/rt (Hz), like promised. I wanted to use
it in the chopper and at 10 KHz the required input capacitor
is harmless.

> I am working on a different approach that DC-couples the input signal to an
> op amp, or several in parallel for low noise. There is no couping capacitor
> needed, and the op amp eliminates the risk of blowing out the base due to

beware of such an event:
<
https://www.flickr.com/photos/137684711@N07/51486809047/in/dateposted-public/
>

I noted a rise of the noise voltage above 500 KHz and in search of the
culprit I removed the wet slug tantalum. I was not the tan, it was the
routing:
<
https://www.flickr.com/photos/137684711@N07/51488299629/in/dateposted-public/
>

Substituting a mesh removed the effect.

Then I forgot until next week end that the tantalum was removed, and I
measured the voltage noise of a lithium battery. Ouch! That stench!
And the amplifier was in a alu-cast Hammond box that was closed.
6 screws.

> the high surge current from a power supply. The capacitors in the feedback
> network need to be insulated or covered to minimize low frequency voltage
> shift due to random air currents affecting the tempco of the capacitors.
> Thanks to P. Hobbs for this observation.
>
> I am also working on high attenuation ripple filters for the op amp power
> supplies. I'd appreciate your comments.

When the op amp has 100 dB pssr, you do not need such extreme external
suppression. I'd be ashamed to admit that I need such suppression.
It would say that I have a huge source of dirt on my board.

Run such an amplifier from Li batteries. The easiest way to get rid of
some ground loops.

Cheers, Gerhard

[Phil Hobbs]

Thing is, the op amp PSR ain't any 100 dB up at SMPS frequencies. Also
of course that 100 dB number is input-referred--if you're running a gain
of 100, the output-referred PSR becomes 60 dB even near DC.

Op amps don't care much about somewhat-drifty supply rails, but can use
some PSR help above about 1 kHz, which makes cap multipliers an
excellent match.

>
> Run such an amplifier from Li batteries. The easiest way to get rid of
> some ground loops.

For lab use, I agree.

Cheers

Phil Hobbs

[Rich S]

"the father of the AD797"
who was that, Gerhard?
I'd love to thank him.
Popped one of those into the front end of an
already-quiet test equipment was the easiest &
quickest and substantial improvement I had.
This was ca. 1986, not long after it appeared.
cheers, RS