EDAboard.com | EDAboard.de | EDAboard.co.uk | WTWH Media

Composite amps

Ask a question - edaboard.com

elektroda.net NewsGroups Forum Index - Electronics Design - Composite amps

Goto page Previous  1, 2, 3, 4, 5, 6, 7  Next


Guest

Thu Feb 27, 2020 3:45 am   



On Wed, 26 Feb 2020 16:57:25 -0500, Phil Hobbs
<pcdhSpamMeSenseless_at_electrooptical.net> wrote:

Quote:
On 2020-02-26 13:36, George Herold wrote:
On Monday, February 24, 2020 at 2:47:45 PM UTC-5, Phil Hobbs wrote:
On 2020-02-24 11:47, George Herold wrote:
On Monday, February 24, 2020 at 6:55:54 AM UTC-5, plastco...@gmail.com wrote:
1) grounded detector, grounded source/emitter HF transistor
2) true zero-bias operation of detector
3) my circuit is simpler
I like to read books))

OK, what do you find better about zero bias operation?
I should admit that for many years I ran all my PD's at
zero bias. I thought this gave me better 'zero' light detection.
(No DC offset with no light... but the dark current from
PDs is generally pretty low.)
Running with some bias has two main advantages.
1.) reduced C.. faster
2.) Higher saturation current (light intensity) without bias the
electrons build up in the junction and it saturates.. more light
gives no more electrons.

George H.
(who is addicted to reading... I need to find a few new fiction writers)


Zero bias is better in one respect: you can get zero leakage current.
For jobs such as very wide range, very slow photometers, that's a win.
Garry Epeldauer et al. wrote a beautiful paper about getting 14 orders
of magnitude in photocurrent, if you don't mind being stuck with
millihertz bandwidths:

https://electrooptical.net/www/optics/eppeldauer14decadephotocurrent.pdf
Hi Phil, I downloaded the above and was chewing through it last night.
Great stuff!
1.) Rs (PD shunt resistance... I've always just treated this as
infinite. Can I measure leakage current and get Rs?
2.) pg 3094 has a nice discussion of 1/f noise.
3.) Are there even better low current opamps these days?
4.) Nice effective BW calcs in App A.
I would add to that, the ENBW for a two pole filter, f_3dB and Q,
is
ENBW = f_3dB * Q *pi/2 = ~1.11 f_3dB (Q=0.707.. Butterworth)


Re: noise BW
Yeah, it's like 1.22x for two noninteracting RC poles, so 1.11 for
Butterworth sounds roughly right.

Re: shunt resistance

For bias voltages << kT/e, both the forward and reverse diffusion
currents are contributing to the conductance--it's just dI/dV, and so is
fairly far from zero for a large-area diode run at zero bias.

Ideally the effective shunt resistance goes up by a factor of 2 or so
with 50 mV of reverse bias, because you shut off the reverse diffusion
current without introducing significant additional leakage. That's a
super useful trick with InGaAs diodes in dim light.

re: 1/f noise
Haven't read it recently, but in photodiodes you actually can get
significant 1/f noise at zero bias, unlike in the case of resistors.

re: low current op amps

BITD I used to really like the OPA111. Its performance was easy to
remember: 1 MHz bandwidth, 1 mV offset, 1 uV/K drift, 1 pA input bias.
(See? I haven't used one in 30 years and I still remember.) ;)

It was one of the primo op amps used in early tunnelling and atomic
force microscopy. Of course it's noisy, but not nearly as bad as the
other popular super-high-Z op amp of the day, namely the LM11.

Nowadays there are much better choices, e.g. JL's fave OPA197.

Cheers

Phil Hobbs


I don't use that as a low noise high-performance amp, but as a
general-purpose gumdrop. It's stable with a big output cap, 3.3u film
or 100u polymer.



--

John Larkin Highland Technology, Inc

The cork popped merrily, and Lord Peter rose to his feet.
"Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"

George Herold
Guest

Thu Feb 27, 2020 3:45 pm   



On Wednesday, February 26, 2020 at 9:18:30 PM UTC-5, jla...@highlandsniptechnology.com wrote:
Quote:
On Wed, 26 Feb 2020 16:57:25 -0500, Phil Hobbs
pcdhSpamMeSenseless_at_electrooptical.net> wrote:

On 2020-02-26 13:36, George Herold wrote:
On Monday, February 24, 2020 at 2:47:45 PM UTC-5, Phil Hobbs wrote:
On 2020-02-24 11:47, George Herold wrote:
On Monday, February 24, 2020 at 6:55:54 AM UTC-5, plastco...@gmail.com wrote:
1) grounded detector, grounded source/emitter HF transistor
2) true zero-bias operation of detector
3) my circuit is simpler
I like to read books))

OK, what do you find better about zero bias operation?
I should admit that for many years I ran all my PD's at
zero bias. I thought this gave me better 'zero' light detection.
(No DC offset with no light... but the dark current from
PDs is generally pretty low.)
Running with some bias has two main advantages.
1.) reduced C.. faster
2.) Higher saturation current (light intensity) without bias the
electrons build up in the junction and it saturates.. more light
gives no more electrons.

George H.
(who is addicted to reading... I need to find a few new fiction writers)


Zero bias is better in one respect: you can get zero leakage current.
For jobs such as very wide range, very slow photometers, that's a win.
Garry Epeldauer et al. wrote a beautiful paper about getting 14 orders
of magnitude in photocurrent, if you don't mind being stuck with
millihertz bandwidths:

https://electrooptical.net/www/optics/eppeldauer14decadephotocurrent.pdf
Hi Phil, I downloaded the above and was chewing through it last night.
Great stuff!
1.) Rs (PD shunt resistance... I've always just treated this as
infinite. Can I measure leakage current and get Rs?
2.) pg 3094 has a nice discussion of 1/f noise.
3.) Are there even better low current opamps these days?
4.) Nice effective BW calcs in App A.
I would add to that, the ENBW for a two pole filter, f_3dB and Q,
is
ENBW = f_3dB * Q *pi/2 = ~1.11 f_3dB (Q=0.707.. Butterworth)


Re: noise BW
Yeah, it's like 1.22x for two noninteracting RC poles, so 1.11 for
Butterworth sounds roughly right.

Re: shunt resistance

For bias voltages << kT/e, both the forward and reverse diffusion
currents are contributing to the conductance--it's just dI/dV, and so is
fairly far from zero for a large-area diode run at zero bias.

Ideally the effective shunt resistance goes up by a factor of 2 or so
with 50 mV of reverse bias, because you shut off the reverse diffusion
current without introducing significant additional leakage. That's a
super useful trick with InGaAs diodes in dim light.

re: 1/f noise
Haven't read it recently, but in photodiodes you actually can get
significant 1/f noise at zero bias, unlike in the case of resistors.

re: low current op amps

BITD I used to really like the OPA111. Its performance was easy to
remember: 1 MHz bandwidth, 1 mV offset, 1 uV/K drift, 1 pA input bias.
(See? I haven't used one in 30 years and I still remember.) ;)

It was one of the primo op amps used in early tunnelling and atomic
force microscopy. Of course it's noisy, but not nearly as bad as the
other popular super-high-Z op amp of the day, namely the LM11.

Nowadays there are much better choices, e.g. JL's fave OPA197.

Cheers

Phil Hobbs

I don't use that as a low noise high-performance amp, but as a
general-purpose gumdrop. It's stable with a big output cap, 3.3u film
or 100u polymer.

OK Thanks. Is there some nice opamp like this (opa197 or 192) that
has a little more GBW? I'm looking at table 4X.2 (AoEx) High spped
VFB's... there are a lot to choose from. 50-100 MHz would be nice.

opa1611 looks OK

George H.
Quote:


--

John Larkin Highland Technology, Inc

The cork popped merrily, and Lord Peter rose to his feet.
"Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"


Lasse Langwadt Christense
Guest

Thu Feb 27, 2020 8:45 pm   



torsdag den 27. februar 2020 kl. 19.59.42 UTC+1 skrev whit3rd:
Quote:
On Tuesday, February 25, 2020 at 3:55:22 PM UTC-8, George Herold wrote:
On Tuesday, February 25, 2020 at 2:13:52 PM UTC-5, plastco...@gmail.com wrote:

Russian photodiode for 13 nm, is expensive.
Example: 4keV photon,...

.. easier to measure x-ray
energy from a pulse height or total current or something.)

At 4 keV, I'd want a proportional counter with (Xenon?) gas,
or a fairly large volume ion chamber (leakage current in biased
parallel-plate capacitor). Neither is a tiny solid detector.

A phosphor, of course, can generate light flashes when hit by X-rays,
and photodiodes can be efficient at detecting the secondary radiation.


years ago I did X-ray fluorescence spectrometry with an Fe55 source (5.9keV)
and a CCD as detector, image processed to find all pixels with empty neighbours


Guest

Thu Feb 27, 2020 8:45 pm   



For gas sensor :
4 keV / 30 = 133ē
For scintillator - SiPM !

http://ixbt.photo/?id=photo:1330019

whit3rd
Guest

Thu Feb 27, 2020 8:45 pm   



On Tuesday, February 25, 2020 at 3:55:22 PM UTC-8, George Herold wrote:
Quote:
On Tuesday, February 25, 2020 at 2:13:52 PM UTC-5, plastco...@gmail.com wrote:

Russian photodiode for 13 nm, is expensive.
Example: 4keV photon,...

.. easier to measure x-ray
energy from a pulse height or total current or something.)


At 4 keV, I'd want a proportional counter with (Xenon?) gas,
or a fairly large volume ion chamber (leakage current in biased
parallel-plate capacitor). Neither is a tiny solid detector.

A phosphor, of course, can generate light flashes when hit by X-rays,
and photodiodes can be efficient at detecting the secondary radiation.


Guest

Fri Feb 28, 2020 12:45 am   



On Thursday, 27 February 2020 19:20:03 UTC, plastco...@gmail.com wrote:
Quote:
For gas sensor :
4 keV / 30 = 133ē
For scintillator - SiPM !

http://ixbt.photo/?id=photo:1330019


eh?


Guest

Fri Feb 28, 2020 12:45 am   



Best result OPA140+BF862 (L=1mH+1mH drain load):
http://ixbt.photo/?id=photo:1329969

Rf=91k and Rf=1M
http://ixbt.photo/?id=photo:1330021

Pushmi-Pullyu
dual accelerator - input JFET, output pnp BJT
DC precision
power supply +-3.8V
active probe - bootstrapped dual JFET half-bridge CPH6904 (out=in+0.86V)
(mediocre, better jfet+ADA4860 or FVF (Flipped Voltage Follower) with BF998)
http://ixbt.photo/?id=photo:1330027

Tortoise outrun Achilles !
time: 1.25us
Сircuit speed indifferent to input capacitance !
http://ixbt.photo/?id=photo:1330025

Сheetah enters the arena:
http://ixbt.photo/?id=photo:1330022
only 1M + 2200pF + OPA656
poor substitute, OPA140+jfet better

The final, a curtain.
OPA656 + BF862
http://ixbt.photo/?id=photo:1330023
GBW=31 GHz ? ((

George Herold
Guest

Fri Feb 28, 2020 1:45 am   



On Thursday, February 27, 2020 at 6:10:34 PM UTC-5, plastco...@gmail.com wrote:
Quote:
Best result OPA140+BF862 (L=1mH+1mH drain load):
http://ixbt.photo/?id=photo:1329969

I'm confused about the two traces are here.
The BF862 is unfortunately no more...
You can only surf so long on the trailing edge of technology.

Quote:

Is this for me? Hard to tell the shape of the response
for the 91k. But the 1M looks more like a two pole response
now anyway.
Quote:

Pushmi-Pullyu
dual accelerator - input JFET, output pnp BJT
DC precision
power supply +-3.8V
active probe - bootstrapped dual JFET half-bridge CPH6904 (out=in+0.86V)
(mediocre, better jfet+ADA4860 or FVF (Flipped Voltage Follower) with BF998)
http://ixbt.photo/?id=photo:1330027


So all the stuff on the left is an active probe?
To look at the inverting node without loading down with capacitance?
So is this right? The jfet is taking the fast part of the
current signal and sending it through to the non-inverting input?

I don't know what the pnp is doing on the output. Unless
it's a driver for some coax cable output?
It looks the the 'output' (your arrow between collector
of pnp and 1k R to -V) should be near -V... are you
now biasing PD via the output?
Quote:

Tortoise outrun Achilles !
time: 1.25us
Сircuit speed indifferent to input capacitance !
http://ixbt.photo/?id=photo:1330025
Well at least to C on the inverting node.

Сheetah enters the arena:
http://ixbt.photo/?id=photo:1330022
only 1M + 2200pF + OPA656
poor substitute, OPA140+jfet better

The final, a curtain.
OPA656 + BF862
http://ixbt.photo/?id=photo:1330023
GBW=31 GHz ? ((

Huh the opa656 looks nice.. +/- 5V.
Thanks

Have you been lurking here long... I'm just reminded of
Jan Pantelje (sp) Which is a good thing. I like Jan.

George H.


Guest

Fri Feb 28, 2020 8:45 am   



Quote:
Rf=91k and Rf=1M
http://ixbt.photo/?id=photo:1330021
Is this for me? Hard to tell the shape of the response
for the 91k. But the 1M looks more like a two pole response
now anyway.
With Rf=91k now Cf>0pF must be present


Quote:

Pushmi-Pullyu
dual accelerator - input JFET, output pnp BJT
DC precision
power supply +-3.8V
active probe - bootstrapped dual JFET half-bridge CPH6904 (out=in+0.86V)
(mediocre, better jfet+ADA4860 or FVF (Flipped Voltage Follower) with BF998)
http://ixbt.photo/?id=photo:1330027

So all the stuff on the left is an active probe?
To look at the inverting node without loading down with capacitance?

Yes, the circuit is very sensitive to the probe capacitance.
Non-inverting node !

Quote:
So is this right? The jfet is taking the fast part of the
current signal and sending it through to the non-inverting input?
JFET is input AC amplifier with feedback through Op Amp TIA))


Quote:
I don't know what the pnp is doing on the output. Unless
it's a driver for some coax cable output?
It looks the the 'output' (your arrow between collector
of pnp and 1k R to -V) should be near -V... are you
now biasing PD via the output?

MMBTH81 is amplifier with gain~10
look in the upper right corner of the book AoE))
no photodiode offset !
slew rate only doubles((
overall gain increases

Capacitance SiPM 60035 = 3400pF
Dmitriy P.


Guest

Fri Feb 28, 2020 8:45 am   



On Friday, 28 February 2020 00:44:11 UTC, George Herold wrote:
Quote:
On Thursday, February 27, 2020 at 6:10:34 PM UTC-5, plastco...@gmail.com wrote:

Best result OPA140+BF862 (L=1mH+1mH drain load):
http://ixbt.photo/?id=photo:1329969

I'm confused about the two traces are here.
The BF862 is unfortunately no more...
You can only surf so long on the trailing edge of technology.


Russia is expert at that game.


NT

Phil Hobbs
Guest

Fri Feb 28, 2020 8:45 pm   



On 2020-02-26 21:18, jlarkin_at_highlandsniptechnology.com wrote:
Quote:
On Wed, 26 Feb 2020 16:57:25 -0500, Phil Hobbs
pcdhSpamMeSenseless_at_electrooptical.net> wrote:

On 2020-02-26 13:36, George Herold wrote:
On Monday, February 24, 2020 at 2:47:45 PM UTC-5, Phil Hobbs wrote:
On 2020-02-24 11:47, George Herold wrote:
On Monday, February 24, 2020 at 6:55:54 AM UTC-5, plastco...@gmail.com wrote:
1) grounded detector, grounded source/emitter HF transistor
2) true zero-bias operation of detector
3) my circuit is simpler
I like to read books))

OK, what do you find better about zero bias operation?
I should admit that for many years I ran all my PD's at
zero bias. I thought this gave me better 'zero' light detection.
(No DC offset with no light... but the dark current from
PDs is generally pretty low.)
Running with some bias has two main advantages.
1.) reduced C.. faster
2.) Higher saturation current (light intensity) without bias the
electrons build up in the junction and it saturates.. more light
gives no more electrons.

George H.
(who is addicted to reading... I need to find a few new fiction writers)


Zero bias is better in one respect: you can get zero leakage current.
For jobs such as very wide range, very slow photometers, that's a win.
Garry Epeldauer et al. wrote a beautiful paper about getting 14 orders
of magnitude in photocurrent, if you don't mind being stuck with
millihertz bandwidths:

https://electrooptical.net/www/optics/eppeldauer14decadephotocurrent.pdf
Hi Phil, I downloaded the above and was chewing through it last night.
Great stuff!
1.) Rs (PD shunt resistance... I've always just treated this as
infinite. Can I measure leakage current and get Rs?
2.) pg 3094 has a nice discussion of 1/f noise.
3.) Are there even better low current opamps these days?
4.) Nice effective BW calcs in App A.
I would add to that, the ENBW for a two pole filter, f_3dB and Q,
is
ENBW = f_3dB * Q *pi/2 = ~1.11 f_3dB (Q=0.707.. Butterworth)


Re: noise BW
Yeah, it's like 1.22x for two noninteracting RC poles, so 1.11 for
Butterworth sounds roughly right.

Re: shunt resistance

For bias voltages << kT/e, both the forward and reverse diffusion
currents are contributing to the conductance--it's just dI/dV, and so is
fairly far from zero for a large-area diode run at zero bias.

Ideally the effective shunt resistance goes up by a factor of 2 or so
with 50 mV of reverse bias, because you shut off the reverse diffusion
current without introducing significant additional leakage. That's a
super useful trick with InGaAs diodes in dim light.

re: 1/f noise
Haven't read it recently, but in photodiodes you actually can get
significant 1/f noise at zero bias, unlike in the case of resistors.

re: low current op amps

BITD I used to really like the OPA111. Its performance was easy to
remember: 1 MHz bandwidth, 1 mV offset, 1 uV/K drift, 1 pA input bias.
(See? I haven't used one in 30 years and I still remember.) ;)

It was one of the primo op amps used in early tunnelling and atomic
force microscopy. Of course it's noisy, but not nearly as bad as the
other popular super-high-Z op amp of the day, namely the LM11.

Nowadays there are much better choices, e.g. JL's fave OPA197.

Cheers

Phil Hobbs

I don't use that as a low noise high-performance amp, but as a
general-purpose gumdrop. It's stable with a big output cap, 3.3u film
or 100u polymer.


I didn't mean that you thought it was the best one, but it's sure better
than an OPA111. ;)

I'm also partial to the OPA140, which is just about the perfect JFET op
amp: 0.5 pA typical Ibias at 25C, 5 nV 1-Hz flatband noise, 30 Hz 1/f
corner, 11 MHz BW, 220 uV max offset over temperature, 0.25 uV/K typical
drift, $2. Not horrible at all.

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


Guest

Sat Feb 29, 2020 1:45 pm   



I remembered where the zero offset of the photodiodes is still important !
Excellent mid-infrared photodiodes from Russia -
http://www.mirdog.spb.ru/products.htm
Ro = 5 Ω ... 10 k

Dmitriy P.

George Herold
Guest

Sun Mar 01, 2020 1:45 am   



On Friday, February 28, 2020 at 2:15:33 AM UTC-5, plastco...@gmail.com wrote:
Quote:
Rf=91k and Rf=1M
http://ixbt.photo/?id=photo:1330021
Is this for me? Hard to tell the shape of the response
for the 91k. But the 1M looks more like a two pole response
now anyway.
With Rf=91k now Cf>0pF must be present


Pushmi-Pullyu
dual accelerator - input JFET, output pnp BJT
DC precision
power supply +-3.8V
active probe - bootstrapped dual JFET half-bridge CPH6904 (out=in+0.86V)
(mediocre, better jfet+ADA4860 or FVF (Flipped Voltage Follower) with BF998)
http://ixbt.photo/?id=photo:1330027

So all the stuff on the left is an active probe?
To look at the inverting node without loading down with capacitance?
Yes, the circuit is very sensitive to the probe capacitance.
Non-inverting node !

So is this right? The jfet is taking the fast part of the
current signal and sending it through to the non-inverting input?
JFET is input AC amplifier with feedback through Op Amp TIA))

I don't know what the pnp is doing on the output. Unless
it's a driver for some coax cable output?
It looks the the 'output' (your arrow between collector
of pnp and 1k R to -V) should be near -V... are you
now biasing PD via the output?
MMBTH81 is amplifier with gain~10
look in the upper right corner of the book AoE))
no photodiode offset !
slew rate only doubles((
overall gain increases


OK got it, thanks. I'm so-so at single transistor design.

George H.
Quote:

Capacitance SiPM 60035 = 3400pF
Dmitriy P.



Guest

Sat Mar 07, 2020 9:45 am   



Composite voltage follower
CFA ADA4860 + BF862
http://ixbt.photo/?id=photo:1330880

Output signals
yellow - circuit A
white - circuit B (bootstrapped)
http://ixbt.photo/?id=photo:1330881

Rise time circuit B (bootstrapped)
http://ixbt.photo/?id=photo:1330882


Guest

Sat Mar 07, 2020 8:45 pm   



very simple - imagine that a capacitor 2200pF is a photodiode with pulse photocurrent ~ 1 µA

Goto page Previous  1, 2, 3, 4, 5, 6, 7  Next

elektroda.net NewsGroups Forum Index - Electronics Design - Composite amps

Ask a question - edaboard.com

Arabic version Bulgarian version Catalan version Czech version Danish version German version Greek version English version Spanish version Finnish version French version Hindi version Croatian version Indonesian version Italian version Hebrew version Japanese version Korean version Lithuanian version Latvian version Dutch version Norwegian version Polish version Portuguese version Romanian version Russian version Slovak version Slovenian version Serbian version Swedish version Tagalog version Ukrainian version Vietnamese version Chinese version Turkish version
EDAboard.com map